Ultrasonics Sonochemistry 18 (2011) 92–98

Contents lists available at ScienceDirect

Ultrasonics Sonochemistry

journal homepage: www.elsevier .com/locate /u l tsonch

Development and optimization of acoustic bubble structures at high frequencies

Judy Lee a,*, Muthupandian Ashokkumar b, Kyuichi Yasui a, Toru Tuziuti a, Teruyuki Kozuka a,Atsuya Towata a, Yasuo Iida a

a National Institute of Advanced Industrial Science and Technology (AIST), 2266-98 Anagahora, Shimoshidami, Moriyama ku, Nagoya 463-8560, Japanb School of Chemistry, University of Melbourne, VIC 3010, Australia

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

Article history:Received 24 April 2009Received in revised form 31 January 2010Accepted 11 March 2010Available online 17 March 2010

Keywords:Ultrasound frequencyAcoustic bubble structureRadiation forcesAttenuationBubble coalescenceSurfactants

1350-4177/$ - see front matter � 2010 Elsevier B.V. Adoi:10.1016/j.ultsonch.2010.03.004

* Corresponding author. Tel.: +81 52 736 7215; faxE-mail address: judy.lee@aist.go.jp (J. Lee).

At high ultrasound frequencies, active bubble structures are difficult to capture due to the decrease intimescale per acoustic cycle and size of bubbles with increasing frequencies. However the current studydemonstrates an association between the spatial distribution of visible bubbles and that of the activebubble structure established in the path of the propagating acoustic wave. By monitoring the occurrenceof these visible bubbles, the development of active bubbles can be inferred for high frequencies. A seriesof still images depicting the formation of visible bubble structures suggest that a strong standing wavefield exists at early stages of wave propagation and weakens by the increase in the attenuation of theacoustic wave, caused by the formation of large coalesced bubbles. This attenuation is clearly demon-strated by the occurrence of a force which causes bubbles to be driven toward the liquid surface and limitstanding wave fields to near the surface. This force is explained in terms of the acoustic streaming andtraveling wave force. It is found that a strong standing wave field is established at 168 kHz. At448 kHz, large coalesced bubbles can significantly attenuate the acoustic pressure amplitude and weakenthe standing wave field. When the frequency is increased to 726 kHz, acoustic streaming becomes signif-icant and is the dominant force behind the disruption of the standing wave structure. The disruption ofthe standing wave structure can be minimized under certain pulse ON and OFF ratios.

� 2010 Elsevier B.V. All rights reserved.

1. Introduction

Acoustic cavitation is a unique phenomenon where extremepressures and temperatures are generated upon the collapse of mi-cron size bubbles. These conditions induce physical and chemicaleffects that are beneficial for a number of industrial applications[1–5]. However, due to the chaotic nature of acoustic cavitation,the control and optimization of the cavitation systems are difficultand have been the subject of a number of studies [6–13].

There are numerous reports in the literature examining theenhancement in integrated sonoluminescence (SL) intensities bythe addition of a surface active solute [14,15] and various sonica-tion conditions such as frequency [16,17], power [18,19], pulserepetition frequency [13,20], reactor vessel geometry [8] and liquidheight [9,21]. Literature reports [9,22–27] on the spatial dis-tribution of active bubbles and sonochemiluminescence (SCL)structures have demonstrated the importance of acquiring a homo-geneous spatial distribution of active bubbles for the enhancementin the integrated SL or SCL intensities. It is further demonstratedthat a homogeneous distribution of active bubbles is obtained

ll rights reserved.

: +81 52 736 7405.

when a strong standing wave field is established in the system[25,28]. However, these SL and SCL structures reported to dateare equilibrium structures observed when steady-states werereached under specific experimental conditions. An understandinginto the development of the SL and SCL structures is critical for abetter control and employment of ultrasound for industrial appli-cations, but is difficult to capture as a function of time due to theshort time scales (milliseconds) and a relatively low number ofactive bubbles.

In the early stages of cavitation, fast photography have revealedthat bubbles emerge from some point source and streams towardthe antinodes to form a dense cluster and filamentary like struc-tures [19,29,30]. However, these investigations were restricted tolow frequencies for the reason that the bubble size and wavelengthdecrease with increasing frequency, making direct imaging diffi-cult without special high speed imaging systems.

The data presented in this manuscript will demonstrate theresemblance between the visible bubble structure and that of theSL bubble structure under equilibrium (steady-state) conditionsat high frequencies. Using this association, the development of ac-tive bubble structures is indirectly inferred from the developmentof visible bubble structures at 168, 448 and 726 kHz. This manu-script will further show how acoustic streaming and weakening

Transducer 1(V1)

Transducer 2Receiver

J. Lee et al. / Ultrasonics Sonochemistry 18 (2011) 92–98 93

of the standing wave field can cause the spatial distribution of bub-bles to become inhomogeneous and how this can be minimizedunder appropriate pulse ON and OFF ratio.

Transducer 1(V2)

Transducer 2Receiver

Fig. 1. Schematic of the system used to measure attenuation of acoustic waves bybubbles. Transducer 1 is modulated between two output voltages, V1 and V2. Thevoltage for V1 is equivalent to 20 W driving power, sufficient to create cavitationbubbles. V2 is set at 10% of V1, below the cavitation threshold, and is attenuated bythe bubbles created by V1. The acoustic signals is received by transducer 2 andcollected by an oscilloscope.

(i)

(ii)

168 kHz 448 kHz 726 kHz 448 kHz

Air/liquid surface stabilized with silicone foam

Free air/liquid surface

Fig. 2. (i) Visible bubble structure and (ii) SL bubble structure as a function offrequency for saturated water and driving power of 20 W. For 448 kHz, the effect ofstabilizing the surface fluctuations with a sheet of silicone foam is shown. Exposuretimes for bubble structure and SL images were 33 ms and 30 s, respectively.

2. Experimental method

2.1. Solutions and sonication conditions

Sodium dodecylsulfate (SDS) was purchased from Sigma–Al-drich, special purity grade and were used without further purifica-tion. The SDS solutions were made by diluting an appropriatevolume of 100 mM stock solution with distilled water that hasbeen saturated with air.

Three 5 cm in diameter piezo-electric transducers (Kaijo SonicCorp.) at resonance frequencies of 168, 448 and 726 kHz wereused. Signals from the function generator (NF Wavefactory, model1946A) were amplified by a power amplifier (NF High Speed bipo-lar Amplifier, HSA4014). The power output to the transducer wasmeasured using a Megasonic power meter (Towa Electronic, modelTDW 6102U). For all experiments a power output of 20 W (1.1 W/cm2), unless stated otherwise, was used. Degassed water was ob-tained using a membrane degassing unit (Membrana, MiniMod-ule). The level of dissolved air content was determined by adissolved oxygen meter (Horiba, model D-25).

2.2. Bubble structure images

The experimental configuration used was the same as that de-scribed in Ref. [25] where a glass vessel which holds 1 L of solutionwas radiated with ultrasound emitted from the transducer fitted atthe bottom. The initial development of bubble structures were cap-tured using a high speed digital compact camera (Casio, Exilim EX-FH20) at a rate of 210 frames per second (fps). This corresponds toan exposure time of �5 ms per frame. For equilibrium bubblestructures, a slower rate of 30 fps (exposure time �33 ms) wasused. In both cases, the sonicated solution was illuminated witha light source from the side of the vessel.

2.3. Void generation rate

The void generation rate was measured using a capillary tech-nique. In this technique the void volume generated by sonicationas a function of time was measured. This void volume measuredwas assumed to be equivalent to the total volume of large coa-lesced bubbles generated for a given sonication time. Further de-tails of this method can be found else where [31]. Anapproximate volume of 280 mL of solution was used to fill the cap-illary cell and the void volume as a function of time was recorded.This void volume was found to increase linearly with time and thevoid generation rate was obtained from the gradient.

2.4. Percentage of attenuation

The configuration used to measure the attenuation of the acous-tic wave by cavitation bubbles is shown in Fig. 1. The amplitude ofthe signal sent to transducer 1 from the function generator wasmodulated between two voltage outputs, V1 and V2. The ampli-tude of V1 was set to equivalent of 20 W (1.1 W/cm2) drivingpower and V2 was set to 10% of V1, below the cavitation threshold.A 5 MHz transducer (transducer 2) acted as a passive detector ofthe acoustic wave from V1 and V2, and the signals were collectedby an oscilloscope. The percentage of attenuation was taken as thepercentage of decrease in the amplitude of V2 by bubbles createdby V1, relative to V2 when the water was degassed to a concentra-tion of 1.0 mg/L where there were no cavitation bubbles present.

3. Results and discussion

3.1. Comparison between visible bubble structures and invisible SLbubbles structures

Shown in Fig. 2(i) are the equilibrium (steady-state) spatial dis-tribution of bubbles generated at frequencies 168, 448 and726 kHz. The bubbles seen in these images are too large to be sono-luminescing (SL) bubbles, which have been shown theoretically[32] and experimentally [33] to be less than 10 microns in diame-ter at these frequencies. These large visible bubbles are bubblesthat have been expelled from the pressure antinodes and becometrapped at adjacent pressure nodes. This expulsion of the bubblesoccurs when the bubble size becomes larger than the active size.The increase in the bubble size occurs predominately from the coa-lescence of sub-resonance size bubbles at the antinodes by the ac-tions of primary and secondary Bjerknes forces. Light scatteringmeasurements monitoring the development of size distributionof bubbles as a function of time have demonstrated this increasein the size and also number of bubbles [34]. Therefore, the spatialdistribution of these visible bubbles would constitute an approxi-mate indication of the spatial distribution of active bubbles thatare otherwise invisible to the naked eye. The SL images taken bya CCD camera with an exposure time of 30 s are shown inFig. 2(ii). Comparing the SL images with that of the visible bubblesin Fig. 2(i), a correlation between the structures of two differentsize of bubbles can be seen along the path of the propagatingacoustic wave for all three frequencies.

94 J. Lee et al. / Ultrasonics Sonochemistry 18 (2011) 92–98

For 168 kHz the structure of the visible bubbles are gathered atseveral horizontal planes orthogonal to the propagating acousticwave in a standing wave pattern. The density of the visible bubblesis higher near the liquid surface than toward the transducer at thebase of the vessel. The SL image for 168 kHz shows a similar stand-ing wave pattern with a strong SL intensity near the liquid surfaceand a weak SL intensity near the transducer. For higher frequen-cies, the visible bubbles are localized to near the liquid surfaceand this is reflected in the SL bubble structure.

A reduction in the surface fluctuations or altering the reflectiv-ity at the boundary normal to the incident wave have shown to in-crease the standing wave field [28,35,36]. Therefore, to furtherdemonstrate the relationship between the spatial distribution ofvisible bubbles and that of the SL bubbles, a sheet of silicone foamwas floated on the liquid to stablilize the liquid surface and in-crease reflectivity. The results are depicted in Fig. 2 and it showsthat the standing wave pattern for both the visible and SL bubbleswere increased by the addition of the silicon foam.

Although the spatial distribution of the visible bubbles showclose resemblance to the structure of SL bubbles, this is only lim-ited to the paths of the acoustic beam, which has a cross-sectionaldiameter of 5 cm normal to the direction of wave propagation. Inaddition to this, if coalescence is hindered or inhibited, active bub-bles may exist in the system without any visible bubbles. Underthis condition, SL images are needed to determine the nature ofthe SL structure. Nevertheless, provided that coalescence is nothindered and sub-resonance bubbles eventually reach a size thatbecomes detectable via coalescence, it is possible to monitor theoccurrence of these visible bubbles to indirectly deduce the devel-opment of active bubbles.

3.2. Development of visible bubble structure

3.2.1. Continuous sonicationShown in Fig. 3 are a series of selected frames documenting the

initial development of the bubble structure at 168, 448 and726 kHz. The colour intensity has been inverted to improve the

0 0.07 0.08 0.10 0.12

0.13 0.16 0.24 1.19 2.38

0 0.24 0.48 0.71 0.95

1.43 1.91 2.38 4.76 9.52

0 0.10 0.13 0.16 0.24

0.38 0.71 0.95 1.43 2.38

168 kHz

448 kHz

726 kHz

Fig. 3. Development of the bubble structure in saturated water as a function of timefor different frequencies. Time in seconds is denoted above the images.

visibility of the bubbles, which appears dark against the lightbackground.

The random appearance of visible bubbles at earlier times of0.24 s and 0.71 s for 168 kHz reflected the general concept that itis the random pre-existing bubble nuclei that act as weak pointsin the liquid which initiates cavitation [37,38]. As discussed earlier,these visible bubbles are bubbles that have been expelled fromnearby antinodes due to predominately the coalescence of activebubbles by Bjerknes forces. With time, as the population and sizeof the bubbles increases, the standing wave structure becomesmore defined and extends across the entire vessel. What is not de-picted in these still photographs is the rapid translational motionof bubbles from the antinodal plane to the next nodal plane. Thistranslational motion observed is usually in an upward direction.With the slower frame rate of 30 fps, used for the images shownin Fig. 2(i), these motions are seen as faint streaks between thenodes for 168 kHz. A recent report by Mettin et al. [39] simulatedthe translational motion of different bubbles which can exist in ahigh-frequency standing wave. In their study, the initial transla-tional motion of bubbles, starting from the antinode toward thenext nodal or some intermediate position, is shown to be only afew hundred acoustic cycles. This rapid movement can accountfor the blurred streaks that appears to originate from the antinodesin the photographs for 168 kHz.

The development of the bubble structure is rather different forhigher frequencies. For 448 kHz, bubbles appear along the verticalaxis of the propagating acoustic wave in a standing wave patternstretching from the liquid surface down toward the transducer.At time 0.12 s bubbles at the bottom of the vessel appear to expe-rience an upward force causing them to stream toward the liquidsurface. The profile of this force is made apparent by the contourof the bubbles advancing toward the liquid surface observed at0.13 s and 0.16 s and disrupts the standing wave structure. Theprofile suggests that the force is weak at the centre, resulting inthe two leading bubble fronts indicated by the arrows. In addition,the force appears to be slightly focused towards the central axis.The images also show that with time, the bubbles at the liquid sur-face spreads out toward the vessel wall and occasionally a burst ofbubbles appears near the bottom of the vessel but is then forcedtowards the liquid surface. For 726 kHz, visible bubbles firstemerge near the transducer and increases in number and size withtime. Contrary to the other two lower frequencies where the emer-gence of the bubble structure was observed randomly along thepath of the acoustic wave, for 726 kHz the emergence of bubbleswere limited to near the base of the transducer. The developmentof a force is observed at 0.16 s and with time the bubbles are pro-pelled towards the liquid surface, similar to that observed at448 kHz. The profile of the force is also made apparent by the out-line of the bubbles and similar to that observed at 448 kHz, thespatial distribution of the bubble structure shrinks towards thesurface and spreads toward the vessel wall.

3.2.2. Importance of attenuation by large non-resonance bubblesIn this study, the observed streaming of bubbles toward the li-

quid surface is believed to be mainly due to the attenuation of theacoustic energy [40]. This attenuation results in the developmentof an energy gradient or force and causes the fluid to move inthe direction of the propagating acoustic wave. This fluid move-ment is commonly known as acoustic streaming [37] and is a non-linear phenomenon that is dependent on both frequency andnonlinear property of the fluid itself. Higher frequencies attenuatemore than lower frequencies and therefore most reports in the lit-erature on acoustic streaming are in the MHz range [41–44].Increasing nonlinearity of fluids, such as increasing fluid viscosity,have also been shown to enhance acoustic streaming [41,43]. Thesize of bubbles or void fraction can also increase the nonlinearity

0

1

2

3

4

5

6

7

8

9

Water 1 mM SDS 10 mM SDS

Void

Rat

e [µ

L/s

]

168 kHz 448 kHz 726 kHz

Fig. 5. Void generation rate at 168, 448 and 726 kHz for water, 1 mM SDS and10 mM SDS. A driving power of 20 W was used for all systems.

0

10

20

30

40

50

60

70

Saturated Water

1 mM SDS 10 mM SDS

Per

cen

tag

e A

tten

uat

ion

168 kHz 448 kHz 726 kHz

Fig. 6. Percentage of attenuation in saturated water, 1 mM SDS and 10 mM SDS at168, 448 and 726 kHz. A driving power of 20 W was used for all systems.

J. Lee et al. / Ultrasonics Sonochemistry 18 (2011) 92–98 95

of the fluid [45]. It has been shown that for water the ratio of B/A, aparameter used to evaluate the nonlinearity of fluids, can reach avalue of about 104 near a void fraction of 10�4 compare to a B/A va-lue of 5 for a void fraction less than 10�8 [45]. This drastic increasein the nonlinearity of fluids with an increase in the void fraction islargely due to the distortion and attenuation of the acoustic waveby large bubbles.

Bubbles at resonance are usually considered to have the great-est attenuation of the acoustic wave [46–48]. However, these stud-ies are usually performed at frequencies of a few Hz and have awavelength much greater than the diameter of bubbles. For acous-tic frequencies in the hundreds of kHz or MHz, coalescence canlead to the formation of large bubbles in the order of hundredsof microns [34]. These large bubbles are non-resonating and haveshown to distort acoustic waves [49] and scatter high-frequencywave fields strongly [50]. It has been shown that these non-reso-nance bubbles scatter significantly more than the resonating bub-bles by virtue of their size rather than via large amplitudepulsations [50,51]. Therefore, the streaming of the bubbles ob-served as a function of time in Fig. 3 can be attributed to the fluidmovement brought about by the increase in the attenuation of theacoustic energy as the population of large bubbles rises with time.

The size and population of the large non-resonance bubbles as afunction of time can be quantified by the void generation rate. Theattenuation of the acoustic wave can also be measured. In order tofurther substantiate the effect of the large non-resonance bubbles,a surfactant was added to inhibit bubble coalescence and decreasethe population of large non-resonance bubbles. The ability of thesurfactant sodium dodecylsulfate (SDS) at inhibiting bubble coa-lescence in the presence of an acoustic wave has been reported[25,31]. Fig. 4 shows the effect of 1 mM and 10 mM SDS on the for-mation of large visible bubbles. It can be seen that there are no vis-ible bubbles under all three frequencies for 1 mM SDS. For 10 mMSDS, visible bubbles are present but to a lesser extent compare tothe case of water shown in Fig. 2(i). The increase in the size of bub-bles at high SDS concentrations is due to the dissociated SDSmonomers acting as excess electrolyte which in turn lowers theelectrostatic repulsion barrier for coalescence to occur. The resultsfor the void rate and percentage of attenuation under different fre-quencies and SDS concentrations are shown in Figs. 5 and 6,respectively. The strong correlation between the two figures andalso the effect of decreasing bubble coalescence by the additionof SDS further demonstrates that large non-resonance bubblescan significantly attenuate the acoustic wave. The lack of stream-ing observed for 168 kHz is because there is very little attenuationof the acoustic wave. Attenuation increases dramatically at448 kHz and 726 kHz despite an increase of only 2–3 times thevoid generation rate at 168 kHz. Similarly, 10 mM SDS at 448 kHzexhibited stronger streaming of bubbles compared to water at168 kHz despite having a lower void generation rate. This is be-cause higher frequencies attenuate to a greater extent, in the orderof square of the frequency [37]. Dahnke et al. [52] have demon-strated theoretically that at low frequencies, a large void fraction

168 kHz 448 kHz 726 kHz

1 mM SDS

10 mM SDS

Fig. 4. The effect of 1 mM SDS and 10 mM SDS on the formation of visible bubblesat frequencies 168, 448 and 726 kHz, with a driving power of 20 W.

of 10�2 is required to significantly decrease the acoustic pressuredistribution.

The streaming of the bubbles observed may also be attributedto acoustic streaming, a nonlinear acoustic phenomena. However,it was pointed out by Mitome et al. [53] that acoustic radiationpressure can act on bubbles and induce fluid movements whichdiffer from those brought about by acoustic streaming. This acous-tic radiation pressure can drive bubbles under linear conditions[37]. In a standing wave field, the radiation force drives bubbles be-low the resonance size to the antinodes and in a traveling wavefield, the radiation force drives bubbles at resonance in the direc-tion of the propagating wave. It is the radiation force from a trav-eling wave field that can cause the bubbles to stream toward theliquid surface. In a real system, due to the attenuation and reflec-tivity at the boundary, the wave field can be partially standingand partially traveling. Therefore, in order to distinguish whetheracoustic streaming or radiation force from the traveling wave isthe driving force behind the streaming of bubbles toward the li-quid surface, we have altered the proportion of the standing andtraveling wave in the acoustic wave field. The proportion of thesewave fields can be manipulated by altering the reflectivity of theboundary normal to the incident acoustic wave. Leighton et al.[28] and Tuziuti et al. [35] reported an increase in the standingwave field by increasing the reflectivity of the boundary. This isdemonstrated in Fig. 2 for 448 kHz with and without a sheet of sil-icone foam. These two images demonstrate the increase in the spa-tial distribution of the standing wave pattern when the liquidsurface is stabilized with a sheet of silicone foam. The developmentof the bubble structure with the liquid surface stabilized by a sheetof silicone foam for 448 kHz and 726 kHz is shown in Fig. 7. For448 kHz, the development of the bubble structure in Fig. 7 did

0 0.48 0.60 0.62 0.71

0.83 0.95 1.19 1.43 2.38

0 0.19 0.24 0.43 0.48

0.71 0.95 1.43 2.38 4.76 448 kHz

726 kHz

Air/liquid surface stabilized with silicone foam

Fig. 7. The effect of stabilizing the surface fluctuations by a sheet of silicone foamon the development of the bubble structure in saturated water as a function of timefor 448 kHz and 726 kHz. Time in seconds is denoted above the images.

96 J. Lee et al. / Ultrasonics Sonochemistry 18 (2011) 92–98

not display the strong streaming of bubbles as that observed inFig. 3 where the liquid surface was not stabilized. This suggeststhat the wave field in Fig. 3 is mostly traveling wave and it isthe radiation force from the traveling wave field that is drivingthe bubbles toward the liquid surface. For 726 kHz, increasingthe standing wave proportion appears to have weakened thestreaming of the bubbles, but there is sufficient streaming tocause the bubble structure to be isolated to near the liquid surface.The fact that significant streaming of bubbles is still present de-spite reducing the proportion of traveling wave field implies thatthe streaming observed at 726 kHz is probably due to acousticstreaming effects.

3.2.3. Power effectThe effect of driving acoustic power on the spatial structure of

bubbles is shown in Fig. 8. For 168 kHz, no strong streaming ofthe bubbles toward the liquid surface is observed across all drivingpowers investigated. For higher frequencies, as the driving powerincreases there exists a driving power at which streaming of thebubbles toward the liquid surface is observed. Increasing drivingpower can increase the population of large bubbles and increasenonlinearity and attenuation effects. The lack of streaming of thebubbles at 168 kHz is another demonstration that lower frequen-cies attenuate to a lesser extent. However, the onset of the stream-ing of bubbles occurs at a lower acoustic power for 448 kHz than

1W 11W 15W 20W

1W 5W 11W 20W

726 kHz

448 kHz

1W 5W 11W 20W

168 kHz

Fig. 8. Equilibrium bubble structure in saturated water for 168, 448 and 726 kHz asa function of power.

for 726 kHz. As discussed previously, the main force responsiblefor the streaming of bubbles for 448 kHz and 726 kHz is different.It is possible that the cause of this difference in the onset of thestreaming of bubbles is the extent at which the driving power af-fects the radiation force from the traveling wave field and acousticstreaming.

3.2.4. Pulsed sonicationPulsed sonication have shown to effect the SL intensity [11],

sonochemical efficiency [13,20,54] as well as the spatial distribu-tion of active bubbles [25] when compared to a continuous system.In order to explore the effect of pulsing on bubble structure devel-opment, further experiments were carried out under pulsed soni-cation conditions at 448 kHz. Shown in Fig. 9 are a series of stillphotographs illustrating the development of the bubble structureunder the pulse condition of 4000 cycles ON and 20,000 cyclesOFF. The development of the bubble field under pulsed sonicationis slower compare to the continuous case and allows the develop-ment of the bubble field to be fully captured. At 0.83 s visible bub-bles begin to emerge and with time, a standing wave pattern formsthroughout the liquid. At 6.67 s it can be seen that a force is presentand the standing wave pattern is disrupted, and bubbles are forcedtoward the liquid surface. The commencement of this streamingoccured later compared to that of the continuous case which wasless than 1 s. This is because the growth in the size of bubblesvia coalescence or rectified diffusion occurs during sonicationand by pulsing, this growth is reduced and thus lessens the forma-tion of large bubbles [34]. Furthermore, dissolution of bubbleswould occur during the pulse OFF duration and decrease the sizeof the bubbles [20]. This decrease in the coalescence and size ofthe bubbles would reduce the degree in which the acoustic pres-sure is attenuated.

The effect of increasing pulse ON or OFF duration is showngraphically in Fig. 10 with the vertical axis as increasing pulseOFF duration and the horizontal axis as increasing pulse ON dura-tion. It can be seen that for a given pulse OFF duration, as the pulseON duration increases the development of the bubble structureresembles the development observed as a function of time. Similardevelopment in the bubble structure is observed with increasingpulse OFF duration for a given fixed pulse ON duration. As dis-cussed previously, for 448 kHz the force causing the translationalmovements of bubbles toward the liquid surface is believed to bethe radiation force from the traveling wave field. With increasingpulse ON duration or decreasing pulse OFF duration, the formationof large coalesced bubbles increases. This results in the increase inthe traveling wave proportion and hence the rise in the radiationforce.

It is possible to categorize Fig. 10 into three regions accordingto SL images reported in the literature [25,34]. Region A is where SLis isolated to the liquid surface, region B is where the SL is

0 0.33 0.83 1.67

2.50 3.33 6.67 10.0

Fig. 9. Initial bubble field development in saturated water for a pulse setting of4000 cycles ON and 20,000 cycles OFF at 448 kHz and a driving power of 20 W.

40000

20000

10000

5000

500

1000 2000 5000 10000Increasing pulse On cycle

Incr

easi

ng

Pu

lse

Off

Cyc

le

C

A

B

Fig. 10. Visible bubble structure for different pulse ON and OFF cycles captured insaturated water at 448 kHz and a driving power of 20 W. Three regions have beenidentified according to the SL images reported in the literature. Region A is where SLis found isolated to the liquid surface. Region B is where SL is homogeneouslydistributed. Region C is where no SL activity was detected.

J. Lee et al. / Ultrasonics Sonochemistry 18 (2011) 92–98 97

homogeneously distributed in the vessel and region C is where noSL activity is detected. Comparing the SL images to that of the vis-ible bubbles, the bubble structure agrees well with the reported SLstructures [25,34] provided that visible bubbles are observed.Fig. 10 demonstrates that there exists an optimum range of pulseON and OFF ratio for strong standing wave field and homogeneousdistribution of active bubbles, which has been shown to translateto a higher sonochemical efficiency [9,55]. If the pulse ON/OFF ratiois in region A, the pulse ON is sufficiently long or the pulse OFF isconsiderably short such that the system effectively behaves like acontinuous system. If the pulse ON/OFF ratio is in region C, thepulse ON duration is inadequate for the generation of detectablebubbles. At the boundary between C and B where there is veryfew visible bubble, it is not possible to determine the structureof SL bubbles based on optical photographs.

4. Conclusion

This study demonstrates that there is a resemblance betweenthe structure of the SL bubbles at the antinodes and that of the coa-lesced visible bubbles at the nodes. This association between thevisible and SL bubbles made it possible to indirectly infer thedevelopment of active cavitation bubbles by monitoring the devel-opment of large coalesced bubbles. It is thus concluded that at theinitial stages of sonication, the wave field is predominately of astrong standing wave and the bubbles are trapped at the antinodeplanes extending from the liquid surface down to the transducer.As the growth in the population and size of bubbles progresseswith increasing sonication time, attenuation of the acoustic pres-sure increases. This subsequently led to the development of acous-tic streaming and radiation force from the traveling wave fieldwhich disrupts the standing wave field, and drives bubbles to theliquid surface. The disruptive forces were weaker for 168 kHz com-pared to 448 kHz and 726 kHz. This was found to be due to lowervoid volume generated and that higher frequencies attenuate to agreater extent. It is demonstrated in this study that at 448 kHz, thestreaming of the bubbles observed is predominately due to the in-crease in the traveling wave proportion and weakening of thestanding wave field. At 726 kHz, the streaming of bubbles is pre-

dominately caused by acoustic streaming. It is shown that thereis an optimum pulse ON and OFF ratio at which the disruption ofthe standing wave field can be minimized.

Acknowledgment

The authors acknowledge the funding from the JSPS Postdoc-toral Fellowship program for foreign researchers and from the Min-istry of Education, Culture, Sports, Science and Technology of Japan(project number 1907765). The authors would like to thank Les Ga-mel for the capillary cell. MA acknowledges the award of AIST Vis-iting Fellowship.

References

[1] C. Petrier, Y. Jian, M.-F. Lamy, Environ. Sci. Technol. 32 (1998) 1316–1318.[2] T.J. Mason, Phil. Trans. R. Soc. Lond. A 357 (1999) 355–369.[3] K. Okitsu, A. Yue, S. Tanabe, H. Matsumoto, Y. Yobiko, Langmuir 17 (2001)

7717–7720.[4] S. Muthukumaran, K. Yang, A. Seuren, S. Kentish, M. Ashokkumar, G.M. Stevens,

Sep. Purif. Technol. 39 (2004) 99–107.[5] S.E. Kentish, T. Wooster, M. Ashokkumar, S. Balachandran, R. Mawson, L.

Simons, Innov. Food Sci. Emerging Technol. 9 (2008) 170–175.[6] P.R. Birkin, J.F. Power, A.M.L. Vincotte, T.G. Leighton, Phys. Chem. Chem. Phys. 5

(2003) 4170–4174.[7] B. Nanzai, K. Okitsu, N. Takenaka, H. Bandow, N. Tajima, Y. Maeda, Ultrason.

Sonochem. 16 (2009) 163–168.[8] S. Koda, T. Kimura, T. Kondo, H. Mitome, Ultrason. Sonochem. 10 (2003) 149–

156.[9] Y. Asakura, T. Nishida, T. Matsuoka, S. Koda, Ultrason. Sonochem. 15 (2008)

244–250.[10] D.J. Casadonte Jr., M. Flores, C. Petrier, Ultrason. Sonochem. 12 (2005) 147–

152.[11] N.V. Dezhkunov, A. Francescutto, P. Ciuti, T.J. Mason, G. Lernetti, A.I. Kulak,

Ultrason. Sonochem. 7 (2000) 19–24.[12] S. Muthukumaran, S. Kentish, S. Lalchandani, M. Ashokkumar, R. Mawson, G.M.

Stevens, F. Grieser, Ultrason. Sonochem. 12 (2005) 29–35.[13] T. Tuziuti, K. Yasui, J. Lee, T. Kozuka, A. Towata, Y. Iida, J. Phys. Chem. A 112

(2008) 4875–4878.[14] N. Segebarth, O. Eulaerts, J. Reisse, L.A. Crum, T.J. Matula, J. Phys. Chem. B 106

(2002) 9181–9190.[15] M. Ashokkumar, R. Hall, P. Mulvaney, F. Grieser, J. Phys. Chem. B 101 (1997)

10845–10850.[16] G.J. Price, M. Ashokkumar, F. Grieser, J. Am. Chem. Soc. 126 (2004) 2755–2762.[17] R. Tronson, Ph.D, The University of Melbourne, 2004.[18] D. Sunartio, M. Ashokkumar, F. Grieser, J. Phys. Chem. B 109 (2005) 20044–

20050.[19] S. Hatanaka, K. Yasui, T. Tuziuti, T. Kozuka, H. Mitome, Jpn. J. Appl. Phys. 40

(2001) 3856–3860.[20] A. Henglein, R. Ulrich, J. Lilie, J. Am. Chem. Soc. 111 (1989) 1974–1979.[21] P.R. Birkin, T.G. Leighton, J.F. Power, M.D. Simpson, A.M.L. Vincotte, P.F. Joseph,

J. Phys. Chem. A 107 (2003) 306–320.[22] S. Hatanaka, K. Yasui, T. Tuziuti, H. Mitome, Jpn. J. Appl. Phys. 39 (2000) 2962–

2966.[23] T. Tuziuti, K. Yasui, Y. Iida, Ultrason. Sonochem. 12 (2005) 73–77.[24] D. Sunartio, K. Yasui, T. Tuziuti, T. Kozuka, Y. Iida, M. Ashokkumar, F. Grieser,

Chem. Phys. Chem. 8 (2007) 2331–2335.[25] J. Lee, K. Yasui, T. Tuziuti, T. Kozuka, A. Towata, Y. Iida, J. Phys. Chem. B 112

(2008) 15333–15341.[26] V. Renaudin, N. Gondrexon, P. Boldo, C. Petrier, A. Bernis, Y. Gonthier, Ultrason.

Sonochem. 1 (1994) s81–s85.[27] F. Trabelsi, H. Ait-lyazidi, J. Berlan, P.-L. Fabre, H. Delmas, A.M. Wilhelm,

Ultrason. Sonochem. 3 (1996) s125–s130.[28] T.G. Leighton, M.J.W. Pickworth, A.J. Walton, P.P. Dendy, Phys. Med. Biol. 33

(1988) 1239–1248.[29] U. Parlitz, R. Mettin, S. Luther, I. Akhatov, M. Voss, W. Lauterborn, Phil. Trans. R.

Soc. Lond. A 357 (1999) 313–334.[30] R. Mettin, S. Luther, C.D. Ohl, W. Lauterborn, Ultrason. Sonochem. 6 (1999) 25–

29.[31] J. Lee, S.E. Kentish, M. Ashokkumar, J. Phys. Chem. B 109 (2005) 5095–5099.[32] K. Yasui, J. Acoust. Soc. Am. 112 (2002) 1405–1413.[33] A. Brotchie, F. Grieser, M. Ashokkumar, Phys. Rev. Lett. 102 (2009) 084302.[34] Y. Iida, M. Ashokkumar, T. Tuziuti, T. Kozuka, K. Yasui, A. Towata, J. Lee,

Ultrason. Sonochem. 17 (2010) 473–479.[35] T. Tuziuti, K. Yasui, T. Kozuka, A. Towata, Y. Iida, J. Phys. Chem. A 111 (2007)

12093–12098.[36] M. Ashokkumar, J. Lee, Y. Iida, K. Yasui, T. Kozuka, T. Tuziuti, A. Towata, Phys.

Chem. Chem. Phys. 11 (2009) 10118–10121.[37] T.G. Leighton, The Acoustic Bubble, Academic Press Limited, Cambridge, 1994.[38] F.R. Young, Cavitation, Imperial College Press, London, 1999.[39] R. Mettin, A.A. Doinikov, Appl. Acoust. 70 (2009) 1330–1339.

98 J. Lee et al. / Ultrasonics Sonochemistry 18 (2011) 92–98

[40] W.L.J. Nyborg, Acoust. Soc. Am. 25 (1953) 68–75.[41] L.N. Liebermann, Phys. Rev. 75 (1949) 1415–1422.[42] A. Nowicki, T. Kowalewski, W. Secomski, J. Wojcik, Eur. J. Ultrasound 7 (1998)

73–81.[43] G. Madelin, D. Grucker, J.M. Franconi, E. Thiaudiere, Ultrasonics 44 (2006) 272–

278.[44] S. Sakamoto, Y. Watanabe, Jpn. J. Appl. Phys. 38 (1999) 3050–3052.[45] W. Lauterborn, T. Kurz, I. Akhatov, in: T.D. Rossing (Ed.), Handbook of

Acoustics, Springer, New York, 2007.[46] F.E. Fox, S.R. Curley, G.S. Larson, J. Acoust. Soc. Am. 27 (1955) 534–539.[47] H. Medwin, Ultrasonics 15 (1977) 7–13.

[48] P.S. Wilson, R.A. Roy, W.M. Carey, J. Acoust. Soc. Am. 117 (2005) 1895–1910.[49] T. Tuziuti, K. Yasui, J. Lee, T. Kozuka, A. Towata, Y. Iida, J. Phys. Chem. A 113

(2009) 8893–8900.[50] T.G. Leighton, S.D. Meers, P.R. White, Proc. R. Soc. Lond. A. 460 (2009) 2521–

2550.[51] K. Commander, E. Moritz, J. Acoust. Soc. Am. 85 (1989) 2665–2669.[52] S. Dahnke, F. Keil, J. Ind. Eng. Chem. Res. 37 (1998) 848–864.[53] H. Mitome, T. Kozuka, T. Tuziuti, L. Wang, IEEE Ultrason. Symp. Proc.,

Piscataway, NJ (1997) 533–536.[54] M. Gutierrez, A. Henglein, J. Phys. Chem. 94 (1990) 3625–3628.[55] Y. Asakura, M. Maebayashi, S. Koda, J. Chem. Eng. Jpn. 38 (2005) 1008–1014.