Ultrasonics Sonochemistry 27 (2015) 22–29

Contents lists available at ScienceDirect

Ultrasonics Sonochemistry
journal homepage: www.elsevier.com/locate/ultson

Ultrasound pressure distributions generated by high frequency

transducers in large reactors
Thomas Leong a,b, Michael Coventry a, Piotr Swiergon a, Kai Knoerzer a, Pablo Juliano a,⇑
a
CSIRO Food and Nutrition Flagship, 671 Sneydes Road, Werribee 3030, Australia
b
Mechanical and Product Design Engineering, Faculty of Science, Engineering and Technology, Swinburne University of Technology, John Street Hawthorn, Victoria 3122, Australia

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

Article history: The performance of an ultrasound reactor chamber relies on the sound pressure level achieved through-
Received 18 December 2014 out the system. The active volume of a high frequency ultrasound chamber can be determined by the
Received in revised form 13 March 2015 sound pressure penetration and distribution provided by the transducers. This work evaluated the sound
Accepted 20 April 2015
pressure levels and uniformity achieved in water by selected commercial scale high frequency plate
Available online 25 April 2015
transducers without and with reflector plates. Sound pressure produced by ultrasonic plate transducers
vertically operating at frequencies of 400 kHz (120 W) and 2 MHz (128 W) was characterized with hydro-
Keywords:
phones in a 2 m long chamber and their effective operating distance across the chamber’s vertical cross
High frequency ultrasound
Reactor design
section was determined. The 2 MHz transducer produced the highest pressure amplitude near the trans-
Sound penetration ducer surface, with a sharp decline of approximately 40% of the sound pressure occurring in the range
Reflection between 55 and 155 mm from the transducer. The placement of a reflector plate 500 mm from the sur-
Sound pressure face of the transducer was shown to improve the sound pressure uniformity of 2 MHz ultrasound.
Ultrasonic transducers Ultrasound at 400 kHz was found to penetrate the fluid up to 2 m without significant losses.
Furthermore, 400 kHz ultrasound generated a more uniform sound pressure distribution regardless of
the presence or absence of a reflector plate. The choice of the transducer distance to the opposite reactor
wall therefore depends on the transducer plate frequency selected. Based on pressure measurements in
water, large scale 400 kHz reactor designs can consider larger transducer distance to opposite wall and
larger active cross-section, and therefore can reach higher volumes than when using 2 MHz transducer
plates.
Crown Copyright Ó 2015 Published by Elsevier B.V. All rights reserved.

1. Introduction distribution for cavitation in the reactor system [9–11].

Information such as the pressure distribution of transducers posi-
The characterization of pressure within a sonoprocessing vessel tioned in large-scale reactors is neither widely nor systematically
is an important step in the design of industrial scale reactors. documented, and differs between reactors. Further complexity is
Several studies have documented the characterization of sonopro- added when reflector plates are positioned in the reactor system
cessing reactors, with a focus on bath and horn-type reactors [1–3]. so that standing waves are generated, as it gives rise to regions
These systems typically use high power, low frequency ultrasound, of pressure nodes and antinodes.
suitable for applications such as homogenization [4], emulsifica- When considering the propagation of sound from large trans-
tion [5], extraction [6] and sonocrystallization [7]. ducers, it should be noted that plate transducers with large surface
High frequency, low power ultrasound, is typically used for areas rarely consist of a single transducer element, and instead are
applications such as cleaning of sensitive components and, more constructed from an array of piezo elements [12]. As such, the
recently, in the separation of multi-component mixtures [8]. sound waves produced from each active transducer component
Only a limited number of studies have looked at the sound pres- will tend to result in interference when in close proximity to one
sure characterization of high frequency transducers in the range another. Where these waves interact, the sound pressure ampli-
from 400 kHz to 2 MHz. Reported studies often use sonochemilu- tude is the combined sum of the amplitude of the individual waves.
minescence as a means to visualize the spatial and temporal Further away from the plate the waves travel, the more uniform
the sound field should become. These two areas are called the
⇑ Corresponding author. Fresnel and the Fraunhofer zones, or more commonly known as
E-mail address: Pablo.juliano@csiro.au (P. Juliano). the near field and far field, respectively. The sound waves produced

http://dx.doi.org/10.1016/j.ultsonch.2015.04.028
1350-4177/Crown Copyright Ó 2015 Published by Elsevier B.V. All rights reserved.
 T. Leong et al. / Ultrasonics Sonochemistry 27 (2015) 22–29 23

by the transducer are ‘more predictable’ and at their maximum at Transducers were mounted on one end of the chamber as
the area just beyond the near field, also known as the ‘natural shown in Fig. 1a. The top of the chamber was open to the air to
focus’. The length of the near-field can be calculated using the facilitate maneuvering of the hydrophone measurement system.
equation [13]: The transducers utilized in this study were submersible plate
transducers (SONOSYS Ultraschallsysteme GmbH, Neuenbuerg,
D2 f Germany) with nominal frequencies of 400 kHz and 2 MHz. The
N¼ ð1Þ
4u active areas of the transducers are 110 mm  75 mm and
100 mm  100 mm for the 400 kHz and the 2 MHz transducers,
where N is the near-field distance, D is the largest dimension of the respectively. The transducers were operated at a nominal electri-
transducer element (mm), f is the frequency of the transducer cal power load of 50% (120 W for the 400 kHz system, and
(MHz) and u is the velocity of the sound (m s1). 128 W for the 2 MHz system) unless otherwise stated.
As these sound waves pass through a medium, they also expe- Preliminary tests performed at power settings of 50% and 100%
rience a loss in energy due to absorption of the energy by the mate- showed no difference in the trends of sound penetration and uni-
rial, known as attenuation. The attenuation is dependent on the formity for these transducers (results not shown). Note that
frequency of the ultrasound as well as the density and viscosity although the power settings are similar, the actual acoustic energy
of the material, and scales as [14]: delivered and hence sound pressure measured for the 2 different
frequencies are quite different in magnitude. The testing chamber
2
2 lf was filled with tap water at ambient temperature at 20 ± 5 °C and
a/ ð2Þ
3 q left to equilibrate at this temperature for several hours prior to
each experiment.
where a is the attenuation coefficient, f is the frequency of the
Two different hydrophones were used for the two frequencies
applied ultrasound (s1), l is the viscosity (Pa s = kg m1 s1) and
investigated. A needle hydrophone (HNC-1000, Onda, Sunnyvale,
q is the density of the fluid medium (kg m3). CA, USA) was used to measure sound pressure levels for the
When sound waves encounter a boundary such as a steel plate,
2 MHz transducer. The sound pressure level for the lower fre-
part of the energy is reflected back to the transducer. The remain-
quency (i.e., 400 kHz) was measured with an ultra-broad-band
ing energy will pass through the plate. Recent work investigating
spherical hydrophone (TC-4034, Reson, Slangerup, Denmark). The
the transmission of sound pressure through steel plate boundaries
peak-to-peak signals were recorded with an oscilloscope (GDS-
(where water was located on both sides of the boundary) at ultra-
1102, GW Instek, Taipei, Taiwan) and converted to sound pressure
sonic frequencies between 400 kHz and 2 MHz has been reported
levels using Eq. (3).
by Michaud et al. [15]. To date, however, no studies have charac-
terized the sound pressure distributions across the length of a reac- p ¼ 20  log10 ðV rms Þ  OCV ð4Þ
tor system at these frequencies in vessels up to 2 m in length.
where Vrms is the root mean square voltage determined from the
This study therefore focuses on understanding the effective
peak-to-peak values measured, and OCV is the open received cir-
operating distance of plate-type transducers that operate in
cuit voltage (read off a hydrophone calibration chart supplied by
large-scale reactor systems by measuring the sound pressure
the manufacturer). Trials were carried out in a silent environment
decline over long distance, the distribution of sound pressure
with no or minimal interference into the measuring system. Note
within the active area over these distances and the influence of
that both hydrophones used in this study measured a baseline
sound wave reflection on the sound pressure uniformity.
noise of 8 ± 4 mV when no ultrasound was applied, which is much
The performance of transducers operating at mid (400 kHz) and
lower than values measured at the center of the vessel and may
high (2 MHz) ultrasonic frequencies at these distances should be
contribute up to an additional 0.3 kPa (Reson) or 3 kPa (Onda)
evaluated to understand the working limitations in larger reactor
to the calculated pressure data. This value was not subtracted
systems. Information pertaining to sound pressure distributions
from the final values reported. Noise from the surrounding envi-
will enable calculation of maximum reactor transducer to wall dis-
ronment was minimized by averaging a minimum of 16 sweeps
tances in large scale operating volumes, which will help determine
in the oscilloscopes.
the optimum number of transducers required for the design of a
Where required, a stainless steel plate with the same cross sec-
reactor and minimizing capital costs entailed. It will therefore
tional area as the vessel and thicknesses of 3 mm, was positioned
enable improvements in future designs to maximize reactor per-
in the chamber at distances of 500 mm, 1000 mm, 1500 mm and
formance in the effective ultrasound processing regions.
2000 mm from the transducer to determine the influence of reflec-
tance to the sound pressure distribution. In such a setup, water
was located on one side of the reflector plate and air on the other
2. Materials and methods
(Fig. 1a).
2.1. Experimental setup
2.2. Hydrophone measurements
A stainless steel chamber with dimensions of 350 mm 
All hydrophones were soaked for a minimum of 20 min in the
350 mm  2100 mm and a wall thickness of 2 mm was used for
water of the testing tank prior to beginning measurements. The
all experiments unless otherwise stated. The chamber itself may
hydrophone was positioned directly in front of the transducer
exhibit resonance modes at certain frequencies. These resonant
pointing at its active face (Fig. 1b). A minimum of three readings
frequencies can be estimated by assuming that the chamber is a
were recorded at selected locations inside the chamber, with the
‘1-D room’ with closed ends using:
first few readings discarded until stabilized readings were
nc obtained. The hydrophone was positioned at various distances
fR ¼ ð3Þ along the x-axis between 55 and 1955 mm (Fig. 1b). The spacing
2L
between each measurement point was 10 mm between 55 and
where n is an integer corresponding to the mode of vibration, c is 155 mm, 20 mm between 155 and 255 mm, 100 mm between
the speed of sound in the fluid and L is the characteristic length 255 and 955 mm and 400–500 mm between 955 and 1855 mm.
of the chamber. Unless otherwise stated, the sound pressure with distance for each
24 T. Leong et al. / Ultrasonics Sonochemistry 27 (2015) 22–29

(a)

(b)

(c)

Fig. 1. (a) Schematic representation of the stainless steel vessel used for sound pressure penetration and distribution tests. (b) Placement of hydrophone during experimental
measurements. The orientation depicted is of the hydrophone being parallel to the direction of sound propagation (0° directivity). (c) Measurement grid utilized for the
evaluation of acoustic pressure distribution in y-z cross-sections in the chamber at discrete distances from the transducer plate.

transducer was measured with the hydrophone positioned directly individual transducer was normalized by dividing the sound pres-
facing the center of the transducer at a depth of 120 mm (measured sure measured at 55 mm using the following equation:
from the base) and 175 mm along the width (see Fig. 1b).
PN ¼ px =p55 ð5Þ
Furthermore, at each selected distance, 15 points in the z-y
plane, covering a 5  3 matrix, see Fig. 1c) were used to map the where PN is the normalized pressure, p55 is the reference pres-
distribution of the sound pressure as it travels along the test cham- sure located at a distance of 55 mm from the surface of the trans-
ber. Note that as the geometry is symmetrical, only half of the ducer, and px is the measured pressure at a distance of x mm from
interfacial area (i.e. 9 measurements) shown in Fig. 1c is mapped the surface of the transducer. For the sound pressure distribution
(further measurements indicated that the distribution was sym- plots, all pressures were normalized to the highest p55 value
metrical; results not shown). Contour plots were drawn in (within the cross sectional area) for each series.
Matlab 2012c (The MathsWorks Inc., Natick, MA, USA). It should
be noted that the figure outputs are not drawn to scale. 2.3. Statistical analysis
For tests within an acoustic standing wave field, a stainless steel
reflector plate (3 mm thickness) was positioned in the tank at The statistical significance of the results was evaluated using a
selected distances from the transducer (500 mm, 1000 mm, General Linear Model by Analysis of Variance (ANOVA, Matlab
1500 mm and 2000 mm). The same mapping procedure as 2012c, The MathsWorks Inc., Natick, MA, USA) for level of signifi-
described for the experiments without the reflector plate was fol- cance of P < 0.05. Where reported, the error bars are the standard
lowed. The fluid on the opposing side of the reflector was air. deviation across a minimum of 3 replicated measurements.
In all cases, a minimum of 3 measurements were recorded at
each location unless otherwise stated. Error bars shown are the 3. Results and discussion
standard deviations of these replicated measurements within each
experimental test. 3.1. Variation of sound pressure with distance
Note that the absolute acoustic pressures generated by the two
different transducers are quite different in magnitude and hence The normalized change in sound pressure along the center of
difficult to directly compare. Therefore, the pressure data for each the chamber (relative to p55 for each transducer), is shown in
 T. Leong et al. / Ultrasonics Sonochemistry 27 (2015) 22–29 25

Fig. 2. The absolute pressure measured at this point for the 400 kHz It is possible to estimate the attenuation coefficients. For 1 MHz
and 2 MHz transducers are 11.3 ± 1.4 kPa and 57.6 ± 4.0 kPa, ultrasound in Water, Leighton [12] reports a value of 0.0253 Np/m.
respectively. From this, we can use Eq. (2) to estimate values for 400 kHz to be
From Fig. 2 it can be seen, for the higher 2 MHz frequency trans- 0.00405 Np/m and for 2 MHz, 0.1012 Np/m. These values would
ducer, that there is a sharp decline in sound pressure from its ini- likely be representative of a ‘clean’, contaminant and cavitation
tial starting position at 55–200 mm, followed by a more gradual bubble free situation.
decline along the remaining length of the chamber. In contrast, The decline in pressure presented in Fig. 2 can also be used to
the lower frequency transducer (i.e., 400 kHz) does not result in estimate the attenuation coefficients assuming that the pressure
a significant decline in pressure level with distance. The normal- intensity decays as e2bx [12]. From this, we can estimate based
ized pressure remains >80% of the p55 value along the entire length on values located at x  0.255 m for 2 MHz, a coefficient of
of the test chamber. 1.00 Np/m, and at x  0.8 m for 400 kHz, a coefficient of
It can be seen that the lower frequency ultrasound transducer 0.139 Np/m. These values are one to two orders of magnitude
utilized in this study offers higher sound pressure penetration higher than that predicted from Eq. (2). This would suggest that
through the chamber, with little decline observed even at the fur- the attenuation observed in this study is strongly influenced by
thest distance towards the end of the chamber. This is expected as what is likely to be cavitation bubbles within the system and con-
per Eq. (2) since the sound attenuation is lower for the 400 kHz fre- tamination of the water with excess air, dispersed solid particles,
quency. By contrast, the 2 MHz transducer sees a decline in pres- and/or organic contaminants.
sure level to approximately 40% of the p55 value at a distance of In both cases, it was also observed that at certain locations
1000 mm from the transducer surface. This behavior is again con- along the length of the chamber, pressure ‘spikes’ appear. These
sistent with the expected attenuation of higher frequency ultra- ‘spikes’ were reproducible, as confirmed by investigations at lower
sound (Eq. (2)). Similar sound pressure penetration behavior has power inputs and repetitions on different days (results not shown).
recently been reported by Michaud et al. [15] for distances below One possibility is that these spikes may be pressure antinodes gen-
200 mm. erated by establishment of a standing wave from reflections from
the opposite wall of the chamber. It is not likely that a standing
wave can be established along the length of the entire chamber
1.6 (2 m), particularly for the 2 MHz ultrasound since the sound
1.4 pressure is strongly attenuated within the first 255 mm from the
1.2 surface of the transducer. Furthermore, a standing wave would
Relave Pressure

1 result in pressure antinode peaks that should appear much more

0.8
frequently since wavelengths are about 1 mm for 2 MHz ultra-
sound and about 4 mm for 400 kHz transducer. Pressure antinodes
0.6
would be spaced apart by half a wavelength. Although measure-
0.4
ments were performed at discrete points along the length of the
0.2 chamber >10 mm, this possibility can be ruled out; the error that
0 would arise from the hydrophone location and the thickness of
55 155 255 355 455 555 655 755 855 955 1455 1855 1955 the hydrophone itself, would actually be larger than the expected
Distance from transducer (mm) distances between the pressure antinodes of a standing wave.
A more likely explanation for the observations of pressure
400kHz 2MHz
spikes detected in the chamber are instead due to intrinsic reso-
nance modes of sound within this particular chamber, which man-
Fig. 2. Change in pressure relative to the pressure measured at 55 mm from the
surface of the transducer for 400 kHz and 2 MHz transducers. No reflector plate ifest themselves as pressure antinodes that are caused by internal
except for the vessel wall at 2100 mm was used for these measurements. reflections along the side walls of the chamber. The observation

Fig. 3. Sound pressure distributions (expressed relative to p55, see Eq. (5)) within the first 255 mm of the 2 MHz transducer.
26 T. Leong et al. / Ultrasonics Sonochemistry 27 (2015) 22–29

(a)

(b)

(c)

(d)
Fig. 4. Sound pressure distributions (expressed relative to p55, see Eq. (5)) as a function of distance for 2 MHz transducer with reflector plate located at distances of (a)
500 mm (b) 1000 mm (c) 1500 mm and (d) 2000 mm.
 T. Leong et al. / Ultrasonics Sonochemistry 27 (2015) 22–29 27

that these ‘spikes’ occur at different locations for the different fre- The first mode of resonance (n = 1) will have a resonance frequency
quencies studied suggests that they depend on the geometry of the of 357 Hz whereas n = 225 will have a resonance frequency of
test chamber [16]. 80 kHz. Ultrasound driven at 400 kHz would result in approxi-
Another similar possibility is the establishment of resonance mately 5 pressure spikes corresponding to the n = 225 mode within
modes along the long chamber walls, in which case it would be the vessel. The 2 MHz transducer would provide approximately 25
expected that every 5th spike along the chamber would coincide pressure spikes at the same mode. Measuring sound at very short
at the two frequencies investigated (400 kHz and 2 MHz). The res- intervals of 0.5 mm or less is difficult to achieve practically and
onance frequency of the chamber can be estimated using Eq. (3). requires an automated system and more accurate hydrophones

(a)

(b)

(c)
Fig. 5. Sound pressure distributions (expressed relative to p55, see Eq. (5)) as a function of distance for 400 kHz transducer with reflector plate located at distances of (a)
1000 mm (b) 1500 mm and (c) 2000 mm.
28 T. Leong et al. / Ultrasonics Sonochemistry 27 (2015) 22–29

to achieve this level of detail. Nevertheless, every 5th spike the pressure maxima which was observed to transition from the
observed for 2000 and 400 kHz should coincide which appears to bottom portion of the cross sectional area to the top when no
be the case in the measured data. reflector plate was used, remains near the bottom when the reflec-
tor plate was positioned at 1000 mm. This again indicates that
3.2. Sound pressure distributions there is a contribution to the measured pressure from reflections
from the top surface. Another possible explanation for this obser-
3.2.1. Without reflector plate vation is due to a change in the chamber and/or wall length reso-
The sound pressure distributions for the 2 MHz ultrasound (y-i nance modes within the vessel when the plate is positioned at this
cross sections) within the first 55–255 mm from the transducer are distance.
shown in Fig. 3. The spacing between each cross sectional slice is Note that with the reflector plate positioned at 1500 mm and
10 mm between 55 and 155 mm, and 20 mm between 155 and 2000 mm, the sound pressure distributions appear to be uniform
255 mm. Note that only the 2 MHz ultrasound was tested here in but a decline is seen in the normalized pressure to below 60%.
the absence of a reflector plate to gauge the pressure uniformity This observation is consistent with expected attenuation of the
during the sharp decline in sound pressure observed in Fig. 2. sound pressure observed in Fig. 2. Uniform sound pressure distri-
As expected, the sound pressure distributions exhibit the high- butions at greater distances are due to the sound pressure becom-
est overall pressure readings for the cross-sections measured clo- ing more diffuse and spreading out as the sound wave travels along
ser to the transducer and decrease for those further away. It is the chamber (i.e., far field effects).
interesting to note that in the absence of a reflector plate, the In contrast, when using the 400 kHz ultrasound, the sound pres-
sound pressure distributions for 2 MHz ultrasound do not appear sure distribution remains more uniform in magnitude (normalized
to be particularly uniform. Indeed, sound pressure maxima are pressure > 60%) even with the reflector plate positioned at dis-
located within a central band running from top to bottom for the tances of 2000 mm from the transducer (Fig. 5a–c). This was
cross sections measured between 55 and 75 mm. expected, considering that the change in normalized pressure with
Between 85 and 115 mm, the pressure maxima shift towards distance shown in Fig. 2 for the 400 kHz displays little decline.
the bottom portion of the cross sectional area. The sound pressures Again, the spiking behavior seen in Fig. 2, where pressure magni-
within the top two thirds are, although slightly lower in magni- tudes increase and decline, even at long distance, can be observed
tude, more uniformly distributed. This observation appears to be in the sound pressure distributions for the 400 kHz ultrasound.
consistent with the ‘spreading out’ of sound pressure pertaining
to near-field and far-field effects as sound propagates from a 4. Conclusion
source. Also notable is that observable sound pressure maxima
increase and decrease in intensity as we move away from the This study confirms to some extent the expected behavior for
transducer between 85 and 115 mm. This is very similar to the ultrasonic transducers with frequencies of 400 kHz and 2 MHz in
‘spiking’ behavior observed in Fig. 2, and is likely due to resonance terms of their penetrative distance in fluids and acoustic pressure
modes within the chamber. distribution when working with large geometries up to 2 m in
Between 125 and 255 mm, the pressure maxima transition length. Ultrasound with a frequency of 400 kHz was capable of
towards the top third of the cross sectional areas. Further to reso- penetrating through almost the entire length of the vessel, and still
nance modes of the chamber itself, another possible reason to produced a uniform sound pressure distribution with normalized
explain this behavior is due to observed wave motions that occur pressures exceeding 60%. By contrast, high frequency ultrasound
on the surface of the liquid as sound propagates through the fluid at 2 MHz, displayed sharp decline in active sound pressure within
chamber. Locations of higher and lower pressure within the system the first 200 mm from the surface of the transducer. However,
can also drive fluid circulation that may influence the measured 2 MHz ultrasound was able to be rendered more uniform in sound
sound pressure values. We observed strong acoustic streaming in pressure magnitude (with magnitudes exceeding 60%) by place-
the region near the active transducer face where pressure values ment of a stainless steel reflector plate within 500 mm of the trans-
were highest. It is also possible that the liquid–air surface interface ducer surface.
can act as a reflector. For long distance sonoprocessing, it is advantageous to use
We speculate that these motions and reflections near the liquid lower frequency ultrasound since sound pressure is more uni-
surface and hydrostatic pressure in the z-direction may contribute formly distributed with distance. For applications requiring high
to a higher measured sound pressure in the absence of a reflector frequency ultrasound greater than 2 MHz, shorter operating dis-
plate. tances with placement of reflector plates, would enable more uni-
form sound pressure. The sound pressure distributions mapped in
3.2.2. With reflector plate this study provides a guide for reactor designs that will have indus-
A reflector plate was positioned at locations of 500 mm, trial relevance to sonoprocessing and sonochemistry applications
1000 mm, 1500 mm and 2000 mm from the surface of the trans- using high frequency ultrasound.
ducer to evaluate its influence on the sound pressure distribution.
Selected cross sectional slices for 2 MHz ultrasound are shown in Acknowledgments
Fig. 4a–d.
It is apparent that when a reflector plate was positioned at The authors would like to acknowledge the RMIT University
500 mm, the measured sound pressure for the 2 MHz ultrasound final year research program for establishment of Michael
becomes more uniformly distributed, up to a distance of 455 mm Coventry’s studentship at CSIRO.
from the surface of the transducer, with normalized pressures
exceeding 60%. This improvement in uniformity is likely due to
References
sound wave reflection, which increases the sound pressure at loca-
tions further from the transducer. [1] P.R. Gogate, I.Z. Shirgaonkar, M. Sivakumar, P. Senthilkumar, N.P. Vichare, A.B.
However, when the reflector was positioned at 1000 mm, the Pandit, Cavitation reactors: efficiency assessment using a model reaction,
sound pressure distributions become less uniform with distance AIChE J. 47 (2001) 2526–2538.
[2] P.R. Gogate, V.S. Sutkar, A.B. Pandit, Sonochemical reactors: important design
from the transducer. Sound pressure distributions appear similar and scale up considerations with a special emphasis on heterogeneous
to the situation where no reflector plate was used. Interestingly, systems, Chem. Eng. J. 166 (2011) 1066–1082.
 T. Leong et al. / Ultrasonics Sonochemistry 27 (2015) 22–29 29

[3] P.M. Kanthale, P.R. Gogate, A.B. Pandit, A. Marie Wilhelm, Mapping of an [10] H. Cao, M. Wan, Y. Qiao, S. Zhang, R. Li, Spatial distribution of
ultrasonic horn: link primary and secondary effects of ultrasound, Ultrason. sonoluminescence and sonochemiluminescence generated by cavitation
Sonochem. 10 (2003) 331–335. bubbles in 1.2 MHz focused ultrasound field, Ultrason. Sonochem. 19 (2012)
[4] F. Martinez, A. Davidson, J. Anderson, S. Nakai, I. Desai, A. Radcliffe, Effects of 257–263.
ultrasonic homogenization of human milk on lipolysis, IgA, IgG, lactoferrin and [11] H. Yin, Y. Qiao, H. Cao, Z. Li, M. Wan, Cavitation mapping by
bacterial content, Nutr. Res. 12 (1992) 561–568. sonochemiluminescence with less bubble displacement induced by acoustic
[5] T.S.H. Leong, T.J. Wooster, S.E. Kentish, M. Ashokkumar, Minimising oil droplet radiation force in a 1.2 MHz HIFU, Ultrason. Sonochem. 21 (2014) 559–565.
size using ultrasonic emulsification, Ultrason. Sonochem. 16 (2009) 721–727. [12] T.G. Leighton, The Acoustic Bubble, Academic Press, San Diego, 1994.
[6] G. Cravotto, L. Boffa, S. Mantegna, P. Perego, M. Avogadro, P. Cintas, Improved [13] J.A. Jensen, N.B. Svendsen, Calculation of pressure fields from arbitrarily
extraction of vegetable oils under high-intensity ultrasound and/or shaped, apodized, and excited ultrasound transducers, IEEE Trans. Ultrason.
microwaves, Ultrason. Sonochem. 15 (2008) 898–902. Ferroelectr. Freq. Control 39 (1992) 262–267.
[7] M. Luque de Castro, F. Priego-Capote, Ultrasound-assisted crystallization [14] G.G. Stokes, On a difficulty in the theory of sound, Philos. Mag. 33 (1848) 349–
(Sonocrystallization), Ultrason. Sonochem. 14 (2007) 717–724. 356.
[8] T. Leong, L. Johansson, P. Juliano, S.L. McArthur, R. Manasseh, Ultrasonic [15] M. Michaud, T. Leong, P. Swiergon, P. Juliano, K. Knoerzer, Design parameters
separation of particulate fluids in small and large scale systems: a review, Ind. of stainless steel plates for maximizing high frequency ultrasound wave
Eng. Chem. Res. (2013). transmission, Ultrason. Sonochem; 2015.
[9] T. Tuziuti, K. Yasui, Y. Iida, M. Sivakumar, Correlation in spatial intensity [16] R.J. Townsend, M. Hill, N.R. Harris, N.M. White, Investigation of two-
distribution between volumetric bubble oscillations and dimensional acoustic resonant modes in a particle separator, Ultrasonics 44
sonochemiluminescence in a multibubble system, Res. Chem. Intermed. 30 (2006) E467–E471.
(2004) 755–762.