i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

Available online at www.sciencedirect.com

ScienceDirect

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

Development of efficient sonoreactor geometries

for hydrogen production

Sherif S. Rashwan, Atef Mohany*, Ibrahim Dincer

Faculty of Engineering and Applied Science, University of Ontario Institute of Technology, 2000 Simcoe St. North,
Oshawa, Ontario, L1G 0C5, Canada

highlights

Mapping analysis on the performance of various sonoreactor geometries.

Pressure distribution at different ultrasonic frequencies is reported.

Acoustic Model is validated with experimental work in the literature.

Effect of operational, acoustical and geometrical parameters is presented.

Hydrogen yield and energy consumption of the sonohydrogen process is highlighted.

article info abstract

Article history: Ultrasonic water processing is recognized as an emerging technology that has shown
Received 1 December 2020 promising results for hydrogen production. It is also known that numerous parameters
Received in revised form affect the sonication process, such as the ultrasonic efficiency, acoustic pressure distri-
11 February 2021 bution, input power, and the consequent cavitation activity. However, the current effi-
Accepted 14 February 2021 ciency of such a technique limits its scalability for industrial production. Therefore, in this
Available online 13 March 2021 study, different sonoreactor geometries are developed, and a comprehensive assessment is
probed to validate their performance. The study considered the effects of the geometrical
Keywords: parameters, such as the sonoreactor geometry, wall boundary conditions, and the number
Sonic reactor of sonotrodes. The results show that a slight change in the sonoreactors’ vessel geometry
Hydrogen production while maintaining all other parameters constant significantly affects the reactor’s pressure
Energy field distribution. The vessel geometry with a concave bottom wall has recorded the
Efficiency highest magnitude of the negative pressure, leading to more efficient cavitation bubbles. In
Acoustics contrast, the vessel geometry with a conical shape recorded the worst performance. On the
Pressure distribution effect of the number of sonotrodes, an eigenfrequency analysis is performed to check the
excitable acoustic modes and frequencies to trigger resonance condition; the resonance
condition will enhance the sonohydrogen process accordingly. The present study performs
a hydrogen quantification analysis where the effects of increasing input power and
maneuvering the geometrical effects are investigated. The maximum hydrogen production
is recorded at 300 W in the amount of 2.5  109 mol/J when increasing the input power to
one sonotrode mounted in a typical sonoreactor cylindrical shape at an energy conversion
efficiency of 23%. For the study on the multiple sonoreactor, the amount of hydrogen
produced is 308  109 mol/kWh at 180 W with an energy conversion efficiency of 33%.
© 2021 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

* Corresponding author.
E-mail address: Atef.Mohany@uoit.ca (A. Mohany).
https://doi.org/10.1016/j.ijhydene.2021.02.113
0360-3199/© 2021 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
15220 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

bubbles that dissociate water molecules; the process is known

Nomenclature as sonolysis. The pressure difference breaks apart the chem-
ical bond of water, which results in the generation of highly
A Acceleration [m/s2] reactive radicals of OH and H. In addition to the physical
Ap Transducer probe surface area [m2] forces being exerted by acoustic cavitation that can be
C Speed of sound [m/s] measured by aluminum foil, there is also a chemical effect
c0 Speed of sound at the ambient condition [m/s] (sonochemical) occurring that causes the molecular bond
CLY-1P Cylindrical sonoreactor with 1 Probe cleavage of water molecules (sonolysis), releasing powerful
CLY-2P Cylindrical sonoreactor with 2 Probes oxidizing hydroxyl radicals with a sonochemical test. The
CLY-3P Cylindrical sonoreactor with 3 Probes reaction provides a sensitive method of estimating hydroxyl
CLY-4P Cylindrical sonoreactor with 4 Probes radical’s production. The faster the fluorescent change occurs,
CLY-5P Cylindrical sonoreactor with 5 Probes the more intense the cavitation, and the more intense the
Dp Probe diameter [m] fluorescent color, the more intense the radical hydroxyl pro-
HEX-3P Hexagonal sonoreactor with 3 Probes duction. A typical sonoreactor system consisted of multiple
HEX-5P Hexagonal sonoreactor with 5 Probes types of equipment. First, the generator (with amplitude/fre-
I Acoustic intensity [W/m2] quency/power arrangement display) excites vibration in the
IUs Ultrasonic intensity [W/m2] water-cooled piezoelectric transducer; the Barbell horn am-
f Ultrasound frequency [kHz] plifies the vibration amplitude and delivers the ultrasonic
P Acoustic power [W] energy to the liquid premix, which is continuously pumped
p Acoustic pressure [kPa] through the reactor chamber from the premix tank into the
p0 Acoustic amplitude [kPa] product tank. The determination of the electrical power Pelec is
R Radial distance [m] done by measuring the power of standby and the total
SQR-3P Square sonoreactor with 3 Probes consumed power Ptotal . In conclusion, sonochemists world-
SQR-5P Square sonoreactor with 5 Probes wide shared their opinion on “green sonochemistry” and gave
T Time [s] their vision on future trends developed in the field [5].
U Velocity [m/s] High-frequency sound waves have the power to discover
V0 Velocity amplitude [m/s] and extinguish. They have been used in several medical
x Position in x direction [m] extensively, clinical, and engineering applications for moni-
X0 Displacement toring, diagnostics, and cleaning processes [6]. Various studies
Z Acoustic impedance concerned the influence of ultrasound in the cleaning and
Greek symbols removing fouling within a liquid-filled medium [7e9]. Arafa
q Radial angle et al. [10] reviewed some of the recent patents and de-
l Wavelength [m] velopments of thermoacoustic applications. For the water
m Dynamic viscosity [kg/m/s] sonication application, cavitation is generated due to high-
r Density [kg/m3] pressure fluctuations in the liquid medium, leading to
u Angular frequency [rad/s] acoustic cavitation bubbles. In terms of practicality, the
sonoreactors were used a long time ago in many applications.
Subscripts For example, in Chemical and Biochemical Processes, Kim
∞ Uniform state et al. [11] brought a review of the alternative sources and en-
a Acoustic ergy methods for intensifying these processes. Lin et al. [12]
US Ultrasound utilized the benefit of the hydrodynamic cavitation and its
consequence intensification process in designing a sonohy-
drothermal synthesis with enhanced catalytic performance.
Introduction In ultrasonic sonochemistry, the collapse of these micro
acoustic bubbles is used to enhance chemical reaction rates
It has been recognized that sonochemistry opens new op- and thereby increase reactivity.
portunities for green chemistry [1,2]. However, there are still The importance of the geometrical designs of sonoreactors
many questions that need to be answered. Some significant lies in their significant contribution to the efficiency and per-
questions are: how does the use of ultrasound affect chemical formance parameters. Zhang et al. [13] performed numerical
reactions? Can scientific advancement and environmental simulation and experimental investigations on the influence
sustainability go hand in hand?. The application of ultrasound of sound directions of acoustic field characteristics within a
waves to chemical reactions (i.e., sonochemistry) offers rectangle-shaped sonoreactor. They provided useful guide-
enormous potential for innovation in eco-friendly and eco- lines for the cavitation threshold, including percentages on
efficient chemistry, which is what we need in today’s world the regular cavitation and severe cavitation. Son et al. [14]
[3,4]. This article introduces the basics of ultrasonic waves in reported that sonochemical reactors’ design is still developing
sonoreactors and the recent development in mapping and for upscaling to serve in conventional and industrial opera-
designing efficient sonoreactors. This is before moving on to tions. Therefore, the design of sonoreactors requires consid-
look at acoustic cavitation and the estimation of ultrasonic erable attention, reliable modeling, and perdition of the
parameters. It is known that ultrasonic waves are powerful. ultrasound energy absorbed in the liquid medium, which re-
The resultant high energy forces lead to implode cavitation sults in the generation of acoustic cavitation bubbles [2].
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
Several parameters contribute to the performance of the
sonoreactors. For example, the ultrasonic frequency has the
highest contribution to hydrogen production [15]. Several
improvements have been developed so far to obtain the
maximum possibility or probability of generating cavitation
zones. For example, redesigning the transducer probes (i.e.,
sonotrodes) obtain a higher cavitation rate [16,17]. Some
studies used an array of transducer probes to overcome the
issue of energy dissipation [18,19]. For instance, Rahimi et al.
[20] performed an experimental study on transducer position,
and ultrasonic wave interaction on the heat transfer inside a
cylindrical sonoreactor featured with two transducers
mounted on the lateral wall and three transducers mounted at
the bottom wall. The results revealed that a single transducer
mounted at the lateral wall could be more efficient from heat
transfer. Fang et al. [18] studied the effect of different trans-
ducer probe tips on cavitation and acoustic streaming to
improve the cavitation efficiency. They performed experi-
mental and numerical studies, which revealed that the type of
sonotrode tip affects the acoustic streaming. Thus, it will alter
the acoustic pressure distribution, flow direction, and the
generated cavitation zones. They reported that the flat sono-
trode tip provides a more efficient cavitation zone than the
tapered and the triangular shape sonotrodes. Widening the
cavitation zone area will improve cavitation efficiency. Tiong
et al. [21] have studied the acoustic pressure for the emulsion
of oil-in-water using an ultrasonic transducer probe. They
performed several experimental studies concerning the ul-
trasonic horn’s position and the high intensity. The study has
successfully correlated the effect of the horn position on the
acoustic pressure generated. Niazi et al. [22] have studied the
operational parameters of different sonoreactor configura-
tions in water and crude oil. The results showed that the lower
frequency range is optimum to use as it produced the
maximum magnitude of the negative pressure. Son et al. [23]
investigated the acoustic cavitation energy in a large-scale
sonoreactor. They presented a successful attempt to obtain
two empirical correlations to estimate the sonochemical effi-
ciency. The first one is to correlate the cavitation energy with
the input transducer energy in Watt. Simultaneously, the
second correlation is meant to calculate the sonochemical
efficiency by using cavitation energy, which is initially ob-
tained from the first correlation. Although different in-
vestigations discussed here have provided beneficial
information regarding ultrasound water sonication principles,
scaling up the sonoreactor for conventional and industrial
operation remains a considerable challenge. From the chem-
istry point of view, ultrasonic has been used as an assisted
process to enhance chemical reactions and promote the
chemical environment in different applications [24,25]. Wang
et al. [26] reviewed possible processes for intensification of
water electrolysis for hydrogen production. Achaoui et al. [27]
investigated the tunable microbubble sonoreactor using
electrolysis and ultrasound. It has been proven that the effi-
ciency of hydrogen production improved by 4.5%, and the
energy efficiency improved by 1.3% while using ultrasound-
assisted electrochemical electrolysis-process. Additionally,
researchers investigated ultrasonic-assisted wastewater
15221
treatment [28] and ultrasonic assisted-microalgae technology
[29]. Recently, Rashwan et al. [30,31] performed a numerical
simulation analysis for ultrasonic-based hydrogen produc-
tion. These research studies are still developing and need
further investigations. Moreover, ultrasonic proven a signifi-
cant improvement when used in wastewater treatment ap-
plications. Some researchers reported the possibility of
producing hydrogen out of such a process using the power of
ultrasound. Raschitor et al. [32] have improved electrolysis of
colloid-polluted wastes using ultrasounds and electro-
coagulation. Also, Rashwan et al. [28] published a perspective
article regarding an ultrasonic-assisted coagulation and floc-
culation process and a possible hydrogen production out of
such process. From the hydrogen production point of view, Fu
et al. [33] reviewed the principles, challenges, and perspec-
tives of hydrogen production from waste-activated sludge.
Yang and Wang [34] reported an improved hydrogen produc-
tion from sewage sludge while using the water electrolysis
added with ultrasonic.
Some of the issues related to scaling up sonochemical
reactors were addressed by Berlan and Mason [35]. Besides,
the challenges associated with the design of the sonoreactor
are summarized in our recent review article [2]. Undoubt-
edly, further research is needed to overcome such design
issues and explore the opportunities for developing more
efficient sonoreactors. Typical issues to date with sonor-
eactors that require further investigation can be shortened
as follow:

The acoustic pressure distribution inside the reactor is not
yet fully understood. It will need to be addressed in order to
enhance our fundamental understanding of the acoustics
field, which is highly complex due to the inhomogeneous
distribution of bubbles.

The sound waves’ attenuation in the liquid medium due to
the generated acoustic cavitation bubbles needs to be
clarified. As ultrasound waves are emitted through the
liquid medium, the sound waves’ acoustic intensity de-
creases as the sound waves propagate away from the ul-
trasonic probe. However, the attenuation rate is ill-defined
due to an interaction between the sound waves and the
cavitation bubbles.

Scaling-up the sonoreactor and the implementation of in-
dustrial sonoreactors are not materialized yet. The major-
ity of the current research studies are still addressing
issues with lab-scale sonoreactors to identify the optimum
design and operating conditions. Our recent article [36]
provided a review and guidelines for scaling-up different
sonoreactor geometries.
Despite the exciting work conducted up to date to un-
derstand ultrasound water sonication, many challenges
need further investigation to design industrial sonoreactors.
Moreover, reliable numerical strategies and simulation tools
are not obtainable yet, and these issues mainly depend on
experiences and trial-and-error. Apart from these design
issues, it is a matter of experience that even the same
reactor can vary drastically under similar conditions.
15222 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
Therefore, this article aims to provide a comprehensive
model and a tremendous computational approach to tack-
ling these issues.
The present study’s main objective is to investigate the
optimum arrangement for generating acoustic cavitation
bubbles for hydrogen production. Four sonoreactor geome-
tries will be simulated using the finite-element method and
solved by the available computational fluid dynamics tool in
COMSOL Multiphysics. This study’s second objective is to
investigate the influence of the geometrical parameters such
as the sonoreactor geometry, wall boundary, and conditions
on the acoustic pressure distribution inside the sonoreactor.
This includes varying the outer sonoreactor geometries,
namely, the typical cylindrical sonoreactor shape, the hex-
agonal sonoreactor, and the square sonoreactor. Additionally,
the effect of adding multiple sonotrodes is also investigated,
and the acoustic pressure distribution is quantified,
compared, and reported. Finally, the input power effect on the
Fig. 1 e Computational domains of unique sonoreactor geometries with a sonotrode introduced from the top.
cavitation energy and sonochemical hydrogen yield is inves-
tigated while varying the input power through having multi-
ple sonotrodes.
Numerical modeling
In this section, the numerical approach used for solving the
linear acoustic wave propagation is described. The focus is to
investigate the finite element method FEM of wave propaga-
tion in different geometries. The goal is to benchmark the four
proposed geometries based on validation studies and perform
further computational analysis to quantify their acoustic
characteristics and performance. Moreover, the effects of
different boundary conditions and scaling-up the sonoreactor
geometry are of interest. Since the linear analysis is compu-
tationally inexpensive, the COMSOL Multiphysics FEM pack-
age is performed.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
Computational geometry development
The numerical modeling is performed using different 3-D
sonoreactor geometries with their dimensions drawn in 2-
D domains and transformed into 3-D by revolving the 2-D
surfaces 360 , which is shown in Fig. 1. COMSOL Multi-
physics is used to draw all of the geometries. GEO.1 is a
typical sonoreactor configuration inspired by an experi-
mental work reported by Ref. [37] and validated for bench-
marking. While GEO.2 is an inverse cone configuration,
GEO.3 is a typical cylinder sonoreactor with a curved bottom
wall, and GEO.4 is a regular cone shape with a broader
bottom area. The geometries are built and inspired by
Ref. [38], who researched analytical applications for sono-
chemistry. These geometries are customary geometrical
shapes of sonoreactors used in chemical industries. The
geometries developed based on how the sonoreactor wall
and base will influence the computational results, including
a contracting wall and base (GEO.2), a concave base (GEO.3),
and an expanding wall and base (GEO.4).
At an ambient temperature of 25  C and atmospheric
pressure, all sound reactors are filled with water. The ultra-
sound transducer is immersed in the water at a certain dis-
tance d ¼ 20 mm (20 kHz, 36 W, and a tip-diameter,
DP ¼ 20 mm) from the top of the acrylic glass cylinder. The
acoustic pressure-frequency domain COMSOL module is used
mainly to predict the acoustic and flow fields’ characteristics
of the sonoreactor models.
Governing equations
The wave equation for acoustic pressure passing through a
liquid medium is one of the most fundamental acoustics
equations. It defines the space and time properties of the
sound field, which can be specified as:
v P 1 v P
2 2
 ¼0
vx2 c2 vt2
where the acoustic pressure (P ¼ N/m2), the speed of sound
passing through the liquid medium (c ¼ m/s). By extending
equation (1) considering 3-D in a homogenous liquid medium,
the equation can then be specified as:
1 1 v2 P
V2 P  ¼0
r rc2 vt2
For numerical simulation, ultrasound field propagation is
examined. There are certain presumptions for this particular
application that the ultrasound waves can propagate through
the fluid medium, including the linearity of the sound wave,
disregard of shear stresses, incompressible fluid medium, and
the assumption of time-harmonic strain. We can obtain the
following equation based on these assumptions, which is
defined as the Helmholtz equation:
V2 P u2
 2P¼0
r rc
The equation is solved using COMSOL Multiphysics soft-
ware to give access to the acoustic pressure distribution inside
the sonoreactor after selecting the correct physical models
15223
and the suitable boundary conditions. Additionally, the
following equations are used to calculate each of the following
by order: the acoustic intensity, velocity amplitude, sonotrode
displacement, and acceleration [39] as follows:
PUS
 2 
IUS ¼ W m (4)
Ap
P0
V0 ¼ ½m = s (5)
r0 C0
V0
Z0 ¼ ½mm (6)
u

 
a ¼ 4p2 f 2 Z0 m s2 (7)
Lastly, the cavitation energy and cavitation yield are used
to estimate the hydrogen quantity emitted by the sonohy-
drogen sonoreactor. The cavitation yield is defined as the
number of reaction products per unit of irradiation time in
seconds per unit of energy consumption in kWh. The cavita-
tion energy indicates the energy dissipated from the supplied
input/electrical energy, and it can be described in terms of the
input power as [14]:

Cavitation energy:
2
Cavitation Energy ¼  0:0008 ðinput energyÞ
þ 0:4699 ðinput energyÞ (8)

Sonochemical energy consumption:
2
Sonohydrogen Yield ¼ 0:0003 ðcavitation energyÞ
þ 0:0140 ðcavitation energyÞ (9)
The selected physical models and boundary conditions will
be illustrated in the next sections. However, in a later study
(1) concerning the acoustic streaming induced by the sonication
effects, the compressibility effects have been involved in the
study through a comparison study when using three different
densities, namely, the constant density, linear density, and
nonlinear density effects on the acoustics streaming.
Physical models and boundary conditions
(2)
The acoustic module for frequency domain physics has been
selected to model a continuous wave excitation and obtain the
permanent time-harmonic solution [39,40]. The stimulation
amplitude should be dependent on the strength and power of
the acoustic input. Domain with all necessary boundary con-
ditions to simulate the cases is shown in Fig. 1. The simulations
incorporate a linear model with boundary conditions summa-
rized below while assuming infinitely rigid and perfectly
reflective acoustic boundaries in all cases, except one special
case with absorbing boundaries. This assumption is changed
(3) during the study about the effect of different types of bound-
aries. Since most ultrasonic sources emit sinusoidal waves, the
pressure wave can be assumed as harmonic. The conversion
factor of the ultrasonic probe is assumed to be 0.85. Lastly,
during the acoustic streaming validation, it is assumed that the
15224 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

probe’s acoustic streaming pattern is a jet inlet that is used cylindrical container with an internal diameter,
regularly over the years [41,42]. The boundary conditions are D ¼ 135 mm, and a height, H ¼ 170 mm, with an aspect ratio
implemented based on COMSOL 5.4 library modules [43] and (H/D) of 1.26. The operating parameters of the sonoreactor
previously reported studies [22,37,44] and given as follows: are 20 kHz ultrasound frequency and acoustic power of
36 W. The input parameters for the validation study used by
a) The pressure limit condition is applied to the transducer’s Ref. [37] are given as follows density r0 ¼ 998 kg/m3; speed of
tip as [P ¼ P0], based on acoustic energy and strength sound C0 ¼ 1498 m/s; frequency f ¼ 20 kHz; ultrasound
measurements. The acoustic strength is, therefore, given power PUS ¼ 36 W; transducer probe diameter Dp ¼ 20 mm;
as follows: IUS ¼ PUS
¼
P20 intensity I ¼ 114.65 kW/m2, the corresponding acoustic
A 2r0 C0 , where IUS is the acoustic in-
2 amplitude PA ¼ 585:49 kPa. The comparison is drawn with
tensity in [W/m ], PUS is acoustic power in [W], and A is the
the previous experimental study considering the acoustic
transducer probe tip surface area in [m2].
pressure distribution along the centerline of the sonor-
b) The sidewall of the sound-hard border is determined, and
eactor. Noting that they did not measure the speed of sound,
the sides of the ultrasound transducer are described as the
despite its drastic change with the container’s material. The
sound-hard border wall and are as follows: vp= vn ¼ 0. It is
speed of sound determined by measuring the pressure and
presumed that the wall material has a condensed effect
temperature of the water, taking into consideration that the
since obviously the reactor material and compliance with
container’s material will drastically affect the speed of
this presumption would impact.
sound.
c) The water’s surface should be a sound-soft boundary
(P ¼ 0) where the pressure P0 will disappear at that
boundary is connected with a soft obstacle. It is also the so-
called surface for releasing pressure, which applies to Results and discussion
acoustic problems underwater.
A comprehensive numerical analysis of the pressure distri-
For the sake of benchmarking, the model validation bution inside various geometries is done using the FEM.
study for GEO.1 is successfully performed and reported in a Considering the effect of the geometrical parameters such as
recent article published in ultrasonic sonochemistry by different sonoreactor geometries, wall boundary conditions,
Rashwan et al. [36]. GEO.1 consists of an acrylic glass including absorbing/reflecting walls, the effect of the number

Fig. 2 e Velocity streamlines due to acoustic streaming in the present study (left) compared to the experimental results
(right) of laboratory horn at 300 W by Dahlem et al. [45].
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
of sonotrodes, and the effect of the outer sonoreactor geom-
etry are investigated, quantified, and reported in the following
subsections.
Model validation
Validation of the numerical model is successfully con-
ducted with experimental data. The experimental work
available in the literature considering data measured by
the particle image velocimetry PIV scares in the literature.
However, the experimental work by Dahlem et al. [45]
shows the PIV flow characteristics inside a sonoreactor for
given parameters at a high-frequency of 20 kHz and
acoustic power of 300 W. Despite the fact that the power
is different, the velocity field inside the sonoreactor is the
same. The predicted CFD velocity profile and pattern show
an excellent agreement with the experimental data avail-
able in the literature. Moreover, a previous model valida-
tion study for GEO.1 is successfully performed concerning
the pressure field at the reactor axis and reported by
Rashwan et al. [36]. Henceforward, the acoustic streaming
Fig. 3 e Effect of absorbing boundary conditions on wave propagation for an ultrasonic frequency of 20 kHz.
15225
field of the model is successfully validated with the
experimental visualization of the typical flow pattern in a
half vertical plan is reported by Dahlem et al. [45] and as
shown on the left side of Fig. 2.
The velocity pattern consisted of two eddies, the large eddy
allocated at the bottom corner of the sonoreactor. The small
vortex is shown at the top corners of the reactor. The
axisymmetric model for a transducer probe is established
with a no-slip boundary condition. As mentioned before, the
liquid medium was initially stagnant, and the flow induced by
the ultrasound effects because of the acoustics source. The
modeling simulations show an excellent agreement with the
experiments regarding the flow pattern captured by the PIV.
There are two counter-rotating eddies, as seen in the experi-
mental results. These two large eddies are also captured with
the current numerical study, which validated the present
model. These results are also following the results presented
by Trujillo and Knoerzer [41], Schenker [46], and Slama [47].
They reported a similar observation of the characterization of
ultrasound-induced acoustic streaming and the jet-like
streaming velocity profile.
15226 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
Fig. 4 e Effect of reflecting boundary conditions on wave propagation for an ultrasonic frequency of 20 kHz.
Pressure distribution in different sonoreactors
Sonoreactor geometry is one of the important parameters that
significantly affect sonoreactors’ performance. Firstly, in
order to make a reliable comparison, all geometries are
exposed to the same operating conditions, boundary condi-
tions, and acoustic intensity to probe the possibility of cavi-
tation. A comprehensive analysis of the geometries and
pressure distribution is done using the FEM. Changing the
direction of the sonoreactor boundaries affects the acoustic
pressure field of the fluid inside the sonoreactor. This study
investigates the maximum possibility/probability of gener-
ating acoustic cavitation bubbles in each sonoreactor. It can be
revealed that the maximum acoustic pressure value did not
significantly change with the sonoreactor geometries. While
the minimum pressure slightly fluctuates. This is because the
reflected wave interferes with the incident waves leading
either to the building or destruction. The problem with GEO.3
is that it has only two peaks of acoustic pressure, which is not
preferable from the cavitation bubble generation perspective.
In contrast, the issue of GEO.4 is because acoustic pressure
decays with the distance from the sonotrode tip. Acoustic
standing wave means that the wave is dependent on the ge-
ometry (space dependent). It means that at any position, the
amplitude will oscillate at 20 kHz. So, we are showing one
instant in the acoustic cycle. The other instant that is 180 out
of phase will be the same distribution but multiplied by (1).
Hence, the acoustic pressure peaks located at the required
distance at which the reflected wave will make a constructive
interference. Then we adapt the geometry of the reactor based
on those distances. This way, we get the maximum perfor-
mance. In which the acoustic pressure limits are defined by
the max and min acoustic pressure that each reactor can
achieve at the same operating conditions. Exceeding these
limits will lead to unstable acoustic zones. Interestingly, all
cases’ max pressure remains almost constant because of the
same amplitude in all geometries. In contrast, the minimum
pressure is significantly affected by the orientation of sonor-
eactor walls that might reflect off or absorb the incident wave.
Effect of different boundary conditions
The novelty of this study is the variation of the sonoreactor
wall in order to see how this would affect the performance.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0 15227

Fig. 5 e Centerline pressure profile for various geometries with (a) absorbing boundaries and (b) reflecting boundaries.

Table 1 e The acoustic parameters corresponding to different ultrasound probe diameters.

Probe diameter Dp [mm] 20 30 40 50 60 70
Probe Area Ap ½m2  0.000314 0.000707 0.001256 0.001963 0.002826 0.003847
Ultrasound Intensity ½W= m2  114,650 50,955 28,662 18,344 12,739 9359
Acoustic Amplitude P0 ½kPa 585.49 390.33 292.75 234.2 195.16 167.28
Velocity Amplitude ½m= s 0.39 0.26 0.2 0.16 0.13 0.11
Displacement ½mm 3.12 2.08 1.56 1.25 1.04 0.89
Acceleration [39] ½m= s2  49,189.2 32,792.8 24,594.6 19,675.7 16,396.4 14,054.1

The linear wave equation has proven efficient in investigating
different parameters affecting the pressure distribution inside
the sonoreactor. Therefore it is used over again to investigate
the sensitivity of the ultrasonic wave propagation to the
boundary conditions for different sonoreactor geometries.
Therefore, the study considered different boundary condi-
tions for all reactors, including the effect of the absorbing and
reflecting sonoreactors’ walls. Klima et al. [48] performed a
FEM approach to optimize a cylindrical reactor’s geometry by
changing the boundary conditions by either making it
reflecting or absorbing boundaries. The results are shown in
Fig. 3 and Fig. 4 for the absorbing and reflecting boundaries,
respectively. It can be seen in Fig. 3 that the acoustic wave
emitted from the sonotrode has created sound wave layers
with the highest pressure located at the tip of the sonotrode,
and the absorbing boundaries have attenuated the introduced
15228 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

Fig. 6 e Effect of transducer probe diameter, Dp on (a) the acoustic map, maximum and minimum acoustic pressure (b) on the
axial acoustic pressure distribution.

acoustic pressure. Besides, the acoustic pressure peaks near
the transducer probe tip in all geometries remained un-
changed, while a slight change is noticed in the minimum
pressure inside the sonoreactor. Generally, all geometries
have acted the same in the case of absorbing boundaries.
For the sake of comparison, the same simulation is per-
formed by changing the absorbing walls to reflecting walls.
The reflecting boundaries have led to a higher probability of
generating acoustic cavitation bubbles and a higher number of
cavitation zones. Fig. 4 presents the sound-hard boundaries at
the sonoreactor walls representing the reflected boundaries
compared to the sound-soft walls representing the absorbing
boundaries. It is seen that the sound-hard boundaries help in
creating high and low-pressure regions in the reactor with
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0 15229

Fig. 7 e Acoustic shapes and modes for sonoreactors at 20 kHz (left) and corresponding Eigen frequencies (right) with
different ultrasonic probes; the average pressure is taken over the volume from the COMSOL Acoustic Module Solution. Note
that the vapor pressure for cavitation is 3171.47 Pa at 25.0  C

more significant cavitation zones. Generally, altering the wall
properties, higher and lower pressure regions can be gener-
ated all over the sonoreactor.
Furthermore, Fig. 5(a) presents the axial pressure distri-
bution of different sonoreactor geometries concerning the
absorbing boundaries. Generally, the ultrasonic wave decays
as the propagation distance increases. By comparing all ge-
ometries, it is seen that all geometries experience almost the
same pressure distribution with a slight difference in the 2nd
pressure peak location. GEO.4 has shown the maximum 2nd
peak. It is because it has the most extensive area of the bottom
wall. Undoubtedly, this will directly affect the cavitation
possibilities as the bubbles should be introduced to many
consecutive cycles generate, enlarge, and explode to benefit
from the energy produced from the bubble collapse.
Moreover, Fig. 5(b) presents the axial pressure profile for
different geometries. The first two geometries experience
pressure peaks two times of the ultrasound source amplitude
by comparing all geometries. It is concluded that higher and
lower pressure zones can be achieved even at low frequencies
by altering the wall boundary conditions. It will directly affect
the cavitation possibilities as the bubbles should be intro-
duced to many consecutive cycles to generate, enlarge, and
explode to benefit from the energy produced from the bubble
collapse. For this reason, GEO.3 and GEO. 4 are not preferable
as they could be viewed negatively because the bubble won’t
have enough time to execute the collapse process properly.
Effect of probe diameter
The novelty of this study is investigating the cases where we
have larger sonotrodes or ultrasonic transducer diameters.
The question beyond this study is: does the area subjected to
the ultrasonic would have a positive effect? Therefore, the
probe diameter effect on the sonoreactor GEO.1 is simulated,
analyzed, and reported. The transducer probe diameter is an
important geometrical parameter. Undoubtedly, altering the
probe diameter will change the acoustic pressure amplitude,
which can be calculated by a call to mind equation (5). To set
the pressure boundary state, the pressure amplitude P0 at the
probe tip is required concerning different probe diameters and
the same acoustic intensity. The acoustical parameters at
different ultrasonic transducer probe diameters are shown in
Table 1.
The acceleration of the transducer tip in the axial direction
can be given as az ¼ 4p2 f 2 Z0 ½m =s2  which is reported by
15230 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

Fig. 8 e Acoustic shapes and modes for sonoreactors at 20 kHz (left) and corresponding Eigen frequencies (right) with
different ultrasonic probes; the average pressure is taken over the volume from the COMSOL Acoustic Module Solution. Note
that the vapor pressure for cavitation is 3171.47 Pa at 25.0  C.

Rubinetti [39] in their recent numerical modeling study and
validation concept for acoustic streaming induced by the ul-
trasonic transducer. This unique study sets the acoustic
boundary condition for the sonotrode tip as axial acceleration.
The Origin and derivation of such an equation can be found in
Chapter 10, “Modeling energy in power ultrasound trans-
ducers,” of the book entitled “Power Ultrasonics: Applications
of High-Intensity Ultrasound” by Gallego and Graff [49]. The
equation originates from the displacement equation along
with the piezoelectric transducer, which is considered linear
assuming Newtown’s Law. The simulations reported in this
book chapter give the acceleration as a function of the driving
frequency under varying amplitudes levels and up to
80,000 m/s2.
Fig. 6(a) presents the maximum and minimum acoustic
pressure variation at different probe diameters. From the
stable acoustic pressure zone that there is a symmetrical
behavior at the zero acoustic pressure line. The maximum
magnitude of the negative pressure increases while
increasing the transducer diameter. It is attributed to the fact
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0 15231

Fig. 9 e Maximum and minimum acoustic pressure for all geometries at (a) 20 kHz and (b) Eigen frequencies.

Table 2 e Optimization and comparison of the effect of the number of sonotrodes.

#N of Sonotrodes 1 2 3 4 5
Frequency [Hz] f ¼ 20000
Minimum pressure [Pa] 4.43  105 6.99  105 1.27  106 1.30  106 1.22  106
Eigen Frequency [Hz] 19670 20069 19990 20077 20089
Minimum pressure [Pa] 9.31  105 1.18  106 1.21  106 3.48  106 1.39  107

that when the transducer probe dimension alters, this will diameter’s effect on the axial acoustic pressure distribution. It
alter the distance of the sound-hard boundaries and change is well-known that the ultrasound waves should be intro-
the wave point (node or antinode) and change the reflected duced from the transducer probe tip with a diameter smaller
waves’ phase well [22]. Fig. 6(b) presents the transducer probe than the acoustic wavelength; hence the acoustic cavitation
15232 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

Fig. 10 e Acoustic shapes and modes for sonoreactors at 20 kHz (left) and corresponding Eigen frequencies (right) with
different ultrasonic probes; the average pressure is taken over the volume from the COMSOL Acoustic Module Solution. Note
that the vapor pressure for cavitation is 3171.47 Pa at 25.0  C.

bubbles could be generated. As seen, the axial location of the
acoustic pressure nodes and antinodes matched at different
probe diameters and did not alter; however, the acoustic
pressure values did not alter considerably while increasing
the probe diameter.
Geometric optimization
In the following sections, a three-dimensional geometric
optimization study is performed to investigate the optimum
geometry that best benefits the sonohydrogen process. The
study is considered in three phases; the first phase is con-
ducting acoustic simulation on different arrays and geome-
tries considering Eigen frequencies and the most typical
ultrasonic frequency 20 kHz. The second phase is to simulate
the acoustic streaming regarding the flow field distribution.
The third and final phase is the hydrogen production calcu-
lation from such geometries.
Effect of the number of sonotrodes
A novel simulation analysis is conducted on the effect of
having multiple probes in on sonoreactor. First of all, having
studied the type-A sonoreactor contributes to the field as few
studies have reported simulations about this type of sonor-
eactordmost of the studies available in the literature
considered type-b and type-c sonoreactors. Son et al. [50]
performed a geometric-optimization experimentally of
sonoreactors to enhance the cavitation activity using a type-b
bath sonoreactor. The probe is fixed at the lowest side of the
sonoreactor. They considered changing the aspect ratio of the
sonoreactor by varying the liquid height above the ultrasonic
transducer. They concluded that as the liquid height in-
creases, the cavitation yield increases substantially under the
same input power. In this parametric study, we have consid-
ered varying the number of sonotrodes to see how this would
affect the negative pressure’s maximum magnitude. For this
reason, 5 three-dimensional geometries are built with a
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0 15233

Fig. 11 e Acoustic shapes and modes for sonoreactors at 20 kHz (left) and corresponding eigenfrequency (right) with 5
transducers each; the average pressure is taken over the volume from the COMSOL Acoustic Module Solution. Note that the
vapor pressure for cavitation is 3171.47 Pa at 25.0 C.

different number of ultrasonic probes immersed from the top
side, starting from 1 to 5. All geometries have the same di-
mensions and boundary conditions while changing the
number of probes, and all probes were set to operate on the
same phase. We run initial acoustic simulations for all five
geometries using COMSOL to determine the Eigen frequencies
or the natural frequencies and the corresponding mode
shapes for all the suggested geometries. The acoustic simu-
lations took minutes to solve the Helmholtz equation. Once
the acoustic simulations are done, the excitable modes and
frequencies are checked. The ultrasonic probe is then intro-
duced to work simultaneously to trigger a resonance condi-
tion [51,52]. This resonance condition will enhance the
sonohydrogen process and make it more efficient. Fig. 7 and
Fig. 8 consider the differences in the acoustic shapes and
modes for different sonoreactors at 20 kHz, as shown in the
figure’s left column. The right column shows the same
sonoreactors operating at their corresponding Eigen fre-
quencies. In the case of a cylindrical sonoreactor with 3
sonotrodes, the geometry is tested regarding different Eigen
frequencies. We found that, at an Eigen frequency of 20076 Hz,
the sonoreactor would not hold the pressure generated and
accumulated inside the sonoreactor. It may lead to damage to
the sonoreactor. Therefore, another Eigen frequency is
selected from the solution 19990 Hz that gives more reason-
able pressure values. The cylindrical sonoreactor with 3
sonotrodes shows a higher magnitude of negative pressure,
which will promote the acoustic cavitation generation
process.
For the sake of clarification, a comparison is made to
differentiate between all geometries. The comparison is
drawn on the maximum and minimum acoustic pressures
and reported in Fig. 9(a) at 20 kHz & (b) at the Eigen frequencies
corresponds to each geometry. The results show no significant
changes in the maximum and minimum pressures when
using 1, 2, up to 5 probes for both studies at 20 kHz and each
geometry’s eigenfrequency.
The highest sound pressure is usually related to the
acoustic power supplied through the transducer probe, which
has not changed significantly. In contrast, a significant change
is noticed in the maximum magnitude of the negative pres-
sure, which corresponds to the formation of acoustic cavita-
tion bubbles. This is because the higher the magnitude of the
negative pressure, the bigger the minimum pressure and the
15234 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

Fig. 12 e Maximum and minimum acoustic pressure for all geometries at (a) 20 kHz and (b) Eigen frequencies for geometries
featured with 5 probes.

possibly generating more bubbles. Table 2 shows the sum-

Table 3 e Optimization and comparison on the effect of
mary description of the optimization study and compares the
different geometries for the 5 sonotrodes analysis.
effect of the number of sonotrodes.
Geometries Cylinder Hexagon Square
Frequency [Hz] f ¼ 20000 Effect of the outer sonoreactor geometry
Minimum pressure [Pa] 1.22  106 8.28  106 2.29  106 In this section, a novel study is conducted considering the
Eigen Frequency [Hz] 20089 20209 19993
effect of having different outer sonoreactor geometries built in
Minimum pressure [Pa] 1.39  107 2.33  106 2.25  106
three dimensions, namely, the typical cylindrical sonoreactor,
Hexagon reactor, and square sonoreactor. This study is made
water vapor pressure. The optimum performance is recorded in two phases; the first phase is building those unique three-
when the 3-sonotrodes geometry is used as the pressure drops dimensional geometries with 3 ultrasonic transducer probes.
down to 1.21  106, enhancing the sonochemical activity and While the second phase, all unique geometries are featured
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
Fig. 13 e Cavitation percentage over the sonoreactor volume based on Blake Cavitation Analysis.
with 5 ultrasonic transducer probes. In this parametric study,
we have considered varying outer geometry to see how this
would affect the negative pressure’s maximum magnitude.
For this reason, 3 three-dimensional geometries are built. All
geometries have the same diameter length and boundary
conditions. The hexagon and the square are drawn inside the
circular cylinder diameter given earlier. The same simulation
procedure used in the previous section is used herein. We run
initial acoustic simulations for all three geometries using
COMSOL to determine the Eigen frequencies or what is so-
called the natural frequencies and the corresponding mode
shapes for all the suggested geometries. The acoustic simu-
lations took a couple of minutes to solve the Helmholtz
equation. Once the acoustic simulations are done, the excit-
able modes and frequencies are checked. Then, the ultrasonic
probe is introduced to work at the same frequency to trigger a
resonance condition. This resonance condition will enhance
the sonohydrogen process and make it more efficient. Fig. 10
presents the differences in the acoustic shapes and modes
for sonoreactors with 3 sonotrodes at 20 kHz (left) and corre-
sponding Eigen frequencies (right). All geometries have the
same dimensions and boundary conditions while changing
the number of probes, and all probes were set to operate on
the same phase. As seen in the case of hexagonal sonoreactor
with 3 sonotrodes, the magnitude of the negative pressure
increased from 2.54  106 to 2.23  106, thus increase the
possibility of cavitation. The results revealed that, in some
cases, when operating the sonoreactor at the Eigen frequency,
it will trigger resonance that affects the maximum magnitude
of the negative pressure. The hexagonal sonoreactor is tested
at the eigenfrequency of 19903 Hz; however, the results were
not in favor with the safe operation of the sonoreactor and
may lead to collapse. Therefore, a follow-up study is con-
ducted using another eigenfrequency of 20206 Hz showing a
reasonable average pressure distribution as reported in Fig. 10.
15235
For comparison, Fig. 11 presents the differences in the
acoustic shapes and modes for sonoreactors when the
sonoreactors were featured with 5 probes at 20 kHz. At the
same time, the sonoreactors were tested while operating
under the Eigen frequency conditions. The results are sorted,
and it has been revealed that operating a Hexagonal sonor-
eactor at the following conditions: Eigen Frequency of
19331 Hz with 5 sonotrodes may adversely affect the sonor-
eactor body due to the build-up pressure that the sonoreactors
with the following dimensions (HEX sonoreactor, 5 sonotr-
odes, 170 cm long, 50 cm diagonal length) cannot stand and it
may collapse. Therefore, another Eigen frequency is chosen
for the hexagonal and square sonoreactors and reported in
Fig. 11. The main observation from such analysis is that, in
order to generate cavitation, the average pressure has to fall
below the vapor pressure, which is 3171.47 Pa at 25.0  C. The
results showed that the square sonoreactor is considered not
reasonable for generating acoustic cavitation bubbles as its
average pressure is almost atmospheric.
The maximum and minimum pressure magnitudes are
reported here; after conducting the Eigen frequency test for all
geometry in 3-D, which improved the maximum magnitude of
the negative pressure and the cavitation energy and compared
them with the typical ultrasonic frequency of 20 kHz. Fig. 12,
the hexagon reactor showed better understanding at 20 kHz
with no significant difference with its Eigen frequency, while
the square showed much better at its natural frequency.
As reported in the previous studies, we reported the dif-
ferences in the magnitude of the negative pressure between
all geometries are shown in Table 3 regarding the optimiza-
tion and comparison of the effect of different geometries.
Further contrast is made to compare the geometries’ per-
formance with 3 sonotrodes and geometries with 5 sonotr-
odes. At 20 kHz, there are no differences in the negative
pressure in the cylindrical sonoreactor when 3 or 5 sonotrodes
15236 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

Fig. 14 e Input power vs. Cavitation Energy and Energy Conversion Efficiency (a) and Input power vs. Cavitation Energy
Sonohydrogen Efficiency (b).

are applied. However, in the hexagonal sonotrode with 5
probes showed a significant change. This can be attributed to
the constructive interference of the sound waves when having
5 sonotrodes. Opposite observation is found when having the
square sonoreactor. The constructive interference takes place
when having 3 sonotrodes. It does not matter how many
sonotrodes are introduced in the sonoreactor, but the most
important thing is when constructive interference would
occur.
In closing, this analysis is meant to investigate which
sonoreactor has a higher possibility of generating acoustic
cavitation by understanding the physics and the maximum
magnitude of the negative pressure associated with each
sonoreactor. However, the number of bubbles increases
significantly when the ultrasonic frequency increases. This
observation is matching with some reports in the literature. It
is not possible to make a clear comparison since most litera-
ture studies are performed experimentally at different con-
ditions than our study, including the ultrasonic frequency,
acoustic intensity, and acoustic amplitude.
Blake cavitation threshold
The actual limit separating dissolved and inertial cavitation
bubbles are known as the Blake threshold. The Blake
threshold PBL is characterized by the initial bubble radius R0 ,
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0 15237

Fig. 15 e Input power vs. Cavitation Energy vs. Energy Conversion Efficiency (a), Input power vs. Cavitation Energy and
consumption (b).

static pressure P0 , and the water vapor pressure Pv as reported threshold exceeds 0.64 MPa, regarded as the severe cavitation
by Zhang et al. [13]: threshold, which refers to the generated intense bubble clouds
 3=2  1=2 as noted in Fig. 13. In addition, when the initial radius of the
2 2s 2s nuclei is higher than 10 mm, this is corresponding to the mini-
PBL ¼ P0  PV þ pffiffiffi P0  PV þ (10)
3 3 R0 R0 mum cavitation threshold which is around 0.1 MPa. Thus, the
where s is the surface tension of water is 0.0725 N/m, the vapor predicted cavitation volume is quantified and compared to the
pressure is pv ¼ 2:33 kPa. The Blake cavitation threshold is then total volume of the sonoreactors for different sonoreactor
quantified according to equation (10), and the results of the configurations. Fig. 13 presents a comparison between all
cavitation percentage are presented in Fig. 13. The radius of sonoreactors in the present study.
nuclei is assumed to be in the range between 0.1 mm to 10 mm. The figure presents the cavitation volume percentage, and
At the minimum radius of nuclei, R0 ¼ 0:1 mm, the cavitation the severe cavitation volume percentage as compared to the
15238 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
total volume of the sonoreactor corresponds to 100%. The rest
of the volume is water with no predicted cavitation. The
cavitation volume is identified using COMSOL Multiphysics by
allocating the nodes whose pressure higher than the cavita-
tion threshold. In contrast, the severe cavitation volume is
quantified by allocating the number of nodes whose pressure
higher than 0.64 MPa, which corresponds to 0.1 mm. It can be
concluded that, under the cylindrical sonoreactor, the higher
the number of probes, the higher the cavitation volume and
the severe cavitation volume. As seen, the minimum cavita-
tion percentage is recorded by CLY-1P because it has only 1
probe, which made the probability of generating cavitation is
the lowest due to the high volume of the sonoreactor as
compared to the probe volume. On the other hand, between
all the cylindrical sonoreactors, the maximum cavitation
volume generated by CLY-4P and CLY-5P is approximately
55%. However, the maximum cavitation recorded overall is
observed from SQR-3P. Noting that, the case SQR-3P has a
cavitation volume of 61.1% and a severe cavitation volume of
36.9. In contrast, the maximum severe cavitation is recorded
for HEX-3P with 44.1% of its volume covered with bubble radii
lower than 0.1 mm. This study is crucial as it gives an insight
into the cavitation threshold and the predicted cavitation
volume required for efficient hydrogen production.
Hydrogen yield and energy consumption
The cavitation energy and cavitation yield are used to esti-
mate the hydrogen quantity emitted by the sonohydrogen
sonoreactor. The cavitation yield is defined as the number of
reaction products per unit of irradiation time in seconds per
unit of energy consumption in kWh. The cavitation energy
indicates the energy dissipated from the supplied input/elec-
trical energy, and it can be described in terms of the input
power as per equations (10) and (11). The acoustic cavitation
has a threshold. The cavitation threshold can be defined as the
power at which the cavitation bubble forms in a liquid at
which the minimum power amplitude is required from a
minimum cavitation density. Once the cavitation threshold is
reached, low-density cavitation is presented in pressure an-
tinodes. This cavitation is generally not uniform and generally
not useful for any application, either sonohydrogen or ultra-
sonic cleaning. Above the cavitation threshold, cavitation
density increases as power increased. Better cavitation bubble
density is observed and may be seen with increased power
within a certain range. The cavitation energy converted from
the input power is displayed in Fig. 14, along with the energy
conversion efficiency. The conversion efficiency decreases
while increasing the input power due to an increase in the
surroundings’ energy loss. The onset of surface cavitation
occurs when the acoustic cavitation bubbles start to form on
the ultrasonic transducer’s surface at a particular power
amplitude rather than developing within the liquid’s body.
Fig. 14 (right) presents the input power vs. the cavitation en-
ergy and the sonochemical yield. It shows how much input
power is converted to cavitation energy. It also shows that
after 300 W input power, the cavitation energy decreases
because the onset of surface cavitation is reached. Suppose
the acoustic power amplitude exceeded the onset of surface
cavitation. In that case, the forces required to couple
vibrational energy to the liquid are exceeded, and additional
power would not increase cavitation density. However, it
would accelerate the transducer probe tip’s erosion that may
result in premature transducer failure. In multiple sonotrodes,
the analysis is re-conducted using the same methodology to
quantify the amount of hydrogen. The cavitation energy
converted from the input power is displayed in Fig. 15, along
with the energy conversion efficiency. The conversion effi-
ciency decreases while increasing the input power due to an
increase in the surroundings’ energy loss. As seen, the cavi-
tation energy has a linear relation with the input energy with
no drop in the conversion efficiency when using up to 5
sonotrodes with a capacity rate of 180 W or 36 W each. The
trend for energy consumption is also increasing while
increasing the number of sonotrodes and power, as seen in
Fig. 15. The energy conversion can be identified as the ratio
between the cavitation energy output over the input power,
measuring how much input energy is converted to the cavi-
tation energy. This process’s losses are due to the heat dissi-
pated into the water inside the sonoreactor and energy losses
through the sonoreactor body.
Conclusions
Since the linear analysis is computationally inexpensive, it
can simulate the reactors’ different design parameters.
Several sets of numerical investigations are performed. The
first set is a comprehensive analysis of the effects of various
sonoreactor geometries to come up with a benchmark. The
effects of the geometrical parameters such as the sonoreactor
geometry, wall boundary conditions, and upscaling of the
reactor are investigated, analyzed, and reported. The key
findings of this study are given below:

The sonoreactor geometry has the power for creating
multiple maximum and minimum pressure zones required
for the cavitation effects of taking place, and it also has the
power to kill these pressure zones

The boundary conditions have a significant effect on the
wave pattern.

The sound-hard boundaries help in creating high and low-
pressure regions in the reactor with more significant
cavitation zones

GEO.2, which has a concave base, achieves the optimum
sonoreactor performance because it reached the lowest
minimum pressure of 5.65  105.

In contrast, GEO.4, which has a wide base, recorded the
worst performance with a minimum pressure of 2.72 
105.

Changing the diameter will alter the length of the sound
hard boundaries and will change the phase of the reflected
waves.

In that case, the forces required to couple vibrational en-
ergy to the liquid are exceeded, and additional power
would not increase cavitation density.

An eigenfrequency test is made to check the excitable
modes and frequencies to trigger resonance condition; the
resonance condition will enhance the sonohydrogen
process.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0

CLY-4P and CLY-5P generate the maximum cavitation
volume of around 55%

The case of SQR-3P has its sonoreactor volume full by the
cavitation volume 61.1% and severe cavitation volume of
36.9.

The maximum hydrogen production is recorded at 300 W
in the amount of 2.5  109 mol/J when increasing the input
power to one sonotrode mounted in a typical sonoreactor
cylindrical shape at an energy conversion efficiency of 23%.

For the study on the multiple sonoreactor, the amount of
hydrogen produced is 308  109 mol/kWh at 180 W with an
energy conversion efficiency of 33%.
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
Acknowledgements
The authors would like to express gratitude to the Govern-
ment of Ontario, Canada, for providing the Ontario Trillium
Scholarship (OTS).
references
[1] Rashwan SS, Dincer I, Mohany A. Sonication to
hydrogenization: sono-hydro-gen. Int J Energy Res
2019;43:1045e8. https://doi.org/10.1002/er.4339.
[2] Rashwan SS, Dincer I, Mohany A, Pollet BG. The sono-hydro-
gen process (ultrasound induced hydrogen production):
challenges and opportunities. Int J Hydrogen Energy
2019;44(29):14500e26. https://doi.org/10.1016/
j.ijhydene.2019.04.115.
[3] Calado Galva ~ o de Melo S, da Silva MEP, da Silva MEB, et al.
Ultrasound as a tool for reducing energy consumption in
electrocatalytic hydrogenation of aromatic ketones using
graphite as catalyst support. Int J Hydrogen Energy
2020;45(43):22855e72. https://doi.org/10.1016/
j.ijhydene.2020.06.157.
[4] Escobar-Alarco  n L, Iturbe-Garcı́a JL, Gonza
 lez-Zavala F, et al.
Hydrogen production by laser irradiation of metals in water
under an ultrasonic field: a novel approach. Int J Hydrogen
Energy 2019;44(3):1579e85. https://doi.org/10.1016/
j.ijhydene.2018.11.158.
[5] Manickam S, Ashokkumar M. Cavitation: a novel energy-
efficient technique for the generation of nanomaterials. 2014.
[6] Koruk H, Serenli M, Sanliturk KY. Application of ultrasonic
vibrations for minimization of the accumulation of limescale
in steam irons. Appl Acoust 2018;135:22e8. https://doi.org/
10.1016/j.apacoust.2018.01.024.
[7] Qu Z, Yang J, Wu L, et al. Methodology for removing fouling
within liquid-filled pipelines based on ultrasonic guided
waves cavitation effect. Appl Acoust 2020;157:107e18.
https://doi.org/10.1016/j.apacoust.2019.107018.
[8] Verhaagen B, Zanderink T, Fernandez Rivas D. Ultrasonic
cleaning of 3D printed objects and cleaning challenge
devices. Appl Acoust 2016;103:172e81. https://doi.org/
10.1016/j.apacoust.2015.06.010.
15239
[9] Xu H, Tu J, Niu F, Yang P. Cavitation dose in an ultrasonic
cleaner and its dependence on experimental parameters.
Appl Acoust 2016;101:179e84. https://doi.org/10.1016/
j.apacoust.2015.08.020.
[10] Arafa N, Mohany A. Developments and recent patents on
Thermoacoustic devices. Recent Pat Mech Eng
2012;5(2):79e88. https://doi.org/10.2174/
1874477X11205020079.
[11] Kim JO, Choi S, Kim JH. Vibroacoustic characteristics of
ultrasonic cleaners. Appl Acoust 1999;58(2):211e28. https://
doi.org/10.1016/S0003-682X(98)00039-5.
[12] Koo Sin Lin C, Liebler M, Jenderka KV. Investigations on the
sound field between waveguide and counterpart induced by
high-intensity focused ultrasound in thin polymer films.
Appl Acoust 2018;142:114e22. https://doi.org/10.1016/
j.apacoust.2018.08.011.
[13] Zhang Z, Gao T, Liu X, et al. Influence of sound directions on
acoustic field characteristics within a rectangle-shaped
sonoreactor: numerical simulation and experimental study.
Ultrason Sonochem 2018;42:787e94. https://doi.org/10.1016/
j.ultsonch.2017.12.024.
[14] Son Y, Lim M, Ashokkumar M, Khim J. Geometric
optimization of sonoreactors for the enhancement of
sonochemical activity. J Phys Chem C 2011;115(10):4096e103.
https://doi.org/10.1021/jp110319y.
[15] Pflieger R, Ndiaye AA, Chave T, Nikitenko SI. Influence of
ultrasonic frequency on Swan band sonoluminescence and
sonochemical activity in aqueous tert-butyl alcohol
solutions. J Phys Chem B 2015;119(1):284e90. https://doi.org/
10.1021/jp509898p.
[16] Shah YT, Pandit AB, Moholkar VS. Cavitation reactors. 1999.
[17] Eskin GI, Eskin DG. Ultrasonic treatment of light alloy melts.
2014.
[18] Graff KF. Power ultrasonic transducers: principles and
design. In: Power ultrasonics: applications of high-intensity
ultrasound; 2014.
[19] Kumar A, Gogate PR, Pandit AB. Mapping the efficacy of new
designs for large scale sonochemical reactors. Ultrason
Sonochem 2007;14(5):538e44. https://doi.org/10.1016/
j.ultsonch.2006.11.005.
[20] Rahimi M, Dehbani M, Abolhasani M. Experimental study on
the effects of acoustic streaming of high frequency
ultrasonic waves on convective heat transfer: effects of
transducer position and wave interference. Int Commun
Heat Mass Tran 2012;39(5):720e5. https://doi.org/10.1016/
j.icheatmasstransfer.2012.03.013.
[21] Tiong TJ, Chu JK, Lim LY, et al. A computational and
experimental study on acoustic pressure for ultrasonically
formed oil-in-water emulsion. Ultrason Sonochem
2019;56:46e54. https://doi.org/10.1016/j.ultsonch.2019.03.026.
[22] Niazi S, Hashemabadi SH, Noroozi S. Numerical simulation
of operational parameters and sonoreactor configurations
for the highest possibility of acoustic cavitation in crude oil.
Chem Eng Commun 2014;201:1340e59.
[23] Son Y, Lim M, Khim J. Investigation of acoustic cavitation
energy in a large-scale sonoreactor. Ultrason Sonochem
2009;16:552e6. https://doi.org/10.1016/j.ultsonch.2008.12.004.
[24] Zapata-Ferna  ndez JR, Gochi-Ponce Y, Salazar-Gastelum MI,
et al. Ultrasonic-assisted galvanic displacement synthesis of
PtePd/MWCNT for enhanced oxygen reduction reaction:
effect of Pt concentration. Int J Hydrogen Energy
2017;42(15):9806e15. https://doi.org/10.1016/
j.ijhydene.2017.02.057.
[25] Park HU, Lee E, Kwon YU. TiC supported Pt-based
nanoparticles: facile sonochemical synthesis and
electrocatalytic properties for methanol oxidation reaction.
Int J Hydrogen Energy 2017;42(31):19885e93. https://doi.org/
10.1016/j.ijhydene.2017.06.144.
15240 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 6 ( 2 0 2 1 ) 1 5 2 1 9 e1 5 2 4 0
[26] Wang M, Wang Z, Gong X, Guo Z. The intensification
technologies to water electrolysis for hydrogen production -
a review. Renew Sustain Energy Rev 2014;29:573e88.
[27] Achaoui Y, Metwally K, Fouan D, et al. Tunable microbubble
generator using electrolysis and ultrasound. AIP Adv
2017;7(1):015011. https://doi.org/10.1063/1.4973720.
[28] Rashwan SS, Dincer I, Mohany A. A journey of wastewater to
clean hydrogen: a perspective. Int J Energy Res 2020;1e8.
https://doi.org/10.1002/er.6279.
[29] Hao X, Suo H, Peng H, et al. Simulation and exploration of
cavitation process during microalgae oil extracting with
ultrasonic-assisted for hydrogen production. Int J
Hydrogen Energy 2021;46(3):2890e8. https://doi.org/
10.1016/j.ijhydene.2020.06.045.
[30] Rashwan S S, Dincer I, Mohany A. A unique study on the
effect of dissolved gases and bubble temperatures on the
ultrasonic hydrogen (sonohydrogen) production. Int J
Hydrogen Energy 2020;45(41):20808e19.
[31] Rashwan SS, Dincer I, Mohany A. An investigation of
ultrasonic based hydrogen production. Energy
2020;2015:118006. https://doi.org/10.1016/
j.energy.2020.118006. 118006.
[32] Raschitor A, Llanos J, Can ~ izares P, Rodrigo MA. Improved
electrolysis of colloid-polluted wastes using ultrasounds and
electrocoagulation. Separ Purif Technol 2020;231:115926.
https://doi.org/10.1016/j.seppur.2019.115926.
[33] Fu Q, Wang D, Li X, et al. Towards hydrogen production from
waste activated sludge: principles, challenges and
perspectives. Renew Sustain Energy Rev 2021;135:110283.
[34] Wang D, Ji C, Wang S, et al. Experimental investigation on
near wall ignited lean methane/hydrogen/air flame. Energy
2019;168:1094e103. https://doi.org/10.1016/
j.energy.2018.11.115.
[35] Berlan J, Mason TJ. Sonochemistry: from research laboratories
to industrial plants. Ultrasonics 1992;30(4):203e12. https://
doi.org/10.1016/0041-624X(92)90078-Z.
[36] Rashwan SS, Dincer I, Mohany A. Investigation of acoustic
and geometric effects on the sonoreactor performance.
Ultrason Sonochem 2020;68:105174. https://doi.org/10.1016/
j.ultsonch.2020.105174.
[37] Gogate PR, Tatake PA, Kanthale PM, Pandit AB. Mapping of
sonochemical reactors: review, analysis, and experimental
verification. AIChE J 2002;48:1542e60.
[38] Capelo-Martı́nez JL. Ultrasound in chemistry: analytical
applications. 2009.
[39] Rubinetti D, Weiss DA, Wahlen A, Müller J. Numerical
modeling and validation concept for acoustic streaming
induced by ultrasonic treatment. In: 2016 COMSOL Conf
Munich; 2016. p. 12e4.
[40] Niazi S, Hashemabadi SH, Razi MM. CFD simulation of
acoustic cavitation in a crude oil upgrading sonoreactor and
prediction of collapse temperature and pressure of a
cavitation bubble. Chem Eng Res Des 2014;92(1):166e73.
[41] Trujillo FJ, Knoerzer K. A computational modeling approach
of the jet-like acoustic streaming and heat generation
induced by low frequency high power ultrasonic horn
reactors. Ultrason Sonochem 2011;18(6):1263e73. https://
doi.org/10.1016/j.ultsonch.2011.04.004.
[42] Moudjed B, Botton V, Henry D, et al. Near-field acoustic
streaming jet. Phys Rev E - Stat Nonlinear Soft Matter Phys
2015;91(3):033011. https://doi.org/10.1103/
PhysRevE.91.033011.
[43] Comsol Multiphysics software ultrasonic flow meter with
piezoelectric transducers: coupling between FEM and DG.
[44] Sutkar VS, Gogate PR, Csoka L. Theoretical prediction of
cavitational activity distribution in sonochemical reactors.
Chem Eng J 2010;158:290e5. https://doi.org/10.1016/
j.cej.2010.01.049.
[45] Dahlem O, Reisse J, Halloin V. The radially vibrating horn: a
scaling-up possibility for sonochemical reactions. Chem Eng
Sci 1999;54(13e14):2829e38. https://doi.org/10.1016/S0009-
2509(98)00356-X.
[46] Schenker MC, Pourquie  MJBM, Eskin DG, Boersma BJ. PIV
quantification of the flow induced by an ultrasonic horn and
numerical modeling of the flow and related processing
times. Ultrason Sonochem 2013;20(1):502e9. https://doi.org/
10.1016/j.ultsonch.2012.04.014.
[47] Slama RBH, Gilles B, Chiekh M Ben, Bera JC. Characterization
of focused-ultrasound-induced acoustic streaming. Exp
Therm Fluid Sci 2019;101:37e47. https://doi.org/10.1016/
j.expthermflusci.2018.10.001.
[48] Klı́ma J, Frias-Ferrer A, Gonza  lez-Garcı́a J, et al.
Optimisation of 20 kHz sonoreactor geometry on the basis
of numerical simulation of local ultrasonic intensity and
qualitative comparison with experimental results. Ultrason
Sonochem 2007;14:19e28. https://doi.org/10.1016/
j.ultsonch.2006.01.001.
[49] Gallego-Jua  rez JA, Graff KF. Power ultrasonics: applications
of high-intensity ultrasound. 2014.
[50] Son Y. Simple design strategy for bath-type high-frequency
sonoreactors. Chem Eng J 2017;328:654e64. https://doi.org/
10.1016/j.cej.2017.07.012.
[51] Islam MR, Shaaban M, Mohany A. Vortex dynamics and
acoustic sources in the wake of finned cylinders during
resonance excitation. Phys Fluids 2020;32(7):075117. https://
doi.org/10.1063/5.0016076.
[52] Mohany A. Flow-excited acoustic resonance of isolated
cylinders in cross-flow. Nucl Rev 2014;1:45e55.