Ultrasonics Sonochemistry 24 (2015) 184–192

Contents lists available at ScienceDirect

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

Mathematical modeling of a single stage ultrasonically assisted

distillation process
Taha Mahdi a,b,d, Arshad Ahmad a,b,⇑, Adnan Ripin a,b, Tuan Amran Tuan Abdullah a,b, Mohamed M. Nasef a,c,
Mohamad W. Ali a,b
a
Institute of Hydrogen Economy, Universiti Teknologi Malaysia, 81310 Johor Bahru, Malaysia
b
Faculty of Chemical Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Malaysia
c
Malaysia Japan International Institute of Technology, Universiti Teknologi Malaysia, 54100 Kuala Lumpur, Malaysia
d
Midland Refineries Company, Ministry of Oil, Daura, Baghdad, Iraq

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

Article history: The ability of sonication phenomena in facilitating separation of azeotropic mixtures presents a
Received 4 February 2014 promising approach for the development of more intensified and efficient distillation systems than
Received in revised form 31 October 2014 conventional ones. To expedite the much-needed development, a mathematical model of the system
Accepted 5 November 2014
based on conservation principles, vapor–liquid equilibrium and sonochemistry was developed in this
Available online 13 November 2014
study. The model that was founded on a single stage vapor–liquid equilibrium system and enhanced with
ultrasonic waves was coded using MATLAB simulator and validated with experimental data for ethanol–
Keywords:
ethyl acetate mixture. The effects of both ultrasonic frequency and intensity on the relative volatility and
Mathematical modeling
Ultrasound
azeotropic point were examined, and the optimal conditions were obtained using genetic algorithm. The
Azeotrope experimental data validated the model with a reasonable accuracy. The results of this study revealed that
Ethyl acetate/ethanol the azeotropic point of the mixture can be totally eliminated with the right combination of sonication
Vapor–liquid equilibrium parameters and this can be utilized in facilitating design efforts towards establishing a workable
ultrasonically intensified distillation system.
Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction rapidly formed and disappearing hot-spots under nearly ambient

In process industries, distillation is still one of the preferred
processes despite its difficulties in separating mixtures with very
close boiling points and those that form azeotrope. To overcome
this limitation, variety of frontier technologies have been explored
[1]. For example, dividing-wall column has been introduced to
separate more components in a single distillation unit, thereby
offering energy savings along with substantial capital and space
reduction [2,3]. Furthermore, it is also possible to implement this
separation technique with azeotropic distillation [4], or extractive
distillation [5], so that it can be integrated into single column
configuration. However, this process is marred by high pressure
drop and temperature difference caused by the increase in the
boiling point [6].
Another new approach is to intensify the process by adding
ultrasonic equipment to the distillation system. Using cavitation
as a source of energy input for chemical processes to generate
⇑ Corresponding author at: Institute of Hydrogen Economy, Universiti Teknologi
Malaysia, 81310 Johor Bahru, Malaysia.
E-mail address: arshad@cheme.utm.my (A. Ahmad).
conditions offer potential improvements to conventional
distillation. The whole process of generation, growth and collapse
of cavities occurs rapidly, of the order of few microseconds, and
this phenomenon alters physical properties of the mixtures and
enhances the mass [7] and heat [8] transfer, thus offering further
exploitation to intensify vapor–liquid separation.
As a foundation for the development of ultrasonic distillation
process, studies on vapor–liquid equilibrium (VLE) under ultrason-
ically intensified environment have been carried out. These include
experimental works on the VLE of methanol–water [9], MTBE-
methanol [10] and cyclohexane-benzene [11]. In all cases, positive
changes on the VLE characteristics were observed and sonication
effects have been proven to alter the relative volatility of azeotro-
pic mixtures, thus enabling higher purity separation in a single
distillation column.
To facilitate further development, a mathematical model
describing the process is needed so that comprehensive design
study of the ultrasonic distillation system can be carried out. In this
paper, this issue is addressed. A mathematical model that
represents a single stage vapor–liquid equilibrium system with
intensification using ultrasonic waves is derived and validated.

http://dx.doi.org/10.1016/j.ultsonch.2014.11.005
1350-4177/Ó 2014 Elsevier B.V. All rights reserved.
 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192 185

This is followed by a simulation study to investigate the process since the cycle of expansion and compression is slower, larger
characteristics and determination of optimal operating condition bubbles are produced. Moreover, when the population of bubbles
for separation this system. is high, which typically happen at high sonication intensity, some
of the bubbles coalesce to form larger bubbles [24]. As a conse-
quence, the bubbles have larger surface area of contact with the
2. Mechanism of ultrasonic separation
surrounding liquid, thus allowing more light molecules to diffuse
into the bubble, thus increasing its vapor pressure, which in turn
2.1. Mechanisms of bubble collapse
further increase the bubble size. When the bubble is larger than
the resonance size, it will be pushed toward the nodes by primary
Acoustic cavitation is a phenomenon by which ultrasonic waves
Bjerknes forces and becoming inactive [26,27].
induce bubbles formation, growth and collapse [12]. The collapse
The second condition is related to the operating temperature.
normally takes place when the bubble reaches critical size referred
An increase in the bulk liquid temperature leads to a reduction
to as the resonance size. Depending on the operating condition, the
in bubbles’ activities due to two reasons [25–27]. Firstly, dissolved
growth-collapse process may end up in two possible scenarios.
gases in the liquid evaporate to the surface at high temperature,
Firstly, if they are smaller than the resonance size, bubbles tend
thus reducing the bubble population. Secondly, as temperature of
to migrate from the minimum pressure, also known as pressure
the bulk liquid increases, the liquid vapor pressure inside a cavita-
node to the maximum pressure referred to as antinode. This is
tion bubble is increased, leading to an increase of the bubble size.
driven by primary Bjerknes forces, and will lead to a condition
When the size exceeds the resonance size and become ‘‘inactive’’
whereby bubbles are collapsing inside the liquid and generating
as mentioned previously.
high temperature [13]. This causes the formation of radicals and
The third condition is related to decomposition of components
highly reactive intermediates within the bubbles during the
in the presence of hydrocarbons [28,29]. When molar heat of
collapse. For this reason, they are called ‘‘active’’ cavitation
hydrocarbons is much larger than that of gases inside the bubble,
bubbles. This condition facilitates various chemical pathways, thus
the temperature generated during the bubble collapse decreases
enhancing sonochemical reactions [14]. Recently, numerous
monotonously to an extent that it is unable to dissociate hydrocar-
papers have reported enhancement effect of ultrasound on
bons inside the bubble, thus making it inactive. This is contrary to
biodiesel synthesis with basic strength of catalyst [15–17]. They
the observation by Yasui et al. [30,31] when the liquid environ-
established the mechanism of this enhancement by discrimination
ments were aqueous. In their work involving aqueous methanol
of the physical and chemical effects of cavitation bubbles in the
environment, they reported that as the bubbles collapsed metha-
system on transesterification of oil with alcohol using a catalyst.
nol molecules were dissociated inside a bubble. Similarly, when
Secondly, if they are larger than the resonance size, they will be
pure water is used, water vapor dissociated inside the heated bub-
forced to the node to become ‘‘inactive’’. These bubbles eventually
ble and chemical species such as OH radical and H atom are created
float out of the liquid due to buoyancy forces and collapse at the
inside the bubble during the violent collapse of bubbles [31].
liquid surface [18]. Similar observation is reported in a study
Based on the above arguments, ‘‘inactive’’ conditions are estab-
involving ethanol/water mixture, where the bubble travel through
lished when the operating conditions are at low frequency, high
the liquid mixture and collapse in the fountain jet formed at the
temperature, and hydrocarbons. It is also important to note that
liquid surface releasing the alcohol vapor in the bubble [19]. These
the ultrasonic wave generates micro-point vacuum condition
phenomena have significant impact on mechanical and physical
within the liquid during bubbles formation. In this condition,
processes such as cleaning and vapor–liquid separation processes.
azeotrope of the vapor components inside the bubbles is altered,
However, certain combination of operating conditions may also
resulting in changes in vapor liquid equilibrium. This is confirmed
create exceptions. For example, a study by Matula [20] revealed
by a various studies that proved the breaking of azeotrope under
that at 20 kHz and 1.8 bar, the bubble is repelled from the antinode
vacuum pressure condition [32,33]. To understanding the mecha-
even if the size is smaller than the resonance size.
nism of the enhancement separation of the system in this process,
Other researchers [21,22] also proposed an alternative
these bubbles eventually float out of the liquid due to buoyancy
mechanism based on capillary wave. In this hypothesis, a liquid
forces and collapse at the liquid surface in the fountain jet
is parametrically excited by ultrasound waves such that capillary
releasing the vapor in the bubble to the vapor phase. Thus, the
waves are formed on the surface. As the amplitude of these waves
mole fractions of the vapor inside the bubbles are considering
increase, the capillary become instable and small liquid droplets
equal to those in the vapor phase. These are the scenarios consid-
pinch off from the crests (peaks) of the capillary wave causing
ered in this study.
atomization. Oscillation and collapses of the cavitation bubbles
enhance the capillary wave perturbations and thus facilitate the
pinch-off of droplets mist formation [23]. However, the visible mist
3. Mathematical modeling
was produce by mixture droplets which is depend on the physical
properties of a mixture and operation conditions. If the boiling
The mathematical model developed here is focusing on the use
points of the components of a mixture are close, the percentages
of ultrasound in facilitating a distillation process. To simplify
of these components in the droplet mist are also close; and vice
model development efforts, a number of assumptions on the phys-
versa [21,22]. Therefore, in the present system this theory is futile
ical characteristics of the bubble are made. The cavitation bubble is
to break the azeotrope.
assumed to be spherically symmetric and is initially composed of
mixture of gas (air) and liquid vapor. The surrounding liquid is
2.2. Factor influencing the mechanisms of bubble collapse assumed incompressible, with constant and uniform dynamic
viscosity, and is at steady-state condition. The non-equilibrium
Many literatures have examined the effect of physical proper- condition is during the growth of a bubble which is very short time
ties of mixtures and operating conditions on the activity of acoustic (microsecond). During this time, the bubble is unstable due to the
cavitation bubbles [24–30]. In summary, there are three important amount of material that gets into the bubble (during expansion) is
conditions that may cause cavitation bubbles to lose their activi- larger than what comes out of the bubble (during compression).
ties. The first condition is concerned about the influence of ultra- Therefore the final number of molecules inside the bubble will
sonic frequency on the cavitation bubbles. At lower frequencies, be calculated at equilibrium condition. The validity of the model
186 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192

2  2
is also limited to certain conditions such as low frequency, high d R 3 dR PðRÞ  P1
temperature of bulk liquid and this system is of hydrocarbons R þ ¼ ð8Þ
dt
2 2 dt q
mixture.
2
Here dR
dt
is the velocity of the cavity wall of radius R, ddt2R is the wall
3.1. Vapor–liquid equilibrium acceleration, q is the density of the liquid, P1 is the pressure in the
liquid at infinity (far away from the bubble) and P(R) is the
One of the major assumptions made in the present investigation pressure in the liquid at the bubble boundary. With the inclusion
is that that bubbles occur inside the liquid, whose size is uniform, of surface tension r and viscosity l effects, the equation of bubble
and they filled with a vapor in equilibrium with the bulk solution dynamics becomes:
and the bubbles enriched with ethyl acetate due to surface excess, 2  2   
and can be represented by Rault’s law. Thus, the relationships
d R 3 dR 1 2r 4l dR
R þ ¼ Pi  P1   ð9Þ
between vapor compositions in the bubble and the compositions dt
2 2 dt q R R dt
in the bulk liquid, is denoted by: Here, Pi is the pressure in the bubble, and this equation is
P  yi ¼ ci  xi  P oi ð1Þ frequently referred to as the Rayleigh–Plesset equation [36] and
serves as a governing equation that represents the dynamics of
Here, P is the pressure of the system, yi and xi are mole fractions spherical bubbles in an infinite body of liquid.
in vapor and liquid phase respectively, ci is the activity coefficient
and P oi is the vapor pressure of i component. Under ultrasonic influ- 3.3. Bubble expansion
ence, it is the vapor composition that will be seriously impacted,
and when this can be estimated, the liquid composition can be Consider an initial bubble containing a very tiny mass mg of
computed by using Eq. (1). The activity coefficient can be deter- noncondensable gas (air) and liquid vapor at ambient temperature
mined using suitable model depending on the type of liquid system T, in a system shown in Fig. 1.
involved. For the ethanol/ethyl acetate mixture considered here, The pressure of the noncondensible gas in the bubble can be
NRTL model is chosen. Other information needed to determine estimated using the ideal gas law, and is given by [37]:
the VLE compositions are obtained by looking into the details of
mg RT
bubble dynamics and sonication phenomena. Pg ¼ ð10Þ
NRTL model was chosen because of its suitability for liquid– M g 43 pR3o
liquid systems containing alcohols and non-polar hydrocarbon liq-
where Mg is the molecular weight the air and R universal gas
uids such as ETAC [34]. The activity coefficients for the NRTL model
constant. The partial pressure of compositions can be estimated
of binary mixture were determined using the following equations:
by assuming that it is equal to the sum of all vapor pressure of
! P o
 2 individual component i (Pv ¼ P i ) [38].
G21 G12 s12
ln c1 ¼ x22 s21 þ ð2Þ Note that inside the bubble, there are the noncondensible gas,
x1 þ x2 G21 ðx2 þ x1 G12 Þ2 the partial pressure of which is Pg and the vapor with partial
pressure Pv. The combination of these two partial pressures gives
 2 ! the total pressure inside the bubble (i.e., Pb = Pg + Pv). At the bubble
G12 G21 s21 interface, the liquid pressure is lower than the pressure inside the
ln c2 ¼ x21 s12 þ ð3Þ
x2 þ x1 G12 ðx1 þ x2 G21 Þ2 bubble because of surface tension of the liquid phase which plays
an important role in bubble dynamics as shown in Fig. 1 [39]. This
G12 ¼ ea12 s12 ð4Þ phenomenon is known as Laplace pressure, and is given by the
expression:
G21 ¼ ea21 s21 ð5Þ 2r
Pg þ Pv ¼ Po þ ð11Þ
The parameters a12 and a21 are non-randomness parameters, R
and in usual cases, a12 is set equal to a21. In practice, the non- Putting Eq. (10) into Eq. (11) yields:
randomness parameter a12 is set to 0.3 for the case of liquid–liquid
mg RT 2r
nonideality and for non-aqueous systems [34]. Therefore, this Po ¼ Pv o þ  ð12Þ
value was adopted for this case study: M g 43 pR3o Ro

g 12  g 22
s12 ¼ ð6Þ
RT
g 21  g 11
s21 ¼ ð7Þ
RT Po
Here R is the gas constant and T the absolute temperature, and
g21 and g11 are energies of interaction between a 1–2 and 1–1 pair
of molecules, respectively. Again in common practice, the value of
g12 is set equal to g21. Ro
Rmax
3.2. Rayleigh–Plesset equation
Pb=Pg+Pv
The study on bubble dynamics can be traced back to the early
works of Rayleigh [35], in which, the collapse of an empty spherical
bubble from an initial radius, Ro, to a new radius R at time t was
considered. By equating the work accomplished by the hydrostatic
pressure to the kinetic energy of the fluid surrounding the bubble,
the motion of the bubble wall can be represented by Eq. (8): Fig. 1. Micro bubble.
 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192
Eq. (12) gives the initial bubble radius Ro in mechanical equilibrium
for a given liquid ambient pressure Po. A relationship between
ultrasound frequency and the initial bubble radius Ro is given by
Eq. (13):
 1=2
1 3Po
Ro ¼
2p f ql
3.4. Sonication effect
In the system studied here, the liquid medium is subjected to
sonication effect from waves generated by the ultrasonic trans-
ducer. When a wave with pressure amplitude PA and frequency f
passes through a cavitating medium, the pressure P1 in the liquid
at any time t is given as [40]:
P1 ðtÞ ¼ P o  PA sinð2pftÞ
where Po is the ambient pressure, and the pressure amplitude, PA of
the sound field that is a function of the acoustic intensity that is
given by [41]:
pffiffiffiffiffiffiffiffiffiffiffi
PA ¼ 2IqC
where I represents intensity of the ultrasound and C is the velocity
of sound through the liquid. By substituting Eq. (15) into Eq. (14),
and subsequently Eq. (14) into the Rayleigh–Plesset Equation (Eq.
(9)), the effect of ultrasonic intensity and frequency on the bubble
size can be estimated. For the purpose of this study, the condition
of interest is at equilibrium, which is assumed to be established
when the bubble size is largest with bubble radius Rmax and is about
to collapse at the liquid surface.
3.5. Maximum bubble size
During the rarefaction phase of the acoustic cycle, the radius of
a bubble that was initially at Ro will expand to a maximum radius,
Rmax. Solving the Rayleigh–Plesset equation for the maximum R
gives Eq. (16) below [24]:
 1=2  1=3
2 2 2ðP A  Po Þ
Rmax ¼ ðPA  P o Þ 1þ
3pf qPA 3Po
where PA is the pressure amplitude, f is the acoustic frequency and q
is the density of the liquid.
The relationship between the pressure and bubble volume
during compression or expansion can be described analytically.
The expansion process is assumed isothermal and the bubble size
is small enough such that the pressure and temperature inside
the bubble can be assumed uniform. The instantaneous vapor pres-
sure of components Pi is then related to the initial vapor pressure of
components Pio by the following relation [42]:
 3c
Ro
Pi ¼ Pio
Rmax
where c is the ratio of the heat capacities of the vapor, the
isothermal case is accounted for by setting parameter c to 1. Ther-
modynamic process described by the above equation is called a
Polytropic process.
3.6. Ideal gas law
Using equation of state, the number of moles of vapor
components can be computed for a defined bubble size (i.e.
volume), temperature and pressure. In this case, the partial
pressure of components inside the bubble is assumed equal to their
saturation vapor pressures. Since the partial pressure changes as
the bubble is expanding, the number of moles of vapor
187
components would also vary during this isothermal phase of
bubble dynamics. Nevertheless, the condition of interests is at
the maximum bubble size, and by using ideal gas law, the number
of moles of components vapor (ni) inside the bubble at Rmax can be
calculated by Eq. (18) [43]:
ð13Þ
4pPi R3max
ni ¼ ð18Þ
3RT
Knowing the number of moles ni of each species involved in the
system, the mole fractions of components i (yi) inside the bubble at
Rmax can be determined using Eq. (19):
ni
yi ¼ X ð19Þ
ni
Eq. (19) gives the vapor composition at equilibrium, and having
ð14Þ this, the liquid composition can be computed by using Eq. (1), and
the VLE is therefore defined. Due to the higher vapor pressure (Pi)
of ethyl acetate compared with ethanol, we expected the value of
mole fraction in vapor phase for ethyl acetate to be larger than
for ethanol.
ð15Þ
4. Simulation study
The model discussed in the preceding sections is coded in MAT-
LAB programming environment. The system considered is the
ethanol/ethyl acetate mixture, which is known to form homoge-
nous azeotropic mixture 45 mol% of ethanol, and a difference in
boiling points of 1.2 °C, thus making it impossible to achieve higher
purity separation in a single fractional distillation column [44].
4.1. Model validation
As a measure of fitness to validate the model developed in this
study, the percentage of average absolute deviation (AAD) between
the model prediction and the experimental data is used. The AAD is
computed using Eq. (20) below:
qffiffiffiffiffiffiffiffiffiffi2
^Þ
ðyy
y2
ð16Þ AAD ¼  100 ð20Þ
n
where y and ŷ are the experimental data and model prediction
respectively, and n is number of points or observations used in
the analysis. This criterion is applied to conditions described by
Mahdi et al. [45] and comparisons are made on the process behavior
observed.
4.2. VLE behavior without sonication effect
The first comparison examined is the VLE characteristics at
normal atmospheric condition without the influence of sonication.
ð17Þ Here, the model is compared with experimental data from Mahdi
et al. [45], Topphoff et al. [46] and Calvar et al. [47]. The results
are summarized by Fig. 2 below. It is noted that the model agree
well with experimental works with AAD of 2.5% [45] 2014),
1.66% [46] and 1.9% [47]. These results indicate that the model is
of reasonable accuracy in predicting the VLE curve of the system
under atmospheric condition and without the influence of
sonication.
4.3. Effect of sonication on VLE characteristics
The effect of sonication parameters on VLE of ETOH/ETAC
mixture is shown in Fig. 3. In Fig. 3(a), the influence of ultrasonic
intensity on the VLE at a frequency of 25 kHz is shown. The results
show that the equilibrium curve and azeotropic point of the
188 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192
Fig. 2. xy-Diagram of ETOH/ETAC system without sonication and compared with
literature.
mixture are shifted upwards with increasing ultrasonic intensity. It
is also important to note that by adding an ultrasonic cavitation
effect with the frequency of 25 kHz and intensity of 500 W/cm2,
the azeotropic point of ETOH/ETAC is completely eliminated. Sim-
ilar trends are observed for higher intensities. For example,
Fig. 3(b) shows the influence of ultrasonic frequency on the VLE
at an intensity of 300 W/A cm2. This observation indicates that it
is possible to realize high purity distillation separation by the assis-
tance of sonication phenomena, thus opens the opportunity for
separation in a single process unit as opposed to multiple process
units configuration offered by the azeotropic and extractive distil-
lation processes.
4.4. Sensitivity analysis
Azeotropic point of the mixture can be calculated when inter-
sect the VLE of ETOH/ETAC system curve with the standard line
(45°). Each intersection between the VLE curve of the mixture
and the standard line represents one azeotropic point for specific
operation conditions including ultrasound intensity and frequency.
Fig. 3. xy-Diagram of ETOH/ ETAC system with sonication at different intensity, (a) at frequency of 25 kHz, (b) at frequency of 70 kHz.
The model is simulated to study the effect of sonication frequency
and intensity on the azeotropic point and the relative volatility.
The results are compared with experimental works previously
carried out [45], and the AAD is determined to indicate how close
the model agrees with the experimental works.
4.4.1. Effect of intensity
Model predictions of the azeotropic points and relative
volatilities against the ultrasonic intensity at a fixed frequency of
25 kHz are shown in Fig. 4. When the intensity is increased, the
azeotrope point and relative volatility (aij ) are increased. The
relative volatility between component i and j, are computed from
individual vapor and liquid compositions using Eq. (21):
yi =xi
aij ¼ ð21Þ
yj =xj
Since it is a derived value from the compositions, it is expected
that its variations with changing ultrasonic intensity or frequency
follow the same trend of azeotrope point.
The azeotrope point and relative volatility of the model agree
well with the experimental results for the lower range on intensity
up to 300 W/cm2 with AAD of 6.8% and 6.3% respectively. Beyond
this value, i.e. at higher intensity, the deviation is rather significant,
giving overall AAD of 17.3% and 27.5% respectively. This condition
is however not the case when the system is operating at higher fre-
quency, as shown in Fig. 5. In this case, the AAD between the model
and the experimental work are 6.1% for the azeotrope point and
6.6% for the relative volatility.
Note that sonication is a rapid transient process, and during this
very short period, which is in the order of microseconds, heat and
mass transfer processes are very fast. Although the net changes in
the operating conditions of the distillation process is small, the fast
transport processes impact the interface composition of the liquid
and vapor. An increase in ultrasonic intensity leads to an increase
in the amount energy entering the liquid medium, thus producing
more micro bubbles and creating vacuum effects inside the liquid.
However, as the intensity reaches an upper limit, the tendency of
cavitation bubbles to collide becomes high. Since the time avail-
able for bubbles to collapse is insufficient, they combine to form
a bubble ‘cushion’ at the radiating face of the ultrasonic transducer,
which in turn reduces the effect of coupling sound energy to the
liquid system. This phenomenon reduces the transmission of ultra-
sonic energy into the liquid medium and produces less cavitational
 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192
Fig. 4. Comparison of model predictions and experimental observations with increasing ultrasonic intensity at frequency of 25 kHz of ETOH/ETAC system on (a) azeotropic
point, (b) relative volatility.
Fig. 5. Comparison of model predictions and experimental observations with increasing ultrasonic intensity at frequency of 70 kHz of ETOH/ETAC system on (a) azeotropic
point, (b) relative volatility.
and vacuum effect [24]. That explains the downward trend shown
by the experimental results beyond 300 W/cm2. This phenomenon
is however not accounted for in the present model, thus explains
the large deviation in Fig. 4.
At high frequency, the downward deviation observed by exper-
imental works at intensity higher than 300 W/cm2 is reduced. As
shown in Fig. 5 when the system operates at 70 kHz, the down-
ward deviation is not noticeable at intensity of 400 W/cm2. This
is consistent with earlier works [45], where it was concluded that
at high sonication frequency, the time required to create bubbles
may exceed that of the rarefaction cycle. At this condition, the
overall bubbles population is reduced due to higher cavitation rate
compared to bubbles production. Because of this, the ‘‘bubble
cushion’’ phenomenon mentioned earlier will only occur at lower
frequencies. Furthermore, to support the intended separation pro-
cess, at this value and beyond, the sonication intensity must be
increased with increasing frequency to create more bubbles, thus
balancing the effect of frequency on azeotrope point and relative
volatility. For this reason, frequencies in the range 20–50 kHz have
traditionally been used for separation purposes [48].
4.4.2. Effect of frequency
Fig. 6 shows the good agreement between model prediction and
experimental data on the effect of ultrasonic frequency on
azeotrope point and relative volatility. The average absolute devi-
ation of azeotrope point and relative volatility between the model
predictions and experimental results is 5.4% for the azeotrope
point and 4.9% for the relative volatility. The relative volatility
189
decreases with the increase of ultrasonic frequency at constant
ultrasonic intensity. When the intensity is 300 W/cm2, the highest
value of 2.36 is obtained at frequency of 25 kHz, while the lowest
value of 1.62 is obtained at frequency of 70 kHz.
Similarly, as shown in Fig. 7, good agreement between model
prediction and experimental data is established for the operation
at 300 W/cm2. The average absolute deviation of relative volatility
between the model predictions and experimental results is 10% for
the azeotrope point and 9.6% for the relative volatility.
4.4.3. Summary of the effect of ultrasonic parameters on azeotrope
point and relative volatility
Fig. 8 shows the plots of azeotropic point of ETOH/ ETAC mix-
ture against ultrasound intensity from 100 to 800 W/cm2, and fre-
quencies between 20 and 100 kHz, while Fig. 9 shows the relative
volatility. The estimation of azeotropic point in Fig. 8 depends on
the data in Fig. 3 and other ultrasound intensities and frequencies,
when intersect the VLE of the mixture curve with the standard line
(45°). The sonication intensity is found to have stronger effect on
the VLE, compared to the frequency as illustrated by the steeper
gradient along the intensity axis. The model also suggests the use
of lower frequency so that higher relative volatility can be estab-
lished, thus facilitating the intended vapor–liquid separation.
4.5. Optimization
One of the key advantages of having process model is that it
enables sensitivity analyses to be carried out thoroughly and
190 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192

Fig. 6. Comparison of model predictions and experimental observations with increasing ultrasonic frequency at intensity of 100 W/cm2 of ETOH/ETAC system on (a)
azeotropic point, (b) relative volatility.

Fig. 7. Comparison of model predictions and experimental observations with increasing ultrasonic frequency at intensity of 300 W/cm2 of ETOH/ETAC system on (a)
azeotropic point, (b) relative volatility.

Fig. 8. Azeotropic point of ETOH/ ETAC as a function of ultrasonic intensity at different frequencies.

efficiently. It also allows optimum process parameters to be deter- optimum operating conditions, genetic algorithm (GA) is used.
mined conveniently. This can be carried out using one of the many GA is a technique that emulates theories of biological evolution,
optimization algorithms available. In this work, to identify the natural selection and the survival of the fittest. It falls within the
 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192
Fig. 9. Relative volatility of ETOH/ETAC as a function of ultrasonic intensity at different frequencies.
class of random optimization and has the advantage of not being
susceptible to local minima, which affects the gradient-based
optimization techniques.
The optimization study is carried out using GA toolbox available
in MATLAB and is aimed at maximizing the relative volatility. The
result revealed that the optimum operating condition can be
established at ultrasonic intensity of 485 W/cm2 and frequency
of 26.3 kHz. At this condition, the relative volatility is 3.1.
5. Conclusion
In this work, a mathematical model of a single stage vapor–
liquid separation system has been established by assuming that a
VLE is established when the bubble is at its maximum size and is
about to collapse. The model has been validated with experimental
data to a reasonable accuracy, with a limitation of not being able to
represent the anticipated bubble-cushion phenomenon. This
phenomenon produces less cavitational and vacuum effect which
is only occurs at lower frequencies higher intensity. That is because
at this condition, ultrasound waves produces more bubbles and the
time available for bubbles to collapse is insufficient, thus the
tendency of cavitation bubbles to collide becomes high and they
combine to form a ‘bubble cushion’. It is nevertheless very useful
and serves as a good starting point for further work on the design
of ultrasonic-assisted distillation system, which is currently being
carried out.
The model is suited for distillation process that is intensified
with ultrasonic phenomena at lower intensity, if lower frequency
is used. If the model were to be used for a wider range of operation
at lower frequency, some compensation would be needed.
The model demonstrates that the azeotropic point of the
mixture can be totally eliminated with the right combination of
sonication parameters, and the maximum relative volatility of 3.1
can be obtained at intensity of 485 W/cm2 and frequency of
26.3 kHz.
Acknowledgments
The authors are grateful to Universiti Teknologi Malaysia, and
the Ministry of Education for financial supports through Prototype
Research Grant Scheme Vot No. 4L612 and infrastructures
191
provided to carry out this research. Authors are also thankful to
the Midland Refineries Company, Ministry of Oil, Republic of Iraq.
References
[1] T. Mahdi, A. Ahmad, M.M. Nasef, A. Ripin, State-of-the-Art Technologies for
Separation of Azeotropic Mixtures, Sep. Purif. Rev. 44 (4) (2015) 308–330,
http://dx.doi.org/10.1080/15422119.2014.963607.
[2] L.-Y. Sun, X.-W. Chang, C.-X. Qi, Q.-S. Li, Implementation of ethanol
dehydration using dividing-wall heterogeneous azeotropic distillation
column, Sep. Sci. Technol. 46 (2011) 1365–1375.
[3] A.A. Kiss, D.J. Suszwalak, Enhanced bioethanol dehydration by extractive and
azeotropic distillation in dividing-wall columns, Sep. Purif. Technol. 86 (2012)
70–78.
[4] S. Midori, S. Zheng, I. Yamada, Azeotropic distillation process with vertical
divided-wall column, Kagaku Kogaku Ronbunshu 27 (2001) 756–823.
[5] C. Bravo-Bravo, J.G. Segovia-Hernández, C. Gutiérrez-Antonio, A.L. Durán, A.N.
Bonilla-Petriciolet, A. Briones-Ramírez, Extractive dividing wall column:
design and optimization, Ind. Eng. Chem. Res. 49 (2010) 3672–3688.
[6] R. Isopescu, A. Woinaroschy, L. Draghiciu, Energy reduction in a divided wall
distillation column, Rev. Chim. 59 (2008) 812–815.
[7] D. Zhou, D. Liu, X. Hu, C. Ma, Effect of acoustic cavitation on boiling heat
transfer, Exp. Therm. Fluid Sci. 26 (2002) 931–938.
[8] S. Rodrigues, G.A. Pinto, Ultrasound extraction of phenolic compounds from
coconut (Cocos nucifera) shell powder, J. Food Eng. 80 (2007) 869–872.
[9] A. Ripin, S.K. Abdul Mudalip, R. Mohd Yunus, Effects of ultrasonic waves on
enhancement of relative volatilities in methanol–water mixtures, Jurnal
Teknologi 48 (2008) 61–73.
[10] A. Ripin, S.K. Abdul Mudalip, Z. Sukaimi, R.M. Yunus, Z.A. Manan, Effects of
ultrasonic waves on vapor–liquid equilibrium of an azeotropic mixture, Sep.
Sci. Technol. 44 (2009) 2707–2719.
[11] S.K. Abdul Mudalip, A. Ripin, R. Mohd Yunus, S.Z. Sulaiman, R. Che Man, Effects
of ultrasonic waves on vapor–liquid equilibrium of cyclohexane/benzene, Int.
J. Adv. Sci. Eng. Inform. Technol. 1 (2011) 72–76.
[12] R.E. Apfel, Sonic effervescence: a tutorial on acoustic cavitation, J. Acoust. Soc.
Am. 101 (1997) 1227–1237.
[13] M. Ashokkumarb, J. Lee, S. Kentish, F. Grieser, Bubbles in an acoustic field: an
overview, Ultrason. Sonochem. 14 (2007) 470–475.
[14] M. Ashokkumar, T.J. Mason, Sonochemistry, Kirk-Othmer encyclopedia of
chemical technology, 2007, http://dx.doi.org/10.1002/
0471238961.1915141519211912.a01.pub2.
[15] A. Kalva, T. Sivasankar, V.S. Moholkar, Physical mechanism of ultrasound-
assisted synthesis of biodiesel, Ind. Eng. Chem. Res. 48 (2009) 534–544.
[16] P.A. Parkar, H.A. Choudhary, V.S. Moholkar, Mechanistic and kinetic
investigations in ultrasound assisted acid catalyzed biodiesel synthesis,
Chem. Eng. J. 187 (2012) 248–260.
[17] H.A. Choudhury, S. Chakma, V.S. Moholkar, Mechanistic insight into
sonochemical biodiesel synthesis using heterogeneous base catalyst,
Ultrason. Sonochem. 21 (1) (2014) 169–181.
[18] O. Louisnard, J. González-García, Acoustic cavitation, in: Ultrasound
Technologies for Food and Bioprocessing, Springer, New York, 2011, pp. 13–64.
192 T. Mahdi et al. / Ultrasonics Sonochemistry 24 (2015) 184–192
[19] K. Suzuki, K. Arashi, S. Nii, Determination of droplet and vapor ratio of
ultrasonically-atomized aqueous ethanol solution, J. Chem. Eng. Jpn. 45 (2012)
337–342.
[20] T.J. Matula, S.M. Cordry, R.A. Roy, L.A. Crum, Bjerknes force and bubble
levitation under single-bubble sonoluminescence condition, J. Acoust. Soc. Am.
102 (1997) 1522–1527.
[21] M. Sato, K. Matsuura, T. Fujii, Ethanol separation from ethanol-water solution
by ultrasonic atomization and its proposed mechanism based on parametric
decay instability of capillary wave, J. Chem. Phys. 114 (2001) 2382–2386.
[22] K. Suzuki, D.M. Kirpalani, S. Nii, Influence of cavitation on ethanol enrichment
in an ultrasonic atomization system, J. Chem. Eng. Jpn. 44 (2011) 616–622.
[23] D.M. Kirpalani, F. Toll, Revealing the physicochemical mechanism for
ultrasonic separation of alcohol–water mixtures, J. Chem. Phys. 117 (8)
(2002) 3874–3877.
[24] T.J. Mason, J. Phillip, Applied Sonochemistry, Wiley-VCH, Weinheim, 2002.
[25] Y.T. Didenko, D. Nastich, S. Pugach, Y. Polovinka, V. Kvochka, The effect of bulk
solution temperature on the intensity and spectra of water sonoluminescence,
Ultrasonics 32 (1994) 71–76.
[26] H.-S. Son, S.-K. Kim, J.-K. Im, J. Khim, K.-D. Zoh, Effect of bulk temperature and
frequency on the sonolytic degradation of 1,4-dioxane with FeO, Ind. Eng.
Chem. Res. 50 (2011) 5394–5400.
[27] Y. Jiang, C. Petrier, T.D. Waite, Sonolysis of 4-chlorophenol in aqueous solution:
effects of substrate concentration, aqueous temperature and ultrasonic
frequency, Ultrason. Sonochem. 13 (2006) 415–422.
[28] M. Ashokkumar, L.A. Crum, C.A. Frensley, F. Grieser, T.J. Matula, W.B.
McNamara, K.S. Suslick, Effect of solutes on single-bubble sonoluminescence
in water, J. Phys. Chem. A 104 (2000) 8462–8465.
[29] R. Tögel, S. Hilgenfeldt, D. Lohse, Squeezing alcohols into sonoluminescing
bubbles: the universal role of surfactants, Phys. Rev. Lett. 84 (2000) 2509–
2512.
[30] K. Yasui, T. Tuziuti, T. Kozuka, A. Towata, Y. Iida, Relationship between the
bubble temperature and main oxidant created inside an air bubble under
ultrasound, J. Chem. Phys. 127 (2007). 154502-154502.
[31] K. Yasui, Effect of volatile solute on sonocheminescence, J. Chem. Phys. 116
(2002) 2945–2954.
[32] Y. Kamihara, T. Watanabe, M. Hirano, H. Hosono, Iron-based layered
superconductor La [O1x Fx] FeAs (x = 0.05–0.12) with Tc = 26 K, J. Am. Chem.
Soc. 130 (2008) 3296–3297.
[33] J. Wisniak, L. Aharon, Y. Nagar, H. Segura, R. Reich, Effect of pressure on the
vapor–liquid equilibria of the system methanol + ethyl 1,1-dimethylethyl
ether, Phys. Chem. Liq. 39 (2001) 723–737.
[34] H. Renon, J.M. Prausnitz, Local compositions in thermodynamic excess
functions for liquid mixtures, AlChE J. 14 (1968) 135–144.
[35] L. Rayleigh, VIII. On the pressure developed in a liquid during the collapse of a
spherical cavity, J. Sci. 34 (1917) 94–98.
[36] M. Plesset, The dynamics of cavitation bubbles, J. Appl. Mech. 16 (1949) 227–
282.
[37] A. Naji Meidani, M. Hasan, Mathematical and physical modelling of bubble
growth due to ultrasound, Appl. Math. Model. 28 (2004) 333–351.
[38] S. Sochard, A.-M. Wilhelm, H. Delmas, Gas-vapour bubble dynamics and
homogeneous sonochemistry, Chem. Eng. Sci. 53 (1998) 239–254.
[39] J.-P. Franc, Physics and Control of Cavitation, in, DTIC Document, 2006.
[40] V. Moholkar, P. Senthil Kumar, A. Pandit, Hydrodynamic cavitation for
sonochemical effects, Ultrason. Sonochem. 6 (1999) 53–65.
[41] G. Servant, J.P. Caltagirone, A. Gérard, J.L. Laborde, A. Hita, Numerical
simulation of cavitation bubble dynamics induced by ultrasound waves in a
high frequency reactor, Ultrason. Sonochem. 7 (2000) 217–227.
[42] H. Alehossein, Z. Qin, Numerical analysis of Rayleigh–Plesset equation for
cavitating water jets, Int. J. Numer. Meth. Eng. 72 (2007) 780–807.
[43] R. Rajan, R. Kumar, K. Gandhi, Modelling of sonochemical oxidation of the
water-KI-CCl4 system, Chem. Eng. Sci. 53 (1998) 255–271.
[44] Q. Li, J. Zhang, Z. Lei, J. Zhu, J. Zhu, X. Huang, Selection of ionic liquids as
entrainers for the separation of ethyl acetate and ethanol, Ind. Eng. Chem. Res.
48 (2009) 9006–9012.
[45] T. Mahdi, A. Ahmad, A. Ripin, M.M. Nasef, Vapor–liquid equilibrium of ethanol/
ethyl acetate mixture in ultrasonic intensified environment, Korean J. Chem.
Eng. 31 (2014) 875–880.
[46] M. Topphoff, J. Kiepe, J. Gmehling, Effects of lithium nitrate on the vapor–liquid
equilibria of methyl acetate + methanol and ethyl acetate + ethanol, J. Chem.
Eng. Data 46 (2001) 1333–1337.
[47] N. Calvar, A. Dominguez, J. Tojo, Vapor–liquid equilibria for the quaternary
reactive system ethyl acetate + ethanol + water + acetic acid and some of the
constituent binary systems at 101.3 kPa, Fluid Phase Equilib. 235 (2005) 215–
222.
[48] S.L. Peshkovsky, A.S. Peshkovsky, Matching a transducer to water at cavitation:
acoustic horn design principles, Ultrason. Sonochem. 14 (2007) 314–322.