Separation of Azeotropic Mixtures Using Air Microbubbles

Generated by Fluidic Oscillation

Nada Abdulrazzaq
Dept. of Chemical and Biological Engineering, University of Sheffield,
Mappin Street, Sheffield S1 3JD, U.K.
Dept. of Chemical Engineering, University of Baghdad, Baghdad, Iraq
Baseem Al-Sabbagh
Dept. of Chemical Engineering, University of Baghdad, Baghdad, Iraq
Julia M. Rees
School of Mathematics and Statistics, Hicks Building, University of Sheffield,
Hounsfield Road, Sheffield S3 7RH, U.K.
William B. Zimmerman
Dept. of Chemical and Biological Engineering, University of Sheffield,
Mappin Street, Sheffield S1 3JD, U.K.

DOI 10.1002/aic.15097
Published online in Wiley Online Library (wileyonlinelibrary.com)

The feasibility of separating the azeotropic mixture of ethanol-water using microbubble-mediated batch distillation is
presented. The effects of the depth of the liquid mixture in the bubble tank and of the inlet air microbubble temperature
on the process efficiency were investigated. The enrichment of ethanol in the vapor phase was higher than that achieved
at equilibrium conditions for all liquid ethanol mole fractions considered, including the azeotrope. On decreasing the
depth of the liquid mixture and increasing the temperature of the air microbubbles, the separation efficiency of ethanol
was improved. Ethanol with purity of about 98.2 vol % was obtained using the lowest liquid level (3 mm) in conjunction
with the highest air microbubble temperature (908C). Separation was achieved with a small rise in the temperature of
the liquid mixture (48C) at a depth of 3 mm and evaporation time of 90 min making this system suitable for treating
thermally sensitive mixtures. VC 2015 The Authors AIChE Journal published by Wiley Periodicals, Inc. on behalf of

American Institute of Chemical Engineers AIChE J, 00: 000–000, 2015

Keywords: bubble columns, distillation, evaporation, heat transfer, mass transfer

Introduction isopropanol, methyl acetate, vinyl acetate, and tetrahydrofu-

An azeotropic mixture is a liquid mixture comprising two or
more components whose proportion cannot be altered by con-
ventional atmospheric distillation. The separation of liquid
mixtures into their pure components by traditional distillation
exploits the difference between the concentrations of the vapor
and liquid phases at equilibrium. However, this cannot be
attained with azeotropic mixtures because the equilibrium
vapor and liquid concentrations are identical.1
This thermodynamic equilibrium barrier to distillation
makes the purification and recovery of some components diffi-
cult. Azeotropes exist in many common and important indus-
trial chemical processes, including the production of ethanol,
This is an open access article under the terms of the Creative Commons Attribu-
tion License, which permits use, distribution and reproduction in any medium, pro-
vided the original work is properly cited.
Correspondence concerning this article should be addressed to J. Rees at
j.rees@sheffield.ac.uk.
C 2015 The Authors AIChE Journal published by Wiley Periodicals,
V occur.5 Entrainers can alter the relative volatility of the azeo-
Inc. on behalf of American Institute of Chemical Engineers
ran.2 Furthermore, the application of biofuels, the promising
sustainable alternative for fossil-based fuels, can also be
problematic due to azeotropic properties. Biofuel mixtures
contain high amounts of water which cannot be completely
removed due to the formation of azeotropes with the other
organic components.3,4
The separation of azeotropic mixtures is considered to be
one of the most challenging tasks in chemical processes and
bespoke methods are often needed to facilitate their separa-
tion.5 Different techniques have been explored to resolve aze-
otropic systems. These mainly include pressure swing, and
extractive and azeotropic distillation. The first method
involves the use of two or more distillation columns at differ-
ent pressures while in the other two methods, an additional
agent referred to as an entrainer (light entrainer in the case of
azeotropic distillation and heavy solvent in the case of extrac-
tive distillation) is added to enable the separation process to
tropic constituents to facilitate their breaking.1 In azeotropic

AIChE Journal 2015 Vol. 00, No. 00 1

 Figure 1. The fluidic oscillator: (a) The amplifier, (b) The feedback loop.
distillation, the additional agent needs to be carefully selected
to ensure that it forms another azeotrope with one of the mix-
ture constituents, then it could be removed by a further separa-
tion step such as distillation, or any other separation mode,
which means that additional energy is required for its recov-
ery.6,7 The technique of azeotropic distillation is widely
applied in the alcohol industry, particularly for the separation
of water from ethanol (ethanol dehydration).8 Cyclohexane is
the entrainer that is commonly used for separating the water-
ethanol azeotrope. When cyclohexane is added to the mixture,
a second azeotrope that boils at 62.18C is formed with water.
After heating this ternary mixture, the binary azeotropic mix-
ture of water-cyclohexane is evaporated first, leaving nearly
pure ethanol. In extractive distillation, conversely, the
entrainer should have a higher boiling point than any of the
mixture components and should not form any azeotrope with
either of them. Ethylene glycol is a feasible entrainer for the
separation of the ethanol-water azeotrope by extractive distil-
lation.1 These techniques can be effective for breaking azeo-
tropic systems, however, they are associated with high energy
requirements and can present with operational difficulties.9,10
Recently, new membrane technologies, including pervapo-
ration and vapor permeation, have been applied to the task of
separating azeotropic mixtures. To achieve separation, a mem-
brane which is more permeable for one of the azeotropic spe-
cies than it is for the other is used.11 These methods, however,
have been challenged by several factors including those asso-
ciated with scaling up the separation units, high investment
cost and rapid fouling of the membranes which restrict their
application in the industrial field.12,13
Due to the limitations of the existing methods for resolving
the azeotropic mixtures, a simple, reliable, and low cost tech-
nique is sought. In this work, a new separation process that
involves the use of microbubbles has been explored to investi-
gate its feasibility for breaking the azeotrope of the ethanol-
water mixture. The purification of ethanol has been the focus
of many recent studies as ethanol is one of the most widely
used biofuels and any improvements in its production could
potentially lead to significant reductions in environmental
pollution.14,15
The separation principle of the approach proposed in this
work depends on evaporating ethanol selectively from the aze-
otropic mixture. Evaporation can be achieved through inject-
ing hot air microbubbles, generated with the aid of fluidic
oscillation, to flow upward in a laminar regime in the cold liq-
2 DOI 10.1002/aic Published on behalf of the AIChE
uid mixture. The temperature of the air microbubbles was
selected to be higher than the normal boiling point of ethanol
and lower than that of water to ensure the separation of ethanol
from the mixture.
When hot bubbles are injected into a cold liquid they trans-
fer their energy to the surrounding mixture in two ways: either
as latent heat of evaporation, resulting in an evaporation of the
liquid mixture from the surface to the interior of the bubble, or
as sensible heat transfer, causing a rise in the temperature of
the liquid mixture.16 It used to be assumed that the energy
transmitted by bubbles is split evenly between heat and mass
transfer largely due to liquid turbulent mixing.16 However,
recent experiments in this field have shown that on decreasing
the residence time of microbubbles in a liquid mixture, evapo-
ration begins to dominate over heat transfer while at longer
residence times, heat transfer dominates and the initially
evaporated liquid is recondensed and returned from the bubble
to the surrounding mixture.17 The recondensation process is
associated with a release of sensible heat to the liquid mixture
which leads to an increase in its temperature.
The residence time of bubbles in the liquid mixture can be
controlled by altering the depth of the liquid layer through
which the bubbles can ascend. This means that when the depth
of the liquid layer is shallow, evaporation becomes favorable
which leads in turn to a maximum separation efficiency. As
the depth of the liquid layer increases, sensible heat transfer
becomes more significant leading to reduced vaporization as
well as raising the temperature of the liquid mixture.17
Microbubbles used in the current work are generated via a
fluidic oscillator device. The fluidic oscillator, as shown in
Figure 1, is a bistable valve receiving a steady gas flow,
switching it into an oscillatory flow at a regular frequency
ranging between 1 and 100 Hz.18–20 It is a no-moving part
device that works on the Coanda effect.21 The main part of the
fluidic oscillator is the amplifier which features one inlet ter-
minal, two mid-control terminals and two discharging termi-
nals which are connected to the sparger. This fluidic amplifier
was constructed by a CNC machine from 1.2-mm thick poly-
methylmetacrylate plates. However, it could also be made of
other construction materials to suit various operating condi-
tions. It is provided with a feedback loop which consists sim-
ply of a tube connecting its two control ports, X1 and X2, as
shown in Figure 1b.
The characteristics of the fluidic oscillator help to reduce
the size of the generated bubbles by ensuring an early break
2015 Vol. 00, No. 00 AIChE Journal
 Figure 2. Schematic diagram of the experimental rig used for the microbubble mediated separation unit.
off when the bubble formed at the aperture is hemispherical in
shape, offering the smallest possible size. This, in turn, will
increase the surface area to volume ratio of the bubbles which
will lead to higher mass, heat or momentum transfer rates. The
fluidic oscillation offers several benefits for the generation of
microbubbles. It is a cheap device with no moving parts. It is
easy to manufacture, uses no electricity and requires only an
air supply to generate uniformly spaced microbubbles of
approximately the same size as the pores of the sparger. Low
energy consumption is the main feature that distinguishes this
approach from the traditional methods for microbubble gener-
ation that require a significant supply of energy.19,22
This novel device has been applied recently as a cost effec-
tive method for microbubble generation with successful out-
comes in various fields related to separation and mass transfer.
These include: recovery of oil emulsion,23 separation of algae
from culture media,24 promotion of microalgae growth for bio-
fuel production using CO2 microbubbles,25,26 wastewater
treatment,27 yeast harvesting from growth medium,28 and puri-
fication of potable water by ozone microbubbles as a steriliza-
tion agent.29
Materials and Methods
Ethanol solutions with concentrations of 96 vol %/vol %
and >99.99% (HPLC grade) were purchased from Sigma
Aldrich Company, UK. High purity deionized water was used
for making up ethanol-water mixtures.
Experimental Setup and Procedure
Experiments to investigate the separation of the ethanol-
water azeotrope were conducted on a microbubble mediated
batch distillation unit as shown schematically in Figure 2. It
comprises a process air heater, controller, fluidic oscillator,
and bubble tank. As illustrated in Figure 2, dry pressurized air,
supplied via a laboratory air pipe, was fed through a flow
meter to measure its flow rate. The air was then passed through
the process heater where a total power of 750 W was applied
to increase its temperature to the required value. The tempera-
ture of the heater was registered by a thermocouple (type K)
and was controlled using a controller with overheating protec-
tion. The hot air stream then entered the fluidic oscillator. This
was constructed from aluminum to withstand the high temper-
atures applied in the experiment. The length of the feedback
loop connecting the two control terminals was 25 cm. In this
AIChE Journal 2015 Vol. 00, No. 00 Published on behalf of the AIChE
device, the air stream is distributed between the two bleed
valves and the two outlet streams of the oscillator. The func-
tion of the bleed valves is to expel the excess air that is not
required to enter the diffuser. The two outlet air streams from
the fluidic oscillator then entered the diffusers in the bubble
tank where the microbubbles were generated.
The rectangular bubble tank with glass sidewalls had
dimensions of 15 cm in length, 12 cm in width, and 4 cm in
height. The base of the tank, which contained the diffusers,
was constructed of aluminum to withstand the high tempera-
tures of the inlet air stream. Its top was covered with a glass
lid on to which a small funnel was inserted to direct the rising
vapors to the outlet gas tube to minimize their condensation.
The gas distribution system located at the bottom of the bubble
tank consisted of two rectangular ceramic diffusers made from
Point Four microporous sintered alumina plates, 15 cm in
length, 6 cm in width, and 1 cm in thickness. All of the con-
necting pipelines, the fluidic oscillator, the heater, and the bub-
ble column were thoroughly insulated with glass wool to
minimize heat loss. Four thermocouples (type K) at different
locations within the unit were used to obtain temperature
measurements. The two thermocouples, positioned in the con-
necting pipes between the fluidic oscillator and the bubble col-
umn, were used for measuring the inlet air temperature to the
diffuser. The thermocouple near the base of the bubble column
was used to measure the temperature of the liquid mixture
while the other in the outlet gas stream measured the tempera-
ture of the outlet air. The sensors of all these thermocouples
were located on the centerline of the pipes and connected to a
four channel microprocessor thermometer (Testo Model
176T4) which displayed the temperature readings.
The operating conditions for the azeotropic separation
experiments are listed in Table 1. For all experimental runs,
the inlet air flow rate to the fluidic oscillator was fixed at 80 L/
min while the flow rate to the diffusers was kept at 1 L/min,
thus about 79 L/min were lost through the bleed valves of the
Table 1. Operating Conditions for Azeotropic Mixture
Experiments
Depth of Mixture Layer (mm)
Air Microbubble
Temperature (8C) 3 5 10
80 Run 1 Run 2 Run 3
90 Run 4 Run 5 Run 6
DOI 10.1002/aic 3
fluidic oscillator. This low flow rate was applied to facilitate
the generation of small bubbles in a laminar flow.
At the beginning of each experimental run, the valve of the
air supply was opened and the gauge pressure in the air line
was kept at two bars. The temperature controller was then
turned on. Following this, the bleed valves of the fluidic oscil-
lator were opened gradually and adjusted until the outlet air
flow rate from the tank was maintained at 1 L/min and an
oscillatory air flow was achieved. This was recognized through
a continuous vibrating sound. When the air in the two inlet
streams to the bubble tank had reached the required tempera-
ture, a previously measured volume of ethanol-water azeo-
tropic mixture at a temperature of 208C 6 0.5 was poured into
the bubble tank. Immediately after pouring the liquid mixture
into the bubble tank, the stopwatch was started and periodic
readings of the air temperature at the two inlet streams to the
tank, the temperature of the liquid mixture, as well as the tem-
perature of the outlet vapors were recorded at 5-min intervals.
At the end of each experiment, a syringe was used to extract the
remaining liquid mixture from the tank through a vent in the
lid. The volume of this liquid was measured to calculate the
rate of evaporation of the liquid mixture. Samples were taken
from the remaining solution for analysis. The equation for cal-
culating the percentage of evaporation of the liquid mixture can
be expressed by Figure 3. (a) Gas sensor circuit schematic and (b) Gas
volume evaporated V0 2Vf sensor module.
Percentage evaporation ð%Þ5 5 3100%
intial volume V0
The sensing element of the sensors is made from materials
where V0 is the initial volume of liquid mixture (mL) and Vf is that exhibit low conductivity in alcohol free air (e.g., SnO2).
the final volume of liquid mixture after evaporation (mL). However, in the presence of an alcohol or an organic vapor,
the conductivity increases depending on the gas concentration.
Analytical Method In such cases, the change in the sensor’s conductivity can be
Liquid phase measurement used as a measure of the gas concentration in the sample under
test. Figure 3a shows a simple circuit diagram used to drive a
The chemical compositions of the remaining liquid ethanol- MQ-3 gas sensor. This sensor has six pins: 2A’s and 2B’s as
water samples were determined using a Varian 3900 gas sensing element terminals and 2H’s terminals for the heater to
chromatograph (GC) equipped with a thermal conductivity raise the temperature of the sensing element up to its working
detector. The column used was a HAYESEP P which had an temperature. The change in conductivity can be sensed by
internal diameter of 4 mm and total length of 2 m. The GC monitoring the voltage drop over the load resistor RL, or
oven temperature was set at 1508C. The injection was splitless expressed as a change in the sensor resistance. Usually, this
at a temperature of 1808C. The carrier gas was nitrogen at a simple circuit was integrated into a single printed circuit board
flow rate of 400 mL/min and a pressure of 30 psi. Retention and sold as a gas sensor module with some extra features (see
times of ethanol and water were 2 min and 0.8 min, Figure 3b).
respectively. The measurement apparatus for the gas concentration is
shown in Figure 4. A set of two MQ-3 gas sensor modules was
Gas phase measurements used to simultaneously provide two measurements of the gas
Traditional condensation methods of vapors would not concentration. Measurements were made inside the approxi-
reflect the actual concentration at a given time during the mately 1300 cm3 clear acrylic chamber. Sensors were attached
experiment. Instead, they would provide an accumulated con- to the outer wall of the chamber and connections were pro-
centration of the condensed vapor over a period of time since vided to the gases in the bulk via a circular hole for each sen-
it is difficult to condense the vapor instantaneously to get a sor. Both sensors were supplied with a voltage of 5 V from the
useful liquid sample for GC analysis. To overcome this prob- power supply model TTi Ex354D. Received signals from sen-
lem, alcohol gas sensors were used to measure the ethanol sors were fed into a Pico ADC-20 high resolution data logger
concentration in the gas phase instantaneously. Alcohol gas and then relayed to a computer for analysis.
sensors, or more generally volatile organic compound gas sen- A regulated dry air was injected inside the chamber at a
sors, are used in many applications including vehicle and flow rate of 1800 cm3 min21. This air stream removes any
semiconductor industries, organic vapor detectors/alarms, air/ moisture from inside the test chamber and isolates it from
food quality control, security, and as electronic noses to iden- other contaminants that may interfere with the measurements.
tify gases and their concentrations.30–32 Also, alcohol gas sen- In addition, the continuous flow of the air stream provided an
sors offer many useful features, such as high sensitivity, long approximately stable state for the commencement of the meas-
life, miniature size, low cost, easy construction for the drive urements. Outflowing gases exit the chamber at the same flow
circuit, and good repeatability. rate through a hole on the opposite side to the inlet.

4 DOI 10.1002/aic Published on behalf of the AIChE 2015 Vol. 00, No. 00 AIChE Journal
 Figure 4. Gas concentration measurement apparatus.
First, saturated ethanol gas samples were prepared from
known liquid concentrations. Then, depending on the gas con-
centration, 3 cm3 of saturated gas, taken from calibration sam-
ples, were injected into the test chamber by a syringe through
the gas mixing valve. Injected vapors will be mixed with the
dry air stream, enter the chamber, and become distributed
evenly inside the chamber with the aid of the brushless fan.
After injection, signals were detected due to the change in the
sensor’s conductivity. Their conductivity then returned back
to the initial state when the air inside the chamber became free
from all traces of the injected gas. Calibration curves were
made based on the signal peak value of the injected samples.
Generally, all calibration curves showed virtually a linear rela-
tionship, confirming operation in the sensor’s linear response
region.
After injecting calibration samples, ethanol vapor samples
from the bubble tank were injected in the same way as
described above. The unknown concentration of the injected
gas can be determined from the calibration curves. To validate
the calibration, known concentrations were injected during
and at the end of the experiment.
It should be noted that there is an unavoidable drift of a few
millivolts in the base line of the sensors. Consequently, all
measurements were adjusted in line with the base line average
value following the mapping
Figure 5. Variation of the final concentration of ethanol
in the liquid mixture with the liquid mixture
level.
The error bars represent the standard error.
AIChE Journal 2015 Vol. 00, No. 00 Published on behalf of the AIChE


pf pi 2 bpi 2b
where pf is the final value of the signal after correction, pi rep-
resents the highest value of the signal obtained from the
injected sample, or simply signal peak value, bpi is the sensor
base line of the signal under analysis, and b is the average
base line of the sensor along the experiment.
Results and Discussion
Experiments were carried out at three different depths of
ethanol-water liquid mixture (3, 5, and 10 mm) and two differ-
ent air microbubble temperatures (808C and 908C) for the pur-
pose of exploring their effect on breaking the azeotrope. The
variation of the final concentration of ethanol in the liquid
solution with the depth of the liquid mixture corresponding to
each of the air microbubble temperatures investigated is
shown in Figure 5. Experiments were carried out in triplicate
and in a random arrangement to calculate the error bar for
each experiment.
The data show that the depth of the liquid layer has a signif-
icant effect on the separation efficiency of the azeotropic mix-
ture. For the two air microbubble temperatures considered, the
final concentration of ethanol in the liquid mixture increased
with increasing the liquid depth, with almost no azeotropic
separation occurring at the liquid depth of 10 mm, indicating
that the separation efficiency of the azeotrope is better at lower
Figure 6. Variation of the percentage of evaporation of
the liquid mixture vs. liquid mixture height.
The error bars represent the standard error.
DOI 10.1002/aic 5
 oration process was reported. Likewise, as the bubble tempera-
ture increases, the fraction of latent heat of evaporation will
also be increased. Therefore, a greater quantity of ethanol is
expected to evaporate at higher bubble temperatures which
would lead to a reduction in its concentration in the remaining
solution.
To compare the performance of our novel microbubble dis-
tillation process with that of traditional distillation, additional
experiments were carried out with different liquid ethanol
compositions. Experimental conditions such as liquid level,
evaporation time, and inlet microbubble temperature were
kept constant for these experiments at 3 mm, 90 min, and
908C, respectively. The results are presented in Figure 7. The
Figure 7. Comparison between the K values of ethanol liquid composition was measured by GC analysis while the
(Kexp), obtained by the microbubble distilla- vapor concentration in the header space of the bubble tank
tion at initial conditions of liquid level 3 mm was measured directly by the ethanol gas sensor. This figure
and microbubble temperature 908C and the presents a comparison between the values of the vapor–liquid
corresponding equilibrium values (Keq) at dif- distribution coefficient (Keq) of ethanol at equilibrium condi-
ferent liquid compositions. tions and those obtained by the current work (Kexp) at different
ethanol liquid compositions. Figure 7 clearly demonstrates
liquid levels. This behavior can be attributed to the residence that the K values for the separation of the ethanol-water binary
time of the microbubbles in the liquid mixture. When the resi- mixture using the microbubble mediated distillation technique
dence time of the bubbles is long (i.e. deep liquid depths), the is higher than those achieved at equilibrium state for all liquid
compositions considered, including the azeotrope.
bubbles start to cool as they rise causing a recondensation of
The possible explanation for these results is that since hot,
their vapors which then return back to the outer liquid mixture.
bone dry air microbubbles are injected into the cold liquid,
However, at shorter bubble residence times, which are associ-
nonequilibrium driving forces exist for both heat and mass
ated with lower liquid depths, a greater level of separation can
transfers between the contact phases. Moreover, the microbub-
be achieved as recondensation of vapors can be prevented or
bles in the current method are injected in a laminar regime
minimized. This is most likely the reason why the reduction in
which, in turn, acts so as to prevent the liquid and gaseous
ethanol concentration is smaller at deeper liquid levels com-
phases from reaching equilibrium rapidly, thereby ensuring
pared to its reduction at shallower levels as the bubble residence
continuous heat and mass transfer processes. Unlike closed
time is longer in the case of deeper levels leading to an increase
systems in which liquid molecules can vaporize until the equi-
in the likelihood of recondensation occurring. This effect is also
librium is established, the current system behaves like an open
observed in Figure 6 which shows that the evaporation rate of
system with continuous transfer of ethanol molecules from the
the liquid mixture increases with decrease in liquid depth.
liquid side to the bubble side under both diffusion and internal
Experimental evidence in support of the rise in evaporation
convection due to bubble motion. As a result of this, equilib-
rate with the decrease in liquid level has already been
rium is disrupted and according to Le Chatelier’s principle,
reported.17 By injecting hot air microbubbles into pure water,
the equilibrium should be shifted toward more vaporization to
they found that both the evaporation rate of the liquid and the
compensate for the molecules of ethanol removed from the
absolute humidity of the outlet air stream increased as the
liquid phase
height of the water layer decreased.
The data in Figures 5 and 6 also indicate that increasing the C2 H6 OðlÞ 5C2 H6 OðgÞ
temperature of the air microbubbles from 808C to 908C leads
to a decrease in the concentration of ethanol in the liquid mix-
ture as well as to an increase in the evaporation rate at all liq-
uid depths considered in the experiments. This can be
understood by considering the amount of available energy for
the injected air microbubbles. For molecules of the liquid to
evaporate, they should have sufficient kinetic energy to escape
from the liquid phase to the gaseous phase. Since kinetic
energy varies directly with temperature, evaporation is
expected to proceed more quickly at higher liquid tempera-
tures and this is what happens at higher microbubble tempera-
tures. Increasing the temperature of the injected microbubbles
will increase the energy carried by the gas phase, thus more
sensible heat will be transferred to the liquid phase causing an
increase in its temperature. Higher liquid temperatures
increase the kinetic energy of its molecules which conse-
quently lead to an increase in the fraction of the evaporated
liquid solution to the bubble phase. This is in agreement to Figure 8. Temperature of the outlet air from the bubble
previous results obtained from a direct contact evaporation tank against evaporation time for different
system.16 An augmentation in the rate of evaporation of the liquid levels at 908C air microbubble tempera-
liquid phase due to an increase in temperature during the evap- ture and evaporation time of 90 min.

6 DOI 10.1002/aic Published on behalf of the AIChE 2015 Vol. 00, No. 00 AIChE Journal

Figure 9. Temperature of the liquid mixture against
time for different liquid levels at 908C air
microbubble temperature and evaporation
time of 90 min.
In this scenario, the ethanol concentration in the bubble
phase would be higher than the ratios predicted by the equilib-
rium theory for all liquid phase mole fractions and within the
short contact time available during the ascent of the bubbles in
shallow liquid depths as demonstrated in Figure 7. The shal-
lower the liquid depth, the easier it is to maintain nonequili-
brium conditions between the contacted phases, both
chemically and thermally, which is crucial for continuing the
transfer processes. Conversely, deeper liquid layers are associ-
ated with a lesser quantity of vaporization due to thermal equi-
librium with the surrounding liquid. This effect is demonstrated
in Figure 8 which shows that on decreasing the height of liquid
level, the bubbles exiting the system are hotter.
As the carrier gas inside the microbubbles automatically
breaks the azeotrope by the introduction of a third component,
and the microbubble temperature can be held at a different
temperature from the liquid, vapor-liquid equilibrium is never
established. Thus, microbubble distillation should be generally
applicable to any other azeotropic mixture. The shifting of the
azeotropic equilibrium condition, (xi 5 yi) & (Tliq 5 Tvap), into
nonequilibrium conditions, (xi 6¼ yi) & (Tliq 5 Tvap),17 that has
been achieved here, is one of the unique features of this new
technique of microbubble distillation. Here, xi and yi are the
mole fractions of the ith components in the liquid and gas
phases respectively. Tliq and Tvap denote the temperatures of
the liquid and gas phases.
Interestingly, the breaking of the azeotrope barrier was
achieved with just a small temperature rise in the liquid mix-
ture. The experimental temperature profile of the liquid mix-
ture against evaporation time is presented in Figure 9. It is
noticed that there is no substantial augmentation in the temper-
ature of the liquid mixture from its initial value of 208C. The
minimum liquid temperature rise, of around 48C, was obtained
at the lowest liquid depth (i.e., 3 mm).
As mentioned before, if the residence time of the microbub-
bles in the liquid is long, recondensation of bubble vapors will
occur leading to a rise in the liquid temperature due to the
release of sensible heat accompanying the condensation pro-
cess. This is the reason behind the increase in the temperature
of the liquid as the depth of the liquid level increased. These
results suggest that the optimal separation for azeotropic sys-
tems using microbubble mediated distillation technology can
be attained with a minimal liquid temperature rise. The find-
ings of this work demonstrate that this regime is suitable for
AIChE Journal 2015 Vol. 00, No. 00 Published on behalf of the AIChE
the separation of not only typical azeotropic mixtures but also
for those with thermally sensitive compounds.
Conclusion
This work addresses an important industrial issue of break-
ing the azeotrope using the technique of microbubble medi-
ated distillation. The data obtained from this study showed
that it is possible to break the ethanol-water azeotrope using
this process. From the experimental work, it was found that
decreasing the liquid mixture level enhances the separation
efficiency of ethanol from its azeotropic mixture with water.
Ethanol, with a purity of about 98.2 vol %, was obtained under
conditions of liquid mixture depth of 3 mm and inlet micro-
bubble air temperature of 908C. Also, the separation process
was accomplished with just a slight increase in the tempera-
ture of the liquid mixture, indicating that this regime is also
favorable for separating thermochemically unstable solutions.
Microbubble distillation is a new technology that is still
being developed. However, on scale up it promises to contrib-
ute effectively to the purification of numerous chemicals. The
findings of this work demonstrate that this technique has major
advantages over traditional methods and has great potential to
be adapted and applied into any energy industry where distilla-
tion is involved due to:
 Reduction of the energy requirements by applying effi-
ciently generated microbubbles using the novel technology
of fluidic oscillation.
 Increase in the purity of the products and concentrate sol-
utes without having to boil the mixture—it can thus be used
to treat thermally sensitive liquids.
 Elimination of azeotrope. Purity of separation depends on
the depth of the liquid mixture in the bubble tank.
 No high pressures or temperatures are required, no addi-
tional chemicals, neither fouling membranes nor high capital
and maintenance costs.
 Reduction in greenhouse gas emissions by reducing
dependence on fossil-based fuels.
Even though the results of this study demonstrate the prom-
ising efficiency of the current approach, more effort is required
to demonstrate how these advantages and applications are
industrially viable.
Acknowledgments
W.B.Z. thanks the EPSRC for support from the 4CU Pro-
gramme Grant, 4cu.org.uk, (EP/K001329/1). W.B.Z. and
J.M.R .thank Innovate UK and the EPSRC for a grant under
the IB Catalyst industrial research scheme, EP/N011511/1.
N.N.A. thanks the Higher Committee of Education and
Development in Iraq for a doctoral scholarship.
Literature Cited
1. Julka V, Chiplunkar M, Young LO. Selecting entrainers for azeo-
tropic distillation. Chem Eng Prog. 2009;105:47–53.
2. Pereiro AB, Araujo JMM, Esperanca JMSS, Marrucho IM, Rebelo
LPN. Ionic liquids in separations of azeotropic systems—a review.
J Chem Thermodyn. 2012;46:2–28.
3. Naik S, Goud VV, Rout PK, Dalai AK. Supercritical CO2 fractiona-
tion of bio-oil produced from wheat-hemlock biomass. Bioresour
Technol. 2010;101:7605–7613.
4. Pan C, Chen A, Liu Z, Chen P, Lou H, Zheng X. Bioresource tech-
nology aqueous-phase reforming of the low-boiling fraction of rice
husk pyrolyzed bio-oil in the presence of platinum catalyst for
hydrogen production. Bioresour Technol. 2012;125:335–339.
DOI 10.1002/aic 7
 5. Luyben WL. Methanol/trimethoxysilane azeotrope separation using
pressure-swing distillation. Ind Eng Chem Res. 2014;53:5590–5597.
6. Matsuda KM, Huang KH, Iwakabe KI. Separation of binary azeo-
trope mixture via pressure-swing distillation with heat integration.
J Chem Eng Jpn. 2011;44:969–975.
7. Ponce-De-Leon C, Field RW. Comparison of anionic membranes
used to concentrate nitric acid to beyond the azeotropic mixture.
J Membr Sci. 2000;171:67–77.
8. Bastidas PA, Gil ID, Rodriguez G. Comparison of the main ethanol
dehydration technologies through process simulation. In: 20th European
Symposium on Computer Aided Process Engineering (ESCAPE20),
Ischia, Italy, 2010.
9. Oliveira FS, Pereiro AB, Rebelo LPN, Marrucho IM. Deep eutectic
solvents as extraction media for azeotropic mixtures. Green Chem.
2013;15:1326–1330.
10. Corderi S, Gonzalez B, Calvar N, Gomez, E. Ionic liquids as sol-
vents to separate the azeotropic mixture hexane/ethanol. Fluid Phase
Equilib. 2013;337:11–17.
11. Aouinti L, Belbachir M. A maghnite-clay-H/polymer membrane for
separation of ethanol-water azeotrope. Appl Clay Sci. 2008;39:78–85.
12. Tang B, Bi W, Row KH. Dehydration of ethanol by facile synthe-
sized glucose-based silica. Appl Biochem Biotechnol. 2013;169:
1056–1068.
13. Holtbruegge J, Wierschem M, Steinruecken S, Voss D, Parhomenko
L, Lutze P. Experimental investigation, modeling and scale-up of
hydrophilic vapor permeation membranes: separation of azeotropic
dimethyl carbonate/methanol mixtures. Sep Purif Technol. 2013;118:
862–878.
14. Pacheco-Basulto JA, Hernandez-McConville D, Barroso-Munoz FO,
Hernandez S, Segovia-Hernandez JG, Castro-Montoya AJ, Bonilla-
Petriciolet A. Purification of bioethanol using extractive batch distil-
lation: simulation and experimental studies. Chem Eng Process.
2012;61:30–35.
15. Shirsat SP, Dawande SD, Kakade SS. Simulation and optimization
of extractive distillation sequence with pre-separator for the ethanol
dehydration using n-butyl propionate. Korean J Chem Eng. 2013;30:
2163–2169.
16. Riberiro CP, Lage, PLC. Direct-contact evaporation in the homoge-
neous and heterogeneous bubbling regimes. Part I: experimental
analysis. Int J Heat Mass Transfer. 2004;47:3825–3840.
17. Zimmerman WB, Al-Mashhadani MKH, Bandulasena HCH. Evapo-
ration dynamics of microbubbles. Chem Eng Sci. 2013;101:865–877.
18. Tesar, V. “Paradox” of flow reversal caused by protective wall-jet in
a pipe. Chem Eng J. 2007;128:141–154.
8 DOI 10.1002/aic Published on behalf of the AIChE
19. Zimmerman WB, Tesar V, Butler S, Bandulasena HCH. Microbub-
ble Generation. Recent Pat Eng. 2008;2:1–8.
20. Tesar V, Bandalusena HCH. Bistable diverter valve in microfluidics.
Exp Fluids. 2010;50:1225–1233.
21. Tesar V, Hung CH, Zimmerman WB. No-moving-part hybrid-
synthetic jet actuator. Sens Actuators A. 2006;125:159–169.
22. Zimmerman WB, Tesar V, Bandulasena HCH. Towards energy effi-
cient nanobubble generation with fluidic oscillation. Curr Opin Col-
loid Interface Sci. 2011;16:350–356.
23. Hanotu J, Bandulasena HCH, Chiu T, Zimmerman WB. Oil emul-
sion separation with fluidic oscillator generated microbubbles. Int J
Multiphase Flow. 2013.56:119–125.
24. Hanotu J, Bandulasena HCH, Zimmerman WB. Microflotation per-
formance for algal separation. Biotechnol Bioeng. 2012;109:1663–
1673.
25. Zimmerman WB, Zandi M, Bandulasena HCH, Tesar V, James
Gilmour, D, Ying, K. Design of an airlift loop bioreactor and pilot
scales studies with fluidic oscillator induced microbubbles for growth
of a microalgae Dunaliella salina. Appl Energy. 2011;88(10):3357–
3369.
26. Al-Mashhadani MKH, Bandulasena HCH, Zimmerman WB. CO2
Mass transfer induced through an airlift loop by a microbubble cloud
generated by fluidic oscillation. Ind Eng Chem Res. 2011;51:1864–
1877.
27. Rehman F, Medley GJD, Bandulasena HCH, Zimmerman WB. Flu-
idic oscillator-mediated microbubble generation to provide cost
effective mass transfer and mixing efficiency to the wastewater treat-
ment plants. Environ Res. 2015;137:32–39.
28. Hanotu J, Karunakaran E, Bandulasena HCH, Biggs C, Zimmerman
WB. Harvesting and dewatering yeast by microflotation. Biochem
Eng J. 2014;82:174–182.
29. Lozano-Parada JH, Zimmerman WB. The role of kinetics in the
design of plasma microreactors. Chem Eng Sci. 2010;65:4925–4930.
30. Zhang H, Chang M, Wang J, Ye S. Evaluation of peach quality indi-
ces using an electronic nose by MLR, QPST and BP network. Sens
Actuators B. 2008;134:332–338.
31. Lee DS, Lee DD, Ban SW, Lee M, Kim YT. SnO2 gas sensing array
for combustible and explosive gas leakage recognition. IEEE Sens J.
2002;2:140–148.
32. Khalaf W, Pace C, Gaudioso M. Least square regression method for
estimating gas concentration in an electronic nose system. Sensors.
2009;9:1678–1169.
Manuscript received Aug. 18, 2015, and revision received Oct. 21, 2015.
2015 Vol. 00, No. 00 AIChE Journal