Chemical Engineering Journal 213 (2012) 50–61

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

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

Investigation of mixing characteristics in a packed-bed external loop airlift

bioreactor using tomography images
Mian Hamood-ur-Rehman, Yaser Dahman, Farhad Ein-Mozaffari ⇑
Department of Chemical Engineering, Ryerson University, 350 Victoria Street, Toronto, Ontario, Canada M5B 2K3

h i g h l i g h t s

" A new bioreactor, which combines advantages of both packed bed and external loop airlift bioreactors, was introduced.
" 2D and 3D tomograms were utilized to explore the mixing hydrodynamics.
" Mixing efficiency was quantified using mixing time and superficial liquid velocity.
" The effects of design parameters and operating conditions were characterized.
" Installation of the gas redistributor between two rolls of packing had a significant impact on mixing hydrodynamics.

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

Article history: The present research work introduces a recirculated airlift bioreactor (biocontactor), which combines
Received 3 June 2012 advantages of packed bed and external loop airlift bioreactors. This bioreactor is characterized by an
Received in revised form 27 September internal gas redistributor placed between two beds of packing in the riser while working fluid recirculates
2012
through an external downcomer. The main objective of this research work was to characterize the mixing
Accepted 27 September 2012
Available online 5 October 2012
hydrodynamics of this reactor through the non-intrusive flow visualization technique of electrical resis-
tance tomography (ERT). The tomography images were employed to examine the effects of packing, gas
redistributor, sparger configuration, gas flow rate, and liquid height on the mixing time and superficial
Keywords:
Bioreactors
liquid velocity. Results showed that the mixing time in the bioreactor was decreased with an increase
Mixing in the superficial gas velocity in the riser. It was found that at a constant riser superficial gas velocity,
Tomography the mixing time was increased and the superficial liquid velocity decreased when a cross shaped sparger
Packed bed was replaced by a circular sparger, which had the same diameter, and the same number and size of holes.
Superficial liquid velocity Results also showed that mixing time increased and the superficial liquid velocity decreased by reducing
Internal gas redistributor the liquid height in the bioreactor, in addition to installing the internal gas redistributor and packing in
the riser. The presence of the internal gas redistributor and packing reduced the cross-sectional area
available for flow and increased resistance in the liquid flow path. Results, which were obtained from this
study, can be utilized to improve mixing in external loop airlift bioreactors for wider range of
applications.
Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Airlift bioreactors are encountered in the fields of aerobic fermen-

Airlift bioreactors have been extensively used in biotechnology
industries in recent years in a variety of arrangements and applica-
tions. This includes commercial manufacture of pharmaceuticals,
enzymes, fragrances, dyes and antibiotics [1–3]. The main advanta-
ges of this type of reactors are the simple construction and opera-
tion, low investment and operational cost, absence of regions of
high shear, very fine gas dispersion, enhanced mixing and mass
transfer performance, and relatively low power requirements [4].
⇑ Corresponding author. Tel.: +1 416 979 5000x4251; fax: +1 416 979 5083.
E-mail address: fmozaffa@ryerson.ca (F. Ein-Mozaffari).
tations, waste water treatment, and other operations, where low
shear stress is required [5–7]. The two main types of internal and
external loop recirculated airlift bioreactors have been the subject
of interest for different researchers. The main difference between
these two types is that the internal loop airlift bioreactors consist
of two concentric columns while the external loop airlift bioreac-
tors consist of two separate columns connected near the top and
at the bottom by two horizontal sections. This design gives a sep-
arate gas disengagement section to external loop airlift bioreactors,
which helps in maximizing deaeration [5,7–13].
The external loop airlift bioreactors have low friction in riser
and downcomer and thus provide a better opportunity for mea-
surement and control of flow conditions. These bioreactors also

1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cej.2012.09.106
 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61
Nomenclature
Abc cross-sectional area of the bottom connecting section (m2)
Atc cross-sectional area of the top connecting section (m2)
Ad downcomer cross sectional area (m2)
Ar riser cross sectional area (m2)
Dr diameter of riser (m)
Hr circulating liquid height in the riser (m)
Hbc circulating liquid length in the bottom connecting section
(m)
Htc circulating liquid length in the top connecting section (m)
tr liquid residence time in the riser (s)
offer effective heat transfer and temperature control. Furthermore,
they can achieve independent control of the gas input-rate and li-
quid velocity using a throttling device at the bottom connecting
tube between riser and downcomer of the external loop bioreactor.
The external-loop usually reaches nearly total gas disengagement
at the top section (due to the presence of separate gas-disengage-
ment section) giving rise to a higher difference in density or hydro-
static pressure which results in higher circulation velocity as
compared to the internal loop airlift bioreactor and consequently
improves mixing and heat transfer in the reactor. On the other
hand, the external loop offers lower gas recirculation as compared
to internal loop airlift, which results in lower mass transfer
[3,4,14–16]. The applications of external loop airlift bioreactors
have been increasing and making them as one of the most impor-
tant bioreactor configurations. Many investigators have studied
external loop bioreactors experimentally and theoretically
[2,3,17–19].
Mixing time and superficial liquid velocity play significant roles
in the hydrodynamic characterization of the airlift bioreactors.
Mixing quality is a critical parameter, which is widely used in
the design, modeling and operation, as well as scale up of external
loop airlift bioreactors. In some processes, where non-uniform pro-
files of concentration or temperature can cause side reactions or
reduce the chemical reaction rate, good mixing is very important.
In external loop airlift bioreactors, sparged air is used to achieve
the mixing. The mixing of gas phase is usually negligible because
of the high velocity of gas bubbles in the airlift reactor. The super-
ficial liquid velocity is also a key design parameter for the reactors,
which affects the residence time of gas, mass transfer coefficient,
turbulence, heat transfer coefficient, shear forces to which the
microorganisms are exposed, and the mixing time. The cross sec-
tional area ratio of downcomer to riser is a physical parameter,
which strongly affects the superficial liquid velocity. Any
resistance given to the flow path of the liquid in the bioreactor also
affect the liquid velocity in the bioreactor.
Gumery et al. [20] utilized tomography technique to explore the
mixing characteristics in a draft-tube airlift reactor. Jin et al. [21]
studied the oxygen transfer in a pilot plant external air-lift bioreac-
tor as a function of gas holdup, oxygen transfer coefficient, super-
ficial gas velocity, and dissolved oxygen. To investigate the liquid-
phase mixing in the external-loop airlift reactors, Kawase et al. [22]
measured mixing time, circulation time, and downcomer linear li-
quid velocity. Different configurations of the external loop airlift
bioreactors have been designed by the previous researchers. For in-
stance, Xu and Yu [23] designed a multiple airlifting membrane
bioreactor. The vessel was separated into two compartments, a
tube-side aerobic compartment and a shell-side anaerobic com-
partment, by installing four sintered stainless steel filter tubes be-
tween the top and the bottom sections. Han et al. [10] investigated
the hydrodynamic behavior in a new gas–liquid–solid inverse flu-
idization airlift bioreactor. The advantages of external loop reactors
51
tc circulation time (s)
Ulr superficial liquid velocity in the riser (m/s)
Uld superficial liquid velocity in the downcomer (m/s)
Ugr superficial gas velocity in the riser (m/s)
Greek symbols
u dimensionless factor ()
egr gas holdup in the riser ()
egd gas holdup in the downcomer ()
and inverse fluidized beds were combined to construct this novel
external loop bioreactor. Loh and Liu [6] proposed a novel external
loop inversed fluidized bed airlift bioreactor by installing a valve at
the bottom between the riser and downcomer sections for the
treatment of wastewater containing high phenol concentrations.
Zhang et al. [3] designed an internal for the external loop airlift
bioreactor. This internal contained some semicircular holes and
each hole had a tongue-like plate facing the upstream to break
the bubbles and enhance the flow redistribution. Meng et al. [24]
combined the design strategies of the external loop airlift bioreac-
tor and packed bed bioreactor. They used the woven nylon packing
as a packed bed in the riser section of the airlift bioreactor. Mohan-
ty et al. [25] investigated the hydrodynamics of a novel external
loop airlift bioreactor operating in three vertical stages. The verti-
cal cylindrical column was fitted with a total of seven internals
(four contraction disks and three expansion disks).
An external loop recirculated airlift bioreactor (biocontactor),
with an internal gas redistributor, which was installed between
two rolls of packing in the riser, was designed and built. This de-
sign combines advantages of both packed bed and external loop
airlift bioreactor. The presence of two beds of packing allows for
improving residence time and bioconversion when immobilized
enzymes or microorganisms are utilized in variety of bioengineer-
ing applications. To characterize the mixing hydrodynamics in this
bioreactor, mixing time and superficial liquid velocity were mea-
sured through electrical resistance tomography (ERT), which has
been employed as a reliable and non-intrusive tool for the direct
analysis of the hydrodynamics of multiphase flows in recent years
[20,26–31]. This technology manipulates data obtained from the
sensors located on the periphery of the vessel in order to get the
precise information on the flow regime and concentration distribu-
tion inside the process vessels [32]. Results obtained using the ERT
were used to examine the effect of different parameters that
include having two packed beds in the riser; internal gas redistrib-
utor installed between two packed-bed sections; sparger
configuration; gas flow rate; and liquid height on the mixing
hydrodynamics of the bioreactor. The new contributions of this pa-
per are: (i) a new bioreactor, which combines advantages of both
packed and external loop airlift bioreactors, was introduced; (ii)
Tomogrphy, a reliable and non-intrusive technique, was utilized
to explore the mixing performance of this reactor; and (iii) the ef-
fects of the design parameters and operating conditions on the
mixing hydrodynamics were quantified.
2. Experiment
2.1. Experimental setup
A packed bed external loop recirculated airlift bioreactor with an
internal gas redistributor placed in the way between two rolls of
52 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61

packing in the riser was designed and built. Polyvinyl chloride
material was used to construct this bioreactor. Optimal geometries
in the design of the bioreactor were considered based on results re-
ported previously in the literatures [4,6,10,33–36]. Detailed sche-
matic diagram of the experimental setup is shown in Fig. 1. As
shown in this figure, the sparger was installed at the bottom of
the riser and the air flow was controlled using a rotameter. The dif-
ferent sparger configurations that were examined include the cross
shaped and circular sparger. The superficial gas velocity in the riser
was varied between 0.011 and 0.033 m/s. An internal perforated
plate was utilized as the gas redistributor. The function of this gas
redistributor was to break and regenerate the gas bubbles exiting
the lower packing and before entering the upper packing in the riser.
This gas redistributor was designed to have 415 holes, each of
0.002 m, providing total open area about 3% and was installed be-
tween the two beds of packing in the riser (see Fig. 1). Two rolls of
fiber glass were used as the packing. Each roll was initially a sheet
of 0.234 m width; 6.3 m length and 0.0254 m thickness, which
was tightly rolled and maintained in the cylindrical shape with suf-
ficient distance between each layer (1 cm). The roll of packing had
a 0.248 m diameter, 0.234 m height, and the porosity of about 99%.
2.2. Electrical resistance tomography (ERT) technique
Electrical resistance tomography (ERT) technique was used in
the present study to explore the mixing characteristics in the

Fig. 1. Schematic diagram of the external loop airlift bioreactor utilized in this study.
 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61
packed bed external loop airlift bioreactor. This instrument creates
tomograms showing the cross-sectional distribution of the electri-
cal conductivity of the contents within the sensing zones [37]. This
technology involves the installation of a number of electrodes on
the periphery of the vessel to be imaged [32]. The main compo-
nents of the ERT are electrodes, a data acquisition system, and a
host computer for image reconstruction.
As shown in Fig. 2, sixteen equally spaced stainless steel rectan-
gular electrodes were fitted on the periphery of the bioreactor
(riser and downcomer) in one array (plane) and there were four
planes in the downcomer and four planes in the riser. Fig. 2 also
shows the locations and dimensions of electrodes that were used.
Plane 5 was installed in the middle of the axial height of the lower
packing, while plane 6 was located in the middle of the axial dis-
tance between the top of lower packing and the bottom of the
gas redistributor. Plane 7 was installed in the middle of the axial
distance between the bottom of the upper packing and the top of
the perforated plate while plane 8 was located in the middle of
the axial height of the upper packing. The distance between plane
5 and plane 6 was 0.175 m, the distance between plane 6 and plane
Fig. 2. Position and size of the tomography electrodes.
53
7 was 0.117 m, and the distance between plane 7 and plane 8 was
0.175 m. The distance between plane 5 and the bottom of the bio-
reactor was 0.764 m. Plane 1, plane 2, plane 3, and plane 4 were in-
stalled in the downcomer at the same level from the bottom of the
bioreactor as of plane 8, plane 7, plane 6, and plane 5, respectively.
All these electrodes were connected to the ITS P2000 data acquisi-
tion system (Manchester, UK) via co-axial cables. The data acquisi-
tion system (DAS) was used to send current to the electrodes fitted
on the internal walls of the bioreactor. The adjacent measurement
protocol was utilized in this study [26,38], in which the electrical
current was applied between an adjacent pair of electrodes and
the voltage was measured between all other pairs of adjacent elec-
trodes. Other combinations of adjacent electrode pairs went
through the same procedure until the full rotation was obtained
on the bioreactor. The frequency of 9,600 kHz and injection current
of 1.5 mA were employed for all experiments in this study. In this
study, a non-iterative linear back-projection algorithm [39] was
employed to create the tomograms, which showed the cross-sec-
tional distribution of the electrical conductivity of the contents
within the same measurement plane.
54 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61
2.3. Experimental procedure
The tracer response technique was used to measure the mixing
time in the bioreactor. Mixing time (tm) was calculated by subtract-
ing the time at which the tracer was injected from the time when
the variation in mean conductivity values for all eight tomography
planes simultaneously became ±1% of the steady-state value. In
fact, the conductivity values remained stable at the steady-state
due to complete dispersion of the tracer throughout the reactor.
Before the injection of the tracer, air was purged into the bioreactor
using a sparger at the bottom of the riser with constant flow rate.
The conductivity of the air–water mixture was monitored using
ERT system. After 50 s of supplying air, the steady state was
reached and consequently the conductivity of air–water mixture
became stable. After reaching the steady state, the tracer was in-
jected in the downcomer above the liquid level. In this study,
100 ml of 4% saline solution (NaCl) was used as a tracer. The vari-
ations in mean conductivity signals generated by the ERT system
were monitored on the host computer. The data collection was
stopped when the variations in mean conductivity values taken
from all four planes in the downcomer and all four planes in the
riser simultaneously became ±1% of the steady-state value [20].
The air was then discontinued and water was drained out. Same
procedure was repeated for different gas flow rates. Fig. 3 shows
the change in the mean conductivity of air–water mixture with
time at a riser superficial gas velocity of 1.1  102 m/s following
the injection of tracer for eight ERT planes. The mixing time was
30 s for the data shown in Fig. 3.
Same experimental procedure was used to find the superficial
liquid velocity in the riser as was employed for the mixing time
[6,10,40]. The riser superficial liquid velocity was calculated using
the ratio of the riser liquid circulation height to the riser liquid res-
idence time [14,41]. The following formula was used to find the
superficial liquid velocity in the riser:
Hr
U lr ¼ ð1  egr Þ
tr
where Hr (measured with reference to the top of the connecting
tube) is the circulating liquid height in the riser, tr is the liquid res-
idence time in the riser and egr is the gas holdup in the riser. The
residence time of liquid in the riser (tr) was determined by the fol-
lowing formula:
tm = 30 s
Conductivity (mS/cm)
Time (s)
Fig. 3. Measurement of the mixing time at Ugr = 1.1  102 m/s for cross shaped
sparger with liquid height = 1.63 m.
tc
tr ¼ ð2Þ
u
where tc is the circulation time and it is defined as the time required
for an elemental volume of liquid to circulate through the whole
path-length of the bioreactor. The dimensionless factor u can be
calculated using Eq. (3) and its derivation can be found in Bello [41]:
 
1  egd Ad Abc Hbc Atc Htc
u¼1þ þ þ ð3Þ
1  egr Ar Ar H r Ar H r
where egd is the gas holdup in the downcomer, egr is the gas holdup
in the riser, Ad is the cross-sectional area of the downcomer, Ar is the
cross-sectional area of the riser, Abc is the cross-sectional area of the
bottom connecting section, Atc is the cross-sectional area of the top
connecting section, Hr is the circulating liquid height in the riser, Hbc
is the circulating liquid length in the bottom connecting section and
Htc is the circulating liquid length in the top connecting section.
Substituting values of Ad, Abc, Atc, Hr, Hbc, and Htc in Eq. (3), the gen-
eral form of the dimensionless factor u for the bioreactor used in
this study becomes:
1  egd
u¼1þ ½0:224 þ 0:0551 ð4Þ
1  egr
Values for gas holdup in this bioreactor were measured and
published elsewhere [42]. The superficial liquid velocity in the
downcomer (Uld) was calculated from a mass balance on the liquid
phase [43,44] using Eq. (5):
Ar
U ld ¼ U lr ð5Þ
Ad
where Ulr is the superficial liquid velocity in the riser, Ad is the
cross-sectional area of the downcomer and Ar is the cross-sectional
area of the riser. Fig. 4 represents changes in conductivity with time
for the cross shaped sparger at superficial gas velocity of
1.1  102 m/s in the riser and liquid height of 1.63 m for plane 2
ð1Þ (located in the downcomer) following the injection of saline solu-
tion. The circulation time for the tracer was measured as the time
between two consecutive peaks (i.e., tc = 20 s from the data shown
in Fig. 4). Since the conductivity value of the second peak was much
lower compared to that for the first peak, the second peak cannot be
observed clearly for the given scale on y-axis in Fig. 4. Thus, to
obtain the circulation time accurately, we found the conductivity
Fig. 4. Measurement of the circulation time at Ugr = 1.1  102 m/s using ERT plane
2 (located in downcomer) for the cross shaped sparger with liquid height = 1.63 m.
 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61 55

values at these two peaks by zooming in to get a close-up view of
the peaks using Excel.
3. Results and discussion
Time series of tomographic images that were recorded from 8
measurement planes of the ERT system (Fig. 2) are shown in
Fig. 5. These images were recorded after the injection of the tracer
(i.e., 100 ml of 4% saline solution) from the top of the downcomer
near plane 1 at a constant riser superficial gas velocity (i.e.,
Ugr = 3.3  102 m/s) using the cross shaped sparger with a con-
stant liquid height (i.e., H = 1.63 m). It is a known fact that the con-
ductivity of saline solution is higher than the conductivity of water
in the bioreactor, thus the region with higher conductivity depicts
the higher concentration of saline solution in the bioreactor
[20,38,45]. The blue and red colors in Fig. 5 represent the lowest
and highest conductivity values, respectively.
It can be seen from Fig. 5 that the color of the tomographic
images of planes 1, 2, and 3 located in the downcomer partially
changed to red (higher conductivity value) at t = 2 s since the tracer
was injected near plane 1. The injected tracer reached plane 4 at
t = 3 s and the color of the tomogram of plane 4 located in the
downcomer changed to the red color. These changes to the red col-
or were due to the increase in conductivity of the fluid in the
downcomer. At t = 6 s, tomographic images recorded in planes 1,
2, 3, and 4 changed to green, which represents decrease in conduc-
tivity of the solution. This can be due to the dispersion of saline

Fig. 5. Time series of tomograms collected from 8 measurement planes of ERT system (plane 1, plane 2, plane 3 and plane 4 located on downcomer and plane 5, plane 6, plane
7 and plane 8 located on riser; see Fig. 2) following the injection of saline solution into the air–water mixture using the cross shaped sparger with Ugr = 3.3  102 m/s and
liquid height = 1.63 m (without packing). Blue and red colors represent lowest and highest conductivity values. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
56 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61

Fig. 5. (continued)

solution in the downcomer. After passing through the downcomer,
the saline solution entered into the riser using bottom connecting
tube. However, since the gas holdup in the riser is much higher, the
red regions on the tomograms of planes 5, 6, 7, and 8, located in the
riser from the bottom to top, were not observed. This variation in
conductivity indicates the presence of saline solution in the biore-
actor. The ERT images show that the mixing time under these cir-
cumstances was about 17 s.
Fig. 6 demonstrates the vertical slice of the tomographic images
and the 3D tomograms of air–water mixture in the riser and down-
comer of the bioreactor. It must be mentioned that these vertical
slices represent the vertical area of the riser from planes 5 to 8
and the vertical area of the downcomer from planes 1 to 4. Simi-
larly, images were recorded following the injection of the saline
solution from the top of the downcomer at a constant riser super-
ficial gas velocity of 3.3  102 m/s using the cross shaped sparger
at a constant liquid height of 1.63 m in the bioreactor. The green
color in the downcomer at t = 0 s was due to the lower gas holdup
in the downcomer than that in the riser. Thus, the blue region was
not observed in the downcomer. At t = 0 s, the saline solution was
injected into the top of the downcomer and the color in the down-
comer turned red at t = 2 s, which represents highest conductivity.
At t = 10 s, the color in the downcomer changed to green, which
represents lower conductivity. This can be due to the fact that
the saline solution had left the downcomer and entered into the ri-
ser through the bottom connecting tube. The blue color (lower con-
ductivity due to air bubbles) in the riser was slightly changed to
green color in Fig. 6 after 10 s, which indicates the presence of sal-
ine solution in the riser. The tracer was uniformly dispersed
throughout the bioreactor after 17 s. The blue colors observed in
 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61 57

Fig. 6. Vertical slice tomographic images and 3D tomograms of the air–water mixture in riser and downcomer of the bioreactor following the injection of the saline solution
at a constant riser superficial gas velocity of 3.3  102 m/s using the cross shaped sparger at a constant liquid height of 1.63 m in the bioreactor (without packing). Blue and
red colors represent lowest and highest conductivity values. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

Fig. 7. Effect of gas flow rate and liquid height on mixing time using cross shaped Fig. 8. Effect of sparger configuration on the mixing time with a constant liquid
sparger (without packing). height of 1.63 m in the bioreactor (without packing).
58 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61
Fig. 9. Effect of the packing in the riser (packing height = 0.234 m) on the mixing
time with a constant liquid height of 1.63 m using the cross shaped sparger.
Fig. 10. Effect of the gas redistributor installed in the riser on the mixing time with
a constant liquid height of 1.63 m using the cross shaped sparger.
Fig. 11. Effect of a gas redistributor installed between two beds of packing in the
riser (height of each packing = 0.234 m) on the mixing time with a constant liquid
height of 1.63 m using the cross shaped sparger.
planes 5–8 (Fig. 6) were due to the gas holdup in the riser. In fact,
the regions with blue color (regions of lower conductivity) repre-
sent the higher gas holdup in those areas.
Fig. 7 shows the effects of superficial gas velocity and liquid
height on the mixing time in the bioreactor using the cross shaped
sparger configuration without packing and gas redistributor. It can
be seen from this figure that mixing time in the bioreactor de-
Fig. 12. Effect of the gas flow rate and liquid height on the superficial liquid velocity
in the riser using the cross shaped sparger.
creased with increasing superficial gas velocity in the riser.
Increasing the superficial gas velocity in the riser tends to drive
an increase in the superficial liquid velocity in the riser, which re-
duces the mixing time in the bioreactor. Fig. 7 also shows that mix-
ing time in the riser increased with decreasing liquid height in the
bioreactor. Bentifraouine et al. [35] reported that the superficial li-
quid velocity decreased when the liquid height decreased due to a
decrease in the difference between the hydrostatic pressures in the
riser and the downcomer. Liu et al. [40] also found that with
decreasing the liquid height in the bioreactor, the flowing resis-
tance in the upper connecting tube increased, which resulted in
the reduction of the liquid superficial velocity. This decrease in
the liquid superficial velocity led to a higher mixing time in the
bioreactor.
Profiles for the mixing time of the cross shaped and the circular
sparger (Fig. 1) using different riser superficial gas velocities with
constant liquid height in the bioreactor are shown in Fig. 8. These
results demonstrate that at a constant riser superficial gas velocity,
the mixing time was increased when a circular sparger was used
over a cross shaped sparger of same diameter and with same num-
ber and size of holes. The increase in mixing time was due to the
presence of more resistance in the flow. The cross shaped sparger
consists of six circular tubes of small diameter. The liquid can eas-
ily pass through the space between these circular tubes. While a
circular sparger consists of a circular plate which offers less space
between the riser and the sparger for the liquid to pass and hence
offers more resistance to flow path of the liquid. In fact, the
Fig. 13. Effect of the sparger configuration on the superficial liquid velocity in the
riser with a constant liquid height of 1.63 m in the bioreactor.
 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61
Fig. 14. Effect of the packing installed in the riser (packing height = 0.234 m) on the
superficial liquid velocity in the riser with a constant liquid height of 1.63 m using
the cross shaped sparger configuration.
presence of the cross shaped sparger reduced the cross sectional
area available for flow by about 17% while the presence of the cir-
cular sparger reduced the cross sectional area available for flow by
about 65%. It is a known fact that the increase in resistance results
in a lower liquid superficial velocity [10,46,47], which in turn in-
creases the mixing time in the bioreactor.
The effect of packing in the riser on the mixing time was exam-
ined with the cross shaped sparger and constant liquid height of
1.63 m. It was observed from results in Fig. 9 that at a constant ri-
ser superficial gas velocity, the mixing time was increased when a
packing was installed in the riser. The installation of packing in the
riser offers more resistance to the flow path of the liquid in the ri-
ser which results in the decrease in the liquid superficial velocity
[24,36]. This decrease in liquid superficial velocity in turn increases
the mixing time in the bioreactor. However, it is still important to
install packing in a bioreactor for processes where packing is used
as a support for immobilized microorganisms or enzymes [48].
Mixing time in the bioreactor was measured at different super-
ficial gas velocities in the presence and absence of the gas redis-
tributor (without packing). It can be noticed from Fig. 10 that at
a constant riser superficial gas velocity, the mixing time in the bio-
reactor equipped with an internal gas redistributor was higher
than that in the bioreactor without the gas redistributor. Some
investigators studied the liquid velocity in the fluidized bed and in-
verse fluidized bed external loop airlift bioreactors [10,46,47]. They
observed that the liquid velocity in the bioreactor with fluidized
bed was lower than that without any fluidized bed due to the resis-
Fig. 15. Effect of the gas redistributor installed in the riser on the superficial liquid
velocity in the riser with a constant liquid height of 1.63 m for the cross shaped
sparger configuration.
59
Fig. 16. Effect of using a gas redistributor between two beds of packing in the riser
(height of each packing = 0.234 m) on the riser superficial liquid velocity with a
constant liquid height of 1.63 m using a cross shaped sparger configuration.
tance in the flow path of the liquid. When an internal gas redistrib-
utor is installed in the riser of the bioreactor, it enhances the flow
resistance, which in turn decreases the liquid superficial velocity.
This decrease in liquid superficial velocity delays the time required
for mixing in the bioreactor. This delay results in an increase in
mixing time in the bioreactor.
The effect of the gas redistributor on the mixing time in the bio-
reactor in the presence of the packing in the riser is summarized in
Fig. 11 (with a constant liquid height and using the cross shaped
sparger). Clearly, the results show an increase in mixing time when
the gas redistributor was installed between the two rolls of pack-
ing at a constant superficial gas velocity in the riser. This increase
in mixing time was due to the presence of more resistance in the
form of packing and a gas redistributor in the flow path of the li-
quid in the riser. The presence of this resistance in the liquid flow
contributes to lower liquid superficial velocity [10,24,36,46,47],
which in turn increases the mixing time in the bioreactor.
Fig. 12 represents the effects of the superficial gas velocity and
liquid height on the superficial liquid velocity in the riser using the
cross shaped sparger. It can be seen that the superficial liquid
velocity in the riser was increased with increasing the superficial
gas velocity in the riser. The superficial liquid velocity in the biore-
actor depends on the hydrostatic pressure difference between the
riser and the downcomer of the bioreactor. When the gas flow rate
is increased in the riser, the hydrostatic pressure difference be-
tween the riser and the downcomer increases, which in turn in-
creases the superficial liquid velocity in the riser. In the earlier
investigations conducted by Han et al. [10] and Kemblowski
et al. [16], Loh and Liu [6], Mohanty et al. [25], they all agreed on
that when the gas flow rate in the bioreactor is increased, the
superficial liquid velocity also increases. It can also be seen from
Fig. 12 that the superficial liquid velocity in the riser was decreased
with decreasing liquid height in the bioreactor. At constant super-
ficial gas velocity, when the liquid height in the bioreactor is de-
creased, the resistance in the path of the liquid flowing in the
upper connecting tube of the bioreactor is increased. This increased
resistance, in turn, decreases the liquid superficial velocity in the
bioreactor [33–35,40].
At different superficial gas velocities, Fig. 13 illustrates the ef-
fect of gas sparger type on the superficial liquid velocity in the riser
at a constant liquid height in the bioreactor. It was noticed that the
riser superficial liquid velocity was decreased at constant riser
superficial gas velocity when a circular sparger was used over a
cross shaped sparger of same diameter and with same number
and size of holes on it. This decrease in superficial liquid velocity
in the riser is due to the fact that the circular sparger consists of
a circular solid plate, which offers more resistance than its
60 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61
counterpart (cross shaped sparger). A cross shaped sparger consists
of six small diameter circular tubes. These small diameter tubes
have much more space between each other for the water to pass
easily without much resistance. In fact, the presence of the cross
shaped sparger reduced the cross sectional area available for flow
by about 17% while the presence of the circular sparger reduced
the cross sectional area available for flow by about 65%. Increasing
resistance in the flow path of the liquid contributes to lower super-
ficial liquid velocity in the bioreactor [10,46,47]. Moreover, the
holes in the cross shaped sparger are located throughout the cross
section of the sparger; however, in a circular sparger, they are lo-
cated only at the periphery of the plate. Therefore, the cross shaped
sparger can disperse air more evenly across column of the riser
than that of the circular sparger.
Fig. 14 demonstrates the effect of packing on the superficial li-
quid velocity in the riser using the cross shaped sparger with a con-
stant liquid height of 1.63 m in the bioreactor. It was observed that
the superficial liquid velocity in the riser decreased when the pack-
ing was installed in the riser at a constant riser superficial gas
velocity. The installation of packing in the riser reduces the cross
section area available for flow and offers more resistance to the
flow path of liquid, which results in the decrease in liquid superfi-
cial velocity [24,36].
Fig. 15 is a typical plot showing the effect of the superficial gas
velocity on the superficial liquid velocity in the riser, with and
without the gas redistributor. It can be seen that the superficial li-
quid velocity in the riser of the bioreactor fitted with the internal
gas redistributor at a constant superficial gas velocity was lower
than that measured for the bioreactor without internal gas redis-
tributor. The lower superficial liquid velocity in the riser was due
to the presence of internal gas redisitributor, which reduced the
cross-sectional area available for flow and increased flow resis-
tance in the bioreactor equipped with the internal gas redistributor
[10,46,47].
Fig. 16 illustrates the effect of the internal gas redistributor in-
stalled in the riser between the two rolls of packing on the super-
ficial liquid velocity in the riser with a constant liquid height of
1.63 m for the cross shaped sparger. It was noticed that the riser
superficial liquid velocity decreased, at constant riser superficial
gas velocity, when an internal gas redistributor was installed be-
tween two rolls of packing in the riser. A bioreactor with an inter-
nal gas redistributor between two beds of packing provides less
cross sectional area available for flow and offers much more resis-
tance in the flow path of the liquid than that without the internal
gas redistributor and packing. This increase in resistance in the
flow path of the liquid contributes to lower liquid superficial veloc-
ity in the bioreactor [10,36,46,47].
In general, the superficial liquid velocity in the downcomer was
higher than the superficial liquid velocity in the riser because of
the smaller diameter of the downcomer. The range of superficial li-
quid velocity found in the downcomer was 0.103–0.785 m/s.
Experimental results of this study indicated that the gas redistrib-
utor significantly decreased the superficial liquid velocity and in-
creased the mixing time. However, higher gas holdup can be
achieved with the installation of the gas redistributor between
two beds of packing in the bioreactor [42]. Higher gas holdup
improves the mass transfer inside the reactor by increasing the
interfacial mass transfer area. The installation of the gas redistrib-
utor also enhances the uniform distribution of the smaller gas bub-
bles in the riser.
4. Conclusion
In this study, electrical resistance tomography (ERT) was suc-
cessfully employed to explore the hydrodynamics of mixing in an
external loop airlift bioreactor equipped with an internal gas redis-
tributor between two beds of packing in the riser. To characterize
the mixing performance of this bioreactor, ERT data were collected
after the injection of the tracer to measure the mixing time and the
superficial liquid velocity in the bioreactor. To measure the distri-
bution of the tracer within the riser and the downcomer, the adja-
cent measurement strategy and linear back-projection algorithm
were employed. The experimental results demonstrated that the
mixing time in the bioreactor decreased with an increase in the
superficial gas velocity in the riser. When the gas flow rate was in-
creased in the riser, the hydrostatic pressure difference between
the riser and the downcomer was increased, which in turn in-
creased the superficial liquid velocity in the riser. This increase in
the superficial liquid velocity in the riser decreased the mixing
time in the bioreactor. The ERT results also revealed that the mix-
ing time was increased and the superficial liquid velocity was de-
creased by reducing the liquid height in the bioreactor, installing
the internal gas redistributor and/or installation of packing in the
riser, and by using the circular sparger. This was due to the in-
crease in the resistance in the liquid flow path.
Acknowledgments
The financial supports of the Natural Science and Engineering
Research Council of Canada (NSERC), the Agriculture and Agri-Food
Canada through the Agricultural Bioproducts Innovation Network
(ABIN), and the Ryerson University are gratefully acknowledged.
References
[1] Z. Al-Qodah, W. Lafi, Modelling of antibiotics production in magneto three-
phase airlift fermenter, Biochem. Eng. J. 7 (2001) 7–16.
[2] S. Sarkar, K. Mohanty, B.C. Meikap, Hydrodynamic modeling of a novel multi-
stage gas–liquid external loop airlift reactor, Chem. Eng. J. 145 (2008) 69–77.
[3] T. Zhang, J. Wang, T. Wang, J. Lin, Y. Jin, Effect of internal on the hydrodynamics
in external-loop airlift reactors, Chem. Eng. Process. 44 (2005) 81–87.
[4] F. Benyahia, L. Jones, Scale effects on hydrodynamic and mass transfer
characteristics of external loop airlift reactors, J. Chem. Technol. Biotechnol.
69 (3) (1997) 301–308.
[5] J. Klein, J. Maia, A.A. Vicente, L. Domingues, J.A. Teixeira, M. Jurascik,
Relationships between hydrodynamics and rheology of flocculating yeast
suspensions in a high-cell-density airlift bioreactor, Biotechnol. Bioeng. 89 (4)
(2005) 393–399.
[6] K. Loh, J. Liu, External loop inversed fluidized bed airlift bioreactor (EIFBAB) for
treating high strength phenolic wastewater, Chem. Eng. Sci. 56 (2001) 6171–
6176.
[7] K. Mohanty, D. Das, M.N. Biswas, Treatment of phenolic wastewater in a novel
multi-stage external loop airlift reactor using activated carbon, Sep. Purif.
Technol. 58 (2008) 311–319.
[8] M. Gavrilescu, R.Z. Tudose, Mixing studies in external-loop airlift reactors,
Chem. Eng. J. 66 (1997) 97–104.
[9] F. Gumery, F. Ein-Mozaffari, Y. Dahman, Characteristics of local flow dynamics
and macro-mixing in airlift column reactors for reliable design and scale-up,
Int. J. Chem. Reac. Eng. 7 (2009). Review R4.
[10] S.J. Han, R.B.H. Tan, K.C. Loh, Hydrodynamic behaviour in a new gas liquid solid
inverse fluidization airlift bioreactor, Trans. IChemE Part C 78 (2000) 207–215.
[11] P.M. Kilonzo, A. Margaritis, M.A. Bergougnou, Hydrodynamics and mass
transfer characteristics in an inverse internal loop airlift-driven fibrous-bed
bioreactor, Chem. Eng. J. 157 (2010) 146–160.
[12] P.M. Kilonzo, A. Margaritis, M.A. Bergougnou, Hydrodynamic characteristics in
an inverse internal-loop airlift-driven fibrous-bed bioreactor, Chem. Eng. Sci.
65 (2010) 692–707.
[13] W. Tang, L. Fan, Steady state phenol degradation in a draft-tube, gas-liquid-
solid fluidized-bed bioreactor, AIChE J. 33 (2) (1987) 239–249.
[14] R.A. Bello, C.W. Robinson, M. Moo-Young, Liquid circulation and mixing
characteristics of airlift contactors, Can. J. Chem. Eng. 62 (1984) 573–577.
[15] H. Dhaouadi, S. Poncin, J.M. Hornut, G. Wild, Solid effects on hydrodynamics
and heat transfer in an external loop airlift reactor, Chem. Eng. Sci. 61 (2006)
1300–1311.
[16] Z. Kemblowski, J. Przywarski, A. Diab, An average gas holdup and liquid
circulation velocity in airlift reactors with external loop, Chem. Eng. Sci. 48
(1993) 4023–4035.
[17] K. Mohanty, D. Das, M.N. Biswas, Mass transfer characteristics of a novel multi-
stage external loop airlift reactor, Chem. Eng. J. 133 (2007) 257–264.
[18] K. Shimizu, S. Takada, T. Takahashi, Y. Kawase, Phenomenological simulation
model for gas holdups and volumetric mass transfer coefficients in external-
loop airlift reactors, Chem. Eng. J. 84 (2001) 599–603.
 M. Hamood-ur-Rehman et al. / Chemical Engineering Journal 213 (2012) 50–61
[19] F. Yazdian, S.A. Shojaosadati, M. Nosrati, M.P. Hajiabbas, E. Vasheghani-
Farahani, Investigation of gas properties, design, and operational parameters
on hydrodynamic characteristics, mass transfer, and biomass production from
natural gas in an external airlift loop bioreactor, Chem. Eng. Sci. 64 (2009)
2455–2465.
[20] F. Gumery, F. Ein-Mozaffari, Y. Dahman, Macromixing hydrodynamic study in
draft-tube airlift reactors using electrical resistance tomography, Bioprocess.
Biosyst. Eng. 34 (2011) 135–144.
[21] B. Jin, Q. Yu, Q.X. Yan, J. Leeuwen, Characterization and improvement of
oxygen transfer in pilot plant external air-lift bioreactor for mycelia biomass
production, World J. Microbiol. Biotechnol. 17 (2001) 265–272.
[22] Y. Kawase, N. Omori, M. Tsujimura, Liquid-phase mixing in external-loop airlift
bioreactors, J. Chem. Technol. Biotechnol. 61 (1994) 49–55.
[23] Z. Xu, J. Yu, Hydrodynamics and mass transfer in a novel multi-airlifting
membrane bioreactor, Chem. Eng. Sci. 63 (2008) 1941–1949.
[24] A.X. Meng, G.A. Hill, A.K. Dalai, Hydrodynamic characteristics in an external
loop airlift bioreactor containing a spinning sparger and a packed bed, Ind. Eng.
Chem. Res. 41 (2002) 2124–2128.
[25] K. Mohanty, D. Das, M.N. Biswas, Hydrodynamics of a novel multi-stage
external loop airlift reactor, Chem. Eng. Sci. 61 (2006) 4617–4624.
[26] S. Hosseini, D. Patel, F. Ein-Mozaffari, M. Mehrvar, Study of solid–liquid mixing
in agitated tanks through electrical resistance tomography, Chem. Eng. Sci. 65
(2010) 1374–1384.
[27] L.K. Ishkintana, C.P.J. Bennington, Gas holdup in pulp fibre suspensions: gas
voidage profiles in a batch-operated sparged tower, Chem. Eng. Sci. 65 (2010)
2569–2578.
[28] H. Jin, M. Wang, R.A. Williams, The effect of sparger geometry on gas bubble
flow behaviors using electrical resistance tomography, Chin. J. Chem. Eng. 14
(1) (2006) 127–131.
[29] H. Jin, S. Yang, M. Wang, R.A. Williams, Measurement of gas holdup profiles in
a gas liquid cocurrent bubble column using electrical resistance tomography,
Flow Meas. Instrum. 18 (2007) 191–196.
[30] H. Jin, M. Wang, R.A. Williams, Measurement of gas holdup profiles in a gas
liquid cocurrent bubble column using electrical resistance tomography, Chem.
Eng. J. 130 (2007) 179–185.
[31] P. Tahvildarian, H. Ng, M. D’Amato, S. Drappel, F. Ein-Mozaffari, S.R. Upreti,
Using electrical resistance tomography images to characterize the mixing of
micron-sized polymeric particles in a slurry reactor, Chem. Eng. J. 172 (2011)
517–525.
[32] R.A. Williams, M.S. Beck, Process Tomography: Principles, Techniques and
Applications, Butterworth-Heinemann, Boston, 1995.
[33] W.A. Al-Masry, Effect of liquid volume in the gas-separator on the
hydrodynamics of airlift reactors, J. Chem. Technol. Biotechnol. 74 (1999)
931–936.
[34] W.A. A1-Masry, Influence of gas-separator and scale up on the hydrodynamics
of external loop circulating bubble columns, Chem. Eng. Res. Des. 82 (A3)
(2004) 381–389.
61
[35] C. Bentifraouine, C. Xuereb, J. Riba, Effect of gas liquid separator and liquid
height on the global hydrodynamic parameters of an external loop airlift
contactor, Chem. Eng. J. 66 (1997) 91–95.
[36] H. Nikakhtari, G.A. Hill, Hydrodynamic and oxygen mass transfer in an
external loop airlift bioreactor with a packed bed, Biochem. Eng. J. 27 (2005)
138–145.
[37] Z.F. Zhao, M. Mehrab, F. Ein-Mozaffari, Mixing time in an agitated multi-lamp
cylindrical photoreactor using electrical resistance tomography, J. Chem.
Technol. Biotechnol. 83 (2008) 1676–1688.
[38] L. Pakzad, F. Ein-Mozaffari, P. Chan, Using electrical resistance tomography and
computational fluid dynamics modeling to study the formation of cavern in
the mixing of pseudoplastic fluids possessing yield stress, Chem. Eng. Sci. 63
(2008) 2508–2522.
[39] A. Madupu, A. Mazumdar, Z. Jinsong, D. Roelant, R. Srivastava, Electrical
resistance tomography for real-time mapping of the solid-liquid interface in
tanks containing optically opaque fluids, Proc. SPIE – Int. Soc. Opt. Eng. 5674
(2005) 36–46.
[40] M. Liu, T. Zhang, T. Wang, W. Yu, J. Wang, Experimental study and modeling on
liquid dispersion in external-loop airlift slurry reactors, Chem. Eng. J. 139
(2008) 523–531.
[41] R.A. Bello, A characterization study of airlift contactors for application to
fermentations. PhD Thesis, Department of Chemical Engineering, University of
Waterloo, Canada, 1981.
[42] M. Hamood-ur-Rehman, F. Ein-Mozaffari, Y. Dahman, Dynamic and local gas
holdup studies in external loop recirculating airlift reactor with two rolls of
fibreglass packing using electrical resistance tomography, J. Chem. Technol.
Biotechnol., in press doi: http://dx.doi.org/10.1002/jctb.3917.
[43] A.H. Essadki, M. Bennajah, B. Gourich, C. Vial, M. Azzi, H. Delmas,
Electrocoagulation/electroflotation in an external-loop airlift reactor—
application to the decolorization of textile dye wastewater: a case study,
Chem. Eng. Process. 47 (2008) 1211–1223.
[44] M. Popovic, C.W. Robinson, Mixing characteristics of external-loop airlifts:
non-newtonian systems, Chem. Eng. Sci. 48 (8) (1993) 1405–1413.
[45] G.T. Bolton, C.W. Hooper, R. Mann, E.H. Stitt, Flow distribution and velocity
measurement in a radial flow fixed bed reactor using electrical resistance
tomography, Chem. Eng. Sci. 59 (2004) 1989–1997.
[46] Y.X. Guo, M.N. Rather, H.C. Ti, Hydrodynamics and mass transfer studies in a
novel external loop airlift reactor, Chem. Eng. J. 67 (1997) 205–214.
[47] K. Loh, S. Ranganath, External-loop fluidized bed airlift bioreactor (EFBAB) for
the cometabolic biotransformation of 4-chlorophenol (4-cp) in the presence of
phenol, Chem. Eng. Sci. 60 (2005) 6313–6319.
[48] J. Horiuchi, K. Tabata, T. Kanno, M. Kobayashi, Continuous acetic acid
production by a packed bed bioreactor employing charcoal pellets derived
from waste mushroom medium, J. Biosci. Bioeng. 89 (2) (2000) 126–130.