Advances in Chemical Engi neering and Science , 2011, 1, 239-244
doi:10.4236/aces.2011.14034 Published Online October 2011 (http://www.SciRP.org/journal/aces)
Evaluation of the Inverse Fluidized Bed Biological Reactor
for Treating High-Strength Industrial Wastewaters
Włodzimierz Sokół, Belay Woldeyes
Department of C hemi c al Engineering, Addis Abab a U ni vers i t y , Addis Ababa, Ethiopia
Recieved July 7, 2011; revised September 16, 2011; accepted Septembe r 22, 2011
Abstract
The aim of this work was to investigate the aerobic degradation of high-strength industrial (refinery) waste-
waters in the inverse fluidized bed biological reactor, in which polypropylene particles of density 910 kg/m3
were fluidized by an upward flow of gas through a bed. Measurements of chemical oxygen demand (COD)
versus residence time t were performed for various ratios of settled bed volume to reactor volume (Vb/VR)
and air velocities u. The largest COD reduction, namely, from 54,840 to 2190 mg/l, i.e. a 96% COD de-
crease, was achieved when the reactor was operated at the ratio (Vb/VR) = 0.55, air velocity u = 0.046 m/s and
t = 65 h. Thus, these values of (Vb/VR), u and t can be considered as the optimal operating parameters for a
reactor when used in treatment of high-strength refinery wastewaters. In the treatment operation conducted in
a reactor optimally controlled at (Vb/VR) = 0.55, u = 0.046 m/s and t = 65 h, the conversions obtained for all
phenolic constituents of the wastewater were larger than 95%. The conversions of about 90% were attained
for other hydrocarbons.
Keywords: Biological Wastewater Treatment, Aerobic Wastewater Treatment, High-Strength Industrial
Wastewaters, Inverse Biological Reactor, Fluidized Bed Bioreactor, Low-Density Biomass
Support
1. Introduction
The application of a fluidized bed technology to biologi-
cal wastewater treatment has brought a remarkable
breaktrough. The technology owes its high-rate success
to much higher surface area and biomass concentration
than those that can be achieved in the conventional
treatment processes. A fluidized bed biological reactor
(FBBR) has attracted considerable interest as an alterna-
tive to the conventional suspended growth and fixed-film
wastewater treatment processes due to its high efficiency
performance.
Treatment of industrial wastewaters requires a great
deal of space when using systems based on activated
sludge in which the retention time is many days [1]. On
the other hand, a FBBR is capable of achieving treatment
in low retention time because of the high biomass con-
centrations that can be achieved in the apparatus [2,3].
Among the compounds of refinery wastewaters that
are the greatest biohazards to the environment are phe-
nols (monohydric and polyhydric), derivatized phenols,
polycyclic organics (polynuclear aromatic hydrocarbons),
thiocyanates, ammonia and cyanides [4].
A three-phase (gas-liquid-solid) FBBR has been suc-
cessfully applied in aerobic biological treatment of in-
dustrial and municipal wastewaters [5-8]. The reactor
outperforms other reactor configurations used in waste-
water treatment such as the activated sludge system and
packed-bed (or trickling-filter) reactor [9-12]. The supe-
rior performance of the FBBR stems from the very high
biomass concentration (up to 30 - 40 kg/m3) that can be
achieved due to immobilisation of cells onto or into the
solid particles.
However, the excessive growth of biomass on support
media can lead to the channelling of bioparticles (support
particles covered by biomass) in fluidized bed since the
biomass loading can increase to such an extent that the
bioparticles began to be carried over from a reactor. The
application of a low density (matrix particle density
smaller than that of liquid) biomass support in a reactor
allows the control of biomass loading and provides the
high oxygen concentration in the reacting liquid media
[2,13].
In a FBBR containing low-density particles, fluidiza-

240 W. SOKÓŁ ET AL.
tion can be conducted either by an upward co-current
flow of gas and liquid through a bed (Figure 1) [2,3] or
by a downward flow of liquid and countercurrent upward
flow of gas [5,14]. In the former, fluidization is achieved
by an upward flow of gas whereby the gas bubbles make
the bed expanding downwards into the less dense mix-
ture of gas and liquid. In the latter, the bed is fluidized by
a downward flow of a liquid counter to the net buoyancy
force of the particles. Such type of fluidization is termed
the inverse fluidization.
Sokół and Halfani [13] have reported that a steady-
state biomass loading was obtained in a FBBR, in which
polypropylene particles of density 910 kg/m3 were fluid-
ized by an upward co-current flow of gas and liquid.
Karamanev et al. [14] have achieved a constant biomass
loading in a FFBR, in which low density particles were
fluidized by downflow of the liquid. Rusten et al. [15]
have demonstrated that a steady-state biomass loading
was attained in a FFBR containing low-density particles
made of polyethylene.
The aim of this work was to investigate the aerobic
degradation of high strength wastewaters in a FFBR, in
which polypropylene particles of density 910 kg/m3 were
fluidized by an upward flow of gas through a bed. Ex-
periments on COD reduction were performed for various
ratios of settled bed volume to reactor volume (Vb/VR),
air velocities u and residence times t.
Figure 1. Scheme of the inverse fluidized bed biological re-
actor.
2. Experimentation
2.1. Experimental Set-up
Experiments were conducted in the reactor shown in
Figure 2. A growing medium, stored in a reservoir 1,
was pumped into the bottom of the reactor by a centrifu-
gal pump 5. Before entering the bed, the liquid was
mixed with air by means of a sparger. The air was intro-
duced to the bed through a distributor 7 whose plate had
200 holes of 4 mm diameter on a triangular pitch.
Figure 2. Schematic diagram of the experimental apparatus: 1, reservoir; 2, temperature control system; 3, pH control system; 4,
liquid rotameter; 5, pump; 6, intermediate reservoir; 7, air distributor; 8, sampling; 9, fluidized section; 10, disengaging section;
11, air rotameter.

W. SOKÓŁ ET AL.
241
The fluidised bed section 9, made of Duran glass, had a
20 cm internal diameter and was 6 m high. It was ended
by a disengaging cap 10 with a 60 cm internal diameter
and a height of 80 cm. The biomass sloughed off from
the particles was separated from the effluent in a vessel 6
and removed from the system. The flow rate of the liquid
was measured by a rotameter 4 and controlled by a ball
valve. The air flow rate was measured using a rotameter
11 and controlled by a needle valve. The pH was ad-
justed by a control system 3, consisting of a pH-meter
and micro-pumps supplying base or acid; as required.
The temperature control system 2 consisted of a coil with
cold water and an electric heater coupled with a contact
thermometer.
The biomass support was the polypropylene particles
of density 910 kg/m3 whose dimensions are given in
Figure 3.
2.2. Feed and Microorganisms
The growing medium was the wastewater whose com-
position is given in Table 1. The wastewater was en-
riched in mineral salts by adding the following (mg/l):
(NH4)2SO4: 500; KH2PO4: 200; MgCl2: 30; NaCl: 30;
CaCl2: 20; and FeCl3: 7 [1].
The inoculum was the activated sludge taken from the
biological treatment unit operated at the refinery from
which wastewater was used in this research.
2.3. Methodology
Sokół and Korpal [1] have established that the optimal
ratio (Vb/VR) for a FBBR when used in biological waste-
water treatment was equal to 0.55. Therefore, in this
study experiments were performed for the ratios (Vb/VR)
equal to 0.50, 0.55 and 0.60. This was to cover the
searched range of (Vb/VR) from 0.50 to 0.60 in step 0.05
which is sufficient accuracy for industrial practice.
The air velocities u applied in experiments are given in
Figures 4-6.
2.3.1. Biomass Culturing
The particles and the growing medium were introduced
into the reactor to give a ratio (Vb/VR) = 0.50. To start
growth of the microorganisms on the particles, a batch
culture was first initiated by introducing about 15 l of the
inoculum into the reactor. Then the culture was incu-
bated for approximately 48 h to encourage cell growth
and the adhesion of freely suspended biomass on the
particles. The air was supplied at a flow rate of 0.025
m3/s and this was found to be sufficient for biomass
growth [1,2]. The pH was controlled in the range 6.5 -
7.0 and the temperature was maintained at 28˚C - 30˚C.
Figure 3. Dimensions (in mm) of biomass support.
Table 1. Composition of feed (wastewater) and effluent
from a reactor optimally controlled at (Vb/VR)m = 0.55, um =
0.046 m/s and t = 65 h.
Constituent Concentration
×103 mg/l Fractional
conversion
Feed Effluent
o-Cresol
m-Cresol
3,5 Dimethylphenol
Phenol
2,4 Dimethylphenol
Benzene
Toluene
3,4-Dimethylphenol
Isopropylphenol
o-Xylene
2,6-Dimethylphenol
C3-Phenyl
Ethylphenol
C4-Phenyl
15250
8240
7050
5740
4760
3450
2972
2190
2180
1130
980
240
190
70
597.8
326.2
421.4
174.6
286.2
102.5
88.6
107.2
87.4
57.9
67.3
16.4
5.6
6.3
0.96
0.96
0.94
0.97
0.94
0.97
0.97
0.95
0.96
0.95
0.93
0.93
0.97
0.91
Figure 4. Relationship between chemical oxygen demand
COD and residence time t for ratio (Vb/VR) = 0.50 and
various air velocities u.
When the biofilm had begun to grow on the particles,
the growing medium was started to be pumped into the
reactor at a dilution rate D = 0.15 h–1. This value of D
corresponded to the smallest time t applied for the ratio
(Vb/VR) = 0.50 (t = 1/D = 6.67 h in Figure 4). Next, the
air velocity u was set at the smallest value applied for the
(Vb/VR) = 0.50 (u = 0.022 m/s in Figure 4) and the culti-
vation was continued until the constant biomass loading

242 W. SOKÓŁ ET AL.
was achieved in a reactor. The occurrence of the steady-
state biomass loading was established by weighting the
mass of cells grown on the support. The biomass was
scraped from the sample particles and dried at tempera-
ture 105˚C for 1 h. It was considered that the steady state
occurred when the weight of biomass in two consecutive
samples differed less than 5%. The steady-state biomass
loading was attained in a reactor after the cultivation for
approximately two weeks.
2.3.2 Treatment Operation
When the steady-state biomass loading was achieved, a
sample liquid was withdrawn from the reactor and COD
was measured by the procedure recommended by Ver-
straete and van Vaerenbergh[16]. It was established that
once the constant biomass loading occurred in a reactor,
the value of COD was practically at steady state.
Next, the air velocity u was increased stepwise to its
next value applied for (Vb/VR) = 0.50 (u = 0.028 m/s in
Figure 4) and the cultivation was continued until the
new steady-state biomass loading was obtained. When
this was attained, a value of COD was measured by the
method mentioned earlier [16]. These experiments for
(Vb/VR) = 0.50 were conducted for all values of u shown
in Figure 4.
Then the dilution rate D was decreased stepwise to its
next value applied for (Vb/VR) = 0.50 (t = 1/D = 13.34 h
in Figure 4) and the air velocity u was re-set to its
smallest value applied for the (Vb/VR) = 0.50 (u = 0.022
m/s in Figure 4). The cultivation was continued until the
steady-state biomass loading was achieved. When this
occurred, a value of COD was measured following the
procedure mentioned earlier [16]. These experiments
were conducted for all air velocities u and residence
times t shown in Figure 4. The results of the experiments
are given in Figure 4.
The above experiments were also performed for the
ratios (Vb/VR) equal to 0.55 and 0.60. In order to get the
ratio (Vb/VR) = 0.55, an adequate volume of biomass-free
particles was added to a reactor at the end of experiment-
tation for (Vb/VR) = 0.50. Similarly, the ratio (Vb/VR) =
0.60 was obtained by the addition of fresh particles to a
reactor at the end of experimentation for (Vb/VR) = 0.55.
The results of the experiments are shown in Figures 5
and 6.
In order to determine the largest COD removal, ex-
periments on COD reduction were performed for those
values of um for which the largest COD reductions were
achieved in runs shown in Figures 4-6. The results of the
experiments are given in Figure 7.
It should be pointed out that the air velocities u ap-
plied in the experiments were several times larger than
the minimum fluidization velocity uf. This was possible
Figure 5. Dependence of chemical oxygen demand COD on
residence time t for ratio (Vb/VR) = 0.55 and various air
velocities u.
Figure 6. Relationship between chemical oxygen demand
COD and residence time t for ratio (Vb/VR) = 0.60 and
various air velocities u.
Figure 7. Dependence of chemical oxygen demand COD on
residence time t for treatment operation conducted in a
reactor controlled at values of (Vb/VR)m and um for which
the greatest COD removals were obtained in runs shown in
Figures 4-6.
because the reactor was operated at the ratios (Vb/VR)
smaller than the critical values of (Vb/VR)cr [13]. At the
ratios (Vb/VR) equal to, or larger than, (Vb/VR)cr, the

W. SOKÓŁ ET AL.
243
movement of the whole bed was impossible: the particles
either remained at the top of the reactor or they settled at
its bottom. On the other hand, the air velocities u were
smaller than the critical velocity ucr at which the entire
bed settled at the reactor bottom.
Stratification of the particles coated with the biomass
led to their movement to the base of the bed where con-
centrations of constituents of the wastewater were the
highest. This was desirable since the constituents could
penetrate far into the biofilm so that most of the biomass
was active [1,2].
3. Results and Discussion
It can be seen in Figures 4-6 that, for a set time t and
ratio (Vb/VR), a concentration of COD in effluent de-
pended on the air velocity u. A reduction in COD ini-
tially increased, and then decreased with an increase in u.
It can be noticed, for example, in Figure 4 that for a set t
the values of COD were decreasing with an increase in u
up to 0.041 m/s. The smallest value of COD was at-
tained for u = 0.041 m/s. For the air velocities u larger
than 0.041 m/s, the values of COD were increasing with
an increase in u. This can be explained by the fact that
with an increase in u up to 0.041 m/s, an interfacial
(air-liquid) area increased, and consequently the amount
of the oxygen supplied for biomass growth increased [1].
Thus, for the u smaller than 0.041 m/s, oxygen was the
limiting factor for biomass growth. On the other hand,
for the air velocities greater than 0.041 m/s, the degrada-
tion rate of the constituents of the wastewaters was the
controlling factor of the treatment process [3].
The value of um for which the largest decrease in COD
was obtained for a set t, depended on the ratio (Vb/VR),
and hence on volume Vb of the particles applied in the
reactor (Figures 4-6). With increasing Vb, the value of um
increased. Thus, a large volume of the particles can lead
to an increase in the amount of the air required for bio-
mass growth, and consequently to an increase in the re-
sulting energy cost [13].
It can be noted in Figures 4-6 that for set t and u, a
decrease in COD values depended on the ratio (Vb/VR).
The largest reduction in COD was attained at (Vb/VR) =
0.55. An increase in COD removal with an increase in
the (Vb/VR) from 0.50 to 0.55 can be attributed to the fact
that for increasing (Vb/VR), more biomass grown on the
particles participated in degradation of the constituents of
the wastewater. On the other hand, a decrease in COD
removal observed with an increase in (Vb/VR) from 0.55
to 0.60 was due to the fact that in this case, a significant
volume of the reactor was occupied by the particles, and
consequently the aeration characteristics of the bed had
worsened [2].
It can be noticed in Figure 7 that the values of COD
were practically at steady state for times t greater than 65
h. The largest COD removal occurred when the reactor
was operated at (Vb/VR)m = 0.55 and um = 0.046 m/s. A
decrease in COD from 54,840 to 2190 mg/l, that is, a
96% COD removal, was achieved when a reactor was
optimally controlled at (Vb/VR)m = 0.55, um = 0.046 m/s
and t = 65 h.
The biomass loading was successfully controlled in a
reactor containing low density particles used as biomass
support. This was due to particle geometry and particu-
larly availability of the internal surface and the grooves
on external surface of the particles for biomass growth.
With such geometry of the particles, shear forces occur
ring between the particles and the liquid sloughed off
excess of biomass mainly from the external, and to less
extend from the internal, surface of the particles. Fur-
thermore, the attrition, associated with particle- particle
and particle-wall collisions, sloughed off biomass grown
in the grooves and on the internal surface was less abrupt
than the cells grown on the external surface of the parti-
cles.
In the first continuous culture, conducted at (Vb/VR) =
0.50 after switching from batch to continuous culture, the
steady-state biomass loading was achieved after culture-
ing for about 12 days. In the continuous cultures per-
formed after change in (Vb/VR) at a set u, the constant
mass of cells grown on the support media was achieved
after approximately 6 days of operation. With change in
u at a set (Vb/VR), the new steady-state biomass loading
occurred after the culturing for about 4 days.
4. Conclusions
The largest COD decrease, namely, from 54,840 to 2190
mg/l, i.e. a 96% COD reduction, was achieved when the
reactor was operated at the ratio (Vb/VR) = 0.55, air ve-
locity u = 0.046 m/s and t = 65 h. Thus, these values of
(Vb/VR), u and t can be considered as the optimal operat-
ing parameters for a reactor when used in treatment of
high-strength refinery wastewaters.
In the treatment operation conducted in a reactor op-
timally controlled at (Vb/VR) = 0.55, u = 0.046 m/s and t
= 65 h, the conversions obtained for all phenolic con-
stituents of the wastewater were larger than 95%. The
conversions of about 90% were attained for other hydro-
carbons.
5. References
[1] W. Sokół and W. Korpal, “Aerobic Treatment of Waste-
waters in the Inverse Fluidised Bed Biofilm Reactor,”
Chemical Engineering Journal, Vol. 118, No. 3, 2006, pp.
199-205. doi:10.1016/j.cej.2005.11.013

W. SOKÓŁ ET AL.
244
[2] W. Sokół, “Operational Range for a Gas-Liquid-Solid
Fluidized Bed Aerobic Biofilm Reactor with a Low-
Density Biomass Support,” International Journal of
Chemical Reaction Engineering, Vol. 8, 2010, Article ID:
A111. http://www.bepress.com/ijcre/vol8/A111
[3] W. Sokół, A. Ambaw and B. Woldeyes, “Biological Wa-
stewater Treatment in the Inverse Fluidised Bed Reac-
tor,” Chemical Engineering Journal, Vol. 150, No. 1,
2009, pp. 63-68. doi:10.1016/j.cej.2008.12.021
[4] P. Hüppe, H. Hoke and D. C. Hempel, “Biological Treat-
ment of Effluents from a Coal Tar Refinery Using Immo-
bilized Biomass,” Chemical Engineering Technology,
Vol. 13, No. 1, 1990, pp. 73-79.
doi:10.1002/ceat.270130110
[5] A. Alvarado-Lassman, E. Rustrian, M. A. Garcia- Alva-
rado, G. C. Rodriguez-Jimenez and E. Houbron, “Brew-
ery Wastewater Treatment Using Anaerobic Inverse Flu-
idized Bed Reactors,” Bioresource Technology, Vol. 99,
2008, pp. 3009-3015. doi:10.1016/j.biortech.2007.06.022
[6] M. Bajaj, C. Gallert and J. Winter, “Biodegradation of
High Phenol Containing Synthetic Wastewater by an
Aerobic Fixed Bed Reactor,” Bioresource Technology,
Vol. 99, No. 17, 2008, pp. 8376-8381.
doi:10.1016/j.biortech.2008.02.057
[7] A. Lohi, M. Aivarez-Cuenca, G. Anania, S. R. Upreti and
L. Wan, “Biodegradation of Diesel Fuel-Contaminated
Wastewater Using a Three-Phase Fluidized Bed Reactor,”
Journal of Hazardous Materials, Vol. 154, No. 1-3, 2008,
pp. 105-111. doi:10.1016/j.jhazmat.2007.10.001
[8] M. Rajasimman and C. Karthikeyan, “Aerobic Digestion
of Starch Wastewater in a Fluidized Bed Bioreactor with
Low Density Biomass Support,” Journal of Hazardous
Materials, Vol. 143, No. 1-2, 2007, pp. 82-86.
doi:10.1016/j.jhazmat.2006.08.071
[9] N. Fernandez, S. Montalvo, R. Borja, L. Guerrero, E.
Sanchez, I. Cortes, M. F. Comenarejo, L. Traviso and F.
Raposo, “Performance Evaluation of an Anaerobic Flu-
idized Bed Reactor with Natural Zeolite as Support Ma-
terial When Treating High-Strength Distillery Waste-
water,” Renewable Energy, Vol. 33, No. 11, 2008, pp. 2458-
2466. doi:10.1016/j.renene.2008.02.002
[10] P. A. Fitzgerald, “Comprehensive Monitoring of a Fluid-
ized Bed Reactor for Anaerobic Treatment of High
Strength Wastewater,” Chemical Engineering Science,
Vol. 51, No. 11, 1996, pp. 2829-2834.
doi:10.1016/0009-2509(96)00160-1
[11] R. Sowmeyan and G. Swaminathan, “Evaluation of In-
verse Anaerobic Fluidized Bed Reactor for Treating High
Strength Organic Wastewater,” Bioresource Technology,
Vol. 99, No. 9, 2008, pp. 3877-3880.
doi:10.1016/j.biortech.2007.08.021
[12] R. Sowmeyan and G. Swaminathan, “Performance of In-
verse Anaerobic Fluidized Bed Reactor for Treating High
Strength Organic Wastewater during Start-Up Phase,” Bio-
resource Technology, Vol. 99, No. 14, 2008, pp. 6280-
6284. doi:10.1016/j.biortech.2007.12.001
[13] W. Sokół and M. R. Halfani, “Hydrodynamics of a
Gas-Liquid-Solid Fluidized Bed Bioreactor with a Low
Density Biomass Support,” Biochemical Engineering
Journal, Vol. 3, No. 3, 1999, pp. 185-192.
doi:10.1016/S1369-703X(99)00016-9
[14] D. G. Karamanev, T. Nagamune and K. Endo, “Hydro-
dynamics and Mass Transfer Study of a Gas-Liquid-Solid
Draft Tube Spouted Bed Bioreactor,” Chemical Engi-
neering Science, Vol. 47, No. 13-14, 1992, pp. 3581-
3588. doi:10.1016/0009-2509(92)85073-K
[15] B. Rusten, H. Odegaard and A. Lundar, “Aerobic Treat-
ment of Wastewaters in a Novel Biological Reactor,”
Water Science and Technology, Vol. 26, 1992, pp. 703-
708.
[16] W. Verstraete and E. van Vaerenbergh, “Aerobic Acti-
vated Sludge,” In: W. Schonborn, Ed., Biotechnology,
Vol. 8, VCH Verlagessellschaft mbH, Weinheim, 1986,
pp. 43-112.
Notation
COD Chemical oxygen demand (kg/m3)
D Dilution rate (h–1)
t Mean residence time (h)
u Superficial upflow air velocity (m/s)
uf Minimum fluidization air velocity (m/s)
ucr Critical air velocity (m/s)
Vb Volume of settled bed (m3)
VR Reactor volume (m3)
Subscript
m denotes values giving the greatest COD reduction