Chemical Engineering Science 61 (2006) 5612 – 5619
www.elsevier.com/locate/ces
Mass transfer coefficients of styrene and oxygen into silicone oil emulsions in
a bubble reactor
E. Dumont ∗ , Y. Andrès, P. Le Cloirec
Ecole des Mines de Nantes, UMR CNRS 6144 GEPEA, 4 rue Alfred Kastler, BP 20722, 44307 Nantes Cedex 03, France
Received 15 November 2005; received in revised form 13 March 2006; accepted 5 April 2006
Available online 3 May 2006
Abstract
The absorption of oxygen and styrene in water–silicone oil emulsions was independently studied in laboratory-scale bubble reactors at a
constant gas flow rate for the whole range of emulsion compositions (0–10% v/v). The volumetric mass transfer coefficients to the emulsions
were experimentally measured using a dynamic absorption method. It was assumed that the gas phase contacts preferentially the water phase.
In the case of oxygen absorption, it was found that the addition of silicone oil hinders oxygen mass transfer compared to an air–water system.
Decreases in kL aoxygen of up to 25% were noted. Such decreases in the oxygen mass transfer coefficient, which imply longer aeration times
to transfer oxygen, could represent a limiting step in biotechnological processes strongly dependent on oxygen concentration. Nevertheless, as
the large affinity of silicone oil for oxygen enables greater amounts of oxygen to be transferred from the gas phase, it appears that the addition
of more than 5% silicone oil should be beneficial to increase the oxygen transfer rate. In the case of styrene absorption, it was established that
the volumetric mass transfer coefficient based on the emulsion volume is roughly constant with the increase in the emulsion composition. In
spite of the relatively high cost of silicone oil, water–silicone oil emulsions remain relevant to treat low-solubility volatile organic compounds,
such as styrene, in low-concentration gas streams.
䉷 2006 Elsevier Ltd. All rights reserved.
Keywords: Absorption; Mass transfer; Multiphase reactors; VOC; Silicone oil; Oxygen
1. Introduction
Styrene, a xenobiotic volatile organic compound, is widely
used in industrialised countries in the production of resins and
polymers such as polystyrene. A large amount of styrene is
emitted via low-concentration gas exhaust, which represents a
severe environmental hazard requiring treatment. Conventional
technologies, including combustion (thermal and catalytic oxidation), adsorption and condensation, which have traditionally
been used to treat volatile organic compounds, are not economically viable to treat low-concentration exhaust streams. Biological technologies, which have been recognised for many years
as a cost-effective method for purifying air contaminated with
low concentrations of odorous compounds (Deshusses, 1997;
Cox and Deshusses, 1998; Kennes and Thalasso, 1998;
Groenestijn van and Kraakman, 2005), are also limited by
∗ Corresponding author. Tel.: +33 (0)2 51 85 82 66;
fax: +33 (0)2 51 85 82 99.
E-mail addresses: [email protected] (E. Dumont),
[email protected] (Y. Andrès), [email protected] (P. Le Cloirec).
0009-2509/$ - see front matter 䉷 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ces.2006.04.026
the low solubility of styrene in water and its toxicity towards biomass. To overcome these physical constraints for
bioreactor operation, a water-immiscible, biocompatible and
non-biodegradable organic solvent can be added to the culture
systems to enhance the process (Kennes and Veiga, 2001).
For instance, some hydrophobic phases, like silicone oil,
with a large affinity for styrene can be used for its degradation to harmless substances such as carbon dioxide, water
and cell components. Silicone oil acts as a reservoir for the
controlled delivery of styrene to the aqueous phase, where
biodegradation occurs. At the same time, silicone oil is assumed to enhance the oxygen mass transfer by acting as a
surface-active agent to lower the surface tension of water
and increase the gaseous specific interfacial area. However,
the influence of the addition of an organic phase to water
on the mass transfer of a solute (styrene or oxygen) from the
gas to the water phase is not well understood.
Although there are numerous papers studying the biodegradation (or the mass transfer rate) of toxic and poorly watersoluble compounds improved by silicone oils (Déziel et al.,
E. Dumont et al. / Chemical Engineering Science 61 (2006) 5612 – 5619
5613
1999), only a few studies are devoted to the determination of
kL a. Peeva et al. (2001) experimentally measuring the kL a of
n-decane in 0–90% water–silicone oil emulsions, found that
the volumetric mass transfer coefficient was independent of the
emulsion composition. In this context, it appears that a small
fraction of silicone oil is enough to provide a strong enhancement of volatile organic compounds (VOC). In a large study
devoted to the biodegradation of styrene in multiphase systems,
the ability of silicone oil to improve the removal of a pollutant,
such as styrene, and to enhance the transfer of oxygen has been
tested. Therefore, the aim of the present paper is to measure experimentally the volumetric mass transfer coefficients (kL a) of
styrene and oxygen in emulsions of water–silicone oils using a
dynamic method.
2.2. Equipment
2. Materials and methods
2.3. Oxygen equipment
2.1. Chemicals
A schematic overview of the experimental set-up is given
in Fig. 1. The reactor used has an 11.5 L total volume (height
0.33 m, diameter 0.21 m). In the experiments, air was supplied from a compressor and sparged through an elliptical distributor (75 × 150 mm) with 50 holes (1 mm diameter). The
superficial velocity of gas (uG ) was 0.01 m/s (gas flow rate
QG = 3.3 × 10−4 m3 /s). All experiments were carried out at a
constant temperature of 20 ◦ C maintained by the jacket. The total volumes of gas and liquid were 2 and 10 L, respectively. The
oxygen fraction in the gas phase was determined using a paramagnetic oxygen analyser (Cosma Cristal 300). The flow rate
of gas to the oxygen analyser was quite negligible compared to
the total gas flow rate in the reactor (0.2%). Each experiment
was carried out according to the following procedure:
Silicone oil Rhodorsil䉸 fluids 47V5 (dimethylpolysiloxane:
(CH3 )3 SiO[SiO(CH3 )2 ]n Si(CH3 )3 ) with a viscosity of 5 mPa s
was purchased from the Rhodia Company and styrene from
Aldrich. A series of laboratory experiments showed that the
solubility of styrene in silicone oil is 38000 ± 1000 mg/L. The
solubility of styrene in water, measured during absorption experiments, was 60 mg/L (25 ◦ C), which is much lower than
that given in the literature (320 mg/L at 25 ◦ C, Kirk-Othmer,
1983; 300 mg/L at 20 ◦ C, Verschueren, 1996). Similar results
were obtained by Osswald et al. (1996), who measured 35 mg/L
(20 ◦ C) in synthetic seawater, and by the CEDRE (French
Centre of Documentation, Research and Experimentation on
accidental water pollution) who measured 40 mg/L in pure water and 20 mg/L in seawater. In the case of oxygen, its solubility in silicone oil is seven times higher than in water (solubility
ratio: moxygen = 7).
V4
(i) the dissolved oxygen in the emulsion was removed during 10 min by sparging nitrogen supplied from a nitrogen
outlet
air inlet
V5
The same dynamic method (discussed in Section 3) was used
to determine the volumetric mass transfer coefficients of oxygen and styrene during absorption. In this method, a known
gas volume, loaded with oxygen or styrene, was continuously
flowed, via a circulating loop, through the deoxygenated (without styrene, respectively) liquid system. The operation was
batchwise with respect to the liquid system and the decrease
in solute concentration (oxygen or styrene) in the gas phase
was measured against time. Oxygen and styrene absorption experiments were carried out independently. The composition of
water–silicone oil emulsions ranged from 0 to 10% (v/v) for
both.
V6
150 mm
V7
50 mm
V3
oxygen
analyser
N2 inlet
TIC
data recorder
v v v v
Fig. 1. Schematic overview of the experimental set-up used to determine kL a during oxygen absorption (TIC: Temperature indicator control).
5614
E. Dumont et al. / Chemical Engineering Science 61 (2006) 5612 – 5619
cylinder (V1 , V2 , V7 opened; V6 , V4 , V3 closed). Desorbed
oxygen and nitrogen left through valve V7 ,
(ii) the volume above the liquid and all the other volumes
occupied by the gas were renewed with atmospheric air
by means of the compressor without disturbing the liquid
(V3 , V4 , V5 opened; V1 , V2 , V6 closed). After confirming
from the reading of the oxygen sensor that the volume
occupied by the gas was totally filled with air, valves V3 ,
V5 and V7 were closed and valves V2 and V6 were opened
simultaneously,
(iii) the finite volume of air was sparged through the reactor and
the decrease in oxygen concentration in air was monitored
and recorded as a function of time for further analysis.
Between experiments for each emulsion, the reactor was
washed with alcohol then distilled water.
further analysis. The emulsion volumes ranged from 0.3 to 1.5 L
according to the emulsion composition in order to have the same
styrene driving force for each experiment between the beginning and the end of the absorption. Hence, the same decrease
in the styrene concentration in the gas phase was monitored
during experiments. Changes in emulsion volume were taken
into account in order to calculate kL a as will be shown later.
All experiments were carried out at room temperature, which
varied slightly around 25 ◦ C for the tank, and at a constant temperature of 25 ◦ C maintained by the jacket for the reactor. The
superficial velocity of gas (uG ) was 0.066 m/s corresponding
to a gas flow rate of 3.3 × 10−4 m3 /s. Between experiments
for each emulsion, the reactor was washed with alcohol then
distilled water.
3. kL a calculation
2.4. Styrene equipment
Fig. 2 presents a schematic overview of the experimental
set-up used for styrene absorption. A reactor (1 or 2 litres total
volume according to the emulsion composition) filled with the
water–silicone oil emulsions was connected to a 216-litre tank
allowing a known volume of gas polluted with styrene to be
prepared. First, the airflow passed through a styrene gas generator, where styrene was evaporated at 25 ◦ C (valves V3 and V4
opened; valves V1 , V2 and V5 closed). The styrene concentration was determined in the gas phase using a Flame Ionisation
Detector (Combustion HFR 400 FFID) calibrated before each
experiment from standards. Next, valves V3 and V4 were closed
and valve V1 was opened in order to mix the gas phase thoroughly. Finally, after confirming from the reading of the FID
that the styrene concentration was at a constant value, valve V1
was closed and valves V2 and V5 were opened simultaneously
in order to flow the polluted gas through the water–silicone oil
emulsion for absorption. The decrease in styrene concentration
in air was monitored and recorded as a function of time for
V1
FID
V2
V4
V5
V3
styrene gas
generator
TIC
tank filled with
air polluted by
styrene
water-silicone
oil emulsion
v v v v
Fig. 2. Schematic overview of the experimental set-up used to determine kL a
during styrene absorption (TIC: Temperature indicator control; FID: signal
from the Flame Ionisation Detector proportional to the styrene concentration
in the gas phase).
First, the kL a calculation is fully developed for oxygen absorption. Secondly, only those changes allowing the kL a calculation to be adapted for styrene absorption are presented.
3.1. Assumptions and derivation of kL a for oxygen absorption
The dynamic method for kL a measurement in air–water–
silicone oil systems is based on a gas phase mass balance and
on the off-gas analysis technique, which has been proven as a
relevant measurement method (Gillot et al., 2005; Capela et al.,
2004; Fujie et al., 1994). kL a measurement is based on some
assumptions.
1. In the reactor, the ideal gas law is applicable to calculate
the number of moles of oxygen absorbed by the liquid.
The operating conditions justify such an assumption, since
temperature and pressure are low.
2. The resistance of oxygen transfer in the gas phase is neglected, which is nearly always permitted (Van’t Riet,
1979). Thus, the overall mass transfer coefficient is considered as the liquid phase mass transfer coefficient kL a.
3. The liquid phase can be reasonably estimated as perfectly
mixed. This may be assumed considering the convective
recirculation and turbulence caused by the rising gas bubbles leading to a mixing time shorter than the characteristic
mass transfer time 1/kL a.
4. The gas phase is considered as plug flow in the closed
flowing circuit, as in the oxygen analyser circuit (Gogate
and Pandit, 1999).
5. The presence of silicone oil in the emulsion does not change
the Henry constant for oxygen in water.
6. There is no contact between the gas and the organic liquid
phase.
7. As liquid–liquid mass transfer is very fast compared to
gas–liquid mass transfer, a liquid–liquid equilibrium, in
terms of oxygen concentration, can be assumed during oxygen absorption.
8. The response time for the paramagnetic oxygen analyser is
less than 5 s, which is less than the mass transfer response
time of the system: 1/kL a.
E. Dumont et al. / Chemical Engineering Science 61 (2006) 5612 – 5619
In a system including three phases, the total change in concentration of oxygen in air is due to the amount of oxygen absorbed
in the water phase (continuous liquid phase) and the amount of
oxygen absorbed in the organic phase (dispersed liquid phase).
∗
water and C water in Eq. (7)
Introducing the expression of Coxygen
oxygen
gives
−
gas
−VG
water
oil
dCoxygen
dCoxygen
dCoxygen
= VL (1 − )
+ VL .
dt
dt
dt
(1)
As the water phase and the organic phase are assumed to be in
equilibrium at any time (assumption 7)
oil
Coxygen
water
= moxygen Coxygen
·
(2)
P
VG P dyoxygen
= kL a
yoxygen
VL RT
dt
H
1
VG
P
+
(yoxygen(t) − yoxygen(t=0) ) ,
VL [1 + (moxygen − 1)] RT
(10)
whose integration between t = 0 and t gives
⎡
⎣
Eqs. (1) and (2) lead to
⎤
1
+
gas
water
dCoxygen
dCoxygen
VL
−
=
.
(1 + (moxygen − 1))
dt
VG
dt
(3)
The oxygen concentration in the constant volume of gas is
obtained from the mass balance:
gas
Coxygen =
P
yoxygen .
RT
(4)
After differentiating with respect to time t, Eq. (5) is obtained
from Eq. (4):
5615
1
⎦
1+(moxygen −1)
× ln
⎧
⎨
yoxygen(t=0)
⎩ [ +
1
1+(moxygen −1)
]yoxygen(t) −
⎫
⎬
yoxygen(t=0)
1+(moxygen −1)
= kL at
⎭
(11)
with = VL RT /VG H . By plotting the left member of Eq. (11)
versus time, a straight line is obtained whose slope gives the
volumetric mass transfer coefficient kL a for the aqueous phase
in the presence of organic phase. Introducing = 0 in Eq. (11)
allows the determination of kL a in a binary air–water system.
gas
dCoxygen
P dyoxygen
=
.
dt
RT
dt
(5)
Combining Eqs. (5) and (3) gives, after integrating between
water = 0 and y
t = 0 (Coxygen
oxygen = yoxygen(t=0) ) and t
1
VG
P
VL (1 + (moxygen − 1)) RT
× yoxygen(t) − yoxygen(t=0) .
water
Coxygen
= −
(6)
According to assumptions 3 (liquid-phase perfectly mixed) and
6 (no contact between the gas and the organic phase), the mass
transfer rate in the gas liquid system is given by
−VG
P dyoxygen
water ∗
water
− Coxygen
,
= kL aV L Coxygen
RT
dt
(7)
kL a is based on the total volume of the liquid (aqueous and
organic phases) instead of the aqueous volume to avoid an apparent increase in value solely due to a decrease in the aquewater ∗ , defined according to Eq. (8), is the
ous phase volume. Coxygen
equilibrium concentration between water and the absorbed gas
phase:
∗
water
Coxygen
=
Poxygen
.
H
(8)
Dividing by the total pressure P , Eq. (8) becomes
water ∗
Coxygen
P poxygen
P
=
= yoxygen .
H P
H
3.2. Styrene absorption
1. There is no contact between the gas and the organic liquid
phase, which implies that the resistance of styrene transfer
in the gas phase can be neglected. The overall mass transfer
coefficient is considered as the liquid-phase mass transfer
coefficient, kL a.
2. The liquid phase is assumed to be perfectly mixed and the
gas phase is assumed to be plug flow.
3. A liquid–liquid equilibrium in terms of styrene concentration is assumed during styrene absorption (liquid–liquid
mass transfer is very fast compared to gas–liquid mass
transfer).
4. The response time for the FID is less than the mass transfer
response time of the system: 1/kL a.
In the case of styrene absorption, Eqs. (1)–(3) are unchanged
(replacing subscripts oxygen by styrene):
1
VG
VL (1 + (mstyrene − 1))
gas
gas
× Cstyrene(t) − Cstyrene(t=0) .
water
= −
Cstyrene
The mass transfer rate in the gas–liquid system is given by
water ∗ is the styrene equilibrium concentration
Eq. (13) where Cstyrene
between water and the absorbed gas phase:
gas
(9)
(12)
−VG
dCstyrene
dt
water ∗
water
− Cstyrene
.
= kL aV L Cstyrene
(13)
E. Dumont et al. / Chemical Engineering Science 61 (2006) 5612 – 5619
kLa oxygen / kLaref
oxygen
5616
Introducing the expression of Eq. (12) in Eq. (13) gives, after
integrating between t = 0 and t
(1 + (mstyrene − 1))
gas
gas
VG (Cstyrene(t) − Cstyrene(t=0) )
= kL at.
× ln 1 +
water ∗
VL [1 + (mstyrene − 1)]Cstyrene
no silicone oil
1.0
0.9
0.8
0.7
0.6
0
2
4
6
8
By plotting the left member of Eq. (14) versus time, a straight
line is obtained whose slope gives the volumetric mass transfer
coefficient kL a for the aqueous phase in the presence of organic
phase.
4. Experimental results and discussion
4.1. Oxygen results
Fig. 3 shows a typical example of the monitored decrease in
the mole fraction of oxygen (yoxygen ) in the gas phase versus
time. From experimental data, the determination of kL aoxygen
ref
(called kL aoxygen
for the binary system without silicone oil) is
obtained (Fig. 4) by using Eq. (11). Thirty-nine measurements
were carried out in the case of the reference system without
silicone oil. For the apparatus and the hydrodynamic condiref
tions (uG = 0.01 m/s), kL aoxygen
values ranged from 0.0132 to
−1
0.0178 s (mean value: 0.0156 s−1 ; standard deviation 15%).
yoxygen (x 100)
21
20
19
18
17
100
200
time (s)
300
400
Fig. 3. Typical dynamic response of the paramagnetic oxygen analyser corresponding to the decrease in the mole fraction of oxygen in the gas phase
versus time for a binary air–water system.
0.5
y = 0.0156x
0.4
R2 = 0.99
f (t)
0.3
0.2
y = 0.0117x
0.1
R2 = 0.99
0
0
5
10
15
time (s)
20
25
30
Fig. 4. Examples of kL a determination according to Eq. (11). (+) air–water
reference system; (◦) air–water-1% silicone oil 47V5 (f (t) represents the
left member of Eq. (11)).
kLa (s-1)
(14)
0
10
Silicone oil fraction (%)
kLa without silicone
oil + 15%
Silicone oil 47V5
0.020
0.018
0.016
0.014
0.012
0.010
0.008
0.006
0.004
0.002
0.000
kLa without silicone
oil - 15%
0
2
4
6
8
10
12
% silicone oil
Fig. 5. Absolute and relative kL a values for oxygen absorption in emulsions
of water–silicone oil as a function of silicone oil fraction (uG = 0.01 m/s).
In the presence of silicone oil, experiments were carried out in
triplicate. All the experimental data of f (t)oxygen versus time
fitted satisfactorily on a straight line.
Fig. 5 presents relative kL a measurements (kL aoxygen /
ref
kL aoxygen
) for the whole of the experiment as a function of
emulsion composition. The addition of silicone oil 47V5 hinders oxygen mass transfer compared to an air–water system.
The mass transfer coefficient initially decreases rapidly and
then rises rather slowly with increasing oil volume fraction.
The effects are particularly marked for low volume fractions
of silicone oil (1–2%). To our knowledge, only a few direct
kL a measurements in gas–liquid–liquid systems using silicone
oil are available in the literature (Kawase and Moo-Young,
1990; Morao et al., 1999). Morao et al. (1999) found similar
changes in kL a with silicone oil addition and discussed the
influence of the organic liquid on both kL and (a). According
to these authors, silicone oil, acting as an antifoam agent, has
two opposing effects on the interfacial area (a). On the one
hand, silicone oil lowers the surface tension between gas and
liquid, thus decreasing the bubble diameter and consequently
increasing (a). On the other hand, silicone oil tends to promote coalescence, thus decreasing (a). Morao et al. (1999)
and Kawase and Moo-Young (1990) suggested that these opposite effects more or less cancel each other out, and hence
the most important effect is on the film coefficient kL . According to Morao et al. (1999), the variation in the relative kL a
measurements (decreasing then increasing with oil addition;
Fig. 5) can be explained according to three effects: as oil concentration increases, (1) bubble surface mobility decreases, (2)
coalescence is enhanced, (3) surface tension decreases. The
first two effects, which depress kL a, may however reach a
limit, namely, when all contacting bubbles coalesce and mobility has been suppressed. Meanwhile, surface tension keeps
decreasing, reducing bubble size, and hence kL a starts to rise.
E. Dumont et al. / Chemical Engineering Science 61 (2006) 5612 – 5619
8.5
8.0
FID (volt)
7.5
7.0
6.5
3% oil
6.0
4% oil
5.5
5.0
0
500
1000
Time (s)
1500
2000
Fig. 6. Examples of styrene concentration decrease in the gas phase versus
time for two water–silicone oil emulsions (emulsion volume: 0.7 L).
5617
oxygen concentration. Moreover, opposing effects of silicone
oil addition must be considered. On the one hand, the presence
of silicone oil, with a large affinity for oxygen (moxygen = 7),
enables greater amounts of oxygen to be transferred from the
gas phase. On the other hand, the decrease in kL aoxygen with
silicone oil addition implies longer aeration times to reach saturation. If the decrease in kL aoxygen could be explained as a
function of the silicone volume fraction, it would be possible,
as shown by Nielsen et al. (2003), to propose an optimal fraction of oil to maximize the rate of oxygen mass transfer to
the system. In our study, assuming the maximum decrease in
kL aoxygen is about 25% for a system without oil, 5% silicone
oil addition would be enough to begin to enhance the oxygen
mass transfer rate.
4.2. Styrene results
According to this explanation, there is a critical concentration
at which the gas–liquid mass transfer is minimal.
Similar results have been reported in the literature for
other organic liquids. Yoshida et al. (1970) observed such a
variation in relative kL a when using toluene and oleic acid.
By adding n-dodecane, Hassan and Robinson (1977) showed
ref
that (kL aoxygen /kL aoxygen
) was always less than unity for oil
fractions in the range 0% < < 10%. They reported that, as
ref
organic liquid increased from 0 to 1%, (kL aoxygen /kL aoxygen
)
decreased sharply to about 0.8, increased slightly as increased to 6%, and levelled off at an average value of up to
10%. These results described by Hassan and Robinson (1977)
are much closer to those presented in Fig. 6. More recently,
Nielsen et al. (2003) clearly indicated that, as the fraction of
hexadecane in the bioreactor increased, the experimental value
of kL aoxygen decreased. Generally, the experimental oxygen
absorption results reported in the literature show that the value
of kL aoxygen can decrease, remain unaffected or increase upon
addition of the organic phase (Dumont and Delmas, 2003). As
for silicone oil effects, several explanations, based on changes
in the physical properties of the emulsion (density, viscosity,
gas solubility, gas diffusivity) due to the organic phase addition, have been proposed. The interfacial properties of the
emulsion, particularly the spreading coefficient, have been used
to explain the variations in the volumetric mass transfer coefficient with the fraction of the organic phase. However, studies
devoted to the influence of the spreading coefficient have led
to contradictory results (see Dumont and Delmas, 2003). To
date, a satisfactory description of the phenomenon is still lacking. Nonetheless, a recent model allowing an assessment of
ref
the influence of oil addition on the (kL aoxygen /kL aoxygen
) ratio
has been proposed (Zhang et al., 2006). This model takes both
shuttle effects and hydrodynamic effects into account to calculate the enhancement factor due to the presence of droplets
of organic liquids. According to this model, a decrease in kL a
with the addition of oil is possible and could be due to hydrodynamic effects (viscosity increase). However, these effects
alone cannot explain the experimental results of this study.
Clearly, a decrease in the oxygen mass transfer coefficient
due to silicone oil addition could be crucial in many biotechnological processes, particularly those which depend strongly on
Experiments were carried out in triplicate. Fig. 6 shows typical examples of the monitored decrease in the styrene concentration in the gas phase versus time (the FID signal being
proportional to the styrene concentration). The emulsion volume chosen as a reference for the present results is 0.7 litres
(in a one-litre reactor). To account for the low solubility of
styrene in water, a special experiment was carried out to measure kL astyrene for a gas–water system without silicone oil. The
experimental set-up presented in Fig. 2 was used, replacing the
216-litre tank by a 12-litre reactor to allow satisfactory absorption. The mean kL astyrene value measured without silicone oil
was 0.0150 s−1 (±15%).
In the presence of silicone oil, experimental values of
kL astyrene were first determined using Eq. (14) then recalculated from Eq. (15) to account for the change in emulsion
volume between each experiment. This change, varying from
0.3 to 1.5 L, was necessary to have the same driving force for
each experiment between the beginning and the end of the
absorption. Eq. (15) represents a typical correlation for estimating kL a values in an agitated reactor (Nielsen et al., 2003).
Eq. (15) is assumed to apply to our experimental set-up
Pw uG ,
(15)
kL a = VL
where , and represent empirical constants, Pw represents
the power requirement of the aerated reactor (assumed constant
between each experiment), VL is the reactor emulsion volume
and uG represents the superficial gas velocity through the reactor. Eq. (15) can be transformed into the linear Eq. (16) with
= ln() + ln(Pw) + ln(uG ) whose slope gives the influence of VL on kL a values:
ln(kL a) = − ln(VL ).
(16)
Fig. 7 shows the change in ln(kL a) versus ln(VL ) according
to the silicone oil fraction of the emulsion (v/v). Experimental
data allowed an estimation of =1.2. Using =1.2 and VL =0.7
litres, kL astyrene values were recalculated in order to compare
them (Fig. 8). It appears that the volumetric mass transfer coefficient of styrene, contained in a gas phase, to water–silicone oil
5618
E. Dumont et al. / Chemical Engineering Science 61 (2006) 5612 – 5619
5. Conclusion
In (V)
-9
-8.5
-7.5
-8
-6.5
-7
-6
-3
10%
In (kLa )
-3.5
8%
5%
4%
-4
6%
2%
3%
-4.5
1%
0%
y = -1.2x - 13.1
-5
R2 = 0.99
-5.5
Fig. 7. Representation of Eq. (16) for the determination of the empirical
parameter .
0.035
KL astyrene (s-1)
0.030
0.025
0.020
0.015
0.010
0.005
0.000
0
2
4
6
8
10
Silicone oil fraction (%)
Volumetric mass transfer coefficients (kL a) of oxygen and
styrene in gas–water–silicone oil systems have been experimentally measured using a dynamic method. On the one hand, it was
found that addition of silicone oil hinders oxygen mass transfer compared to an air–water system. Decreases in kL aoxygen
of up to 25% have been noted. Such decreases in the oxygen
mass transfer coefficient, which imply longer aeration times to
transfer oxygen, could represent a limiting step in biotechnological processes strongly dependent on oxygen concentration.
Nevertheless, as silicone oil presents a large affinity for oxygen (seven times higher than water), enabling larger amounts
of oxygen to be transferred from the gas phase, the addition of
more than 5% silicone oil should be beneficial to increase the
oxygen transfer rate. Moreover, it was found that the volumetric mass transfer coefficient of styrene from the gas phase to a
water–silicone oil emulsion remains constant with the emulsion
composition. Despite the relatively high cost of silicone oil, it
appears that water–silicone oil emulsions should be relevant to
treat low-concentration styrene gas streams. Future investigations devoted to styrene biodegradation, taking into account the
toxicity of this xenobiotic to microorganisms, will be carried
out in order to quantify both the influence of silicone oil addition on organic biodegradation and the oxygen transfer rate
(OTR) to the biomass.
Fig. 8. kL astyrene values for styrene absorption in emulsions of water–silicone
oil as a function of silicone oil fraction (uG = 0.066 m/s).
Notation
emulsions remains roughly constant with the increasing silicone
oil fraction. Such a result is consistent with the assumption that
the gas phase contacts preferentially the continuous aqueous
phase. Studying the absorption of n-decane in water–silicone
oil emulsions, Peeva et al. (2001) similarly concluded that kL a
was independent of the volume fraction of silicone oil.
Because of the high solubility of oxygen and styrene in silicone oil, the experimental results show that water–silicone oil
emulsions are suitable for the biodegradation of styrene emitted
via low-concentration gas exhaust. For instance, a 10% addition
of silicone oil to water could represent a satisfactory emulsion
composition to enhance the mass transfer rate of both oxygen
and styrene, given that silicone oil is expensive. Biodegradation results reported in the literature seem to confirm the potential of silicone oil for xenobiotic volatile organic compound
treatment. Hekmat and Vortmeyer (2000), measuring the stationary degradation kinetics of volatile aromatic compounds in
a laboratory-scale trickle-bed bioreactor, showed that the addition of 10% silicone oil resulted in a 2.4-fold increase in
the degradation rate. Aldric et al. (2004) reported the value of
using silicone oil at a level of 10% in a biphasic bioreactor.
These authors showed that silicone oil enables large quantities
of isopropyl-benzene to be absorbed within the medium of biological abatement. Budwill and Coleman (1997), developing
a biofilter with 20% silicone oil, observed an increase in nhexane removal compared with the untreated control.
a
C
FID
H
kL
kL a
m
poxygen
P
Pw
QG
R
T
uG
VG
VL
yoxygen
interfacial area, m−1
concentration, mol/m3
signal from the Flame Ionisation Detector
proportional to the styrene concentration in
the gas phase, volt
Henry constant, Pa m3 /mol
mass transfer coefficient, m/s
volumetric mass transfer coefficient, s−1
solubility ratio; solubility of solute in oil
versus solubility of solute in water, dimensionless
partial pressure of oxygen in the gas phase
total pressure of the gas phase, Pa
power, W
gas flow rate, m3 /s
universal gas constant, J/(mol K)
temperature, K
gas velocity, m/s
volume of the gas phase, m3
volume of the emulsion m3
oxygen mole fraction, dimensionless
Greek letters
, , , , , constants
dispersed liquid-phase hold-up, dimensionless
E. Dumont et al. / Chemical Engineering Science 61 (2006) 5612 – 5619
Subscripts
oxygen
styrene
relative to oxygen measurement
relative to styrene measurement
Superscripts
∗
gas
oil
ref
water
equilibrium concentration of solute between water
and gas phase
relative to the gas phase
relative to the silicone oil phase
relative to a gas–water system of reference without
silicone oil
relative to the water phase
Acknowledgement
We thank Chu Zhu for his valuable contribution to this work.
References
Aldric, J.M., Destain, J., Thonart, P., 2004. The two-phase reactor
water/silicone oil: prospects in the off-gas treatment. In: Verstraete, (Ed.),
European Symposium on Environmental Biotechnology, pp. 855–859.
Budwill, K., Coleman, R.N., 1997. Effect of silicone oil on biofiltration of
n-hexane vapours. Mededelingen van de faculteit landbouwwetenschappen.
rijksununiversiteit gent 62/4b, 1521–1528.
Capela, S., Gillot, S., Héduit, A., 2004. Comparison of oxygen-transfer
measurement methods under process conditions. Water Environment
Research 76, 183–188.
Cox, H.H.J., Deshusses, M.A., 1998. Biological waste air treatment in
biotrickling filters. Environmental Biotechnology 9, 256–262.
Deshusses, M.A., 1997. Biological waste air treatment in biofilter.
Environmental Biotechnology 8, 335–339.
Déziel, E., Comeau, Y., Villemur, R., 1999. Two-liquid-phase bioreactors for
enhanced degradation of hydrophobic/toxic compounds. Biodegradation
10, 219–233.
Dumont, E., Delmas, H., 2003. Mass transfer enhancement of gas absorption
in oil-in-water systems: a review. Chemical Engineering Proceedings 42,
419–438.
Fujie, K., Tsuchiya, K., Fan, L.-S., 1994. Determination of volumetric
oxygen transfer coefficient by off-gas analysis. Journal of Fermentation
and Bioengineering 77, 522–527.
5619
Gillot, S., Kies, F., Amiel, C., Roustan, M., Héduit, A., 2005. Application of
the off-gas method of the measurement of oxygen transfer in biofilters.
Chemical Engineering Science 60, 6336–6345.
Gogate, P.R., Pandit, A.B., 1999. Survey of measurement techniques
for gas–liquid mass transfer coefficient in bioreactors. Biochemical
Engineering Journal 4, 7–15.
Groenestijn van, J.W., Kraakman, N.J.R., 2005. Recent developments in
biological waste gas purification in Europe. Chemical Engineering Journal
113, 85–91.
Hassan, I.T.M., Robinson, C.W., 1977. Oxygen transfer in mechanically
agitated aqueous systems containing dispersed hydrocarbon. Biotechnology
and Bioengineering 19, 661–682.
Hekmat, D., Vortmeyer, D., 2000. Biodegradation of poorly water-soluble
volatile aromatic compounds from waste air. Chemical Engineering
Technology 23, 315–318.
Kawase, Y., Moo-Young, M., 1990. The effect of antifoam agents on mass
transfer in bioreactors. Bioprocess Engineering 5, 169–173.
Kennes, C., Thalasso, F., 1998. Waste gas biotreatment technology. Journal
of Chemical Technology and Biotechnology 72, 303–319.
Kennes, C., Veiga, M.C., 2001. Bioreactors for Waste Gas Treatment. Kluwer
Academic Publishers, Dordrecht, pp. 148–151.
Kirk-Othmer, 1983. Encyclopedia of Chemical Technology, third ed., vol. 21.
Wiley, New York.
Morao, A., Maia, C.I., Fonseca, M.M.R., Vasconcelos, J.M.T., Alves, S.S.,
1999. Effect of antifoam addition on gas–liquid mass transfer in stirred
fermenters. Bioprocess Engineering 20, 165–172.
Nielsen, D.R., Daugulis, A.J., McLellan, P.J., 2003. A novel method of
simulating oxygen mass transfer in two-phase partitioning bioreactors.
Biotechnololgoy and Bioengineering 83 (6), 735–742.
Osswald, P., Baveye, P., Block, J.C., 1996. Bacterial influence on partitioning
rate during the biodegradation of styrene in a biphasic aqueous-organic
system. Biodegradation 7, 297–302.
Peeva, L., Ben-zvi Yona, S., Merchuk, J.C., 2001. Mass transfer coefficients
of decane to emulsions in a bubble column reactor. Chemical Engineering
Science 56, 5201–5206.
Van’t Riet, K., 1979. Review of measuring methods and results in nonviscous
gas–liquid mass transfer in stirred vessels. Industrial and Engineering
Chemistry Process Design and Development 18 (3), 357–364.
Verschueren, K., 1996. Handbook of environmental data on organic chemicals,
third ed. Van Nostrand Reinhold Co, New York.
Yoshida, F., Yamane, T., Miyamoto, Y., 1970. Oxygen absorption into
oil-in-water emulsions. A study of hydrocarbon fermentors. Industrial
Engineering Chemistry Process Design and Development 9 (4), 570–577.
Zhang, G.D., Cai, W.F., Xu, C.J., Zhou, M., 2006. A general enhancement
factor model of the physical absorption of gases in multiphase systems.
Chemical Engineering Science 61, 558–568.