Environ. Sci. Technol. 2004, 38, 2271-2280
Degradation of Waste Gas Containing Toluene in an Airlift Bioreactor CHENG-SHING LO AND SHYH-JYE HWANG* Department of Chemical Engineering, National Tsing Hua University, Hsinchu, Taiwan, Republic of China
Suspension microorganisms in an internal-loop airlift bioreactor were utilized to treat waste gas containing toluene. The working volume of the reactor was 35 L, and the biomass concentration was 3 kg/m3. The gas pollutant flowed into the reactor from the bottom; it then transferred from the gas phase to the liquid phase and was degraded by the microorganisms suspended in the liquid phase. The microorganisms were able to degrade 50-90% of the inlet toluene when its concentration was from 0.5 to 10 g/m3, and the superficial gas velocity ranged from 0.15 to 1.23 cm/s. A comprehensive mathematical model was also developed to describe the overall degradation process of toluene in the internal-loop airlift bioreactor. The overall degradation process included gas flow, gas-liquid mass transfer, flow and dispersion of the liquid phase, and microbial kinetics. The hydrodynamic properties including the gas and liquid superficial velocities, the gas holdup, the volumetric mass transfer coefficients of toluene and oxygen, and the microbial kinetics were obtained for this model. The substrate inhibition theory was used to simulate the microorganism growth kinetics, and its kinetic constants were obtained experimentally. The penetration theory was used to predict the volumetric mass transfer coefficient. And the drift-flux theory was used to predict the hydrodynamic properties in each section (riser, gasliquid separator, downcomer, and bottom) of the airlift bioreactor. The concentration distributions of toluene and oxygen in the airlift bioreactor and the removal efficiency of toluene predicted by the mathematical model agreed well with the experimental data.
Introduction There are respective advantages and disadvantages of various biochemical process systems for air treatment (e.g., biofilter, bioscrubber, or bubbling system). The biofilter system is widly used in industries at the present. However, if the biofilter system was used to treat high concentrations of volatile organic carbons (VOCs) or was operated for a long time, it would be prone to many problems (e.g., excess biomass growth, clogging, high pressure drop, dry zone formation, and anaerobic zone formation). These would make the removal efficiencies of the contaminants decrease drastically. Although there are no such problems for the bioscrubber, this system allows only the treatment of highly water-soluble compounds because of the gas-liquid mass transfer limitation. However, many VOCs possess the properties of low * Corresponding author telephone: +886-35723221; fax: +88635715408; e-mail:
[email protected]. 10.1021/es030455z CCC: $27.50 Published on Web 03/05/2004
2004 American Chemical Society
solubility and high volatility (e.g., toluene used in this study). Moreover, it has the disadvantage of sprayer clogging. As a result, an airlift bioreactor (one of the gas bubbling systems) was used to treat the waste gas containing toluene in this study because it provided a high ratio of liquid to gas volume and a high volumetric gas-liquid mass transfer rate. The airlift bioreactor consisted of four sections: riser, gas-liquid separator, downcomer, and bottom. In addition, depending on the superficial velocity, there were three flow regimes in the downcomer: I, no bubble; II, bubbles remained stationary; and III, bubbles were carried downward in the downcomer and then upward into the riser (6, 7). When the superficial gas velocity was very low, the circulating liquid velocity was too low to carry any bubble to the downcomer. Hence, there was no bubble in the downcomer. This was regime I. When the superficial gas velocity was increased, the liquid velocity was high enough to carry some small bubbles to the downcomer. Moreover, the small bubbles could coalesce to form large bubbles, and the circulating liquid velocity was about equal to the terminal velocity of these large bubbles in the downcomer. Hence, the bubbles remained stationary. This was regime II. When the superficial gas velocity was further increased, the circulating liquid velocity was high enough to carry more bubbles to the downcomer and then circulated them into the riser. This was regime III. The schematic diagram of the flow regimes is shown in Figure S1 in the Supporting Information. Numerous literature have reported on hydrodynamics and gas-liquid mass transfer on the airlift reactor (1-8). In addition, there have been some research on the application of the airlift reactor to the treatment of waste gas or wastewater containing VOCs (9-12). Livingston (9) used microorganisms immobilized onto Celite diatomaceous earth particles to degrade wastewater containing 3,4-dichloroaniline (34DCA) in a three-phase draft tube fluidized-bed bioreactor. Degradation efficiency of 95% was obtained at a reactor retention time of 1.25 h. Livingston assumed that the reactor was a perfectly mixed system and then developed a reactor model. Freitas dos Santos and Livingston (10) treated wastewater containing 1,2-dichloroethane (DCE) in a enclosed airlift bioreactor system. It was found that over 99% of DCE supplied was biodegraded by this system. Ensley and Kurisko (11) found that microorganisms in an airlift reactor were able to degrade 90-95% of TCE in air for TCE inlet concentration of 300-4000 µg/L. Hecht et al. (12) used bubble column and airlift bioreactor to treat waste gas containing trichloroethylene (TCE). Conversions between 30% and 80% were obtained depending on the operating superficial gas velocity. Hecht et al. (12) also established a simplified reactor model using plug flow for the gas phase, well-mixed flow for the liquid phase, and a pseudo-first-order reaction kinetics for the conversion of TCE. The above mathematical models that describe the degradation of VOCs in the airlift reactor were oversimplified. Therefore, we attempted to develop a mathematical model that considered every aspect. In this study, experiments were conducted in an internalloop airlift reactor. Effects of the operating parameters including hydrodynamic properties, gas-liquid mass transfer coefficient, and microbial kinetic constants on the removal efficiency of the target pollutant (i.e., toluene) were investigated. In addition, a mathematical model was developed to describe the overall degradation process of toluene in the internal-loop airlift bioreactor. The airlift bioreactor was assumed to consist of four sections in series (i.e., the riser, gas-liquid separator, downcomer, and bottom). There were different flow patterns in these sections. The gas-liquid VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2271
FIGURE 1. Schematic diagram of the experimental apparatus. separator and bottom were assumed to be CSTRs, while the flow patterns in the riser and downcomer were simulated by the axial dispersion theory. The substrate inhibition model (13) was used to simulate the microorganism growth kinetics in this study. Comparisons were then made between the concentration distributions of toluene and oxygen and the removal efficiency of toluene obtained by this mathematical model and that obtained experimentally.
Experimental Section The schematic diagram of the experimental apparatus was shown in Figure 1. The airlift bioreactor consisted of a cylindrical vessel 18 cm i.d. and 170 cm in height and a concentric draft tube 12 cm i.d. and 110 cm in height. The draft tube was located 10 cm above the base of the reactor. The unaerated liquid height was kept at 140 cm, and the top clearance (the distance between the liquid surface and the upper end of the draft tube) was 20 cm. The gas was dispersed by a ring sparger located at the bottom of the draft tube with a ring diameter of 10 cm. There were 1-mm holes on the sparger with 1 cm between two holes. The operating gas volumetric flow rate was regulated by flowmeters. Pure liquid toluene was injected into the influent airstream using a syringe pump and evaporated. The synthetic waste gas then passed through a static mixer in order to disperse 2272
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004
toluene uniformly before entering the bioreactor. The desired toluene concentration at a certain superficial gas velocity was achieved by adjusting the injection rate of toluene. As the waste gas was fed into the airlift bioreactor, toluene was transferred to the liquid phase and then degraded by the suspended microorganisms. Finally, the treated gas stream was discharged to the atmosphere from the top of the bioreactor. The source of the microorganisms was activated sludge from a wastewater treatment plant. The temperature of the liquid in the reactor was controlled at 25 °C by a heater, and the pH was maintained at about 7.0 by the addition of 1.0 N NaOH solution. The composition of the mineral salt medium is listed in Table S1 in the Supporting Information and in ref 14. Gas sampling ports were located at the inlet and outlet of the airlift bioreactor. The concentration of toluene in the gas phase was analyzed by a gas chromatograph (Shimadzu GC-14A/FID). Dissolved oxygen was measured by a DO electrode (Ingold) and a DO analyzer (Ideal DOA-100P), while the concentration of oxygen in the gas phase was measured by a portable oxygen analyzer (Cosmos XPO-318). Microbial growth kinetic constants were also measured experimentally in this study. Nine sealed flasks containing the same quantity of the biomass from the bioreactor were
prepared. The same volume of water and mineral salt medium, sufficient air, but various amounts of toluene were added to the flasks. The flasks were then shaken in an incubator maintained at 25 °C. The concentration of toluene in the gas phase in each flask was analyzed by the gas chromatograph every 30 min till the toluene was exhausted. The concentration of the biomass in each flask was then measured. Biomass concentration was determined by filtering the samples through a preweighed 0.45-µm pore size membrane filter, drying the filter paper at 105 °C for 30 min, and then measuring the increase of mass due to the filtered biomass.
Theory Hydrodynamic Properties of the Airlift Bioreactor. Before developing the reactor model, the hydrodynamic properties had to be obtained. The hydrodynamic properties including the superficial gas velocities in the riser and downcomer (Ugr and Ugw), the superficial liquid velocities in the riser and downcomer (Ulr and Ulw), and the gas holdups in the riser and downcomer (gr and gw) were obtained from the driftflux model (15) and energy balance equations (16, 17). The axial dispersion coefficients in the riser and downcomer (Ezr and Ezw) were calculated by Lu’s empirical equations (18):
Ezr ) 48.46Qgi
0.385
( ) ( )
Ezw ) 16.21Qgi0.495
Hd Ht
0.217
Hd Ht
0.093
2 π1/2
x
DLi te
Ol ‚ KS + Tl + Tl /KI KO + Ol µmTl
2
dTl X ) µ dt Yx/s
(5)
where X is the biomass concentration and YX/S is the yield coefficient. Moreover, the biomass mass concentration was given by
X ) X0 + (Tl - Tl0)(-YX/S)
(6)
where X0 is the initial biomass concentration and Tl0 is the initial dissolved toluene concentration in the batch experiment. The following equation was obtained by combining eqs 4-6:
-
[
]
dTl Ol X0 µm Tl ) - (Tl - Tl0) ‚ dt Yx/s KS + Tl + Tl2/KI KO + Ol
(7)
Integrating eq 7:
(2)
∫ dt ) -∫ t
0
(3)
where DLi is the liquid-phase diffusivity of species i and te is the exposure time. Moreover, aL and te were invariable for all species at the same hydrodynamic condition. Therefore, the value of the volumetric mass transfer coefficient of species i (kLiaL) could be determined as long as we had the values of the liquid-phase diffusivity of species i and the oxygen volumetric mass transfer coefficient. The diffusivities of oxygen (DLo) and toluene (DLt) in water at 25 °C were 2.29 × 10-5 and 0.95 × 10-5 (cm2/s), respectively (22). The values of the volumetric mass transfer coefficients of oxygen and toluene in each experimental run are listed in Table S3 in the Supporting Information. Microbial Kinetics. It was assumed that the biomass possessed a substrate-inhibited response to the concentration of the dissolved toluene (13) and a Monod-type response to the dissolved oxygen concentration. Therefore, the expression for the specific growth rate as a function of the dissolved concentrations of toluene and oxygen was (9)
µ)
RTl ) -
(1)
where Qgi is the input gas flow rate, and Hd and Ht are the length of draft tube and the unaerated liquid height, respectively. The values of the hydrodynamic properties described above are listed in Table S2 in the Supporting Information. Gas-Liquid Mass Transfer Characteristic of the Bioreactor. The volumetric mass transfer coefficient of oxygen (kLOaL) was determined by the dynamic method (19). The distribution coefficients of oxygen and toluene (mo and mt) were needed in this method, and their values obtained from vapor-liquid equilibrium data for the pollutant/water system were 37.2 and 0.27, respectively (20). According to the penetration theory, the mass transfer coefficient of species i (kLi) was represented by the following equation (21):
kLi )
where µ is the specific growth rate, µm is the maximum specific growth rate, Ks is the half-saturation constant for toluene, Tl is the dissolved toluene concentration, Ol is the dissolved oxygen concentration, KI is the inhibition coefficient for toluene, and KO is the Monod rate constant for oxygen. In addition, the dissolved toluene reaction rate (RTl) could be expressed by the following equation:
(4)
Tl
Tl0
{[
]
X0 µmTl Ol - (Tl - Tl0) 2 Yx/s K KS + Tl + Tl /KI O + Ol
}
-1
dTl (8)
The value of KO was obtained from Livingston (9), while those of X0, Tl0, and YX/S were obtained from the shake flask tests in this study. When toluene in a certain flask was exhausted, the biomass concentration in that flask became the maximum. Therefore, the relationship between the biomass concentration and the dissolved toluene could be represented by
Xmax - X0 ) YX/STl0
(9)
where Xmax is the maximum biomass concentration. The plot of Xmax versus Tl0 is shown in Figure S2 in the Supporting Information, and the value of YX/S was then obtained from the slope of this plot. The following equation was obtained by substituting the values of X0, Tl0, YX/S, Ol, and KO into eq 8:
t)
1 {-121.78 - 3.14KI - 0.0131KIKS + 1.034Tl + KIµm [35.51 + KI(1.034 + 0.0301KS)] log[34.3 - 1.1Tl] 0.0301KIKS log[Tl]} (10)
A typical plot of Tl versus t is shown in Figure S3 in the Supporting Information, and the kinetic constants (KS, KI, and µm) were then obtained by using the nonlinear leastsquares technique. The values of the kinetic constants are shown in Table S4 in the Supporting Information. Finally, the microkinetic model represented by eq 4 is shown in Figure S4 in the Supporting Information. It was found that the value of Tlmax was about 6.2 g/m3. Note that, when Tl was larger than Tlmax, an increase in the dissolved toluene concentration caused a slight decrease in the specific growth rate. Material Balance for Microbial Growth. It was assumed that the products of the biodegradation of toluene were cells, CO2, and H2O, and the empirical cell formula was assumed to be CH1.84O0.55N0.2 (23). In addition, YX/S was 1.2 (Table S4), VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2273
FIGURE 2. Profiles of reactant concentration or mass rate for regime I (Ug ) 0.15 cm/s, Tgin ) 0.5 g/m3). thus the growth reaction equation was
C7H8 + 4.51O2 + 0.87NH3 f 4.34CH1.84O0.55N0.2(biomass) + 2.66CO2 + 1.31H2O (11) On the basis of the above equation, the consumption rate of the dissolved oxygen was 1.57 times that of the dissolved toluene, for example:
ROl ) 1.57RTl
(12)
where ROl was the dissolved oxygen reaction rate. Material Balance. The detailed derivation of the material balance equations and boundary conditions in the airlift bioreactor are provided in the Supporting Information. In addition, a computer program was developed to obtain the numerical solutions of these material balance equations. The 2274
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004
detailed description of the algorithm is given in the Supporting Information.
Results and Discussion Regime I. When the superficial gas velocity was 0.15 cm/s, the flow regime was in regime I. The results for the input toluene concentrations of 0.5, 5.5, and 7.0 g/m3 are shown in Figures 2-4, respectively. Note that ROr, ROs, ROw, and ROb represent the oxygen reaction rates at a certain location in the riser, gas-liquid separator, downcomer, and bottom, respectively. DOr, DOs, DOr, and DOs represent the oxygen gas-liquid mass transfer rates at a certain location in the riser, gas-liquid separator, downcomer, and bottom, respectively. EOr and EOw represent the liquid-phase axial dispersion rates for oxygen at a certain location in the riser and downcomer, respectively. FOr and FOw represent the liquid-phase mass flow rates for oxygen at a certain location in the riser and downcomer, respectively.
FIGURE 3. Profiles of reactant concentration or mass rate for regime I (Ug ) 0.15 cm/s, Tgin ) 5.5 g/m3). The terms described above could be represented by the following equations:
ROr ) ROlrlr Ogr DOr ) kLoraLr - Olr lr mo
(
EOr ) -Ezrlr
)
∂2Olr ∂z
(13) (14)
(15)
2
∂Olr FOr ) Ulr ∂z
(16)
ROs ) ROlsls
(17)
DOs ) kLosaLs
(
)
Ogsout - Olsout ls mo
(18)
ROw ) ROlwlw DOw ) kLowaLw
(
(19)
)
Ogw - Olw lw mo
EOw ) -Ezwlw
(20)
∂2Olw
(21)
∂zw2
∂Olw FOw ) Ulw ∂zw
(22)
ROb ) ROlblb
(23)
DOb ) kLobaLb
(
)
Ogbout - Olbout lb mo
(24)
Similarly, the corresponding terms for toluene were RTr, RTs, RTw, and RTb; DTr, DTs, DTr, and DTs; ETr and ETw; FTr VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2275
FIGURE 4. Profiles of reactant concentration or mass rate for regime I (Ug ) 0.15 cm/s, Tgin ) 7.0 g/m3). and FTw. The relationships among some of the terms mentioned above were as follows:
DOr ) ROr + FOr + EOr
(25)
DOw ) ROw + FOw + EOw
(26)
DTr ) RTr + FTr + ETr
(27)
DTw ) RTw + FTw + ETw
(28)
Figures 2a, 3a, and 4a show that the dissolved oxygen concentration increased gradually with increasing height in the riser. This was because the oxygen gas-liquid mass transfer rate was greater than its reaction rate all over the riser as shown in Figures 2e, 3e, and 4e. It should be noted that the dissolved oxygen concentration at high input toluene 2276
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004
concentrations was quite low due to high oxygen consumption by toluene degradation. Figure 2b shows that, for Tgin of 0.5 g/m3, the dissolved toluene concentration first increased with increasing height in the riser and then leveled off. This was because the toluene gas-liquid mass transfer rate was higher than its reaction rate, but the difference between these two terms decreased upward along the riser (Figure 2f). However, as shown in Figures 3b and 4b when the input toluene concentrations were 5.5 and 7.0 g/m3, the dissolved toluene concentration increased initially, reached a maximum, and then decreased as the height of the riser increased. The decrease near the upper end of the riser was due to a higher reaction rate of toluene as compared to its gas-liquid mass transfer rate (Figures 3f and 4f). Figures 2d, 3d, and 4d showed that the gas-phase toluene concentration decreased with increasing height of the riser.
This was due to mass transfer of toluene from the gas phase to the liquid phase. Similarly, as shown in Figures 2c, 3c, and 4c, the gas-phase oxygen concentration decreased upward along the riser. However, its decrease was slight because oxygen was in excess as compared to toluene. It is seen in Figures 2e, 3e, and 4e that the gas-liquid mass transfer rate of oxygen was greater than its reaction rate in the gas-liquid separator, whereas the opposite was true in the bottom. Therefore, the dissolved oxygen concentration in the gas-liquid separator was higher than that at the outlet of the riser, and the dissolved oxygen concentration in the bottom was lower than that at the outlet of the downcomer as shown in Figures 2a, 3a, and 4a. In addition, the gas-phase oxygen concentration in the gas-liquid separator was lower than that in the riser, and those in the downcomer and bottom were close to zero. It is seen in Figures 2f, 3f, and 4f that the gas-liquid transfer rate of toluene was less than its reaction rate in the gas-liquid separator and the bottom. Therefore, the dissolved toluene concentration in the gas-liquid separator was lower than that at the outlet of the riser, and the dissolved toluene concentration in the bottom was also lower than that at the outlet of the downcomer as shown in Figures 2b, 3b, and 4b. Note that the trends of the gas-phase toluene concentrations in the gas-liquid separator, downcomer, and bottom were similar to that of the gas-phase oxygen concentrations. As shown in Figure 2a,b, Tl was much lower than Ol when Tgin was 0.5 g/m3. Therefore, the limited reactant of this aerobic biochemical reaction was the dissolved toluene when inputting low toluene concentrations. Consequently, the profiles of oxygen and toluene reaction rates seen in Figure 2e,f were similar to that of the dissolved toluene concentration seen in Figure 2b. However, when Tgin was 5.5 (or 7.0 g/m3), Ol was much lower than Tl as shown in Figure 3a,b (or Figure 4a,b). Therefore, the limited reactant was the dissolved oxygen when inputting high toluene concentrations. Consequently, the profiles of oxygen and toluene reaction rates seen in Figure 3e,f (or Figure 4e,f) were similar to that of the dissolved oxygen concentration seen in Figure 3a (or Figure 4a). The effect of axial dispersion could not be neglected as demonstrated in Figures 2e,f, 3e,f, and 4e,f. These figures show that a significant part of the mass transfer was contributed by the axial dispersion. In addition, Figures 2a,c,d, 3a,c,d, and 4a,c,d also showed the comparison between the experiment data and the predicted values. The average deviation was 10%. Regime II. When the superficial gas velocity was 0.31 cm/ s, the flow regime was in regime II. The results for Tgin of 5.5 g/m3 are shown in Figure S6 in the Supporting Information. It is seen in Figures S6c and S5d in the Supporting Information that, in contrast to regime I, the gas-phase oxygen and toluene concentrations in the upper part of the downcomer where bubbles were present were no longer zero in regime II. However, bubbles were absent in the lower part of the downcomer, so the gas-phase concentrations of oxygen and toluene in this part were still zero. By comparing Figure S6a with Figure S6b in the Supporting Information, it was found that Tl was still lower than Ol. Therefore, the dissolved oxygen was enough for the aerobic biochemical reaction. Consequently, it is seen in Figure S6e,f in the Supporting Information that the profiles of oxygen and toluene reaction rates were similar to the dissolved toluene concentration profile shown in Figure S6b in the Supporting Information. Regime III. When the superficial gas velocity was 0.92 cm/s, the flow regime was in regime III. The results for the input toluene concentrations of 0.5, 7.0, and 10.0 g/m3 are shown in Figures S7-S9 in the Supporting Information, respectively. It is seen in Figures S7d, S8d, and S9d in the Supporting Information that the gas-phase concentration of toluene in the downcomer remained unchanged all over the
downcomer except at the inlet of the downcomer. Thus, the gas-liquid mass transfer rate of toluene in the downcomer was negligible all over the downcomer except at the inlet of the downcomer as shown in Figures S7f, S8f, and S9f in the Supporting Information. This might be because the downward gas velocity in the downcomer was slow enough to result in an equilibrium for toluene between the gas and liquid phases. However, the oxygen gas-liquid mass transfer rate in the downcomer was not negligible in regime III as shown in Figures S7e, S8e, and S9e in the Supporting Information (i.e., oxygen was transferred from the gas phase to the liquid-phase all over the downcomer in regime III). Since the transfer of oxygen from the gas phase to the liquid phase was much more difficult than toluene, an equilibrium for oxygen between the gas and liquid phases could not be achieved in the downcomer. In addition, the transfer of oxygen from the gas phase to the liquid phase in the downcomer provided extra dissolved oxygen for the degradation of toluene in the downcomer, and it also resulted in a higher dissolved oxygen concentration in the riser. Therefore, the degradation of toluene was higher in regime III than in regimes II or I. As a result, regime III was superior to regimes I or II for this aerobic biochemical reaction. Reactor Performance. The average dissolved toluene and oxygen concentrations (Tla and Ola) were defined as follows:
Tla )
Ola )
∑T ∆V ) + T V
+(
∑O ∆V ) + O V
+(
(
lr
lr
∑T
ls ls
lw∆Vlw)
+ TlbVlb
(29)
Vlr + Vls + Vlw + Vlb
(
lr
lr
ls ls
∑O
lw∆Vlw)
+ OlbVlb
Vlr + Vls + Vlw + Vlb
(30)
where ∆Vlr and ∆Vlw are the infinitesimal liquid volume of the riser and the downcomer, respectively; and Vlr, Vls, Vlw, and Vlb are the liquid volume of the riser, gas-liquid separator, downcomer, and bottom, respectively. The elimination capacity (EC) and the removal efficiency (RE) were defined as
EC )
(
Qgin (T - Tgout) Vl gin
(31)
)
(32)
RE ) 1 -
Tgout × 100% Tgin
where Vl is the liquid volume of the reactor, which is equal to (Vlr + Vls + Vlw + Vlb). It is seen in Figure 5a that Tla gradually increased with Tgin at low Tgin. However, Tla increased drastically with Tgin when Tgin was larger than 5 g/m3 for regimes I and II (i.e., a turning point occurred at a Tgin of 5 g/m3). This was because Ola was lower than 2KO when Tgin was larger than 5 g/m3 for regimes I and II as shown in Figure 5b. It is seen in Figure S4 in the Supporting Information that the reaction rate decreased drastically when Ola was lower than 2KO. As a result, the toluene mass transfer rate from the gas phase to the liquid phase was substantially larger than its utilization rate in the liquid phase. Consequently, Tla increased drastically with increasing Tgin when Ola was smaller than 2KO. In addition, since the circulating bubbles were present in the downcomer in regime III, Ola for regime III was higher than that for regimes I or II as shown in Figure 5b. The higher Ola for regime III caused the turning point to appear at a Tgin of 7.5 g/m3 as shown in Figure 5a. Figure 5a also shows that Tla was always lower than the inhibition concentration except for high input toluene concentrations. Moreover, it is seen in Figure S4 in the Supporting Information that, if Tl is larger than Tlmax, there VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2277
FIGURE 5. Reactor performance. was only a slightly negative effect on the reaction rate. Therefore, the inhibition effect caused by high concentrations of toluene was negligible in this study. It is seen in Figure 5c that EC increased with increasing Ug. This was because the volumetric mass transfer coefficient increased with increasing Ug, thus more reactants (toluene and oxygen) were transferred from the gas phase to the liquid phase, and then the reaction rate increased. In addition, since the dissolved oxygen concentration in regime III was larger than that in regimes II or I at the same Tgin (Figure 5b), EC in regime III was substantially higher than that in regimes I or II as shown in Figure 5c. Furthermore, it is seen in Figure 5c that EC first increased linearly with increasing Tgin and then leveled off after a turning point. It even decreased at high Tgin and Ug. This was because the dissolved oxygen concentration was insufficient to support the aerobic reaction at high Tgin. Note that the turning points for EC under various operating conditions were the same as those for Tla (i.e., they appeared when Ola was equal to 2KO). Figure 5d shows that RE generally decreased with increasing Ug due to decreasing residence time. However, RE for the superficial gas velocity of 0.31 cm/s was higher than that for the superficial gas velocity of 0.15 cm/s. This might be due to different flow regimes for 0.15 cm/s (regime I) and 0.31 cm/s (regime II). Figure 5d also shows that RE 2278
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004
kept constant when Tgin was smaller than the turning point (5 g/m3 for regimes I and II and 7.5 g/m3 for regime III), but it decreased when Tgin was larger than the turning point. The relationship between EC and RE was
RE )
Vl EC × × 100% Tgin Qgin
(33)
It is seen in Figure 5c that EC/Tgin was constant when Tgin was smaller than the turning point, but it decreased when Tgin was larger than the turning point. As a result, RE was constant when Tgin was smaller than the turning point, whereas it decreased when Tgin was larger than the turning point. Furthermore, by comparing Figure 5, panels b and d, it was found that RE mainly depended on the residence time (gas velocity) when the dissolved oxygen concentration was larger than 2KO. If the dissolved oxygen concentration was lower than 2KO, RE mainly depended on the magnitude of the dissolved oxygen concentration. Figure 5 also shows that the comparison between the experiment data and the predicted values was satisfactory. A simple model was also developed by assuming completely mixed for the liquid phase and plug flow for the gas phase. The gas phase was in the form of rising bubbles, and they would not backmix. The removal efficiency predicted
FIGURE 6. Comparison between the simple model and the comprehensive model for regimes I and II.
EOj
liquid-phase axial dispersion rate for oxygen at a certain position in j section (g/cm3 s)
ETj
liquid-phase axial dispersion rate for toluene at a certain position in j section (g/cm3 s)
Ezj
axial dispersion coefficient in j section (cm2/s)
FOj
liquid-phase mass flow rate for oxygen at a certain position in j section (g/cm3 s)
FTj
liquid-phase mass flow rate for toluene at a certain position in j section (g/cm3 s)
Hd
length of the draft tube (cm)
Ht
unaerated liquid height (cm)
KI
inhibition coefficient (g/cm3)
KO
Monod rate constant for oxygen (g/cm3)
Ks
half-saturation constant (g/cm3)
kLi
mass transfer coefficient of species i (cm/s)
kLij
mass transfer coefficient of species i in j section (cm/ s)
kLijaLj volumetric mass transfer coefficient of species i in j section (1/s) mi
distribution coefficient of species i (-)
Og
oxygen concentration in the gas phase (g/cm3)
Ogin
input oxygen concentration in the gas phase (g/cm3)
Ogj
oxygen concentration in the gas phase in j section (g/cm3)
Ogjin oxygen concentration in the gas phase at the inlet of j section (g/cm3) Ogjout oxygen concentration in the gas phase at the outlet of j section (g/cm3)
FIGURE 7. Comparison between the simple model and the comprehensive model for regime III. by the simple model was substantially different from that by the comprehensive model as shown in Figures 6 and 7. Moreover, the deviation of the simple model was always much higher than that of the comprehensive model.
Ol
dissolved oxygen concentration (g/cm3)
Ola
average dissolved oxygen concentration (g/cm3)
Olj
dissolved oxygen concentration in j section (g/cm3)
Oljin
dissolved oxygen concentration at the inlet of j section (g/cm3)
Oljout dissolved oxygen concentration at the outlet of j section (g/cm3) Qgin
input gas volumetric rate (cm3/s)
ROlj
dissolved oxygen reaction rate in j section (g/cm3 s)
ROj
oxygen reaction rate at a certain position in j section (g/cm3 s)
RTlj
dissolved toluene reaction rate in j section (g/cm3 s)
RTj
toluene reaction rate at a certain position in j section (g/cm3 s)
Supporting Information Available A more detailed description of the model including Tables S1-S4 and Figures S1-S9. This material is available free of charge via the Internet at http://pubs.acs.org.
Nomenclature
Tg
toluene concentration in the gas phase (g/cm3)
aL
gas-liquid interfacial area per unit liquid volume (1/cm)
Tgj
toluene concentration in the gas phase in j section (g/cm3)
aLj
gas-liquid interfacial area per unit liquid volume in j section (1/cm)
Tgin
DLi
liquid-phase diffusivity of i species (cm2/s)
Tgjin
DOj
oxygen gas-liquid mass transfer rate at a certain position in j section (g/cm3 s)
Tgjout
DTj
toluene gas-liquid mass transfer rate at a certain position in j section (g/cm3 s)
input toluene concentration in the gas phase (g/ cm3) toluene concentration in the gas phase at the inlet of j section (g/cm3) toluene gas concentration in the gas phase at the outlet of j section (g/cm3)
Tl
dissolved toluene concentration (g/cm3)
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2279
Tla
average dissolved toluene concentration (g/cm3)
l
liquid phase
Tl0
initial dissolved toluene concentration in the flask tests (g/cm3)
o
oxygen
out
output or outlet
Tl*
saturation dissolved toluene concentrations (g/cm3)
r
riser
Tlj
dissolved toluene concentration in j section (g/cm3 s)
s
gas-liquid separator
dissolved toluene concentration at the inlet of j section (g/cm3)
t
toluene
w
downcomer
Tljin
Tljout dissolved toluene concentration at the outlet of j section (g/cm3) Tlmax toluene inhibition concentration (g/cm3) te
exposure time (s)
Ug
superficial gas velocity (cm/s)
Ugj
superficial gas velocity in j section (cm/s)
Ulj
superficial liquid velocity in j section (cm/s)
Vj
working volume in j section (cm3)
Vl
liquid volume in the airlift reactor (cm3)
Vlj
liquid volume in j section (cm3)
X
biomass concentration (g/cm3)
X0
initial biomass concentration (g/cm3)
Xmax biomass concentration when the toluene was exhausted in a certain flask (g/cm3) YX/S
yield coefficient (-)
z
axial position in the riser (cm)
zw
axial position in the downcomer (cm)
Greek Symbols
gj
gas holdup in j section (-)
lb
liquid holdup in j section (-)
µ
specific growth rate (1/s)
µm
maximum specific growth rate (1/s)
Subscripts
Literature Cited (1) Bello, R. A.; Robinson, C. W.; Moo-Young, M. Biotechnol. Bioeng. 1985, 27, 369. (2) Siegle, M. H.; Merchuk, J. C.; Schugerl, K. AIChE J. 1986, 32, 1585. (3) Smith, B. C.; Skidmore, D. R. Biotechnol. Bioeng. 1990, 35, 483. (4) Lu, W. J.; Hwang, S. J.; Chang, C. M. Chem. Eng. Sci. 1994, 49, 1465. (5) Ayazi Shamlou, P.; Pollard D. J.; Ison, A. P. Chem. Eng. Sci. 1995, 50, 1579. (6) Heijnen, J. J.; Hols, J.; van der Lans, R. G. J. M.; van Leeuwen, H. L. J. M.; Mulder, A.; Weltevrede, R. Chem. Eng. Sci. 1997, 52, 2527. (7) van Benthum, W. A. J.; van der Lans, R. G. J. M.; van Loosdrecht, M. C. M.; Heijnen, J. J.Chem. Eng. Sci. 1999, 54, 3995. (8) Lo, C. S.; Hwang, S. J. Chem. Eng. J. 2003, 91, 3. (9) Livingston, A. G. Biotechnol. Bioeng. 1991, 38, 260. (10) Freitas dos Santos, L. M.; Livingston, A. G. Appl. Microbiol. Biotechnol. 1993, 40, 151. (11) Ensley, B. D.; Kurisko, P. R. Appl. Environ. Microbiol. 1994, 60, 285. (12) Hecht, V.; Brebbermann, D.; Bremer, P.; Deckwer, W. D. Biotechnol. Bioeng. 1995, 47, 461. (13) Andrews, G. F. Biotechnol. Bioeng. 1968, 44, 707. (14) Tang, H. M. Ph.D. Dissertation, National Tsing Hua University, Taiwan, 1995, p 39. (15) Zuber, N.; Findlay, J. A. J. Heat Transfer 1965, November, 453. (16) Lu, W. J.; Hwang, S. J. Chem. Eng. Sci. 1995, 50, 1301. (17) Hwang, S. J.; Cheng, Y. L. Chem. Eng. Sci. 1997, 52, 3949. (18) Lu, W. J. Ph.D. Dissertation, National Tsing Hua University, Taiwan, 1994, p 126. (19) Chisti, M. Y. Airlift Bioreactors; Elsevier: New York, 1989. (20) Gmehling, J.; Onken, U. Vapor-Liquid Equilibrium Data Collection: Aqueous-Organic Systems; Dechema: Frankfurt, 1997. (21) Higbie, R. Trans. AIChE. 1935, 31, 365. (22) Hayduk, W.; Laudie, H. AIChE J. 1974, 20, 611. (23) Bailey, J. E.; Ollis, D. F. Biochemical Engineering Fundamentals; McGraw-Hill: New York, 1986.
b
bottom
g
gas phase
i
i ) o, t
in
input or inlet
Received for review May 2, 2003. Revised manuscript received January 16, 2004. Accepted January 16, 2004.
j
j ) r, s, w, b
ES030455Z
2280
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004
