Bioresource Technology 102 (2011) 10585–10591
Contents lists available at SciVerse ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Analysis of petroleum biodesulfurization in an airlift bioreactor using response surface methodology Zahra Azimzadeh Irani a,b, Mohammad Reza Mehrnia a,⇑, Fatemeh Yazdian c, Majid Soheily b, Ghasemali Mohebali b, Behnam Rasekh b a b c
School of Chemical Engineering, University College of Engineering, University of Tehran, P.O. Box 11155-4563, Tehran, Iran Research Institute of Petroleum Industry, P.O. Box 14665-1998, Tehran, Iran Department of Life Science Engineering, Faculty of Disciplinary New Science and Technologies, University of Tehran, P.O. Box 14395-1561, Tehran, Iran
a r t i c l e
i n f o
Article history: Received 29 May 2011 Received in revised form 26 August 2011 Accepted 29 August 2011 Available online 5 September 2011 Keywords: Petroleum biodesulfurization Gordonia alkanivorans RIPI90A Airlift bioreactor Volumetric gas liquid mass transfer coefficient Response Surface Methodology
a b s t r a c t For the first time, growing cells of Gordonia alkanivorans RIPI90A were used for biodesulfurization (BDS) of diesel. This process was carried out in an internal airlift bioreactor. BDS parameters (oil/water phase ratio and initial sulfur concentration) were optimized in flasks using response surface methodology. Predicted results were found to be in good agreement with experimental results. Initial sulfur concentration had a remarkable effect on BDS process. Maximum removal of sulfur (21 mg/l) can be achieved at oil/ water phase ratio of 25% (v/v) and initial sulfur concentration of 28 mg/l. Moreover, effect of superficial gas velocity (Ug) and working volume (v) on volumetric gas liquid mass transfer coefficient was studied in an airlift bioreactor for BDS of diesel. The best results were achieved at Ug and v of 2.5 l/min and 6.6 l, respectively. Subsequently, BDS of diesel was investigated in an airlift bioreactor under optimized conditions. Sulfur reduction after 30 h was 14 mg/l. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Sulfur oxide emissions to the atmosphere due to the combustion of sulfur containing compounds in fossil fuels such as diesel and gasoline can cause acid rain and air pollution (EPA, 2011). Sulfur free diesel is ordered by strict current environmental regulations (McFarland, 1998; Oshiro et al., 1999; Monticello, 2000; Maghsoudi et al., 2001). Diesel can be treated by hydrodesulfurization (HDS) method (Chowdhury et al., 2002; Soleimani et al., 2007). HDS is carried out at high temperature and high pressure on a catalyst (CoMo, NiMo, etc.) in the presence of hydrogen. Most sulfur in diesel can be treated easily in this process but recalcitrant organic sulfur part can be removed under invasive and costly conditions (McHale, 1981; Ma et al., 1994; Monticello, 1998; Singh et al., 2001). Other methods such as biodesulfurization (BDS) remove recalcitrant molecules under ambient pressures and temperatures (Arenskötter et al., 2004; Caro et al., 2008). The remained recalcitrant molecules are mainly from the dibenzothiphene (DBT) chemical group (Monticello, 1998). It has been previously reported that Gordonia alkanivorans RIPI90A can convert DBT and DBT containing hexadecane to 2-hydroxybiphenyl (2-HBP) via 4S pathway during both the growth and
⇑ Corresponding author. Tel.: +98 21 66112184; fax: +98 21 66954041. E-mail address:
[email protected] (M.R. Mehrnia). 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2011.08.120
resting stages in a biphasic system (Mohebali et al., 2007a,b). Dimethyl sulfoxide (DMSO) was also reported as an appropriate sulfur source for the mass production of G. alkanivorans RIPI90A and the effect of DMSO concentrations was investigated on the growth rate of this strain (Mohebali et al., 2008). Sulfur containing feedstock and the biocatalyst, usually suspended in the aqueous phase have to be contacted with each other in a bioreactor. In order to avoid mechanical stirring, which puts stress on capability of microorganism, different types of bioreactors have been proposed, the majority of which are loop bioreactors (Mehrnia et al., 2004a,b; Yazdian et al., 2009, 2010). Loop bioreactors are characterized by a definitely directed circulation flow, which can be driven in fluid or fluidized systems by a propeller or jet drive and most typically in gas–liquid systems by an airlift drive or liquid pump. They are particularly suitable for fluid systems requiring high dispersion priority. On the other hand, their simple constructions and operation result in low investment and operational costs (Mehrnia et al., 2004a,b; Yazdian et al., 2009, 2010). The use of multistage airlift reactors may reduce the cost of mixing and overcome poor reaction kinetics and to achieve continues growth and regeneration of the biocatalyst in the same system rather than in a separate reactor (Monticello, 1998). In the design of the draft tube airlift reactor, higher oxygen utilization is an advantage over mechanically stirred bioreactors. Therefore, if whole cells were used as biocatalyst, then less shear damage would
10586
Z.A. Irani et al. / Bioresource Technology 102 (2011) 10585–10591
occur. The absence of mechanical stirring systems also translates into lower capital and operating costs (Mehrnia et al., 2004b). The draft tube airlift bioreactor was studied using water in kerosene micro emulsions (Mehrnia et al., 2004a). The effect of draft tube area versus the top section area on various parameters was studied. Also, the effect of gas flow rate on recirculation and gas carry over due to incomplete gas disengagement was considered (Mehrnia et al., 2004b). The oil/water ratio and viscosity effects on gas hold-up and mass transfer coefficient was tested as well (Mehrnia et al., 2005; Yazdian et al., 2009, 2010). In addition, the significance of riser to down comer cross sectional area, volume of gas–liquid separator value and superficial gas velocity amount on mixing time, gas hold-up and volumetric gas–liquid mass transfer coefficients in an external airlift loop bioreactor were investigated (Yazdian et al., 2009). BDS of water–kerosene emulsions with resting cells of different strains was studied at 100 ml scale and in an airlift bioreactor (Sanchez et al., 2008). Moreover, an experimental and modeling study of BDS of hydrodesulfurized diesel in a trickle bed reactor (Mukhopadhyaya et al., 2006) and an airlift reactor (Nandi, 2010) was also performed. Response Surface Methodology (RSM) is used to examine the relationship between one or more response variables and a set of quantitative experimental factors. This method is often employed after a ‘‘vital few’’ controllable factors have been identified in order to find the factor settings that optimize the response. Central Composite Design (CCD) and Box-Behnken Design (BBD) of RSM are fractional factorial designs for optimization of variables with a limited number of experiments (Hasan et al., 2010; Wu et al., 2010). Investigations on BDS processes have been mostly devoted to the physiology of organisms, metabolic pathways of sulfur reduction and kinetic parameters of the reaction of both model and diesel oils (Folsom et al., 1999; Chang et al., 2001; Guchhait et al., 2005; Rashtchi et al., 2006; Guobin et al., 2006; Mohebali et al., 2007, 2008; Hewitt and Nienow, 2007; Caro et al., 2008; Azimzadeh Irani et al., 2011). Moreover, in the previous works for BDS processes of fossil fuels in the airlift bioreactors (Folsom et al., 1999; Guobin et al., 2006; Yang et al., 2007; Sanchez et al., 2008; Nandi, 2010), there is no data for significant factors and especially optimization of operational parameters based on statistical analysis. In this study, the growing cells of G. alkanivorans RIPI90A were used for BDS of diesel in an internal airlift bioreactor. The effect of initial sulfur concentration on growth and sulfur reduction during the growth phase were investigated. The RSM technique was used to optimize the BDS of diesel by RIPI90A cells between two variables (oil/water phase ratio and initial sulfur concentration) in batch shake flasks. Moreover, superficial gas velocity (Ug) and working volume (v; related to the liquid level above the riser section) were used to optimize gas liquid mass transfer coefficient (kLa) in an airlift bioreactor for an emulsion of 30:70 diesel/water. Finally, BDS process was investigated under the optimum conditions in the airlift bioreactor as well.
2. Methods
2.2. Chemicals All chemicals were of analytical grade. Normal hexadecane and DBT were purchased from Merck. N,N0 -dimethylformamide (DMF) were obtained from Riedel- de Haën. 2.3. Diesel Diesel oil having the following specification was used; initial boiling point: 250 °C final boiling point: 385 °C specific gravity: 8370 kg/m3 and sulfur content: 30 mg/l. 2.4. Micro-organism and culture conditions The culture was initially grown in 250 ml Erlenmeyer flask containing 100 ml of minimal salt (MS) medium having the composition (gram per liter of deionized water) KH2 PO4 (6), Na2 HPO4 (4), NH4 NO3 (1.2), MgCl2 6H2 O (0.75), MnCl2 4H2 O (0.004), CaCl2 2H2 O (0.001), FeCl3 (0.001), containing 2 g/L sodium benzoate as carbon source (MS-SB) medium. DMF was used as co-solvent in which DBT was dissolved and then added to the medium as sulfur source. The pH was adjusted prior to autoclaving to 7.08. All inoculated liquid media were incubated at 30 °C on a rotary shaker operated at 120 rpm for 72 h. 2.5. Batch growth and biodesulfurization study In this study, diesel with different initial sulfur concentrations was used. It is assumed that all the sulfur compounds in diesel oil were lumped into a pseudo-compound and then diluted at different dilution ratios using hexadecane to give final sulfur concentrations. Batch experiments were performed in 100 ml Erlenmeyer flasks containing MS-SB supplemented with diesel oil at desired oil/water phase ratios and initial sulfur concentrations. The reaction broths were always prepared with 10% (v/v) of initial biomass concentration. The flasks were incubated at 30 °C on a rotary shaker operated at 120 rpm for a period of 30 h (Azimzadeh Irani et al., 2011). Following incubation, the content of each flask was centrifuged (12000 rpm, 5 min). The organic phase was analyzed for remaining sulfur concentration in the sample by UV-fluorescence detection (vario TRACE) for quantification of total sulfur concentration. 2.6. Bioreactor description Hydrodynamic characterization and BDS assays were evaluated in an internal airlift bioreactor. Reactor vessel was 0.12 m in diameter and its overall height was 0.7 m. The draft-tube, 0.07 m in internal diameter and 0.35 m tall, was located 0.06 m above the bottom of tank. The vessel was sparged in the concentric zone through a 0.0006 m diameter sparger. The working volume and the overall volume of the reactor were 5 and 8 L, respectively. A dissolved oxygen electrode (Metteler Toledo, Switzerland) and a pH-meter (Metteler Toledo, Switzerland) were placed in the bioreactor. Air was supplied to the reactor through a filter and a rotameter.
2.1. Biocatalyst The biocatalyst used was growing cells of G. alkanivorans RIPI90A. This strain was selected among 23 strains that were gathered on the basis of their growth rate and extent of desulfurization from all around Iran. The detailed description of the biocatalyst is reported by Mohebali et al. (2007a,b). This aerobic Gram (+) strain was maintained as a suspension in nutrient broth containing 20% (v/v) glycerol, at 70 °C and also on glass beads under the same conditions.
2.7. Gas–liquid volumetric mass transfer coefficient (kLa) measurement Determination of kLa was performed with loop sparger in an emulsion of 30:70 diesel/water. This parameter was evaluated for desired superficial gas velocity (Ug) and working volume (v). The kLa was measured with dynamic gassing in method (Shariati et al., 2007; Yazdian et al., 2009). For each test, the fluid was deaerated by bubbling nitrogen. After stopping the nitrogen flow and
10587
Z.A. Irani et al. / Bioresource Technology 102 (2011) 10585–10591
2.9. Software used
Table 1 Coded and non coded values of the variables. Coded and non coded values for optimization of BDS process Independent variables Oil–water phase ratio (%, v/v) X1 Initial sulfur concentration (mg/l) X2
1.41 – –
1 30 7
Coded and non coded values for optimization of kLa Superficial gas velocity (l/min) X1 0.17 1.00 Working volume (l) X2 5.80 6.00
0 40 17.5
+1 50 28
+1.41
3.00 6.50
5.00 7.00
5.83 7.20
ð1Þ
In Eq. (1), E is the fractional approach to equilibrium and can be estimated using following equation:
E ¼ ðC C 0 Þ=ðC CÞ
ð2Þ
⁄
where C is the saturation concentration of dissolved oxygen, C0 is the initial concentration of dissolved oxygen at time t0 when a hydrodynamic steady-state has been reestablished upon the beginning of aeration and C is the dissolved oxygen concentration at any time t. 2.8. Experimental design Optimization of BDS process of diesel oil by growing cells of G. alkanivorans RIPI90A was carried out in 100 ml scale. Oil/water phase ratio (X1) and initial sulfur concentration (X2) that were selected due to several preliminary tests, were regarded as effective factors for this process and the influence of these parameters on sulfur reduction was studied. Optimization of kLa in the airlift reactor for loop sparger was carried out in an emulsion of 30:70 diesel/ water. Superficial gas velocity, Ug(X1) and working volume, v(X2) were selected as effective factors. CCD and RSM were used in order to investigate the relationship between these variables and the optimum levels of them. For this purpose, 9 experimental runs were required as per three-level two-factor fractional factorial CCD. The coded and non coded values of the variables are shown in Table 1 while the results and predicted responses are presented in Table 2. Data from CCD were subjected to a second-order multiple regression analysis to explain the behavior of the system using the least squares regression methodology to obtain the parameter estimators of the mathematical model. The result may be expressed as follow:
Y ¼ b0 þ
X
bi X i þ
X
bii X 2i þ
X
bij X ij
3. Results and discussion 3.1. Behavior of growing cells in biphasic media
allowing the exit of its bubbles, the air started to flow until the fluid became nearly saturated with oxygen. The kLa was calculated as the slope of the linear equation:
lnð1 EÞ ¼ kL aðt t0 Þ
MINITAB 16 was used for this regression analysis of the data obtained and to estimate the coefficients of the regression equation.
ð3Þ
where Y is the response, Xi is the independent variable, b0 is a constant, bi is the slope or linear effect of the input factor, bii is the quadratic effect of input factor, bij is the linear by linear interaction effect between the input factor.
It has been reported previously that resting cells of G. alkanivorans RIPI90A can convert DBT to 2-HBP via 4S pathway in a biphasic system (Mohebali et al., 2007a). It is assumed that in biphasic media, the transfer of DBT from the oil to the aqueous phase is one of the most determinant parameter in the BDS process. Therefore, using hydrophobic biocatalysts such as G. alkanivorans RIPI90A cells are preferable because they are capable to be joined at the oil/water interface for the uptake of DBT there. It was shown that, resting and also growing cells of this strain have an affinity for organic sulfur substrates and their ability of stabilizing the oil/ water emulsion is related to cell surface hydrophobicity that may be related to its cell surface long mycolic acids (Mohebali et al., 2007b). In order to investigate the effect of sulfur on growth of the organism, the growth of bacterium was monitored periodically for different initial sulfur concentrations. The evolutions of the biomass concentration are shown in Fig. 1. As shown in this figure, the amount of maximum growth increases with increase of initial sulfur concentration and maximum growth is obtained at 28 mg/l. This may be due to the increase of sulfur compounds availability that causes better growth of biomass in biphasic media. Fig. 2 shows sulfur reduction of diesel by growing cells of G. alkanivorans RIPI90A during the growth phase for different initial sulfur concentrations. As shown in this figure, as the availability of initial sulfur substrate increases, better growth of biomass causing higher rate of sulfur consumption. The capability of G. alkanivorans RIPI90A for BDS process via 4s pathway and inhibition effects of 2-HBP production (Mohebali et al., 2007a) and transition from late exponential to stationary phase, may cause the sulfur reduction rate becomes constant after 30 h (Azimzadeh Irani et al., 2011). 3.2. Regression analysis A quadratic polynomial model was established on the experimental results of CCD to identify the relationship between responses and variables. The proposed model based on the regression coefficients for the sulfur reduction can be stated as:
Y ¼ 8:83 þ 0:17X 1 þ 10:33X 2 þ 0:5X 21 þ 1:5X 22 0:25X 1 X 2
ð4Þ
where Y is the response value of sulfur reduction (mg/l), X1 is the oil/water phase ratio (%) and X2 is the initial sulfur concentration (mg/l). The linear and square effects of X2 were found to be significant (P < 0.1). A positive sign of the coefficient represents a synergistic effect, while a negative sign indicates an antagonistic effect.
Table 2 Results of Central Composite Design (CCD) for BDS and kLa optimization. Std
Oil/water phase ratio (%)
Initial sulfur concentration (mg/l)
Sulfur reduction (mg/l)
Predict (mg/l)
Ug (l/min)
v (l)
kLa (s1)
Predict (s1)
1 2 3 4 5 6 7 8 9
1 1 1 1 1 1 0 0 0
1 1 1 1 0 0 1 1 0
0.0 1.0 21.0 21.0 9.5 9.5 0.0 21.0 8.5
0.083 0.92 21.25 21.08 9.17 9.50 0.00 20.67 8.83
1 1 1 1 1.41 1.41 0 0 0
1 1 1 1 0 0 1.41 1.41 0
0.0025 0.0066 0.0150 0.0058 0.0119 0.0056 0.0011 0.0110 0.020
0.003 0.007 0.016 0.006 0.011 0.006 0.002 0.011 0.020
10588
Z.A. Irani et al. / Bioresource Technology 102 (2011) 10585–10591
1.8
Dry Cell Weight ( g/l )
1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0
5
10
15
20
25
30
35
Time (h)
: 12 mg/l,
: 14 mg/l,
: 21mg/l,
: 28mg/l
Fig. 1. Biomass growth of Gordonia alkanivorans RIPI90A for different initial sulfur concentrations.
Fig. 2. Sulfur reduction of diesel by growing cells of Gordonia alkanivorans RIPI90A during the growth phase for different initial sulfur concentrations. Table 4 ANOVA of the model for kLa (s1).
Table 3 ANOVA of the model for sulfur reduction (mg/l). Term
Degree of freedom
Mean square
F
P
Model X1 X2
5 1 1 1
129.217 0.167 640.667 0.500
930.36 1.20 4612.80 3.60
0.000 0.353 0.000 0.154
1
4.500
32.40
0.011
1 3
0.250 0.139
1.80
0.272 –
X 21 X 22 X1X2 Residual 2
–
Term
Degree of freedom
Mean square
F-value
Probe
Constant X1 X2 X1X2
5 1 1 1 1
0.000060 0.000025 0.000083 0.000044 0.000091
72.11 29.59 99.60 110.29 169.81
0.003 0.012 0.002 0.002 0.001
1
0.000141
53.34
3
0.000001
X 21 X 22 Residual
–
0.005 –
R2 = 0.9917, R2 (adjusted) = 0.9780.
2
R = 0.9994, R (adjusted) = 0.9983.
The interaction term of X1 and X2 had a negative relationship with the BDS process while the variables X1, X2 and the quadratic term of X1 and X2 had positive effects. High value of the parameter estimate for X2 indicates the importance of initial sulfur concentration in sulfur reduction. Regression analysis revealed a coefficient of determination (R2) value of 0.9994 and adjusted coefficient of determination (AdjR2) of 0.99 that indicates high dependence between the observed and the predicted values of response (Table 3).The relationship between mass transfer coefficient (kLa) and two variables in coded units can be expressed as:
Y ¼ 0:02 0:001751X 1 þ 0:003213X 2 0:005606X 21 0:006956X 22 0:03325X 1 X 2
ð5Þ
where Y is the response value of kLa, X1 is the superficial gas velocity (l/min) and X2 is working volume (l). The effect of all the linear,
square and interaction terms of X1 and X2 (P < 0.05) were found to be significant on the kLa. A relatively high value of the parameter estimate for X2 and high values of coefficients of quadratic term of X2 and interaction term of X1 and X2 show a high value of significance, indicating the importance of these variables on the kLa. The variable X2 had a positive significant relationship with the kLa while the variable X1 and quadratic terms of X1 and X2 and interaction terms of X1 and X2 had negative significant effects. Regression analysis revealed a coefficient of determination (R2) value of 0.9917 and adjusted coefficient of determination (AdjR2) of 0.9780 that indicates high dependence between the observed and the predicted values of response (Table 4). 3.3. Effect of parameters The entire relationships between factors and response can be better understood by examining the planned series of contour
Z.A. Irani et al. / Bioresource Technology 102 (2011) 10585–10591
10589
Fig. 3. Contour plot for sulfur reduction (mg/l).
Fig. 5. Process optimization curve for sulfur reduction.
Fig. 4. Contour plot for kLa (s1).
plots (Fig. 3 and Fig. 4) generated from the predicted models (Eqs. (4) and (5)). Fig. 3 represents contour plot between the responses i.e. sulfur reduction (mg/l) and the combined effect of initial sulfur concentration (mg/l) and oil/water phase ratio (%) on the removal of sulfur. The figure shows that the uptake was increased with the increase in initial sulfur substrate concentration and decrease of oil/water phase ratio. Also, Fig. 4 represents contour plot between the responses i.e. kLa (s1) and the combined effect of superficial gas velocity, Ug (l/min) and working volume, v (l) on the kLa (s1). The figure shows that the value of kLa increased in the beginning and decreased after increasing from middle values.
3.4. Interpretation of process optimization curve Responses optimization helps to identify the factor settings that optimize a single response or a set of responses. It is useful in determining the operating conditions that will result in a desirable response. In the present study, the goal for sulfur reduction using growing cells of G. alkanivorans RIPI90A was to obtain a value at or near the target value of 21 mg/l. Uptake values less than 0 mg/ l and greater than 28 mg/l were unacceptable. Both weight and
Fig. 6. Process optimization curve for kLa.
importance were set at 1. The optimum condition which is defined as the best combination of factor setting for achieving the optimum response, was found to be oil/water phase ratio (30%) and initial sulfur concentration (28 mg/l) for a predicted response of 21.25 and desirability score of 1 (Fig. 5). In the airlift bioreactor, the goal for kLa was to obtain a value at or near the target value of 0.02. Moreover, kLa values less than 0.001 and greater than 0.02 were unacceptable. Both weight and importance were set at 1. The optimum condition which is defined as the best combination of factor setting for achieving the optimum response, was found to be superficial gas velocity (2.5 l/min) and working volume (6.6 l) for a predicted response of 0.0206 and desirability score of 1 (Fig. 6).
10590
Z.A. Irani et al. / Bioresource Technology 102 (2011) 10585–10591
Table 5 The sulfur removal comparison for different airlift bioreactors after 30 h. Microorganism
Substrate
% Sulfur removal
Remarks
RIPI 90A Rhodococcus sp. ATCC 39327 Strain No. 06
Diesel Diesel Kerosene
50 0 34 24
This workGrowing cells30:70 diesel:aqueous Nandi (2010)Growing cells50:50 diesel:aqueous Sanchez et al. (2008)Resting cells 50:50 kerosene:aqueous
Table 6 The sulfur removal comparison for different bioreactors. Microorganism
Bioreactor
Substrate
% Sulfur removal
Remarks
RIPI 90A Rhodococcus sp.
Airlift Airlift
Diesel Diesel
50 100
ATCC 39327 Strain No. 06 R. globerulus DAQ3
Airlift
Kerosene
Two-layer continues
Diesel
64 53 12
Pseudomonas delafieldii R-8
5-L reactor
Diesel
47
R. erythropolis I-19
Energy Bio systems Corp. 2 L reactor Continues stirred tank
Middle distillate 1850 Diesel
67
This work growing cells 30:70 diesel:aqueous After 30 h Nandi (2010) Growing cells50:50 diesel:aqueous After 120 h Sanchez et al. (2008) Resting cells 50:50 kerosene:aqueous After 7 days Yang et al. (2007) Growing cells 20:80 diesel:aqueous After 120 h Guobin et al. (2006) High cell concentration culture 16:84 diesel:aqueous After 20 h Folsom et al. (1999) Resting cells 25:75 diesel:aqueous
50–70
McFarland et al. (1998) Resting cells
Rhodococcus sp.
3.5. Sulfur removal of diesel oil in airlift bioreactor under optimized conditions In order to investigate the G. alkanivorans RIPI90A sulfur removal under optimum conditions in the airlift bioreactor, the superficial gas velocity (Ug) and working volume (v) was set at 2.5 l/min, 6.6 l respectively, and oil/water phase ratio of 30% and initial sulfur concentration of 28 mg/l. The sulfur reduction in the airlift bioreactor was 14 mg/l after 30 h. The sulfur removal comparison with previous works for different airlift bioreactors after 30 h was shown in Table 5.As it is reported in Table 5, we obtained better result. Although sulfur removal for ATCC 39327 and strain No. 06 after 7 days were 64 and 53% respectively, high sulfur reduction rate is related to the first 24 h and no major changes is observed in the last 48 h (Sanchez et al., 2008). Moreover, Rhodococcus sp. was used effectively to reduce sulfur level from 500 mg/l to zero in 120 h but there is almost no change at first 24 h and high sulfur reduction rate is related to last 48 h (Nandi, 2010). The sulfur removal comparison with previous works for different bioreactors was shown in Table 6. Although growing cells of G. alkanivorans RIPI90A was used, sulfur removal of our work was almost similar to other works that used resting cells (Sanchez et al., 2008; Guobin et al., 2006; McFarland et al., 1998) or better (Yang et al., 2007). There are also better results than our work (Nandi, 2010; Sanchez et al., 2008; Folsom et al., 1999). This is because of the difference of microorganisms and the operating conditions. Therefore, airlift bioreactors can be used as a potential for BDS processes of diesel or other fuels and the designed airlift bioreactor is well-organized for this purpose.
4. Conclusion RSM technique was used for optimization of BDS of diesel by RIPI90A cells in an airlift bioreactor. The initial sulfur concentration was selected as effective factor for optimization of sulfur reduction while superficial gas velocity (Ug) and working volume (v) were selected as effective factors for optimization of liquid mass transfer
coefficient (kLa). The results of BDS under the optimum conditions (initial sulfur concentration, 28 mg/l; oil/water phase ratio, 30:70; superficial gas velocity, 2.51 l/min and working volume, 6.65 l) show that RSM is a helpful method to maximize sulfur reduction and the designed bioreactor is efficient for BDS of diesel. Acknowledgements The Iranian Research Institute of Petroleum Industry financed this research. We thank Khalaf Faryadin and Rohollah Nikbakht for their contribution. References Arenskötter, M., Bröker, D., Steinbüchel, A., 2004. Biology of the metabolically diverse genus Gordonia. Appl. Environ. Microbiol. 70, 3195–3204. Azimzadeh Irani, Z., Yazdian, F., Mohebali, G., Soheili, M., Mehrnia, M.R., 2011. Determination of growth kinetic parameters of a desulfurizing bacterium, Gordonia alkanivorans RIPI90A. Chem. Eng. Trans. 24, 937–942. Caro, A., Boltes, K., Leton, P., García-Calvo, E., 2008. Biodesulfurization of dibenzothiophene by growing cells of Psedomonas putida CECT 5279 in biphasic media. Chempsphere 73, 663–669. Chang, J.H., Kim, Y.J., Lee, B.H., Cho, K.S., Ryu, H.W., Chang, Y.K., Chang, H.N., 2001. Production of a desulfurization biocatalyst by two-stage fermentation and its application for the treatment of model and diesel oils. Biotechnology 17, 876– 880. Chowdhury, R., Pedernera, E., Reimert, R., 2002. Trickle-bed reactor model for desulfurization and dearomatization of diesel. AIChE J. 48, 126–135. EPA US. Environmental Protection Agency, 2011;
. Folsom, B.R., Schieche, D.R., Digrazia, P.M., Werner, J., Palmer, S., 1999. Microbial desulfurization of alkylated dibenzothiophenes from a hydrodesulfurized middle distillate by Rhodococcus erythropolis I-19. Appl. Environ. Microbiol. 65, 4967–4972. Guchhait, S., Biswas, D., Bhattacharya, P., Chowdhury, R., 2005. Bio-desulfurization of model organo-sulfur compounds and hydrotreated diesel-experiments and modelling. Chem. Eng. J. 112, 145–151. Guobin, S., Huaiying, Z., Jianmin, X., Guo, C., Wangliang, L., Huizhou, L., 2006. Biodesulfurization of hydrodesulfurized diesel oil with Pseudomonas delafieldii R-8 from high density culture. Biochem. Eng. J. 27, 305–309. Hasan, S.H., Ranjan, D., Talat, M., 2010. Water Hyacinth Biomass (WHB) for the biosorption of hexavalent chromium: optimization of process parameters. Bioresource 5 (2), 563–575. Hewitt, C., Nienow, A., 2007. The scale-up of microbial batch and fed-batch fermentation processes. J. Biotechnol. 131, 134–135.
Z.A. Irani et al. / Bioresource Technology 102 (2011) 10585–10591 Ma, X., Sakanishi, K., Mochida, I., 1994. Hydrodesulfurization reactivates of various sulfur compounds in diesel fuel. Ind. Eng. Chem. Res. 33, 218–222. Maghsoudi, S., Vossoughi, M., Kheirolomoom, A., Tanaka, E., Katoh, S., 2001. Biodesulfurization of hydrocarbons and diesel fuels by Rhodococcus sp. strain P32C1. Biochem. Eng. J. 8, 151–156. McFarland, B.L., 1998. Biodesulfurization. Curr. Opin. Microbiol. 2, 257–264. McFarland, B., Boron, D., Deber, W., Meyer, J., Jonson, A., Atlas, R., 1998. Biocatalytic sulfur removal from fuels: applicability for producing low sulfur gasoline. Crti. Rev. Microbiol. 24, 99–147. McHale, WD., 1981. Process for removing sulfur from petroleum oils. US patent number 4283, 270. Mehrnia, M.R., Bonakdarpour, B., Towfighi, J., Akbarnejad, M.M., 2004a. Design and operational aspects of airlift bioreactors for petroleum biodesulfurization. Environ. Prog. 23 (3), 206–214. Mehrnia, M.R., Towfighi, J., Bonakdarpour, B., Akbarnegad, M.M., 2004b. Influence of top-Section design and draft-tube height on the performance of airlift bioreactors containing water-in-oil microemulsion. J. Chem. Technol. Biotechnol. 79, 260–267. Mehrnia, M.R., Towfighi, J., Bonakdarpour, B., Akamejad, M.M., 2005. Gas hold-up and oxygen transfer in a draft-tube airlift bioreactor with petroleum-based liquids. Biochem. Eng. J. 22, 105–110. Mohebali, G., Ball, A.S., Rasekh, B., Kaytash, A., 2007a. Biodesulfurization potential of a newly isolated bacterium, Gordonia alkanivorans RIPI90A. Enz. Microb. Technol. 40, 578–584. Mohebali, G., Ball, A., Keytash, A., Rasekh, B., 2007b. Stabilization of water/gas oil emulsions by desulfurizing cells of Gordonia alkanivorans RIPI90A. Microbiology 153, 1573–1581. Mohebali, G., Ball, A., Keytash, A., Rasekh, B., 2008. Dimethyl Sulfoxide (DMSO) as the sulfur source for the production of desulfurizing resting cells of Gordonia alkanivorans RIPI90A. Microbiology 154, 878–885. Monticello, D.J., 1998. Riding the fossil fuel biodesulfurization wave. Chem. Tech. 28, 38–45. Monticello, D.J., 2000. Biodeslufurization and the upgrading of petroleum distillates. Curr. Opin. Biotechnol. 11, 540–546. Mukhopadhyaya, M., Chowdhury, R., Bhattacharya, P., 2006. Biodesulfurization of hydrodesulfurized diesel in a trickle bed reactor-experiments and modelling. J. Sci. Ind. Res. 65, 432–436.
10591
Nandi, S., 2010. Biodesulfurization of hydro-desulfurized diesel in an airlift reactor. J. Sci. Ind. Res. 69, 543–547. Oshiro, T., Kojima, T., Torii, K., Kawasoe, H., Izumi, Y., 1999. Purification and characterization of dibenzothiophene (DBT) sulfone monooxygenase, and enzyme involved in DBT desulfurization, from Rhodococcus erythropolis D-1. J. Biosci. Bioeng. 88, 610–616. Rashtchi, M., Mohebali, G., Akbarnejad, M.M., Towfighi, J., Rasekh, B., Keytash, A., 2006. Analysis of biodesulfurization of model oil system by the bacterium, strain RIPI-22. Biochem. Eng. J. 29, 169–173. Sanchez, O.F., Almeciga-Diaz, C.J., Silva, E., Cruz, J.C., Valderrama, J.D., Caicedo, L.A., 2008. Reduction of sulfur levels in kerosene by Pseudomonas sp. strain in an airlift reactor. Latin Am. Appl. Res. 38, 329–335. Shariati, F.P., Bonakdarpour, B., Mehrnia, M.R., 2007. Hydrodynamics and oxygen transfer behaviour of water in diesel microemulsions in a draft tube airlift bioreactor. Chem. Eng. Proc. 46, 334–342. Singh, H., Malhotra, M., George, N., Ghosh, S.K., 2001. Biocatalytic sulfur removal from fossil fuel, PetroTech. 75–90. Soleimani, M., Bassi, A., Margaritis, A., 2007. Biodesulfurization of refractory organic sulfur compounds in fossil fuels. Biotechnol. Adv. 25, 570–596. Wu, J., Wang, J.L., Li, M.H., Lin, J.P., Wei, D.Z., 2010. Optimization of immobilization for selective oxidation of benzyl alcohol by Gluconobacter oxydans using response surface methodology. Biores. Technol. 101, 8936–8941. Yang, J., Hu, Y., Zhao, D., Wang, S., Lau, P., Marison, I., 2007. Two-layer continuesprocess design for the biodesulfurization of diesel oils under bacterial growth conditions. Biochem. Eng. J. 37, 212–218. Yazdian, F., Shojaosadati, S.A., Nosrati, M., Pesaran Hajiabbas, M., VasheganiFarahani, E., 2009. Investigation of gas properties, design and operational parameters on hydrodynamic charachteristics, mass transfer and biomass production from natural gas in an external airlift loop bioreactor. Chem. Eng. Sci. 64, 2455–2465. Yazdian, F., Pesaran Hajiabbas, M., Shojaosadati, S.A., Nosrati, M., VasheganiFarahani, E., Mehrnia, M.R., 2010. Study of hydrodynamics, mass transfer, energy consumption and biomass production from natural gas in a forcedliquid vertical tubular loop bioreactor. Biochem. Eng. J. 49, 192–200.