Biofiltration of Methyl tert-Butyl Ether Vapors by ... - ACS Publications

Figure 1 Biofilter system: 1, air compressor; 2, pentane tank; 3, biofilter; 4, mass flow controller; 5, unidirectional valve; 6, humidifier; 7, cyclo...
0 downloads 0 Views 100KB Size
Environ. Sci. Technol. 2002, 36, 247-253

Biofiltration of Methyl tert-Butyl Ether Vapors by Cometabolism with Pentane: Modeling and Experimental Approach D A V I D D U P A S Q U I E R , †,‡ S E R G I O R E V A H , * ,† A N D R I C H A R D A U R I A †,‡ Departmento de Ingenierı´a Quı´mica, Universidad Auto´noma Metropolitana, UAM-Iztapalapa, Apdo Postal 55-534, CP 09340, Me´xico City, Me´xico, and Institut de Recherche pour le De´veloppement (IRD), Calle Cicero´n 609, Los Morales 11530, Me´xico City, Me´xico

Degradation of methyl tert-butyl ether (MTBE) vapors by cometabolism with pentane using a culture of pentaneoxidizing bacteria (Pseudomonas aeruginosa) was studied in a 2.4-L biofilter packed with vermiculite, an inert mineral support. Experimental pentane elimination capacity (EC) of approximately 12 g m-3 h-1 was obtained for an empty bed residence time (EBRT) of 1.1 h and inlet concentration of 18.6 g m-3. For these experimental conditions, EC of MTBE between 0.3 and 1.8 g m-3 h-1 were measured with inlet MTBE concentration ranging from 1.1 to 12.3 g m-3. The process was modeled with general mass balance equations that consider a kinetic model describing cross-competitive inhibition between MTBE (cosubstrate) and pentane (substrate). The experimental data of pentane and MTBE removal efficiencies were compared to the theoretical predictions of the model. The predicted pentane and MTBE concentration profiles agreed with the experimental data for steady-state operation. Inhibition by MTBE of the pentane EC was demonstrated. Increasing the inlet pentane concentration improved the EC of MTBE but did not significantly change the EC of pentane. MTBE degradation rates obtained in this study were much lower than those using consortia or pure strains that can mineralize MTBE. Nevertheless, the system can be improved by increasing the active biomass.

Introduction In many countries (United States, Mexico, France, etc.), fuel oxygenates are added to gasoline in order to enhance its octane number. In the United States, they have been used since 1988 to improve air quality in some metropolitan areas (1). These additives allow a better gasoline combustion and, consequently, reduce the resulting concentrations of carbon monoxide and unburned hydrocarbons. Among the oxygenates (ethanol, ethyl tert-butyl ether, tert-amyl methyl ether, etc.), methyl tert-butyl ether (MTBE) is being used the most by the industry because of its low cost, ease of production at refineries, favorable blending characteristics with other fuel components, and lack of phase separation in the presence * Corresponding author phone: +52-55-58-04-65-38; fax: +5255-58-04-64-07; e-mail: [email protected]. † Universidad Auto ´ noma Metropolitana. ‡ Institut de Recherche pour le De ´ veloppement. 10.1021/es010942j CCC: $22.00 Published on Web 12/13/2001

 2002 American Chemical Society

of water. Between 1984 and 1994, MTBE production rates increased by 26% annually in the United States. In 1993, its production ranked second among all organic chemicals manufactured (2). Consequently, the use of MTBE as an additive in gasoline has been increasing and has extended to other countries (Mexico, China, Saudi Arabia, etc.). However, due to this massive production combined with its mobility, persistence, toxicity, and high solubility in water (50 g L-1 at ambient temperature), it is an important pollutant of groundwater (3, 4). It has generally been considered that the persistence of MTBE is strongly related to its biological recalcitrance, although recent studies have shown that there is strong potential for natural attenuation (5). Because of the presence of MTBE in the environment, the U.S. EPA has announced its intention to reduce or eliminate this oxygenate in domestic fuels over the next 3 yr. To date, very few studies have been conducted to find microorganisms able to degrade MTBE and to utilize them in biological treatment. Consortia and pure microorganisms have been studied for their ability to degrade MTBE as the sole source of carbon and energy in aerobic (6, 7) and anaerobic conditions (8). However, a limited amount of work has been performed on the cometabolic degradation of MTBE by pure cultures. In one recent study, degradation of MTBE by filamentous fungus in the presence of diethyl ether (DEE) was reported (9). Biodegradation of MTBE by three propaneoxidizing strains (ENV421, ENV425, and Pseudomonas putida) was studied (10). Garnier et al. (11) showed that a soil consortium was able to degrade to completion gasoline containing MTBE. However, when MTBE was tested alone, no degradation was observed. This study demonstrated that MTBE was degraded by cometabolism with n-alkanes (pentane, hexane, and heptane) present in the gasoline. From the consortium, a pentane-oxidizing bacteria (P. aeruginosa) isolated from this consortium was able to degrade MTBE by cometabolism with pentane (12). Biofiltration of MTBE vapors released from a wastewater treatment was reported (13). Acclimation of the microorganisms for more than 1 yr was necessary to observe MTBE degradation. In this case, the biofilter performance was approximately 7 g m-3 h-1, which corresponded to a removal efficiency of nearly 97%. After 6 months of acclimation of aerobic microbial consortium able to degrade MTBE, Fortin and Deshusses (14) obtained elimination capacity of MTBE around of 50 g m-3 h-1 (90% of removal efficiency) in laboratory-scale biotrickling filters. A pure strain, with high MTBE mineralization rate, was reported (15). This strain nevertheless shows some inhibition when BTEX is present (16). For more information, the reader will be able to refer to the paper by Deeb et al. (17) on aerobic MTBE biodegradation. Biological waste air treatment has proven to be a good alternative as compared with conventional treatments (stripping, incineration, adsorption, absorption, ...). Until 1980, biofiltration was mainly used to reduce odor in off-gases, but in the early 1980s, the field of application was extended to the removal of many other volatile organic compounds. Biofiltration is a cost-effective method to treat large volumes of contaminated air with moderate concentrations of volatile organic compounds (18). This technology has low power requirements, and the process equipment is simple and generally easy to operate. In biofiltration, the gas to be treated is forced through a bed packed with material on which microorganisms are attached as a biofilm. Biodegradable volatile compounds and oxygen diffuse into the biofilm where they are subsequently biologically oxidized into less harmful substances such as CO2 and H2O. VOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

247

The aim of this study is to investigate the degradation of MTBE vapors by cometabolism with pentane using a culture of pentane-oxidizing bacteria. The study was carried out with a biofilter packed with a mineral support (vermiculite). A mathematical model was proposed to understand and predict cometabolic degradation of MTBE in the presence of pentane in the reactor. Its predictions were compared with experimental results obtained from the biofilter, and a sensitivity study of the different model parameters was conducted.

gbiomass-1 s-1, Xfa is the active biomass concentration in g mbiolayer-3, and Ks(i) is the half-saturation constant of compound i in g m-3. MTBE is not consumed as a source of carbon for cell synthesis (11); however, its degradation is possible when pentane is supplied in the reactor. Therefore, the biodegradation of MTBE occurs only in the presence of pentane. MTBE degradation is described by eq 3, considering the competitive inhibition and the stimulating effect of pentane:

Theory

Sm rm ) kx(MTBE)Xfa Sm + Ks(MTBE) 1 +

The theoretical model presented in this section was based on that proposed by Ottengraf and van den Oever (19). Crosscompetitive inhibition between MTBE (cosubstrate) and pentane (substrate) was described using a kinetic model similar to the one presented by Arcangeli and Arvin (20). To predict the degradation of MTBE and pentane in the biofilter, the following assumptions were made: (i) The airstream passes through the biofilter bed in plug flow mode. (ii) The gas pollutants at the biofilm/air interface are in equilibrium as dictated by Henry’s law. (iii) The thickness of the biofilm (δ) is small as compared to the radius of the particles, and thus, planar geometry can be used. (iv) Pollutants are transported by diffusion in the biofilm. (v) The biofilm is homogeneous, and its density, defined as the amount of dry biomass per unit volume of biolayer, is constant. (vi) Only pentane (substrate) and MTBE (cosubstrate) are assumed to be rate limiting. The differential equation describing the concentration of the compound i (Si) inside the biolayer at steady-state conditions is

d2Si Dei 2 ) ri dx

(1)

where Dei is the diffusion coefficient of compound i into the biofilm (m2 s-1); ri is the reaction rate of product i (g m-3 s-1); and i is pentane (p) or MTBE (m). The following boundary conditions are employed:

[

x ) 0 Si ) Cgi/mi x ) λ dSi/dx ) 0

]

(

Sm Ks(MTBE)

)

(2)

where kx(i) is the maximum substrate utilization rate in gi 248

9

(

Ks(pent) Sp

)

×

Sp + Ks(pent)

)

(3)

The kinetic model considers that the total biomass (Xf) is comprised of two parts: an active biomass (Xfa) capable of degrading pentane and MTBE and an inactive biomass, which includes dead cells and exopolymeric substances (EPS) (18, 19). From a differential mass balance in the gas phase, Cgi may be calculated as a function of the height (h) in the filter bed according to

-Ug

dCgi ) NiAs dh

(4)

where Ni (g m-2 s-1) is the substrate flux into the biofilm, As is the interfacial area per unit of reactor volume (m2 m-3); h is the distance from the biofilter entrance (m); and Ug is the superficial gas flow rate (m s-1). Finally, the expression for Ni is given by the Fick’s law:

( )

Ni ) - Dei

dSi dx

x)0

(5)

A Runge-Kutta integration algorithm was used to solve the system of eqs 1-5. The integration step in biofilter thickness (∆δ) and height (∆h) was equal to 0.1. This model estimated the different concentration profiles for each pollutant in the biofilm and along the biofilter.

Materials and Methods

where Cgi is the gas concentration of compound i (g m-3), λ is the active biofilm thickness (m), and mi is the distribution coefficient for compound i in the air/water system. The second boundary condition follows from the consideration that somewhere in the biofilm the concentration gradient equals zero. In the reaction-limiting case, the condition is met at x ) λ ) δ. When diffusion limitation exists, the depth of penetration λ in the biofilm may be smaller than its thickness δ; however, the condition is still valid. The reaction rate of pentane, which supports biomass growth, can be described by eq 2, assuming a competitive inhibitory effect of the cosubstrate. The inhibition coefficient of the competitive inhibitor is approximated by its singlesubstrate half-saturation coefficient. Although an inhibitory effect of pentane has been found for this system (12), it was not considered as the gas-phase concentrations used in this study were much lower than those used in ref 12:

Sp rp ) kx(pent)Xfa Sp + Ks(pent) 1 +

(

Sp

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 2, 2002

Inoculum and Mineral Medium. A bacteria (P. aeruginosa), isolated from soil samples contaminated with gasoline (12), was used for the experiments. This bacteria was able to degrade MTBE when cometabolized in the presence of pentane. The isolated strain was enriched, and 2 L of medium was prepared for the inoculation. The medium was a mineral salt solution consisting of the following components: MgSO4‚ 7H2O, 1.0 g L-1; KNO3, 1.0 g L-1; CaCl2, 0.2 g L-1; Fe-EDTA, 0.4 g L-1; K2HPO4, 74 mg L-1; KH2PO4, 26 mg L-1; FeSO4, 0.2 mg L-1; ZnSO4‚7H2O, 0.01 mg L-1; H3BO3, 0.03 mg L-1; CoCl2‚ 2H2O, 0.02 mg L-1; MnCl2‚4H2O, 0.003 mg L-1; NiCl2‚2H2O, 0.002 mg L-1; CaCl2, 0.001 mg L-1; NaMoO4, 0.03 mg L-1. Inoculation was made by flooding the entire reactor twice and keeping each time the solution for 30 min in contact with the support before pumping it out from the reactor. Chemicals. MTBE (98%, d ) 0.740) was from Sigma (St. Louis, MO). A pressurized cylinder containing gaseous 1.0% pentane in air (Praxair, Me´xico) was used to feed the substrate. Biofilter and Packing Material. A three-stage biofilter consisted of a cylindrical glass column with i.d. of 0.078 m and 1 m height. It was equipped with a number of sampling ports along its length (Figure 1). Each filter stage was packed with a mineral support (vermiculite) and supported by a plate to ensure a homogeneous packing distribution. Particles of vermiculite were sieved through -4+5 mesh screens to

TABLE 1. Operating Conditions of the Biofilter inlet MTBE concn (g m-3) inlet pentane concn (g m-3) air flow (L h-1) reactor vol (L) temp (°C) EBRT (τ) (h)

1.1-12.3 0.7-19.2 0.8-39.5 2.4 30 0.06-2.85

tration difference (Cgin,i - Cgout,i) and the inlet concentration (Cgin,i).

Determination of Model Parameters

FIGURE 1. Biofilter system: 1, air compressor; 2, pentane tank; 3, biofilter; 4, mass flow controller; 5, unidirectional valve; 6, humidifier; 7, cyclone; 8, mixing bottle; 9, MTBE bottle in a refrigerated system; 10, peristaltic pump. obtain an average particle size of 4.1 mm. The active filter bed height was 0.5 m, and the packing density was 110 g of dry vermiculite L-1. The initial water content and pH were 70% and 7.0, respectively. The biofilter was placed in a controlled chamber at 30 ( 2 °C. Addition of MTBE and Pentane. A stream containing 1% pentane in air from a pressurized cylinder was supplied directly to the reactor after being saturated with water vapor by sparging through a column. Pentane concentrations lower than 1% were obtained by mixing pentane with compressed air. The main airstream was controlled by an electronic mass flow sensor (33116-20 Cole Parmer, USA). A small air flow was sparged through a 0.5-L flask containing liquid MTBE and mixed with the air contaminated by pentane before entering at the top of the reactor. The MTBE flask was placed into a refrigerated system to control temperature variations. Analysis. Pentane and MTBE concentrations in the gas phase were determined by gas chromatography. A 250-µL airtight syringe was used to sample at different heights in the biofilter. These samples were injected into a FID gas chromatograph (Hewlett-Packard 5890, USA) equipped with a silica capillary column (30 m CP WAX 52 CB, USA). The operating conditions were as follows: injector, 225 °C; oven, 40 °C; and detector, 225 °C. The detection limit was approximately 5 mg of pollutant/m3 of air. The Lowry method was used to determine total biomass concentration assuming that 50% of total biomass is protein. Active biomass concentration was estimated from the work of Pineda et al. (21). Experimental Procedure. The first series of experiments were conducted feeding only pentane to the biofilter. These experiments were performed with different pentane inlet concentrations (Cgin(pent)) ranging from 0.72 to 17.4 g m-3, and empty bed residence times (EBRT ) τ) range from 0.06 to 2.85 h. After steady state was attained, air samples were taken from different levels of the column. The second series of experiments were conducted in the presence of pentane and MTBE. These experiments were carried out by maintaining Cgin(pent) (18 g m-3) and τ (1.1 h) constant. Only, the inlet MTBE concentration (Cgin(MTBE)) was varied from 1.1 to 12.3 g m-3. Operating conditions of the biofilter are presented in Table 1. Definitions. The elimination capacity (EC) of compound i is defined as the difference between inlet (Cgin,i) and outlet (Cgout,i) concentrations divided by the EBRT (τ). EC is expressed as gpollutant m-3biofilter h-1. Removal efficiency (%Ef) of compound i is defined as the ratio between the concen-

To determine As, δ, Xf, and Xfa, the model of biolayer growing on the vermiculite support and the experimental methods presented by Pineda et al. (21) were used. To determine As, the form of the vermiculite particles is assumed to be cubic. These cubes are formed by layered and sandwiched thin mineral sheets. Observations by scanning electronic microscope (SEM) showed that the formation of the biofilm takes place between these layers and that the biofilm only grew on four faces of the particles. The volume occupied by the vermiculite was Vv ) (1 - )Vt, where  is the porosity of the packing material ( ) 0.71) and Vt is the volume of the reactor (2.4 L). Since the vermiculite particles are cubic, the volume of one particle was V1 ) a3 (where a ) 4.1 mm). Therefore, the approximate number of particles (n) constituting the packing material is equal to Vv/V1. Considering that the exchange area for one particle is A1 ) 4a2, the total exchange area was At ) nA1. Finally, As ) At/Vt was equivalent to As ) 284 m2 m-3 of reactor for this study. Pineda et al. (21) estimated the volume of superficial water by differential thermogravimetry. They considered that this superficial water is the first to evaporate and therefore corresponds to the first peak of the differential thermogravimetry spectrum of wet vermiculite. The volume of superficial water was considered equal to the biofilm volume. The superficial water represented about 25% of the total water, which is equal to 1 g of water/g of dry vermiculite. Given the dry packing material density, the total volume of water (Vw) constituting the biolayer was determined (Vw ) 259.6 cm3). Since the thickness of the biolayer is defined by δ ) Vw/(AsVt), the numerical value was found to be δ ) 387 µm. The total biomass density of the biofilm (Xf) was equal to 28 300 g m-3. The active biomass was considered equal to 8% of total biomass (21). Thus, the active biomass density of the biofilm (Xfa) was calculated to be 2264 g m-3. In the literature, a wide range of values for As (25-1000 m-1) is found for different supports as reviewed by Pineda et al. (21). In the case of vermiculite particles, an As value of 675 m-1 was reported for a size distribution ranging from 0.52 to 5.4 mm (21). The value of As reported here for the same packing material is lower since a larger mean particle size (4.1 mm) was used. Values of δ (387 µm) and Xf (28 300 g m-3), calculated in this study, are in the same order of magnitude as those reported by different authors (21). From these values, a total active biomass in the reactor can be calculated as δAsXfa ) 250 g mreactor-3, which was lower than values that have been reported elsewhere (2, 18). Diffusivities of pentane and MTBE (Dei) in the biolayer were assumed to be equal to their diffusivities in water (Di), corrected by a factor depending on the Xf, according to the expression of Fan et al. (22). Kinetic parameters, Ks(pent), Ks(MTBE), kx(pent), and kx(MTBE) were previously determined from batch experiments (12). The half-saturation constant of pentane (Ks(pent)) was equal to the apparent saturation constant deduced from the integrated form of the first-order growth equation coupled with Monod kinetics. This value (0.019 g m-3) was obtained VOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

249

TABLE 2. Values of the Model Parametersa De(pent) (m2 s-1) De(MTBE) (m2 s-1) kx(pent) (g g-1db s-1) kx(MTBE) (g g-1db s-1) Ks(pent) (g m-3) Ks(MTBE) (g m-3) mpent at 25 °C mMTBE at 25 °C h (m) δ (m) As (m2 m-3) Xf (g db m-3biofilm) Xfa (g db m-3biofilm) T (°C) a

1 × 10-9 0.9 × 10-9 58.4 × 10-6 2.87 × 10-6 0.019 185 44.4 0.022 0.5 0.000387 284 28,300 2,264 30

db, dry biomass.

for a pentane concentration of 2.9 µg L-1 and represents the real saturation constant of pentane. The pentane maximum substrate utilization rate (kx(pent)) was obtained by dividing the maximum growth rate (µmax ) 0.19 h-1) by the growth yield coefficient (Yg ) 0.9 g of biomass/g of pentane). MTBE maximum substrate utilization rate (kx(MTBE)) was calculated from the value of the MTBE degradation rate (3.9 nmol min-1 mg-1cell protein) using the MTBE molecular weight and percentage of protein of dry biomass, 88.15 g mol-1 and 50%, respectively. Table 2 presents the parameter values used for solving the model equations. Few studies have addressed the cometabolic biodegradability of MTBE by consortia and pure cultures. Hardison et al. (9) reported that MTBE was degraded by a filamentous fungus (Graphium sp.) in the presence of butane at a ratio varying between 0.03 × 10-6 and 0.252 × 10-6 g g-1db s-1. Biodegradation of MTBE by three propane-oxidizing strains with rates between 0.3 × 10-6 and 6.7 × 10-6 g g-1db s-1 were reported (10). These values are in the same order of magnitude as found in this study (kx(MTBE) ) 2.87 × 10-6 g g-1db s-1). However, Hanson et al. (15) obtained higher kx(MTBE) values (up to 1.8 × 10-3 g g-1db s-1) with a bacterial pure culture that utilized MTBE as its sole carbon and energy source. No data of Ks(pent), Ks(MTBE), and kx(pent) pertaining to the cometabolism of MTBE with alkanes are available in the literature. Nevertheless, values of similar parameters for TCE (trichloroethylene) degraders by cometabolism growing on toluene as the sole carbon sources were reported (20). The values of Ks(toluene) and kx(toluene) varied between 0.027 and 13.8 g m-3 and between 2.4 × 10-6 and 1.9 × 10-4 g g-1db s-1, respectively. These values are in agreement with these obtained in this study for Ks(pent) and kx(pent) (Table 2). Arcangeli and Arvin (20) reported values of Ks(TCE) (0.173 g m-3 < Ks(TCE) < 19.6 g m-3) lower than Ks(MTBE) (185 g m-3). This difference probably is due to the lower affinity for MTBE as compared to TCE for the respective enzyme produced when pentane and toluene are used as carbon sources.

Results and Discussion Degradation of Pentane without MTBE Addition. Figure 2a-c represents the model and experimental profiles of pentane vapor concentrations along the column obtained for different conditions of EBRT and inlet pentane concentrations (Cgin(pent)). These experiments show that the model gives a good representation of pentane degradation. Better agreement was found for Figure 2b,c where the experimental errors were reduced. For Figure 2a,b, the model predicts a linear decrease in pentane concentration in the filter bed, which suggests that the biofilm was fully active. In this case, the degradation rate is independent of the substrate concentration and the degradation in the biofilm follows zero250

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 2, 2002

FIGURE 2. Predicted and experimental profiles of the pentane concentration in the biofilter for pentane inlet concentration of 0.72 g m-3 and τ of 0.077 h (a); 17.4 g m-3 and τ of 0.58 h (b); 13.5 g m-3 and τ of 2.85 h (c). Experimental values (b) and model predictions (s). Error bars (|).

TABLE 3. Experimentally Obtained and Model-Predicted Steady-State Elimination Capacity for Pentane Vapors Alonea τ in h

Cgin(pent) (g m-3)

ECexp (g m-3 h-1)

ECpred (g m-3 h-1)

error (%)

0.0583 0.077 0.58 2.11 2.85

2.4 0.7 17.4 14.9 13.5

6.8 1.3 12.0 7.0 4.7

2.7 0.8 12.2 6.5 4.6

-60.3 -36.4 +1.3 -7.3 -1.0

a EC exp, experimental elimination capacity; ECpred, predicted elimination capacity.

order kinetics (23). At high EBRT (τ ) 2.85 h) (Figure 2c), the biofilter was not completely active. The model predicts a non-linear decrease in pentane concentration in the biofilter that suggests that the biofilter zones are limited by diffusion. Table 3 presents the percentage of error obtained from the comparison between the mathematical model and the experimental data for different Cgin(pent) and EBRT. The highest percentage of error, ranging from - 60 to - 36%, were obtained for the low Cgin(pent) and EBRT while the errors (-7.3 to + 1.3%) decreased when Cgin(pent) and EBRT are higher. In these experiments, the highest pentane %Ef (40%) was obtained at the maximum EC of 12 g m-3 h-1. n-Alkanes are strongly hydrophobic, having Henry’s coefficients (m) in the range of 10-100 (g m-3gas/g m-3water). Because of poor mass transfer from the gas to the aqueous phase, low ECs for n-alkanes were measured in biofilters. For example, low hexane ECs of (0.2-5.4 g of methane equiv. m-3 h-1) were attained using biofilter treating fuel vapors (24). In a full-scale biofilter treating hexane from waste gases of an oil refinery, 68% was eliminated at an EC of 2.5 g m-3 h-1 (25). Higher ECs of hexane were reported in laboratoryscale biofilters: 32 g m-3 h-1 (%Ef ) 39%) (25) and 21 g m-3 h-1 (%Ef ) 99%) (26).

TABLE 4. Elimination Capacities of Pentane and MTBE Obtained from Mathematical Model and Experimentsa τ in h

Cgin,p (g m-3)

Cgin,m (g m-3)

ECexp,p (g m-3 h-1)

ECpred,p (g m-3 h-1)

error,p (%)

ECexp,m (g m-3 h-1)

ECpred,m (g m-3 h-1)

error,m (%)

1.1 1.1 1.1 1.1 1.1

17.9 19.2 18.8 18.8 18.7

1.1 4.8 9.6 11.0 12.3

9.1 7.2 7.5 5.6 4.4

9.7 7.5 5.3 5.1 4.5

+7.6 +3.7 -29.5 -8.5 + 2.3

0.3 0.7 1.5 0.7 1.8

0.2 0.6 0.8 0.9 0.9

-45.4 -13.2 -45.4 +23.7 -47.8

a

ECexp, experimental elimination capacity; ECpred, predicted elimination capacity.

FIGURE 4. Influence of the inlet MTBE concentration (Cgin) on the elimination capacities (EC) of pentane and MTBE. Experimental values: (b) pentane, (O) MTBE. Model predictions: (s) pentane, (- -) MTBE.

FIGURE 3. Predicted and experimental profiles of the pentane and MTBE concentrations in the biofilter. (a) Pentane inlet concentration of 19 g m-3 and MTBE inlet concentration of 4.8 g m-3. (b) Pentane inlet concentration of 18 g m-3 and MTBE inlet concentration of 11 g m-3. Experimental values: (b) pentane, (O) MTBE. Model predictions: (s) pentane, (- -) MTBE. Error bars (|). This first series of experiments showed that the model predicted well the experimental profiles of pentane concentration along the biofilter and, consequently, that the hypotheses used for the determination of the As, δ, Xf, Xfa, and Ks(pent) were appropriate. It should be noted that a preliminary fitting of these parameters with the mathematical model was not used. Degradation of MTBE by Cometabolism with Pentane. A second series of experiments were conducted by maintaining the inlet pentane concentration (18 g m-3) and the EBRT (1.1 h) constant. Only the inlet MTBE concentration was varied as shown in Table 1. Figure 3a,b represents pentane and MTBE concentration profiles along the biofilter for inlet MTBE concentrations (Cgin(MTBE)) of 4.8 and 11 g m-3. At the steady-state condition, a low MTBE EC of 0.7 g m-3 h-1 was found for both experiments. Experiments in the biofilter (not shown) showed that the MTBE was not degraded in the absence of pentane. The EC of MTBE was much lower than the ECs reported in biofilter (ECmax ) 15 g m-3 h-1) (13) and trickling biofilter (ECmax ) 50 g m-3 h-1) (14). Both works have reported that MTBE is completely mineralized by their consortia. For pentane, the EC was maintained at approximately 7 g m-3 h-1. For the two compounds, the relation between the concentration and the biofilter height was linear. This behavior indicates that rates of MTBE and pentane degradation were limited by biological reaction. This can be further

supported by the fact that the system operated with low biomass as mentioned earlier. Improved volumetric rates can be obtained by increasing the biomass in the reactor through nutrient medium addition or by the use or more efficient strains. The mathematical model predicts, with reasonable error (Table 4), the experimental profiles of the pentane and MTBE concentration. Errors in the pentane and MTBE EC prediction ranged from 3.7% to -13.2% (Figure 3a) and from -8.5% to 23.7% (Figure 3b). The studies of Garnier et al. (12) showed in microcosms the effect of the concentration of both compounds on the elimination rates. The effect of the Cgin(MTBE) on the EC of pentane and MTBE was investigated (Figure 4). Higher EC of MTBE were found as the inlet MTBE concentration increases. On the other hand, the marked inhibition of MTBE on the EC of pentane appears very clearly in this figure (50% at around 10 g m-3). Moreover, as the ratio of pentane/MTBE decreases, the EC of MTBE increases slowly. The inhibition of MTBE on pentane degradation is confirmed from the results of the Figure 5, which represents %Ef of pentane and MTBE as a function of the MTBE inlet concentration. When the concentration of MTBE increases from 0 to 20 g m-3, %Ef of pentane and MTBE decrease from 60 to 20% and from 20 to 6%, respectively. Similarly, the %Ef of MTBE decreases but more slowly. The competitive inhibition of MTBE on the pentane uptake was detected even at the lowest MTBE concentrations. A small inhibitory effect of liquid MTBE concentrations of 200 mg L-1 on the biodegradation of aromatic hydrocarbons (benzene, toluene, ethylbenzene, and xylenes) was reported (27). Arcangeli and Arvin (20) investigated the cometabolic degradation of TCE with toluene as the primary source of carbon in a continuously fed biofilm reactor. These authors showed that TCE inhibits toluene degradation for TCE concentrations above 50 µg/L. These results were confirmed from an investigation of TCE degVOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

251

TABLE 5. Model Sensitivity to Values of Pentane Half-Saturation Constant (Ks(pent)), MTBE Half-Saturation Constant (Ks(MTBE)), Pentane Maximum Substrate Elimination (kx(pent)), MTBE Maximum Substrate Elimination (kx(MTBE)), Interfacial Area Per Unit of Reactor Volume (As), Total Biomass Concentration (Xf), and Biofilm Thickness (δ)a E (%) error on R pentane (%) error on R MTBE (%)

Ks(pent)

Ks(MTBE)

kx(pent)

kx(MTBE)

As

Xf

δ

(20 (+) -13.5 (-) +16.7 (+) -9.2 (-) +10.6

(20 (+) +6.3 (-) -9.0 (+) -10.0 (-) +12.2

(20 (+) +15.1 (-) -16.5 (+) -2.8 (-) +2.8

(20 (+) +0.5 (-) -0.5 (+) +20 (-) -20

(20 (+) +15.0 (-) -18.0 (+) +15.0 (-) -18.0

(20 (+) +15.0 (-) -18.0 (+) +15.0 (-) -18.0

(20 (+) +15.0 (-) -18.0 (+) +15.0 (-) -18.0

a R is the relative value of elimination capacity. The reference values of pentane and MTBE EC are 5.95 and 0.52 g m-3 h-1, respectively. These values were obtained with a Cgin(pent) ) 15 g m-3, Cgin(MTBE) ) 5 g m-3, and τ ) 1 h. The reference value for Ks(pent), Ks(MTBE), kx(pent), kx(MTBE), As, Xf, and δ are those reported in Table 2. (+)/(-), represents the sign of the percentage variation of Ks(pent), Ks(MTBE), kx(pent), kx(MTBE), As, Xf, and δ.

FIGURE 5. Influence of the inlet MTBE concentration (Cgin) on the removal efficiencies (%Ef) of pentane and MTBE. Experimental values: (b) pentane, (O) MTBE. Model predictions: (s) pentane, (- -) MTBE. radation with P. putida F1 in the presence of toluene as the primary carbon source (28). In contrast with MTBE, the degradation of TCE may bring about the accumulation of toxic intermediates. For the case of MTBE and pentane, Garnier et al. (12) found for this bacterium that the ratio between the real saturation constants, considering the Henry’s coefficient, of MTBE and pentane was approximately 64 000. The much higher affinity for the pentane by the enzymes involved in the degradation process was confirmed in the biofiltration experiments. Table 4 presents a comparison between the predicted EC (ECpred) and experimental EC (ECexp) of pentane and MTBE. Except for the Cgin(MTBE) of 9.6 g m-3, the percentage of error of pentane and MTBE EC predicted ranged from -8.5 to +7.6%. These errors are higher (-45.4 to + 23.7%) in the case of MTBE due to its low EC, which increases the experimental error. Figure 6 represents the influence of the Cgin(pent) on the relative EC of pentane and MTBE obtained from the mathematical model. These results were obtained by maintaining Cgin(MTBE) and the EBRT equal to 5 g m-3 and 1.1 h, respectively, and are reported as a percentage of the EC (ECr) that would be obtained without MTBE (i.e., no competitive inhibition). It is predicted that the ECr of pentane varies little with the Cgin(pent) and decreases slowly when its concentration increases. Up to a pentane concentration in air of 20 g m-3 (0.45 mg L-1), no inhibition on MTBE degradation was predicted (Figure 6) due to the low solubility of pentane in water. This result is similar to that obtained from batch experiments (11). These authors demonstrated that the degradation of MTBE was stimulated in the presence of pentane and proportional to its concentration. In the case 252

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 2, 2002

FIGURE 6. Model predictions of the influence of the inlet pentane concentration (Cgin(pent)) on the relative elimination capacities (ECr) of pentane (s) and MTBE (- -) for Cgin, MTBE of 5 g m-3, and τ ) 1.1 h. ECr is reported as percent to the EC pent or MTBE obtained without MTBE (i.e., no competitive inhibition). of TCE degradation in cometabolism with toluene, Arcangeli and Arvin (20) observed that degradation of TCE was inhibited when toluene was present in concentrations above 1 mg L-1. By differentiation of eq 3 with respect to Sp, an optimum primary substrate concentration Sp opt, at which the cosubstrate (MTBE) degradation is maximized, can be calculated (20). For a Cgin(MTBE) of 5 g m-3, which corresponds to a concentration of MTBE in the biofilm of 227.3 mg L-1 (mMTBE ) 0.022 at 25 °C), a Sp opt of 1.27 mg.L-1 of pentane in the biofilm was found. Then, in this case, a maximum efficiency of MTBE degradation should be attained when the concentration of pentane in the gas phase is approximately to 56 g m-3 (for mpent ) 44.4 at 25 °C). For this concentration of pentane, a low inhibition effect on the EC of pentane of 15% is expected. Sensitivity of Model. A sensitivity test was performed on this model by modifying each parameter independently with a (20% variation. Table 5 shows the results of the sensitivity studies to the kinetics (Ks(pent), Ks(MTBE), kx(pent), and kx(MTBE)) and physical (As, Xf, and δ) parameters. All of these sensitivity studies were conducted following the same procedure. In the Table 5, the error (E) corresponds to the relative values of the kinetics and physical parameter divided by its corresponding value (Table 2). Relative values (R) of the EC of pentane and MTBE are defined as the EC corresponding to the new values of the parameters divided by the EC obtained from the mathematical model when Cgin(pent) ) 15 g m-3, Cgin(MTBE) ) 5 g m-3, and EBRT ) 1 h. For these conditions, the values obtained from the model for EC of pentane and EC of MTBE were 5.95 and 0.52 g m-3 h-1, respectively. It should be noted that since the comparisons were based on the same EBRT and on the same inlet

concentrations, the relative EC (R) can be also viewed as the relative percent of %Ef. For a variation of (20%, the values of the affinity constants affect both compounds similarly. On the contrary, variations in the rate values (kx(pent) and kx(MTBE)) have a stronger impact on their respective compounds. The result of the sensitivity studies with three physical parameters (As, Xf, and δ) have the same effect on the removal rate of each compound. This result could be explained by the fact that these three parameters were not characteristics of either of the contaminants but were characteristics of the biofilter in general. The relationship appeared quasi linear, and so an increase of one parameter produces the increase of the removal rate, and vice versa. This result would appear to be normal and is confirmed by the form of the expression of EC established by Ottengraf et al. (23) for a simpler metabolism in the case of zero-order kinetics with reaction rate limitation (EC ) kxAsδXf). However, in this study, As, δ, and Xf parameters were not determined independently, and a variation of one will affect the others. As shown in the sensitivity analyses, to have appropriate kinetic and physical parameters for biofiltration purposes, relatively accurate determinations are needed to minimize errors.

Acknowledgments D.D. was a visiting student to the UAM from the ENSCP (Ecole Nationale Supe´rieure de Chimie Paris), France. The authors appreciate the assistance of Armando Dreyer, David Campos Santillan, Patrice Garnier, and Sergio Herna´ndez.

Literature Cited (1) Begley, R.; Rotman, D. Chem. Week 1993, 152, 7. (2) Fortin, N.; Deshusses, M. Environ. Sci. Technol. 1999, 33, 29802986. (3) Squillace, P. J.; Zogorski, J. S.; Wilber, W. G.; Price, C. V. Environ. Sci. Technol. 1996, 30, 1721-1730. (4) Dernbach, L. Environ. Sci. Technol. 2000, 516A-521A. (5) Bradley, P.; Landmeyer, J.; Chapelle, F. H. Environ. Sci. Technol. 2001, 35, 658-662. (6) Salanitro, J. P.; Diaz, L. A.; Williams, M. P.; Wisniewski, H. L. Appl. Environ. Microbiol. 1994, 60, 2593-2596. (7) Mo, K.; Lora, C. O.; Wanken, A. E.; Javanmardian, M.; Yang, X.; Kulpa, C. F. Appl. Microbiol. Biotechnol. 1997, 47, 69-72. (8) Mormile, M. R.; Liu, S.; Sulflita, J. M. Environ. Sci. Technol. 1994, 28, 1727-1732.

(9) Hardison, L. K.; Curry, S.; Ciufetti, L. M.; Hyman, M.. Appl. Environ. Microbiol. 1997, 63, 3059-3067. (10) Steffan, R. J.; McClay, K.; Vainberg, S.; Condee, C. W.; Zhang, D. Appl. Environ. Microbiol. 1997, 63, 4216-4222. (11) Garnier, P. M.; Auria, R.; Augur, C.; Revah, S. J. Gen. Appl. Microbiol. 2000, 46, 79-84. (12) Garnier, P. M.; Auria, R.; Augur, C.; Revah, S. Appl. Microbiol. Biotechnol. 1999, 51, 498-503. (13) Eweis, J. B.; Chang, D. P.; Schroeder, E. D.; Scow, K. M.; Morton, R. L.; Caballero, R. C. In Proceedings of the 90th Annual Meeting and Exhibition of the Air Waste Management Association, Pittsburgh, PA, 1997; Paper 97-RA133.06, 12 pp. (14) Fortin, N.; Deshusses, M. Environ. Sci. Technol. 1999, 33, 29872991. (15) Hanson, J. R.; Ackerman, C. E.; Scow, K. M. Appl. Environ. Microbiol. 1999, 65, 4788-4792. (16) Deeb, R. A.; Hu, H. Y., Hanson, J. R.; Scow, K. M.; AlvarezCohen, L.; Environ. Sci. Technol. 2001, 35, 312-317. (17) Deeb, R. A.; Scow, K. M.; Alvarez-Cohen, L. Biodegradation 2000, 11, 171-186. (18) Devinny, J. S.; Deshusses, M. A.; Webster, T. S. Biofiltration for Air Pollution Control. Lewis Publishers: 1998; 229 p. (19) Ottengraf, S. P. P.; van den Oever, A. H. C. Biotechnol. Bioeng. 1983, 25, 3089-3102. (20) Arcangeli, J. P.; Arvin, E. Environ. Sci. Technol. 1997, 31, 30443052. (21) Pineda, J.; Auria, R.; Perez-Guevara, F.; Revah, S. Bioprocess Eng. 2000, 23, 479-486. (22) Fan, L. S.; Leyva-Ramos, R.; Wisecaver, K. D.; Zehner, B. J. Biotechnol. Bioeng. 1990, 35, 279-286. (23) Ottengraf, S. P. P., Rehm, H. J., Reed, G., Eds. Biotechnology, Vol. 8; VCH Verlagsgesellsch: Weinheim, 1986. (24) Hodge, D. S.; Medina, V. F.; Islander, R. L.; Devinny, J. S. Environ. Technol. 1991, 12, 655-662. (25) van Groenestijn, J. W.; Lake, M. E. Environ. Progress. 1999, 18, 151-155. (26) Morgenroth, E.; Schroeder, E. D.; Chang, D. P. Y.; Scow, K. M. J. Air Waste Manage. Assoc. 1996, 46, 300-308. (27) Jensen, H. M.; Arvin, E. In Contaminated soil ’90; Arendt, F., Hinsenveld, M., van den Brink, W. J., Eds.; Kluwer: Dordrecht, The Netherlands, 1990; pp 445-448. (28) Reardon, K. F.; Mosteller, D. C.; Rogers, J. D. Biotechnol. Bioeng. 2000, 69, 385-400.

Received for review May 4, 2001. Revised manuscript received October 4, 2001. Accepted October 15, 2001. ES010942J

VOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

253