Biofiltration Kinetics of a Gaseous Aldehyde Mixture Using a Synthetic Matrix
By Wang, Li Kolar, Praveen; Kastner, James R; Herner, Brian
ABSTRACT Although aldehydes contribute to ozone and particulate matter formation, there has been little research on the biofiltration of these volatile organic compounds (VOCs), especially as mixtures. Biofiltration degradation kinetics of an aldehyde mixture containing hexanal, 2-methylbutanal (2-MB), and 3- methylbutanal (3-MB) was investigated using a bench-scale, synthetic, media-based biofilter. The adsorption capacity of the synthetic media for a model VOC, 3-methylbutanal, was 10 times that of compost. Periodic residence time distribution analysis (over the course of 1 yr) via a tracer study (84-99% recovery), indicated plug flow without channeling in the synthetic media and lack of compaction in the reactor. Simple first-order and zero-order kinetic models both equally fit the experimental data, yet analysis of the measured rate constants versus fractional conversion suggested an overall first-order model was more appropriate. Kinetic analysis indicated that hexanal had a significantly higher reaction rate (k = 0.09 +- 0.005 1/sec; 23 +- 1.3 ppmv) compared with the branched aldehydes (k = 0.04 +- 0.0036 1/sec; 31 +- 1.6 ppmv for 2-MB and 0.03 +- 0.0051 1/sec; 22 +- 1.3 ppmv for 3-MB). After 3 months of operation, all three compounds reached 100% removal (50 sec residence time, 18-46 ppmv inlet). Media samples withdrawn from the biofilter and observed under scanning electron microscopy analysis indicated microbial growth, suggesting removal of the aldehydes could be attributed to biodegradation.
(ProQuest: … denotes formulae omitted.)
INTRODUCTION
Aldehydes are present in the emissions of many industries including animal rendering, wastewater treatment, particle-board and medium density fiberboard manufacturing, cooking operations, and fuel combustion.1,2 Aldehydes are known to contribute to ozone and particulate matter formation, and even low concentrations can cause health problems such as asthma (e.g., formaldehyde, acetaldehyde).3
Increasing concerns about air quality and more stringent national and international regulations have led to the development and improvement of air pollution control processes for volatile organic compounds (VOCs). Traditional methods used to eliminate VOCs from industrial emissions primarily include physical and chemical methods. Physical methods (e.g., absorption, adsorption) have two disadvantages; (1) the VOCs are not eliminated, they are just transferred from one phase to another; and (2) the sorbents have to be regenerated. Thermal oxidation can eliminate a wide range of VOCs, but require high-energy input and emit additional carbon dioxide (for low concentration VOC emissions). Chemical wet scrubbers require costly oxidizing chemicals (e.g., chlorine dioxide [ClO^sub 2^], sodium hypochlorite [NaOCl]) and can produce chlorinated hydrocarbons if not properly controlled. On the other hand, biofiltration is based on the biodegradation of VOCs by microorganisms immobilized on the surface of a media at ambient temperatures.4,5 Compared with the nonbiological processes, the biological technologies are more economical, efficient, and environmentally benign.
Most biofilters use either natural organic media or synthetic media. Organic media typically includes soil beds, peat, and compost, which are abundant and low-cost. 6,7 However, organic media (e.g., compost) are prone to compaction or clogging and thus can cause channeling or an increased pressure drop of the filter bed.8 Compared with the organic media, synthetic packing materials (e.g., activated carbon, ceramic pellets) do not age and undergo compaction, but are more expensive and need inoculation before use. However, synthetic media have many desired physical and chemical properties, such as a higher adsorption capacity, controlled particle size, and strength, which might enhance removal rates and increase reactor longevity. These advantages were verified by Hirai et al.9 who found that NH3 removal capacities highly dependent on the physical and chemical properties of the inorganic matrix, i.e., media with high porosity, maximum water content, and suitable mean pore diameter showed excellent removal capacity.
Theoretical models have been developed for understanding the biodegradation processes in biofilters. Earlier models were developed to explain the removal of only one single contaminant that adopted the Monod-type rate equation.10 Then Ottengraf et al. derived the design equations to predict the fractional removal based on two extreme conditions of a Monod-type rate equation.4 De- shusses et al.11,12 also developed design equations for a contaminant based on a Michaelis-Menten rate equation. Although many experiments have been conducted to study the biofiltration kinetics of single compound removal and the inhibition mechanism of one compound on another,13-15 little research has been performed on biodegradation of multiple contaminants.
The objective of this research was to determine the biodegradation kinetics of an aldehyde mixture containing 3- methylbutanal, 2-methylbutanal, and hex-anal using a synthetic matrix. These VOCs were chosen based on previous analysis indicating that the major compounds identified in the emissions from a poultry rendering plant included hexanal, 2-methylbutanal, and 3- methylbutanal.16 This previous study also found that the two branched aldehydes, 2-methylbutanal and 3-methylbutanal, were the most consistent, appearing in every sample and typically the largest fraction of the VOC mixture.
MATERIALS AND METHODS
Biofilter Packing Material
The matrix used as the biofilter packing material in this experiment was a synthetic matrix (Biorem Technologies Inc.). This media was of a core and shell type and consisted of a porous hydrophilic nucleus and a hydrophobic shell (Figure 1). The shell was made of a metallic material, microorganisms, nutrients, organic carbon (OC), an alkaline buffer, a bonding agent, adsorptive agent, and hydrophobic agent. The media is theorized to have a permanent life span, a high surface area, and to undergo less compaction during operation.17
Microorganisms are applied during the manufacturing process, which incorporates a component of composted wood product and the naturally occurring microorganisms contained therein. Before application of the matrix and during operation, microorganisms are sustained with OC sources and nutrients provided in the coating formulation.
Media Characterization
The physical and chemical characteristics of the synthetic matrix and compost were determined and included pH, surface area (BET using N^sub 2^-Nova 3000 Quantachrome), and bulk density.18 Surface area was calculated from N^sub 2^ adsorption isotherms at -196 [degrees]C using the six-point BET method. Original samples (0.18-0.26 g) were heated to 200 [degrees]C and degassed under vacuum (10^sup -5^ Torr) to constant pressure (12 hr) before surface area analysis. Compost was used as a traditional organic media for the comparison of adsorption capacity, and was obtained from a local composting facility. The property comparison of these two packing materials is shown in Table 1.
Pressure Measurement
Pressure measurements were made using a Dwyer inclined and vertical portable manometer (Dwyer Instruments, Inc.) with 0-in. H2 O and 0- to 2-in. H2 O ranges. The pressure differences between inlet and outlet of the column were measured by connecting the two tees at the inlet and outlet of the biofilter system with the manometer.
Adsorption Capacity Studies
The adsorption capacity of the synthetic matrix and compost were conducted in 120-mL Amber glass serum bottles at room temperature (23 [degrees]C) equipped with a Mininert valves (Supelco Park). The model VOC used was one of the previously identified aldehydes, 3- methylbutanal. The diameter of the matrix chosen was less than 15 mm to fit the opening of the serum bottles. The mass of the matrix or compost used in the equilibrium adsorption experiments ranged from 0.4 to 9 g. The bottles with the matrix were sterilized at 121 [degrees]C for 20 min. The time for equilibrium to occur was first determined by injecting 3-meth-ylbutanal and sampling every hour until the gas-phase concentration did not change; equilibrium was clearly established within 24 hr. Then, various known amounts of 3- methylbutanal neat liquid were injected into the bottles. After 24 hr, 500 [mu]L of gas headspace was sampled for gas chromatography (GC) analysis (with a 2.5-mL Gastight syringe, Hamilton Co.). Adsorption capacity was measured for a series of gas-phase concentrations using synthetic matrix, compost, and an empty bottle or blank was used as a control; each experiment was conducted in triplicate. The adsorption capacity was calculated from a mass balance at equilibrium (eq 1),
M^sub VOC^ = C^sub g^ V^sub g^ + C^sub s^M^sub s^ (1)
where M^sub VOC^ is the mass of the VOC neat liquid added into the bottle, C^sub g^ is the equilibrium gas-phase concentration (g/ m^sup 3^), Vg is the volume of gas in the bottle (m^sup 3^), C^sub s^ is the equilibrium adsorption capacity (mg VOC/g-matrix), and M^sub s^ is the mass of the matrix (g). Retention Time Distribution Analysis
The retention time distribution (RTD) analysis was used to confirm plug flow and identify potential channeling effects in the reactor. RTD analysis was conducted without a VOC present and with just air flowing through the reactor. Helium was used as tracer and 10 mL of 99.999% helium was injected into the column using a pulse injection technique.19 The injection was made via a tee fitting at the inlet of the reactor, 21 cm away from the packing. Immediately after the injection, the outlet concentration was monitored and recorded with a MGD-2002 Multigas Detector (Radiodetection). The sensitivity and range of this instrument for helium was from 25 to 1,000,000 ppmv (in 25-ppmv increments).
Bench-Scale Biofilter
The experiments were conducted in a continuous flow, packed-bed reactor illustrated in Figure 2. The reactor had three sections; a biofilter body, inlet cap, and outlet cap. There were three sample ports in the body, and one in the inlet and outlet cap for a total of five sample ports. The inlet sample port was 21 cm from the packing. The distances from the top of the packing and the other four sample points were 9.5, 21.5, 33.5, and 68.5 cm, respectively. The actual height of the packing was 49 cm and the reactor was 0.1 m in diameter and 0.55 m in total length. The media was initially contacted with water to generate a 60% dry basis water content and the mass of the media was also recorded.
The compressed air was pressure regulated and filtered with a water trap to eliminate oils and water. The flow rate was controlled using a mass flow controller (URS-40, Celerity, Inc.). A 10-L/min flow meter (Dwyer Instruments, Inc.) was used to verify the flow rate. The air was humidified by passing through two bubble columns in series to reach 84.2% relative humidity (RH) at the outlet of the first humidifier and 92% RH at the outlet of the second humidifier. After humidification and VOC introduction, the contaminated air passed through a column filled with small glass beads to provide mixing and was subsequently passed downward across the media as indicated in Figure 2. All columns were sealed by threaded Teflon plugs with O-rings (ACE Glass, Inc.), and tubing was 6.35 mm in diameter.
The addition of the contaminants was accomplished by a syringe pump (Cole-Parmer 74900-30). The contaminants were added to the airstream as a neat liquid through a stainless steel Swagelok T- fitting with septum. The T-fitting and the liquid mixture were heated (Ther-molyne 45500) to accelerate evaporation.
Analytical Techniques
A Hewlett 5890 packed series II gas chromatograph (coupled to a flame-ionization detector [FID]) equipped with an SPB-1 capillary sulfur column (30 m x 0.32 [mu]m; All-tech Associates, Inc.) and helium as the carrier gas was used for measuring the contaminant concentration along the reactor. A split ratio of 30:1 was used with a column head pressure of 9 psi and the flow rates of the purge vent, split vent and the column were 4, 60, and 2 mL/min, respectively. The temperatures of the oven, injection port, and detector were 80, 250, and 250 [degrees]C, respectively. A standard curve was generated before the experiments by generating at least five gas samples with known concentrations in the range from 3 ppmv to 70 ppmv. The samples were analyzed in triplicate by GC/FID and the standard was periodically checked for linearity and drift.
Scanning Electron Microscopy
After 5 months of operation, triplicate samples of the biofilter matrix were collected at different depths of the reactor. The samples were dissected into 1- to 3-mm cubes with a grease-free razor blade and fixed in 2% glutaralde-hyde in a 0.1 M cacodylate buffer (pH 7.2) at 4 [degrees]C for 90 min. After the samples were washed two times with a 0.1 M cacodylate buffer for 15 min each, the samples were fixed secondarily with a 1% osmium tetroxide in 0.2 M cacodylate buffer at 4 [degrees]C for 90 min. The samples were rinsed twice for 15 min with 0.2 M cacodylate buffer before dehydrating with increasing concentrations of ethanol at 30, 50, 70, 85, 95, and 100% (two rinses at 100%) for 15 min each. The dehydrated samples immersed in ethanol were dried with a critical point drier (model 780-A, Tousimis Inc.). The dried samples were subsequently mounted on an aluminum stub with an adhesive carbon sticky tab. The specimen stub was sputter coated with approximately 150 A of gold using sputter coater (model SPI; SPI Supplies). The observations of the samples were carried out in a digital scanning electron microscope (ZEISS 1450EP; Carl Zeiss Micro Imaging). An accelerating voltage of 20 keV was used and a secondary electron detector was used for imaging the samples. The images obtained from the scanning electron microscopy were processed for publication using Adobe Photoshop (version 7).
RESULTS AND DISCUSSION
Comparison of Adsorption Capacity
In biofiltration, for biodegradation to occur, VOCs must be transferred from the gas phase to the biofilm located on the media surface. In theory, use of a porous, highsurface area media with a high adsorption capacity for the contaminants (e.g., outer hydrophobic shell) could create a concentration gradient through which the contaminants would be transported and partition into the biofilm and subsequently be metabolized by the microorganisms. Therefore, if microbial degradation rates are high enough, a high adsorption capacity may enhance VOC removal. Also, media with high adsorption capacities can adsorb high concentrations of substrates and slowly release them for microbial degradation,20 and thus can buffer against inlet shocks or pulses of VOCs. Therefore, before the continuous flow experiments, adsorption studies were performed to compare the adsorption capacity of the synthetic matrix and the compost. The headspace concentration was measured periodically and found to approach equilibrium within 24 hr. Previous adsorption studies using peat indicated similar results.21 Results from this work indicated that the synthetic matrix had an adsorption capacity 10 times higher than that of compost (Figure 3). The high adsorption capacity was potentially due to the high surface area of the synthetic matrix, which was nearly 16 times higher than that of compost (Table 1). These results indicate one potential advantage of using an engineered synthetic matrix as biofilter media.
The equilibrium data were also analyzed using Freund-lich (eq 2) and Langmuir (eq 3) isotherm equations. The Freundlich isotherm is an empirical adsorption isotherm for nonideal adsorption on heterogeneous surfaces, as well as multiplayer adsorption, and is expressed by the equation:
q^sub e^ = K^sub F^C^sub e^^sup 1/n^ (2)
where q^sub e^ is the adsorption capacity (mg of VOC per g of adsorbent), K^sub F^ and 1/n are Freundlich constants, C^sub e^ is the VOC concentration in the fluid (mg/L). Equation 2 is derived by assuming an exponentially decaying adsorption site energy distribution. The limitation is that it does not follow the fundamental thermodynamic basis because it does not reduce to Henry’s law at lower concentrations.
The Langmuir isotherm is a theoretical equilibrium isotherm relating the amount of solute sorbed on a surface to the concentration of solute. Two assumptions are made in its derivation: the forces of interaction between sorbed molecules are negligible, and once a molecule occupies a site no further adsorption takes place. On the basis of these assumptions, in theory, a saturation value is reached beyond which no further adsorption takes place. The saturated monolayer adsorption capacity can be represented by the following equations:
… (3)
where q^sub e^ is the maximum adsorption capacity corresponding to complete monolayer coverage (mg of solute adsorbed per g of adsorbent), and K^sub L^ is the Langmuir constant (liters of adsorbent per milligram of VOC).
At lower VOC levels (<1000 ppmv), the data were fit using the Freundlich equation (Figure 4). The Freundlich constant K^sub F^ was found to be 0.037 for compost (R^sup 2^ = 0.9418) and 1.3 for the synthetic media (R^sup 2^ = 0.8874). The Freundlich constant n was found to be 0.91 for compost and 1.31 for the synthetic media. The Langmuir isotherm was also fit to the entire dataset for the synthetic matrix (Figure 5) and the Langmuir constant was found to be 0.43 L/mg and the maximum adsorption capacity corresponding to complete monolayer coverage was 4.95 mg/g (R^sup 2^ = 0.9895). These results clearly indicate a higher adsorption capacity of 3- methylbutanal for the synthetic matrix compared with compost. For comparison, a Freund-lich constant for toluene adsorption on peat was reported to be 0.459.24 Similarly, toluene adsorption onto activated carbon indicated a Freundlich constant K^sub F^ to be 2.43 and n to be 8.38 at 298.15 K; the Langmuir maximum adsorption capacity was reported to be 510.4 mg/g.22 Again, these isotherm constants obtained from the literature confirm the higher adsorption capacity of the synthetic matrix for the aldehydes, probably due to its hydrophobic nature and high surface area relative to compost.
Pressure Drop Analysis
When the reactor was initially loaded with matrix, glass wool was placed at the bottom of the reactor to prevent media loss. After the experiment was operated for nearly 1 yr, a large pressure drop along the reactor was observed. Specifically, with superficial gas velocities varying from 7.6 to 53.5 m^sup 3^/m^sup 2^ . hr, the pressure loss between the inlet and number 4 port ranged from 14.2 to 71.2 Pa/m, whereas pressure loss between the inlet and outlet ranged from 1743.6 to 22,168.9 Pa/m. Because water was added at the top of the reactor, some small particles of the matrix were probably flushed out of the matrix and collected on the glass wool, causing this large pressure drop. Therefore, a plastic disk with uniformly distributed holes (~5 mm diameter) was placed at the bottom of the reactor to replace the glass wool, which significantly decreased the pressure drop. Within the same range of superficial gas velocities range, the pressure loss between the inlet and outlet ranged from 19.9 to 254.1 Pa/m. The pressure drop through a biofilter bed typically ranges from 20 to 100 Pa/m, but occasionally increases to 980 Pa/m, with typical superficial gas velocities from 5 to 500 m^sup 3^/m^sup 2^ . hr.23 Le-son and Smith24 report that a biofilter system with adequate moisture control and a porous medium containing bulking agents will typically have a pressure loss less than 900- 1700 Pa/m. Comparing our results to these data, it is clear that the synthetic medium used in this research can lead to a low pressure drop even with significant water addition. Verification of Plug Flow (Tracer Analysis)
To determine the biodegradation kinetics of the target compounds, reactor design and rate equations are needed. Because our reactor was assumed to be of the plug flow-type (PFR), RTD analysis was carried out to determine if the flow hydraulics were plug flow. The RTD curves were analyzed by a dispersion model, which is used for nonideal PFRs with axial dispersion. The Peclet number characterizes the level of dispersion in a reactor and is defined as,
… (4)
where U is the superficial molar average velocity through the bed (m/sec), L is the length of the reactor (m), and D^sub e^ is the effective dispersion coefficient. The Peclet number was calculated from the following equations:
… (5)
tau = [integral]^sup [infinity]^^sub 0^ tE(t)dt (6)
sigma^sup 2^ = [integral]^sup [infinity]^^sub 0^ (t – tau)^sup 2^E(t)dt (7)
… (8)
where E(t) is the residence time distribution function, C(t) is the tracer concentration, tau is the mean residence time, sigma is the second moment of the mean, and t is time. In the limiting cases when Pe^sub r^ = 0 (very high dispersion) we have a complete mixing regime and when Pe^sub r^ = [infinity] (no dispersion) we have a perfect PFR.
The RTD experiments were replicated and performed after initially loading the reactor, and after 6, 11, and 15 months of operation (Figure 6). Tests were conducted at airflow rates of 4.5 L/min (initially) and 5 L/min (all other times), resulting in 44.79, 30.77, 35.98, and 33.94-sec mean residence times based on the RTD analysis (Figure 6). The recoveries of the tracer were 99.36, 83.72, 92.28, and 109% respectively. The calculated Peclet numbers were 15.57, 25.26, 37.6, and 24.36 respectively, indicating the assumption of a PFR was reasonable. Usually, a Peclet number of 500 indicates a very small amount of dispersion, 40 indicates intermediate dispersion, and 5 indicates large dispersion.25 The RTD results also demonstrated that channeling or bypassing of the media did not occur (i.e., a bimodal tracer distribution shifted to shorter residence times was not observed), indicative of limited compaction and aging of the synthetic matrix. Contrarily, Morgan Sagastume et al.8 observed channeling in a compost-based biofilter using the RTD technique, which may have been due to the low compression strength of the compost and/or degradation of the organic media. Therefore, these results indicate that lack of compaction and channeling is an advantage of the synthetic media over the traditional media.
Biodegradation Kinetic Analysis
Initially the biofilter was operated without directly adding water and only one humidification reactor. After three days of operation, the biofilter reached more than 80% aldehyde removal, but after the fifth day, the fractional conversion began to decrease (2- and 3-methylbutanal and hexanal removal efficiencies declined to 30, 30, and 40% respectively, after 11 days). A limited number of matrix samples were then taken from the reactor to measure the water content and a significant decrease in moisture content was observed (Table 2). Water content is known to be very critical in biofiltration.26 Acuna et al.21 report that an initial water content between 55 and 70% (wet weight basis) give optimum toluene degradation for peat (an organic media), and Bohn et al.,27 indicate an optimum moisture content ranging between 8 and 20% for inorganic media (e.g., soil; dry weight basis). The results of this experiment indicated a dramatic decrease of water content (lower than 1%, dry basis), which explains the reduced degradation rates and fractional conversion of the aldehydes. Therefore, during all subsequent experiments, water was added from the top of the reactor twice a week (60 mL) and the humidifiers (two humidifiers in series were used at this point) were filled with water twice a week to maintain the moisture level in the biofilter. An increase in aldehyde degradation was observed because of this improvement (a 25% moisture content [dry basis] resulted in 2- and 3-methylbutanal and hexanal removal efficiencies of 95, 98, and 100% respectively, after 6 days).
During the continuous biofiltration experiments, gas samples were taken from the reactor via the five sample ports and analyzed using the GC/FID. Figure 7A shows a typical chromatograph from the different positions along the reactor. Clearly, concentrations of the three contaminants decrease along the reactor because of biodegradation activity. An unknown peak (after 22 days of operation and the new moisture addition campaign) was found to be present in all gas samples from the reactor, except the inlet sample, and the concentration of the unknown increased over the length of the reactor (on the basis of an increase in peak area), which suggested that this unknown was formed from the metabolism of the alde-hydes. With continued operation, this unknown disappeared (Figure 7B). This may have been due to an initial limited number and diversity of microorganisms, such that the biodegradation capacity was low (i.e., any intermediates formed during metabolism of the aldehydes were not degraded before leaving the reactor). After the reactor had been operated for some time (53 days), the microbial population and diversity probably increased as the microorganisms were continuously provided with air (O^sub 2^), water, and carbon sources. Addition of a secondary carbon/energy source (e.g., methanol to a biofilter treating hydrogen sulfide)28 reportedly increased biofilter microbial diversity and high microbial diversity was reported in a biofilter treating rendering emissions containing primarily 2- and 3- methylbutanal, 2-meth-ylpropanal, hexanal, and dimethyl disulfide.29 Therefore, the biodegradation capacity apparently increased toward any potential metabolic intermediates in alde-hyde degradation.
From the first day of start-up of the reactor to approximately 3 months after operation, the fractional removal of hexanal (nearly 100%) was always higher than that of 3-methylbutanal and 2- methylbutanal (Figures 8 and 9). The reason for this preferential pattern may be that the straight chain aldehyde was more easily metabolized by the microorganisms in the biofil-ter than the branched chain aldehydes. Similar metabolic patterns have been observed between straight chain and branched alkanes.30,31
The kinetic analysis was performed by using the plug flow design equation (eq 9) with the appropriate rate law (-r = kC^sup n^, where r is degradation rate of the VOC) and assumes a homogeneous system, with constant volume, pressure, temperature, and O^sub 2^ in excess.
… (9)
F^sub j^ is the mass flow rate (mol/time), V is reactor volume, r^sub j^ is rate of consumption per unit volume, N^sub j^ is the number of moles of j in the system, and j is the substrate. Using an empirical approach to model the reaction rate and substituting the expressions of (10)
… (10)
into eq 9 we obtain a power rate model to predict VOC concentration profiles,32
… (11)
where C^sub A^ is the substrate concentration; t is time or position along the reactor; k, k^sub 1^ and k^sub 2^ are rate constants for VOC disappearance; and n is the overall reaction order. The reaction rate and rate constant were calculated assuming the order of the reaction and fitting the subsequent model to the concentration profile along the reactor. Using a mechanistic model including diffusion and reaction in a biofilm, previous authors found that when the gas-phase concentration is high, the reaction rate is zero-order and the elimination rate becomes reaction- controlled, whereas at low gas-phase concentrations or low water solubility, the reaction rate is first-order and the elimination rate becomes diffusion-controlled.33 During this experiment, gas samples were withdrawn from different positions along the reactor and the concentration profile quantified (Figure 9). Then zero, first-order, nonlinear models were assumed for each dataset, and the reaction rate and rate constants were calculated, as shown in eqs 12- 14.
… (12)
… (13)
… (14)
In a second method to analyze the reaction kinetics, the resultant rate constants determined from eqs 12, 13, and 14 were plotted against the measured fractional VOC conversion to determine if k remained constant. A systematic change in k suggests an incorrect model.25
Both first- and zero-order kinetic models appeared to the fit the experimental data for aldehyde degradation versus position (or V/Q) along the reactor. As demonstrated in Figure 10, both first- and zero-order models resulted in reasonable goodness of fit values. The nonlinear model (eq 14), which should accurately predict the nonlinear portion of the degradation rate versus inlet concentration curve, did not fit the data over the range of VOC concentrations tested. When the calculated first-and zero-order rate constants were plotted against the measured fractional conversion a systematic change in k^sub zero order^ was observed, suggesting the zero-order model did not fit the data either (Figure 11). Thus, the kinetic analysis suggests an overall first-order model is most appropriate to predict the degradation of hexanal, 2-meth-ylbutanal, and 3- methylbutanal from 10 to 50 ppmv. Similar to our results, butanal degradation kinetics appeared to follow first-order kinetics in a wood bark-based biofilter at an inlet concentration of 10 ppmv34 and isobu-tanal degradation was first-order in a compost-based bio- filter up to 300 ppmv.36 Regardless of the model, overall hexanal removal or reaction rates were significantly higher (level of significance = 0.05) compared with the branched aldehydes (Table 3, using lower detection limits to estimate hexanal concentration). Using the first-order model the reaction rate constants of three aldehydes were estimated over a 6-month period (Table 4). Comparing the kinetics of aldehyde degradation, the measured degradation rates in this work were similar to those previously reported for butanal and isobutanal (2-methylpropanal) for wood bark and compost based biofilters. For example, in our work, first-order rate constants for hexanal (23.4 +- 1.3 ppmv), 2-methyl-butanal (31.3 +- 1.6 ppmv), and 3-methylbutanal (22 +- 1.32 ppmv) were 0.091 +- 0.0052 (SE), 0.04 +- 0.0036, and 0.032 +- 0.0051 1/sec (September 17, 2005 to October 7, 2005, n = 14), respectively, compared with 0.09 1/sec for butanal (10 ppmv)34 and 0.033 1/sec for isobutanal (300 ppmv).36 On the basis of the higher adsorption capacity of the synthetic medium, it was anticipated that the reaction rate of our biofilter system would be higher than that of compost-based systems (if the microbial densities and degradation capacity was high enough). This was expected, because the synthetic medium had a much higher surface area than that of compost, as well as a higher adsorption capacity, which may result in a higher contaminant surface concentration. However, compared with the literature, the reaction rates for the synthetic medium and for compost were similar. This may have been due to the lower molecular weight of the aldehydes used in the compost studies (i.e., differences in microbial metabolism), and/or low microbial levels and growth rates in the synthetic matrix, because no nutrients were added and just humidified air and VOCs were added to the reactor.
Perturbation Effect
After prolonged biofilter operation and kinetic analysis (Table 3) of aldehyde concentrations anticipated in industrial operations (Table 3 and Kastner et. al.16) a pulse increase in aldehyde concentration was introduced into the biofilter. A large rise in aldehyde concentrations was observed, but breakthrough (i.e., C^sub out^/C^sub in^ = 1) did not occur in the biofilter, probably because of the high adsorption capacity of the synthetic matrix for the alde- hydes (Figure 12). Continued removal of the aldehydes was observed after passage of the spike and subsequent operation of the biofilter at low inlet concentrations (similar to levels reported in Table 3- data not shown).
Microbial Analysis
Samples of the support were withdrawn from different positions of the biofilter and observed under scanning electron microscopy. The structure of the original synthetic matrix core was of a porous nature with a limited number of microorganisms (Figure 13, A and C). Samples from the biofilter after 4 months of operation treating a mixture of hexanal, 2-methylbutanal, and 3-methylbu-tanal showed evidence of microbial growth and biofilm formation. Qualitatively, the core sample appeared to have a large number of bacteria, compared with the surface in which fungi were primarily observed (Figure 13, B and D). The hydrophilic nature of the core may account for the presence of bacteria, which cannot tolerate low water activity as well as fungi. Also fungi require higher oxygen concentrations than bacteria for completely aerobic growth. Therefore, the presence of fungi on the surface may merely be due to a decreasing oxygen concentration gradient within the synthetic matrix particles.
CONCLUSIONS
It is clear that the branched aldehydes (2-methylbutanal, 3- methylbutanal) were degraded simultaneously, yet at a slower rate when compared with the straight chain aldehyde, hexanal. Biofilter design for such a mixture should be based on the first-order kinetics of degradation for 3-methylbutanal over a concentration range of 0-50 ppmv because it was the rate-limiting compound (for the media and conditions tested in this work).
The engineered synthetic matrix was capable of supporting microbial growth, which resulted in the development of a microbial population capable of degrading a vapor mixture containing five- and six-carbon aldehydes. Evidence for microbial growth and degradation of the aldehydes (VOCs) was based on the scanning electron microscopy results and the long-term, consistently measured decline in VOC concentration along the biofilter with respect to time.
In addition to providing an environment for biofilm development and VOC biodegradation, the synthetic matrix also maintained its structural integrity over the course of more than 1 yr of operation, with little to no change in measured pH (data not shown) and no evidence of compaction or channeling caused by physical deterioration (as evidenced by the RTD analysis). Finally, the synthetic matrix had a significantly higher surface area and resultant adsorption capacity for 3-methylbutanal when compared with a typical organic matrix, which indicates the synthetic matrix can adsorb high concentrations of the aldehydes tested and potentially slowly release them for microbial degradation, and thus can buffer against inlet shocks or pulses of five- and six-carbon aldehydes.
IMPLICATIONS
Sustained biofiltration, an environmentally friendly control method compared with traditional technologies, requires a matrix with multiple physical characteristics that are hard to find in typical organic medium, yet must be capable of supporting microbial growth for degradation of multiple VOCs. An engineered synthetic media with encapsulated microorganisms, a high surface area and adsorption capacity, and resistance to deterioration, rapidly degraded a mixture of aldehydes. The results suggest that the synthetic matrix was capable of supporting microbial growth and subsequent degradation of multiple VOCs, yet could increase reactor longevity and act to buffer VOC perturbations due to its high adsorption capacity.
REFERENCES
1. Baumann, M.G.D.; Lorenz, L.F.; Batterman, S.A.; Zhang, G. Aldehyde Emissions from Particleboard and Medium Density Fiber Board Prod-ucts; Forest Prod. J. 2000, 50, 75-82.
2. Andres, F.; A ngel, C.B.; Sukh, S. Volatile Aldehyde Emissions from Heated Cooking Oils; J. Sci. Food Agric. 2004, 84, 2015-2021.
3. Leikauf, GD. Hazardous Air Pollutants and Asthma; Environ. Health Perspect. 2002, 110 (Suppl. 4), 505-526.
4. Ottengraf, S.P.P.; Van den Oever, A.H.C. Kinetics of Organic Compound Removal from Waste Gases with a Biological Filter; Biotechnol. Bioeng. 1983, 25, 3089-102.
5. Leson, G.; Winer, A.M. Biofiltration: an Innovative Air Pollution Control Technology for VOC Emissions; J. Air & Waste Manage. Assoc. 1991, 41, 1045-1054.
6. Beerli, M.; Rotman, A. Biofilter-a Unique Method to Reduce and/ or Eliminate VOCs. In Proceedings of Envirocon 89: Ist International Conference on Environmental Issues for Converters; Envirocon: Jacksonville, FL, 1989; 1-32.
7. Smet, E.; Chasaya, G.; Van Langenhove, H.; Verstraete, W. The Effect of Inoculation and the Type of Carrier Material Used on the Biofiltration of Methyl Sulphides; Appl. Microbiol. Biotechnol. 1996, 45, 293-298.
8. Morgan-Sagastume, J.M.; Noyola, A.; Revah, S.; Ergas, S.J. Changes in Physical Properties of a Compost Biofilter Treating Hydrogen Sulfide; J. Air & Waste Manage. Assoc. 2003, 53, 1011- 1021.
9. Hirai, M.; Kamamoto, M.; Yani, M.; Shoda, M. Comparing of the Biological NH3 Removal Characteristics among Four Inorganic Packing Materials; J. Biosci. Bioeng. 2003, 91, 428-430.
10. Jennings, P.A.; Snoeyink, V.L.; Chian, E.S.K. Theoretical Model for a Submerged Biological Filter; Biotechnol. Bioeng. 1976, 18, 1249-1273.
11. Deshusses, M.A.; Hamer, G.; Dunn, I.J. Behavior of Biofilters for Waste Air Biotreatment. I. Dynamic Model Development; Environ. Sci. Tech-nol. 1995, 29, 1048-1058.
12. Deshusses, M.A.; Hamer, G.; Dunn, I.J. Behavior of Biofilters for Waste Air Biotreatment. II. Experimental Evaluation of a Dynamic Model; Environ. Sci. Technol. 1995, 29, 1059-1068.
13. Smet, E.; Van Langenhove, H.; Verstraete, W. Isobutyraldehyde as a Competitor of the Dimethyl Sulfide Degrading Activity in Biofilters; Biodegradation 1997, 8, 53-59.
14. Hwang, S.C.J.; Lee, C.M.; Lee, H.C.; Pua, H.F. Biofiltration of Waste Gases Containing Both Ethyl Acetate and Toluene Using Different Combinations of Bacterial Cultures; J. Biotechnol. 2003, 105, 83-94.
15. Mohseni, M.; Allen, D.G. Biofiltration of Mixtures of Hydrophilic and Hydrophobic Volatile Organic Compounds; Chem. Eng. Sci. 2000, 55, 1545-1558.
16. Kastner, J.R.; Das, K.C. Wet Scrubber Analysis of Volatile Organic Compound Removal in the Rendering Industry; J. Air & Waste Manage. Assoc. 2003, 52, 459-69.
17. Shareefdeen, Z.M.; Herner, B.P. Biological Filter. International Patent Classification: B01D53/85; B01D53/84; (IPC1- 7): B01D39/02 (6) Publication info:WO2005037403-2005-04-28.
18. Kastner, J.R.; Das, K.C.; Melear, N.D. Catalytic Oxidation of Gaseous Reduced Sulfur Compounds Using Coal Fly Ash; J. Hazard. Mater. 2002, 95, 81-90.
19. Levenspiel, O. Chemical Reaction Engineering; John Wiley, New York, 1972; pp 253-325.
20. Khaled, A.S.; Shelef, G.; Levanoxf, D.; Armonb, R.; Dosoretza, C.G. Microbial Degradation of Aromatic and Polyaromatic Toxic Compounds Adsorbed on Powdered Activated Carbon; J. Biotechnol. 1996, 51, 265-272.
21. Acuna, M.E.; Perez, F.; Auria, R.; Revah, S. Microbiological and Kinetic Aspects of a Biofilter for the Removal of Toluene from Waste Gases; Biotechnol. Bioeng. 1999, 63, 175-184.
22. Benkhedda, J.; Jaubert, J.N.; Barth, D. Experimental and Modeled Results Describing the Adsorption of Toluene onto Activated Carbon; J. Chem. Eng. Data 2000, 45, 650-653. 23. Devinny, J.S.; Deshusses, M.A.; Webster, T.S. Biofiltration for Air Pollution Control; Lewis: Boca Raton, FL, 1999.
24. Leson, G.; Smith, B.J. Petroleum Environmental Research Forum Field Study on Biofilters for Control for Control of Volatile Hydrocarbons; J. Environ. Eng. 1997, 123, 556-562.
25. Fogler, H.S. Elements of Chemical Reaction Engineering, 4th ed.; Prentice Hall PTR: Indianapolis, 2006.
26. Auria, R.; Aycaguer, A.C.; Devinny, J. Influence of Water Content on the Degradation Capacity of Ethanol in Biofiltration; J Air & Waste Manage. Assoc. 1998, 48, 65-70.
27. Bohn, H.L.; Bohn, K.H. Moisture in Biofilters; Environ. Prog. 1999, 18, 156-161.
28. Ding, Y.; Das, K.C.; Whitman, W.B.; Kastner, J.R. Enhanced Biofiltra-tion of Hydrogen Sulfide in the Presence of Methanol and Resultant Bacterial Diversity; ASABE Trans. 2006, 49, 2051-2059.
29. Friedrich, U.; Prior, K.; Altendorf, K.; Lipski, A. High Bacterial Diversity of a Waste Gas-Degrading Community in an Industrial Biofilter as Shown by a 16s RDNA Clone Library; Environ. Microbiol. 2002, 4, 721-734.
30. Watkinson, R.J.; Morgan, P. Physiology of Aliphatic Hydrocarbon Degrading Microorganisms; Biodegradation 1990, 1, 79- 92.
31. Olson, J.J.; Mills, G.L.; Herbert, B.E.; Morris, P.J. Biodegradation Rates of Separated Diesel Components; Environ. Toxicol. Chem. 1999, 18, 2448-2453.
32. Hamaker, J.W. Organic Chemicals in the Soil Environment; Dekker: New York, 1972; 253-340.
33. Ottengraf, S.P.P. In Biotechnology; Rehm, J.J., Reed, G., Eds.; VCH Verlagsgesellschaft: Weinheim, Germany, 1986; Vol. 8, pp 427-452.
34. Weckhuysen, B.; Vriens, L.; Verachtert, H. The Effect of Nutrient Supplementation on the Biofiltration Removal of Butanal in Contaminated Air; Appl. Microbiol. Biotechnol. 1993, 39, 395-399.
35. Kapetanios, E.G.; Loizidou, M.; Valkanas, G. Compost Production from Greek Domestic Refuse; Bioresource Technol. 1993, 44, 13-16.
36. Sercu, B.; Demeestere, K.; Baillieul, H.; Langenhove, H.V. Degradation of Isobutanal at High Loading Rates in a Compost Biofilter; J. Air & Waste Manage. Assoc. 2005, 55, 1217-1227.
Li Wang, Praveen Kolar, and James R. Kastner
Department of Biological and Agricultural Engineering, University of Georgia, Athens, GA
Brian Herner
BIOREM Technologies Inc., Guelph, Ontario, Canada
About the Authors
Li Wang was an MS graduate student at the Department of Biological and Agriculture Engineering, University of Georgia, Driftmier Engineering Center in Athens, GA, and is currently a biostatistician at the Vanderbilt University School of Medicine. Praveen Kolar is a Ph.D. graduate student, and James Kastner is an associate professor with the Department of Biological and Agricultural Engineering at the University of Georgia Driftmier Engineering Center in Athens, GA. Brian Herner is president of BIOREM Technologies Inc. in Guelph, Ontario, Canada. Please address correspondence to: James Kastner, University of Georgia, Department of Biological and Agricultural Engineering, Driftmier Engineering Center, Athens, GA 30602; phone: +1-706-583-0155; fax: +1-706-542- 8806; e-mail: jkastner@engr.uga.edu.
Copyright Air and Waste Management Association Mar 2008
(c) 2008 Journal of the Air & Waste Management Association. Provided by ProQuest Information and Learning. All rights Reserved.