Environ. Sci. Technol. 2005, 39, 5856-5863
Biodegradation of r-Pinene in Model Biofilms in Biofilters MARTHA J. MILLER AND D. GRANT ALLEN* University of Toronto, Department of Chemical Engineering and Applied Chemistry, 200 College Street, Toronto, Ontario, Canada M5S 3E5
Treatment of air pollutants in a biofilter requires that the compound be effectively transported from the gas phase to the organisms that reside in a biofilm that forms upon a packing material. Models of biofiltration generally treat the biofilm like water by using a Henry’s law constant to predict mass transfer rates into the biofilm where degradation occurs and, hence, predict low rates for hydrophobic compounds. However, some compounds that are virtually insoluble in water are also treated unusually well. The objective of this study was to develop a fundamental understanding of the apparent enhanced degradation of hydrophobic pollutants in biofilms. Specifically, the goals of this study were to experimentally determine transport and reaction rates of hydrophobic pollutants in artificial biofilms. We studied the transport and reaction rates of R-pinene (as a model hydrophobic pollutant) in a headspace in contact with a well-defined biofilm made up of biomass immobilized in low melting point agarose and found that reaction rates were similar in order of magnitude to biofilter rates. The transport rates through these films once deactivated were found to be the same as through agar (diffusion coefficient between 2.6 and 3.4 × 10-6 cm2/s). The degradation rates through model biofilms ranged from 2 to 4 × 10-7 (g/(cm2 min)). A new explanation of high degradation rates was put forth whereby a biologically mediated transformation is taking place in which R-pinene is oxidized into a more soluble, less volatile compound that can then penetrate deeper into the biofilm. The formation of this more soluble byproduct was confirmed with batch kinetics experiments using filtered samples, and its proposed identity is cis-2,8-p-menthadien-1-ol, a menthadienol, a novel metabolite of R-pinene degradation. A simple conceptual model based on these results is also presented.
Introduction Biofiltration of Hydrophobic Pollutants. Biofiltration is an air pollution control strategy that involves passing waste air through a column that contains a media upon which microorganisms capable of degrading pollutants live. The compound to be degraded must pass from the gas to the organisms that reside in a biofilm that forms upon the media. In general, biofilters achieve the highest rates of removal for compounds that are water soluble and biodegradable (e.g., ethanol and methanol). Models of biofiltration generally support this since they neglect the gas phase mass transfer * Corresponding author phone: (416)978-8517; fax: (416)978-8605; e-mail:
[email protected]. 5856
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
resistance and treat the biofilm like water by using a Henry’s law constant to predict mass transfer rates into the biofilm where degradation occurs (1-4). However, some compounds that are virtually insoluble in water are also treated unusually well. For example, recent studies (5-7) show that R-pinene can be completely removed by biofiltration even with loading rates as high as 40 g/(m3 bed h) despite the fact that the solubility of R-pinene ranges between 2 and 22 mg/L (8-13) and that the dimensionless air/water partition coefficient ranges between 4 and 6 (12, 14, 15). Similarly, Spigno et al. (16) studied the biofiltration of hexane, another hydrophobic pollutant, and found the average removal rates to be as high as 150 g/(m3 bed h), again despite the fact that the Merck Index lists it as insoluble in water. Kibazohi (17) also reported hexane removal rates in a perlite biofilter to be up to 120 g/(m3 bed h). From a modeling perspective, our research has shown that if one uses the air/water partition coefficient in the modeling of the biofiltration of R-pinene, negligible rates of R-pinene removal are predicted; only by lowering this parameter several 100-fold is it possible to predict the observed removal rates (5). Furthermore, the results put forth by Deshusses and Johnson (18) suggest that the biodegradation of compounds in biofilters is influenced both by the pollutant availability (related to its Henry’s Law constant) and to a lesser degree by the hydrophobicity of the compounds (related to its octanol/water partition coefficient). The unexpectedly high observed removal rates for some hydrophobic compounds suggest that current models do not accurately describe biofilm transport and reaction processes since based on these models we would predict poor mass transfer of these compounds and hence low removal rates. Our previous research explored the possibility that a biofilm can enhance the transport of R-pinene as compared to that of water. This research showed that the steady-state transport rate of R-pinene through artificial biofilms consisting of irradiated leachate (collected from R-pinene degrading biofilters) immobilized in low melting point agarose is not significantly different from the transport rates through agar (e.g., water-like) films (14). Since there was no enhancement in transport rate in these artificial biofilms, this work was taken further in this study to see if a more realistic film could be formed in situ and then deactivated to study transport rates. The objective of this study was to develop a fundamental understanding of the apparent enhanced degradation of hydrophobic pollutants in biofilms. Specifically, the goals of this study were to experimentally determine transport and reaction rates of hydrophobic pollutants in artificial biofilms. R-Pinene was selected as a model hydrophobic pollutant because of its high volatility, low water solubility, past experience with this pollutant (5, 19-22), and because of its relevance to the application of biofiltration of emissions in the forest products industry. We studied the transport and reaction rates of R-pinene in a headspace in contact with a well-defined model biofilm made up of biomass immobilized in low melting point agarose. On the basis of these results, a new model was proposed in which R-pinene undergoes a biologically mediated transformation into a more soluble, less volatile compound that can then penetrate deeper into the biofilm. The formation of this more soluble byproduct was confirmed with batch kinetics experiments using filtered samples. 10.1021/es048254y CCC: $30.25
2005 American Chemical Society Published on Web 06/22/2005
FIGURE 1. Schematic diagram of the diffusion cell.
Materials and Methods Diffusion Cell Experiments. Two diffusion cells (volume 1.3 L each) were designed and constructed out of Teflon to measure the transport and reaction rates of R-pinene in samples of biomass immobilized in agarose (Figure 1). The diffusion cells consisted of two air chambers separated by a template assembly that contained the biofilm sample. Diffusion Cell Experiments with Biomass and Low Melting Point Agarose. The model biofilms used in the diffusion cell were made using mixtures of biomass in leachate (from biofilters packed with ceramic Raschig rings treating R-pinene and with a dry solid concentration of approximately 8 g of VSS/L), water, and low melting point agarose. The model biofilms (leachate in 0.75% low melting point agarose) were made by boiling 0.75 g of low melting point agarose (ultraPURE low melting point agarose Life Technologies, Grand Island, NY) in deionized (Nanopure) water (50 mL) and then adding it to the leachate (50 mL) after it had cooled enough to allow the temperature after mixing to remain just above the gelling temperature. The mixture was then poured into a plastic Petri dish that contained a steel sample plate (with hole of 1.5 mm thickness and 30 mm diameter). One membrane (polyethersulfone, Supor-450, 47 mm diameter, 6 mil thick) was submerged on the top of the steel sample plate. The sample was allowed to gel overnight. The sample and plate were cut out from the agarose, and another membrane was placed on the now exposed agarose film. The sample plate was sandwiched between two steel washers and then placed in the diffusion cell (14, 15). Parallel diffusion cells, containing samples made with 50% leachate from the R-pinene biofilter, immobilized in low melting point agarose were spiked with R-pinene on the high side to have concentrations ranging from 700 to 1000 ppmv. The liquid R-pinene was allowed to evaporate first in the high side so that when the diffusion cell was screwed together, the initial concentration in the high side would reach a uniform value quickly and would be in the air phase. Water was added to initially humidify the air, and water was present in troughs throughout the experiment to prevent the agar films from drying out. After the concentration on the high sides had dropped, the diffusion cells were dismantled. Additional R-pinene and water were added to the high side and allowed to evaporate for a half to a full hour, and then the diffusion cells were reassembled. This was repeated three times. For the fourth and fifth spike, diffusion cell #2 was purged with nitrogen from a nitrogen gas cylinder. This was in an attempt to deplete oxygen within the diffusion cell to inhibit the aerobic degradation of R-pinene to study transport rates without reaction in a more realistic film as compared to the studies previously by Miller and Allen (14) looking at transport through films made with leachate that was deactivated with radiation. The sample in diffusion cell #1 was sprayed with Amphyl before the sixth spike to inhibit biological activity and study transport through these films.
Amphyl is a disinfectant deodorant spray that is germicidal, tuberculocidal, fungicidal, and virucidal. It contains 78.500% ethanol, 0.136% o-phenylphenol, and 21.364% inert ingredients. For the seventh spike, diffusion cell #2 was purged five times with nitrogen in an AtmosBag (Aldrich Chemical Co., Inc., Milwaukee, WI), a flexible, inflatable, polyethylene chamber, and left overnight. While still in the Atmosbag, the R-pinene was added to the high side about 1 h before the diffusion cell was put back together. Once assembled, the diffusion cell was removed from the Atmosbag to enable sampling from the gas sampling ports. Concentrations of R-pinene were measured with a gas chromatograph (Varian GC Star 3600-CX, Varian Chromatography Systems, Walnut Creek, CA). The conditions of the GC were as follows: 30 mL/min each of H2 and He (as the carrier gas), 300 mL/min air, column temperature of 120 °C, injection temperature of 150 °C, and detection temperature of 250 °C (using an FID, flame ionization detector). The GC was equipped with a 15 m, 0.53 mm i.d. megabore (5%phenyl)-methylpolysiloxane capillary column (J&W Scientific DB-5). A 250 µL Hamilton Gastight #1825 syringe (Hamilton Company, Reno, NV) was used to obtain the air samples. Batch Kinetics Experiments. Batch kinetics experiments were conducted to investigate the degradation rates of R-pinene by liquid samples of leachate and also to investigate for evidence of byproduct formation in filtered samples where most of the initial biomass was removed. Experiments were conducted whereby small amounts (5-10 mL aliquots) of leachate from the R-pinene biofilters were placed in 76 mL Teflon bottles. Experiments were performed in Teflon bottles so that the effect of R-pinene adsorption was negligible (8). After sealing, to acclimate the cells prior to filtering, 0.4 µL of R-pinene was then injected and left to evaporate in the headspace, and the concentration of R-pinene in the headspace was monitored with time. Several of the leachate samples were then filtered with 1.5 µm Whatman glass microfiber filters (934-AH) to remove most of the cells. All of the sample bottles were then spiked with 0.4 µL of R-pinene, and the air phase concentration was measured as a function of time. One of the bottles containing filtered leachate was then refrigerated to later analyze the liquid for byproduct formation, and the remaining bottles were spiked again. This was repeated to acquire filtered leachate samples that had been spiked with varying amounts of R-pinene (from 0.4 to 1.6 µL). The four filtered leachate samples were sent for analysis of byproduct formation. The liquid was extracted with three times the sample volume of ethyl acetate (3 × 5 mL) and then evaporated down to a small volume. A 1 µL sample of this was then injected into a gas chromatograph to look for major peaks (GC HP5890 with a capillary HP-1 column). Several samples were also sent to be analyzed by gas chromatography-mass spectrometry (GC/MS) (HP5890 II gas chromatograph with a DB-1 J&W Scientific capillary column coupled to a VG Trio 1000 quadrupole mass spectrometer that has an EI source) to try and identify major peaks based on their spectra as compared to library spectra. The retention times of known candidate compounds were also compared to the retention time of the byproduct.
Results and Discussion Diffusion Cell Experiments. Diffusion Cell Experiments with Leachate and Low Melting Point Agarose. Various treatments were done during different spikes of the high sides for samples made with leachate from the R-pinene biofilter, immobilized in low melting point agarose. R-Pinene was repeatedly spiked into the high sides of the diffusion cells and measured as a function of time. The next sections present the concentration versus time data for the various sets of spikes or treatments: VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5857
FIGURE 2. Removal of r-pinene from the high concentration side of the diffusion cell vs timestimes are normalized so that time zero corresponds to when the high side was spiked with r-pinene. Spikes 1-3: different symbols for repeating spikes; solid and open symbols for the duplicate diffusion cells; r-pinene concentrations in the receiving chamber remained below 2 ppmv. the acclimation phase, the effect of nitrogen, and the effect of disinfectant (Amphyl). First Three SpikessAcclimation. The degradation rates of R-pinene in both diffusion cells were reproducible. Figure 2 shows the first three spikes normalized so that time zero corresponds to when the diffusion cell was reassembled after adding the R-pinene to the high side. This demonstrates the repeatability of the method. As shown in Figure 2, there is an acclimation phase that lasts about 3000 min (∼2 days) after the first spike before the concentration in the high side drops. There are several explanations for this initial delay. For instance, even though the leachate in the sample has not been deactivated, it has been mixed with low melting point agarose, which results in an environmental change for the degrading organisms. As well, the microorganisms present in the sample may need to be acclimated to a higher concentration of R-pinene than that in the biofilters (1000 ppmv vs 35 ppmv). Also, the number of microorganisms required to appreciably degrade the R-pinene may be larger than the initial number so that there would be an acclimation period before the concentration drop would be observed. Furthermore, the microorganisms may need an acclimation period to induce the production of enzymes that facilitate degradation. After the second and third spike of R-pinene, there was no acclimation period, which indicated that the organisms had adjusted to the conditions in the diffusion cells. The concentration data decreases linearly with time (i.e., constant degradation rate) for the main portion of the experiment. It is only at low concentrations that the rate of degradation slows down, which is typical of first-order biological degradation curves. Effect of Nitrogen Purging (Lack of Oxygen). One of the diffusion cells was purged with nitrogen gas in an attempt to inhibit the aerobic degradation of R-pinene, to study transport rates in the absence of reaction. Figure 3 shows the data from the fourth and fifth spikes where diffusion cell #2 was purged with nitrogen. It also shows the results from the seventh spike, which occurred after this diffusion cell was placed overnight in a nitrogen atmosphere inside the Atmosbag. Figure 3 shows that, contrary to expectations, a nitrogen environment does not prevent the degradation of R-pinene from the high side. One possible explanation for the depletion of R-pinene in a nitrogen environment is that an initial reaction with R-pinene takes place that requires very little oxygen to render the R-pinene soluble and nonvolatile. If this first step is rapid and carried out by enzymes, this could explain higher degradation rates observed in the biofilters and in the diffusion cells. If the R-pinene is degraded aerobically, an estimate of the maximum required oxygen in 5858
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
FIGURE 3. Effect of nitrogen purging on r-pinene removalshigh side r-pinene concentration as a function of time. Spikes 4 and 5: with N2 purge (untreated diffusion cellssolid symbols for comparison). Spike 7: in Atmosbagsr-pinene concentrations in the receiving chamber remained below 4 ppmv.
FIGURE 4. Effect of Amphyl on r-pinene removalshigh side r-pinene concentration as a function of time. Spike 6: with Amphyl (untreated diffusion cellsopen symbols for comparison). Spike 7: with Amphyl residual (remaining from the first spraying)sr-pinene concentrations in the receiving chamber remained below 2 ppmv except for samples treated with Amphyl, which are given in Figure 5. the air is less than 1.4 vol % for the highest R-pinene concentration of 1000 ppmv (assuming the initial R-pinene in the high side completely mineralizes to carbon dioxide and water (i.e., 14 mol of O2 per mol of R-pinene, C10H16)). If only one oxygen atom (O) is required per molecule of R-pinene, this decreases the amount of oxygen required by a factor of 28, and it is possible that there was still sufficient oxygen present due to incomplete purging or diffusion of oxygen into the diffusion cell. Under strict anaerobic conditions, degradation of R-pinene did not take place in leachate samples taken from the biofilters (15). There have been studies reporting that R-pinene can be degraded anaerobically with byproducts being formed (23, 24). Because there was no apparent acclimation of these samples to anaerobic conditions, it is unlikely that the degradation that takes place is anaerobic. The fact that degradation still took place in the diffusion cell but not in anaerobic batch experiments (15) indicates that the diffusion cell was not completely purged of all oxygen. Studying Transport by Inactivation of the Microorganisms. As an alternative method to study transport without reaction in a film acclimated to degrade R-pinene, a disinfectant spray was used. This was also done to confirm that it is a biological phenomenon responsible for the degradation of R-pinene. The model biofilm in diffusion cell #1 was sprayed with Amphyl before the sixth spike. Figure 4 shows the high side concentration measurements from diffusion cell #1 as compared to those of diffusion cell #2 used as a control between nitrogen treatments. Figure 4 shows that except for the first data points for the sixth spike, the concentration on the high side of the diffusion
FIGURE 5. Low side r-pinene concentrations as a function of time in diffusion cell experiments using 50% r-pinene grown leachate immobilized in low melting point agarose. Spike 6: with Amphyl. Spike 7: with Amphyl residual left in the sample from the spraying before spike 6. cell with the model biofilm sprayed with Amphyl drops at a much lower rate than the concentration in the control diffusion cell. This supports the idea that it is microbial degradation that is responsible for the removal of R-pinene from the high side of the diffusion cell. Figure 5 shows that the R-pinene concentration on the low side did rise when the model biofilm was treated with Amphyl, indicating that the R-pinene is able to penetrate through the whole film. The R-pinene in the other spikes not treated with Amphyl was being degraded within the film before it was able to diffuse across the whole sample since the concentration in the low sides did not increase significantly (concentrations of less than 2 ppmv). The concentration in the low side rises linearly during the initial period and then appears to level off for the sixth spike. One possible explanation for this phenomenon is that because the model biofilm was sprayed with Amphyl on the side that was facing the high side, this may have led to incomplete penetration of the disinfecting components into the other side of the film during the first 4000 min. The concentration may stop rising in a linear fashion if there is still a small amount of microbial degradation taking place on the low side of the model biofilm. The disinfecting components may have penetrated the whole film for the seventh spike as the concentration continues to rise throughout. The diffusion coefficient was determined to be 3.4 × 10-6 cm2/s for the sixth spike, which is the same as the diffusion coefficient through agar reported by Miller and Allen (14). This was calculated (see ref 14 for equations) using data from the linear portion of the low side data, a high side concentration of 500 ppmv, and the air/agar mobile phase partition coefficient, which was previously shown to be the same as in water (14). The diffusion coefficient was determined to be 2.6 × 10-6 cm2/s for the seventh spike, which again is on the same order of magnitude as the value through agar. These findings, consistent with those presented in our previous paper (14), are contrary to the initial hypothesis that the transport rate through an R-pinene acclimated biofilm is enhanced due to physicochemical changes to the film induced by the microorganisms (e.g., surfactants, adsorption to solids). It was initially thought that this was achieved by an increase in the overall solubility in the biofilm and that studying the partitioning and diffusion of R-pinene through more realistic films would show an increased transport rate. Results from the previous transport experiments (14) indicate that higher transport rates will only be seen if the solubility of the mobile phase or the transport through the mobile phase increases. If the microorganisms degrading the R-pinene affect this higher mobile phase transport, it would only manifest itself in the active portion of the film.
FIGURE 6. Removal rates on a basis of (g/(cm2 min) × 107) for various spikes in the two diffusion cells as compared with a biofilter estimateserror bars represent the 95% confidence levels based on the linear portion of the slopes. Replicates except where noted. Degradation Rates. Figure 6 summarizes the calculated degradation rates (g/(cm2 min)) for each spike. The rates of degradation were estimated from the slope of the linear portion of the curves for each spike and the model biofilm surface area (i.e., the slope was multiplied by the volume of the high side of the diffusion cell and subsequently divided by the surface area of the model biofilm). Also shown is the rate estimated for a biofilter with a removal rate of 20 g/(m3 bed h) and assuming complete coverage of 1/2 in. Raschig rings with a surface area of 370 m2/m3 bed. As shown in Figure 6, the degradation rates in the diffusion cell are on the same order of magnitude and range from 2 to 4 × 10-7 (g/(cm2 min)) for the acclimated model biofilm samples, excluding those sprayed with Amphyl. The small differences in rates could be due to a variety of reasons. The rates are calculated from the linear data of the degradation curves before the rate slows down. If data points were used that were not in the linear region, it could slightly change the rates. In addition, these are biological degradation rates and so the catalyst, microorganisms or enzymes, may be changing for the different spikes. The diffusion cell rates are generally higher than the rate estimated for the biofilter but still on the same order of magnitude. These slightly higher rates may be due to the fact that the concentration in the diffusion cell is higher than in the biofilters. However, the concentration drop in the diffusion cell, which is proportional to the rate, remains linear until concentrations just below 50 ppmv, and then it drops at a slower rate. Because the biofilters have concentrations between 5 and 160 ppmv, it is expected that the biofilter rate is smaller than the diffusion cell rates. Moreover, the rates in the biofilter may actually be higher on a per unit area basis due to incomplete coverage of all the Raschig rings. Because of the high rates of removal and the lack of increase in transport rates through these films and the low amounts of oxygen required to remove R-pinene from the air phase, a new mechanism of R-pinene transport and reaction is proposed. This new mechanism involves the biologically mediated transformation of R-pinene into a more soluble byproduct, the formation of which was investigated in filtered leachate samples as described in the following section. Batch Kinetic Experiments with Filtered Sampless Evidence of Byproduct Formation. Experiments with filtered leachate in Teflon bottles showed R-pinene removal from the gas phase. The air phase concentration profiles for the samples, which were spiked at varying times with R-pinene, are presented in Figure 7. The air phase R-pinene concentrations in samples containing filtrate drop more slowly than those of the unfiltered VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5859
FIGURE 8. Byproduct formation in filtered leachate samplesscounts of the peak for filtered leachate samples spiked with varying amounts of r-pinene as compared to a water control and an unfiltered leachate sample (with cells). FIGURE 7. Air phase r-pinene concentration profiles in samples containing filtered leachate for various spikes of r-pinene as a function of timeslines shown as visual aids to follow the trend for each spike. Filtrate from filtering with 1.5 µm Whatman glass microfiber filters (934-AH). W. cells are from an unfiltered leachate sample. sample during the first spike (Figure 7). For subsequent spikes, the rates are similar, indicating that there is an initial acclimation phase needed. If an extracellular enzyme was filtered out of the original sample, microorganisms producing the enzyme would need an acclimation phase to multiply if they were present in the sample. These microorganisms could be present in the initial samples since they were not prepared in a sterile manner. The enzyme could also be membrane bound, and so the cells producing the enzyme would need to acclimate. Alternatively, the enzyme was present in the filtered sample, but it may need active cells to regenerate it or cofactors for the oxidation to take place. The hypothesis that the enzymes or cofactors need to regenerate themselves is supported in the literature. In similar experiments, Dybas et al. (25) found that washed Pseudomonas sp. strain KC cells did not significantly transform carbon tetrachloride to carbon dioxide and that cell-free culture supernatant occasionally did, but not reliably. They also found activity when they reconstituted the washed cells with the 500 MW filtrate fraction, which indicated that both intracellular and extracellular factors are normally required. They proposed a possible model that involves reactivation of the factor(s) at the cell membrane. A similar phenomenon could be taking place in these filtered leachate samples where a cofactor needs to be regenerated by cells in order for the oxidation of R-pinene into something more soluble to occur. Consistent with this, it has been reported that the R-pinene monooxygenase that catalyzes the oxygenation of R-pinene to R-pinene epoxide is dependent on NADH as a cofactor (26, 27). The fact that the filtered leachate samples were clear at the beginning of the experiment and were slightly cloudy at the end of the experiment shows that there was some microbial growth. However, it is not known whether these microorganisms were producing the enzyme required for the oxidation of R-pinene to take place or providing the reductive power to regenerate the cofactor. GC analysis of the extracted liquid phase samples showed one major peak, much larger than the other peaks, suggesting that a byproduct was being formed in these samples. The counts of this peak, which eluted from the column at a retention time of 8.5 min, are given in Figure 8 for the filtered leachate samples that had been spiked with varying amounts of R-pinene. The amount of byproduct, given by the counts or peak area on the GC, increases in the filtered leachate samples as 5860
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
FIGURE 9. Structures of cis-2,8-p-mentha-dien-1-ol, trans-carveol, and r-pinene. more R-pinene is added to the samples, as shown in Figure 8. For comparison, the amount of byproduct in the water control is much smaller than in the filtered leachate samples, supporting the hypothesis that a byproduct is being formed as the result of R-pinene oxidation. The leachate sample with the initial amount of cells also has a smaller amount of byproduct, indicating that it is being further broken down by the cells. It is also interesting to note that after the first sample, to which 0.4 µL of R-pinene was added, the amount of byproduct that built up in the filtered samples is almost directly proportional to the amount added, which indicates that the amount of byproduct formed increased with the amount of R-pinene consumed. The initial extra amount of byproduct in the samples may be from byproduct build-up, while the microorganisms were being acclimated to R-pinene prior to the experiment. This would happen if the degradation rate of the byproduct by other cells is slower than the rate at which R-pinene is converted into something more soluble. This would also explain why there is byproduct in the sample that contains cells. In other words, there would still be byproduct in the sample with cells even when all of the R-pinene had been removed from the gas phase. Identification of the Byproduct. The proposed identity of the byproduct found building up in the filtered samples is cis-2,8-p-menthadien-1-ol, a menthadienol. Its structure is given in Figure 9 along with trans-carveol, shown for comparison since it is an isomer of carveol, a product that has been reported to be an R-pinene degradation product (28, 29). The structure of R-pinene is also given to show the reactant structure prior to ring opening and the addition of the OH group. The compound, cis-2,8-p-mentha-dien-1-ol whose structure is shown first in Figure 9, was proposed based on an exhaustive search of mass spectra available in mass spectra libraries and literature as well as GC analysis of potential compounds. The extracted samples containing the byproduct were sent for GC-MS (gas chromatography-mass spectroscopy), and it was this resulting mass spectra (Figure 10) that was compared to library spectra. Some possible compounds were also eliminated based on a comparison of GC retention times of commercial samples to the retention time of the byproduct. These compounds were chosen because of their similarity in reported library mass spectra as well as
FIGURE 10. Mass spectrum of byproduct peak. similarity in structure. These compounds also have higher solubilities and lower volatilities than R-pinene, properties that the byproduct is expected to have. Figure 10 shows that there are significant peaks at relative masses of 91, 119, 134, and 77, which have been reported in other menthadienols. Arata et al. (30) looked at the isomerization of 2- and 3-carene oxides, terpene oxides, over solid acids and bases and examined the byproducts that were formed. They identified one of the byproducts to be cis2,8,(9)-p-mentha-dien-1-ol and reported peaks at 134, 119, 91, 109, 79, 43, 41, and 152 (M+) as compared to a reference sample that had peaks at 91, 119, 134, 41, 79, 77, 43, and 152 (M+). This authentic sample was prepared by isomerization of 2-carene oxide with metatitanic acid according to a reported patent. This sample has an almost identical mass spectrum to the byproduct in our study, which supports the identification of this byproduct as cis-2,8,(9)-p-menthadien1-ol. In addition, Garneau et al. (31) reported the mass spectra of six menthadienols including that of cis-2,8,(9)-p-menthadien-1-ol. The major peaks were found at similar numbers but with an extra peak reported at 109. This peak was present in all of their mass spectra of menthadienols but is not present in all mass spectra of menthadienols; thus, it could depend on the type of machine used to do the analysis. The similarity of the mass spectra reported previously to the mass spectrum of the byproduct supports the identification of the byproduct as cis-2,8-p-mentha-dien-1-ol. The byproduct mass spectrum shown in Figure 10 does not have a peak at 152, the molecular ion peak. Other menthadienols that Arata et al. (30) reported were missing the 152 peak and only had the 134 (M+ - 18) peak. This indicates that a water group would be missing from the mass spectra and may be lost in the injector of the gas chromatograph. Furthermore, the 152 peak is the smallest peak reported by Arata et al. (30) for the compounds that have peaks at 152. This indicates that the molecular mass ion peak may not be present in the mass spectrum. Physical Properties of the Byproduct. Because cis-2,8,(9)p-menthadien-1-ol is similar in structure to carveol, we estimated its solubility to be similar to carveol, and thus much higher than that of R-pinene. The solubility of carveol ranges from 1115 to 2900 mg/L (11, 13), which is much higher than the range given for R-pinene of 2-22 mg/L (8-13). trans-2,8-p-Menthadiene-1-ol as well as many other oxygenated terpenoids have been reported to be found in aqueous aroma components in oranges, indicating that these components are water soluble (32). However, one of the
compounds found in the aqueous aroma extracts, dlimonene, is insoluble in water. Nevertheless, since dlimonene is the main component of oil extracted from citrus rind, it is not surprising that there may be a small amount partitioned into the aqueous phase even though it is not water-soluble. The fact that carveol’s solubility is high and that a similar compound, trans-2,8-p-menthadiene-1-ol, is found in aqueous extracts supports the notion that the proposed compound, cis-2,8,(9)-p-menthadien-1-ol, has a much higher solubility than R-pinene. The vapor pressure, which characterizes the volatility of a compound, was estimated to be lower than R-pinene for this byproduct based on calculated values. Although experimental values of the vapor pressure of this compound were not available, the vapor pressures of oxygenated terpenoids are much lower than terpenes, and so we would expect this compound to have a low vapor pressure. Moreover, there was no other significant peak besides R-pinene when gas phase samples were injected into the GC over the course of the experiments. This indicates that the byproduct did not build up in the gas phase even when R-pinene had been removed from the gas phase and the byproduct was clearly building up in the liquid phase. In addition, estimated vapor pressures of 3.49 and 0.029 Torr for R-pinene and cis-2,8-p-menthadien-1-ol, respectively, were found using the database program SciFinder Scholar, which calculated them using Advanced Chemistry Development (ACD) Software Solaris V4.67. The value calculated for R-pinene is close to the experimental value of 4.75 Torr as reported previously (12). This supports that the value calculated for cis-2,8,-p-menthadien-1-ol may be close to the real value in magnitude. The fact that the relative value estimated for cis-2,8-p-menthadien-1-ol is much smaller than the value estimated for R-pinene shows that, structurally, this compound should have a much smaller vapor pressure. The air/water partition coefficient or dimensionless Henry’s law constant is calculated to be 0.00024 using a solubility of 1000 mg/L and an estimated vapor pressure of 0.029 Torr. This value is 4 orders of magnitude lower than the value of 4.4 for R-pinene, showing that it is much more likely to be found in the aqueous phase. This supports the hypothesis that if a byproduct is being formed via a biologically mediated transformation, it remains in the aqueous phase where it can then diffuse deeper within the film to be further degraded. VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5861
FIGURE 11. Conceptual model of biologically mediated transformation. There is also literature pertaining to the formation of cis2,8-p-menthadien-1-ol from other organic compounds. Busmann and Berger (33) found this compound to be one of numerous conversion products of myrcene by different strains of basidiomycetes which shows that cis-2,8-p-menthadien-1-ol can be produced from another terpene. It has also been reported as a metabolic and oxidation product of limonene (34). In addition, limonene has been reported to be a product in the metabolic pathway of R-pinene degradation (35). Limonene monooxygenase is then able to transform limonene into carveol, another menthadienol. It is possible that R-pinene can undergo a ring opening (to form limonene) and oxygenation (to form cis-2,8-p-menthadien-1-ol) in one step that involves an enzyme. Alternatively, the two steps could be rapid and closely linked to form the byproduct. While cis-2,8-p-menthadien-1-ol has not been identified in previous work as a metabolic product in the R-pinene degradation pathway, the literature suggests that it is a plausible degradation compound. Proposed Conceptual ModelsBiologically Mediated Transformation of r-Pinene. A simple model was developed that described results seen in the diffusion cell using R-pinene leachate immobilized in low melting point agarose (Figure 11). The model was developed based on an enzyme catalyzing the initial partial oxidation of R-pinene into a more soluble and less volatile compound. The model assumes that the film is thick enough that the compound is then further degraded by microorganisms present deeper in the film. The soluble compound given in the previous model is denoted as R-soluble rather than cis-2,8-p-menthadien-1ol. It is possible that if different microorganisms were present in the sample, they could produce different enzymes that could oxidize R-pinene into different oxygenated terpenoids. Several terpenoids have higher solubility than that of R-pinene. Thus, depending on the metabolic pathway of the microorganisms, different soluble compounds could be formed. In addition, the concept of having a more soluble byproduct formed can be generalized to other hydrophobic pollutants. The model presented here is very different from other models of biofiltration or models of transport and reaction 5862
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 15, 2005
of a compound within a biofilm because it assumes that a biologically mediated first step is taking place, changing the R-pinene into a more soluble compound. Conventional models assume that microorganisms degrade the compound into carbon dioxide and water and put this conversion into one rate constant. The model presented here assumes that an initial conversion of R-pinene takes place into something more soluble but that other microorganisms deeper within the film are capable of converting this byproduct eventually into carbon dioxide and water through a series of metabolic steps. This research could also explain why other hydrophobic pollutants may be treated in biofilters at higher than predicted removal rates. Other compounds could have a similar mechanism of degradation within biofilters, whereby they are converted into more soluble and less volatile compounds by an enzyme present within the biofilm. The mathematical description of the conceptual model proposed here and the fitted rates of reaction of the enzyme catalyzing the partial oxidation of R-pinene in the model is the subject of another paper (36). In this paper, reaction rates and zero-order rate constants are calculated and shown to be comparable to literature values of known enzyme systems.
Acknowledgments The authors acknowledge the financial support from the Biofiltration Consortium at the Pulp and Paper Centre at the University of Toronto: Aracruz Celulose S.A. (Brazil), Bowater Pulp and Paper Canada, Inc., Domtar, Inc., Georgia Pacific Corporation, Nippon Paper Industries Co., Ltd. (Japan), and Weyerhaeuser Company as well as the Minimizing Discharges Consortium at the Pulp and Paper Centre at the University of Toronto: Aracruz Celulose, S.A., Domtar, Inc., EKA Chemicals, Inc., Georgia-Pacific Corporation, Irving Pulp and Paper, Ltd., Japan Carlit Co., Ltd., Potlatch Corporation, ERCO Worldwide (formerly Sterling Pulp Chemicals, Ltd.), Carter Holt Harvey Tasman, and Tembec, Inc. The scholarships and support of the Natural Science and Engineering Research Council of Canada (NSERC) and the financial support from the Government of Ontario/DuPont Canada Graduate Schol-
arship in Science and Technology are also gratefully acknowledged.
Literature Cited (1) 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, 3089-3102. (2) Ottengraf, S. P. P. Exhaust gas purification. In Biotechnology; Schoenborn, W., Ed.; VCH: Weinheim, Germany, 1986; pp 425452. (3) Devinny, J. S.; Deshusses, M. A.; Webster, T. S. Biofiltration for Air Pollution Control; Lewis Publishers: Boca Raton, FL, 1999. (4) Deshusses, M. A.; Shareefdeen, Z. Modeling of Biofilters and Biotrickling Filters for Odor and VOC Control Applications. In Biotechnology for Odor and Air Pollution Control; Shareefdeen, Z., Singh, A. Eds.; Springer-Verlag: Berlin, Heidelberg, 2005. (5) Mohseni, M.; Allen, D. G. Biofiltration of Mixtures of Hydrophilic and Hydrophobic Volatile Organic Compounds. Chem. Eng. Sci. 2000, 55, 1545-1558. (6) Mohseni, M.; Allen, D. G. Biofiltration of R-Pinene Using Wood Waste and Activated Carbon Media. In Proceedings of the 1996 Conference on Biofiltration (on Air Pollution Control Technology); Reynolds, F. E., Ed.; The Reynolds Group: Tustin, CA, 1996. (7) van Groenestijn, J. W.; Liu, J. X. Removal of R-Pinene from Gases using Biofilters Containing Fungi. Atmos. Environ. 2002, 36, 5501-5508. (8) Miller, M. J. Determination of Henry’s Law Constant and Aqueous Solubility of R-pinene. B.S. Thesis, University of Toronto, 1997. (9) Li, J.; Perdue, E. M.; Pavlostathis, S. G.; Araujo, R. Physicochemical Properties of Selected Monoterpenes. Environ. Intl. 1998, 24, 353-358. (10) Schmid, C.; Steinbrecher, R.; Ziegler, H. Partition Coefficients of Plant Cuticles for Monoterpenes. Trees 1992, 6, 32-36. (11) Fichan, I.; Larroche, C.; Gros, J. B. Water Solubility, Vapor Pressure, and Activity Coefficients of Terpenes and Terpenoids. J. Chem. Eng. Data. 1999, 44, 56-62. (12) Howard, P. H.; Meylan, W. M. Handbook of Physical Properties of Organic Chemicals; Lewis Publishers: Boca Raton, FL, 1997. (13) Weidenhamer, J. D.; Macias, F. A.; Fischer, N. H.; Williamson, G. B. Just How Insoluble are Monoterpenes? J. Chem. Ecol. 1993, 19, 1799-1807. (14) Miller, M. J.; Allen, D. G. Transport of Hydrophobic Pollutants through Biofilms in Biofilters. Chem. Eng. Sci. 2004, 59, 35153525. (15) Miller, M. J. Transport and Degradation of Hydrophobic Pollutants in Biofilms in Biofilters. Ph.D. Thesis, University of Toronto, 2004. (16) Spigno, G.; Pagella, C.; Fumi, M. D.; Molteni, R.; De Faveri, D. M. VOCs Removal from Waste Gases: Gas-phase Bioreactor for the Abatement of Hexane by Aspergillus niger. Chem. Eng. Sci. 2003, 58, 739-746. (17) Kibazohi, O. Biofiltration of Hexane: Enhancement of Performance by Nitrogen Supplementation. Ph.D. Thesis, University of Waterloo, 2000. (18) Deshusses, M. A.; Johnson, C. T. Development and Validation of a Simple Protocol to Rapidly Determine the Performance of Biofilters for VOC Treatment. Environ. Sci. Technol. 2000, 34, 461-467.
(19) Mohseni, M. Biofiltration of Hydrophilic and Hydrophobic Volatile Organic Compounds Using Wood-Based Media. Ph.D. Thesis, University of Toronto, 1998. (20) Dirk-Faitaikis, C.; Allen, D. G. Biofiltration of Cyclic Air Emissions of R-pinene at Low and High Frequencies. J. Air Waste Manage. Assoc. 2003, 53, 1373-1383. (21) Dirk-Faitakis, C. B. Experimental Studies and Theoretical Modeling of an Unsteady-State Biofilter used to Treat Fluctuating Concentrations of an R-pinene Air Emission. Ph.D. Thesis, University of Toronto, 2000. (22) de Castro, A. M. Fatty Acid Methyl Ester Analysis of Microbial Communities in Biofilters Inoculated With Different Sources. M.S. Thesis, University of Toronto, 1997. (23) Harder, J.; Probian, C. Microbial Degradation of Monoterpenes in the Absence of Molecular Oxygen. Appl. Environ. Microbiol. 1995, 61, 3804-3808. (24) Heyen, U.; Harder, J. Geranic Acid Formation, an Initial Reaction of Anaerobic Monoterpene Metabolism in Denitrifying Alcaligenes defragrans. Appl. Environ. Microbiol. 2000, 66, 30043009. (25) Dybas, M. J.; Tatara, G. M.; Criddle, C. S. Localization and Characterization of the Carbon Tetrachloride Transformation Activity of Pseudomonas sp. Strain KC. Appl. Environ. Microbiol. 1995, 61, 758-762. (26) Best, D. J.; Floyd, N. C.; Magalhaes, A.; Burfield, A.; Rhodes, P. M. Initial enzymic steps in the degradation of R-pinene by Pseudomonas fluorescens NCIMB 11671. Biocatalysis 1987, 1, 147-159. (27) Colocousi, A.; Leak, D. J. Cloning and Expression of the R-pinene Monooxygenase Gene From Pseudomonas fluorescens NCIMB 11671. In Environmental Biotechnology: Principles and Applications; Moo-Young, M., Anderson, W. A., Chakrabarty, A. M., Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1995; pp 144-157. (28) Savithiry, N.; Gage, D.; Fu, W.; Oriel, P. Degradation of Pinene by Bacillus pallidus BR425. Biodegradation 1998, 9, 337-341. (29) Gibbon, G. H.; Pirt, S. J. The Degradation of R-pinene by Pseudomonas PX1. FEBS Lett. 1971, 18, 103-105. (30) Arata, K.; Bledsoe, J. O., Jr.; Tanabe, K. Epoxide Rearrangement. 11. Isomerization of 2- and 3-carene Oxides Over Solid Acids and Bases. J. Org. Chem. 1978, 43, 1660-1664. (31) Garneau, F. X.; Collin, G.; Gagnon, H.; Jean, F. I.; Chalchat, J. C. Mass Spectra and Retention Indices of Six Menthadienols. J. Essent. Oil Res. 1997, 9, 491-493. (32) Moshonas, M. G.; Lund, E. D.; Berry, R. E.; Veldhuis, M. K. Distribution of Aqueous Aroma Components in the Orange. J. Agric. Food Chem. 1972, 20, 668-690. (33) Busmann, D.; Berger, R. G. Conversion of Myrcene by Submerged Cultured Basidiomycetes. J. Biotechnol. 1994, 37, 39-43. (34) Connolly, J. D.; Hill, R. A. Dictionary of Terpenoids; Chapman and Hall: London, 1991. (35) Yoo, S. K.; Day, D. F.; Cadwallader, K. R. Bioconversion of Rand β-pinene by Pseudomonas sp. strain PIN. Process Biochem. 2001, 36, 925-932. (36) Miller, M. J.; Allen, D. G. Modeling Transport and Degradation of Hydrophobic Pollutants in Biofilter Biofilms. Chem. Eng. J. 2005, in press.
Received for review November 9, 2004. Revised manuscript received April 20, 2005. Accepted May 5, 2005. ES048254Y
VOL. 39, NO. 15, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5863
