The Biofiltration of Indoor Air: Air Flux and ... - ACS Publications

Dec 1, 2000 - ALAN B. DARLINGTON,*. JAMES F. DAT, AND MICHAEL A. DIXON. Division of Horticultural Science, Department of Plant. Agriculture ...
0 downloads 0 Views 109KB Size
Environ. Sci. Technol. 2001, 35, 240-246

The Biofiltration of Indoor Air: Air Flux and Temperature Influences the Removal of Toluene, Ethylbenzene, and Xylene ALAN B. DARLINGTON,* JAMES F. DAT, AND MICHAEL A. DIXON Division of Horticultural Science, Department of Plant Agriculture, University of Guelph, Guelph, Ontario, Canada N1G 2W1

An alternative approach to maintaining indoor air quality may be the biofiltration of air circulated within the space. A biofilter with living botanical matter as the packing medium reduced concentrations of toluene, ethylbenzene, and o-xylene concurrently present at parts per billion (volume) in indoor air. The greatest reduction in concentrations per pass was under the slowest influent air flux (0.025 m s-1); however, the maximum amount removed per unit time occurred under the most rapid flux (0.2 m s-1). There was little difference between the different compounds with removal capacities of between 1.3 and 2.4 µmol m-3biofilter s-1 (between 0.5 and 0.9 g m-3biofilter h -1) depending on influent flux and temperature. Contrary to biofilters subjected to higher influent concentrations, the optimal temperatures for removal by this biofilter decreased to less than 20 °C at the most rapid flux for all three compounds. Microbial activity was decreased at these cooler temperatures suggesting the biofilter was not microbially limited but rather was limited by the availability of substrate. The cooler temperatures allowed greater partitioning of the VOCs into the water column which had a greater impact on removal than its reduction in microbial activity.

Introduction Although the concentrations of VOCs typically found indoors are very low by industrial biofiltration standards (i.e. typically less than 200 µg m-3) (1, 2), the fact that North Americans spend in excess of 85% of their time indoors (3) makes these exposure levels an important public health issue. Traditionally, indoor air quality (IAQ) is maintained through ventilation with new outside air, diluting the indoor contaminants. However, conditioning this additional air flow in terms of its temperature can greatly increase the energy consumption of the building and hence the building’s operation cost. An alternative approach may be biofiltration of the air circulated within the space. However, the very low levels and the broad range of VOCs typically found indoors suggest that it would be extremely difficult to develop and maintain the highly specialized microbial population seen in traditional biofilters. It may be more appropriate to utilize phytoremediation approaches developed for contaminated soils (4, 5) where green plants facilitate contaminant removal. More specific to the indoor environment, green plants offer the ability to * Corresponding author phone: (519)824-4120 ext 3589; fax: (519)767-0755; e-mail: [email protected]. 240

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001

sequester the major indoor contaminant CO2 from the indoor environment (6). Biofilters containing plants remove significant amounts of VOCs from the indoor space (7, 8) and do not lower IAQ due to generation of other VOCs and microbial spores (8). However, before this technology can gain acceptance, descriptions of the performance of the system are required. Because of the low concentrations of VOCs involved, complete removal of contaminants is not realistic. But rather, it is important to determine conditions which enable maximum removal of VOCs per unit time. To this end, the air flux through the filter and the temperature were varied. Altering the flux of air (the rate of air flow expressed as m s-1, i.e., m3air m-2biofilter s-1) with a given concentration of VOCs through a biofilter with set dimensions will influence both the VOC loading rate (amount of material per unit filter area or volume per unit time) and the retention time of the material in the filter. Changes in temperature will affect both the physical (the solubility of the VOCs) and biological characteristics of the biofilter. The monoaromatics (toluene, ethylbenzene, and xylene) collectively referred to as TEX were chosen for this study because of their high frequency of occurrence indoors. A survey of 26 houses found TEX in all locations, with average concentrations of 20.2, 2.4, and 5.8 µmol m-3, respectively, with some as high 2400 µmol m-3 (9). For analytical ease, only o-xylene was tested. The behavior of these compounds in industrial biofilters has been well studied (10-12); however, little is known about the ability of biofilters to remove these compounds at “realistic” indoor air concentrations. To be an effective means of maintaining IAQ biofilters must be able to remove compounds such as these. The goal of this study is 3-fold: first, to determine whether the biofilter can remove TEX present at realistic indoor contaminant levels under “real world conditions”; second, to determine the biofilter’s optimum operating conditions in terms of temperature and influent flux for the removal of TEX; and third, to determine the relative roles of VOC partitioning into the water column and microbial degradation rates on removal over the range of examined temperatures and fluxes.

Experimental Methods The Building. The test biofiltration system was located in the ground floor of Canada Life Assurance headquarters building (Toronto, ON, Canada). The configuration of the building, its air handling system and the 160 m2 (640 m3) “environmental room” (the room housing the biofilter) have been described in detail elsewhere (7, 8, 13). Although the room housing the biofilter was designed to treat air from the rest of the building, the room was operated independent of the building to facilitate experimentation. The air was circulated within the room at 15-20 air changes per hour by its own dedicated air handling system. During the experimental period, leakage from the room averaged 0.2 ( 0.1 air changes per hour. The room was maintained at 22 °C during the day and 18 °C at night with relative humidity between 40 and 75%. Natural lighting from south facing windows was supplemented by 12 h of lighting from high-pressure sodium and metal halide lamps located 2.3 m above the aquarium. The environmental control system was provided by L. W. Anderson Software Consultant (Leamington, ON). The Biofiltration System. The system has been described elsewhere (8, 13). Briefly, there were three major functional components of the system: an array of bioscrubbers through which room air was drawn, a region of hydroponically grown 10.1021/es0010507 CCC: $20.00

 2001 American Chemical Society Published on Web 12/01/2000

plant canopy and thus the impact of the higher plants on removal was minimal. At the base of the entire plant community was a 3500 L aquarium containing a variety of aquatic (such as Elodea sp., Cabomba sp., and Vallisneria sp.) and semiaquatic plants (such as Cyperus spp., Myriophyllum prosperinacoides, and Lysimachia sp.). Aquarium water was circulated through the hydroponics and flowed down the surface of the biofilters. Condensate from the airconditioning unit associated with the room’s air handling system was returned to the aquatic system. There was little or no loss of liquid water over the duration of the experiment although about 20 L of water was required per day to compensate for the loss of water vapor.

FIGURE 1. Experimental setup of indoor air biofilter; four biofilter modules (only two are shown) were arranged in parallel in terms of air flow in a relatively sealed indoor space. Air was drawn through the biofilters (a) by a dedicated air handling system (b) and returned it to the influent air mass (c). Fluxes through the biofilters were independently controlled with valves (d). TEX levels of the effluent and influent air streams were automatically measured with a gas chromatograph. A solenoid system (e) interfaced with the chromatograph selected the sampling site. To control influent concentrations, influent readings were transferred to a peripheral computer (f) which activated controlled air flow through one of three impingers of the specific VOC (h) (only one shown). The VOCs were released to the space via the air handling system. higher plants, and an aquarium. The experimental arrangement is presented in Figure 1. The bioscrubbers (designed by Genetron Systems Inc., Toronto, ON, Canada) were four fiberglass air plenums (1.2 × 2 × 0.2 m) faced with porous, constantly wetted lava rock arranged in parallel. The biofilters were separated by 0.7 m. The external rock vertical face of four modules was covered with mosses (Plagiomnium cuspidatum and Taxiphyllum deplanatum) supported by geotextile cloth. The depth of moss on each biofilter (thickness of the moss and support media) was estimated based on 20 subsamples per panel and varied between 1.5 and 2.5 cm. It is reasonable to assume that the mosses had superficial surface areas similar to pure peat, reported as 1.6 m2 g-1 (14). The system was not inoculated with specific microbes. For readily degraded VOCs, the degradation occurs in the scrubbing unit and does not accumulate in the water column (7, 8). Thus, each “bioscrubber” can be considered as an independent trickling “biofilters”. The biofilters were integrated into the air handling system for the room. Air from the room was drawn through a 0.27 m2 subsection of each biofilter by a variable speed centrifugal fan and then mixed with the normal return air from the room, conditioned in terms of its humidity and temperature, and then circulated back to the room. The air fluxes through each biofilter were controlled with valves located between the biofilters and the air handling system and monitored by anemometers located in the air ducts. The air flux rates of the air stream through each of the separate biofilters were varied daily. Four different air fluxes of 0.025, 0.05, 0.100, and 0.200 m s-1 were tested (roughly equivalent to 5-40 ft3 of air per ft2 of biofilter per minute). Air flow through the biofilter could also be forced through a continuous 1 m wide region of hydroponically grown plants in front of the biofilters. The principal plant species in this region were Dracaena godseffiana, Adiantum raddianum, Hedera helix, Spathiphyllum maunahoa, Rhododendron obtusum, Marraya sp., Vriesea splendens, and Dieffenbachia picta. In this study, the biofilters were situated above the

The aquatic system supplied the water (and heat) to the biofilters. In this study, the temperature of the biofilter was varied by altering the temperature of the water used to wet the filters’ surface (set to 20, 25, and 30 °C). The same temperature was applied to all biofilters at the same time. The precise temperature of the biofilter was determined as the effluent air temperature. After changes in temperature, the system was given 7 days to stabilize. The pH was 6.7 ( 0.1, and salinity did not exceed 0.1 mS cm-1. The entire system was maintained in a low nutrient state with little nutrient addition after the initial startup (24 months earlier). Despite the concentrations of macronutrients in the water column being low (FM levels), the foliar analyses of the plant community indicated no mineral deficiencies in the system. Analytical Methods. An SRI 310 gas chromatograph (GC) equipped with a 30 m RESTEK MXT-volatiles column with 0.53 mm ID and a photoionization detector was used to detect the amount of VOC in the air stream. The operating conditions were as follows: oven, 100 °C; detector, 200 °C; carrier gas (He) flow rate of 50 mL min-1. The detector was operated in HIGH GAIN mode. Under these conditions the retention times for toluene, PCE, ethylbenzene, and o-xylene were 2.3, 2.8, 3.5, and 4.1 min, respectively. During the experiments, the VOC levels at seven different locations were constantly monitored. These sites were the air exiting each of the four biofilters (the effluents): the air entering the biofilters (the influent) and two reference locations elsewhere in the building. Air samples were delivered via tubing to a central sampling site relay system and then through a 10 port VICI E36 injection valve with a 1 mL injection loop (Valco Instrument Co. Inc.) to the GC by a MagneTek vacuum pump at a rate of ca. 300 mL min-1. All tubing was either degreased flexible copper or Teflon tubing (1/8” ID) fitted with gastight Swaglock fittings. A Peter Paul Electronic three-way 24VDC solenoid was installed between the Valco injection valve and the vacuum pump. This valve drew air through the tubing (and injector) to the pump for 3.25 min (flushing volume about 1.1 L) and 45 s prior to sample injection, the valve was switched to stop the flow through the injection system. This enabled line pressure to reach near atmospheric levels prior to injection into the GC (there was a slight 1 kPa pressure drop across the biofilter). Control of Influent VOC Levels. To control influent, the GC was interfaced with a peripheral VOC release system composed of customized software operating on a PC, a switching system, and up to four VOC emitters (see Figure 1). Each emitter was composed of a low pressure air pump, an air flow regulator, and an impinger filled with the selected VOC connected to the room’s ventilation system. Briefly, when the influent (room air) was sampled by the GC, its governing computer “printed” the results to the peripheral computer (via a RS232 connection). If the actual level of the individual VOC was below the desired amounts, the peripheral computer activated the appropriate air pump, and a specific amount of air was bubbled through the solvent in the impinger. The resulting air saturated with the desired VOL. 35, NO. 1, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

241

FIGURE 2. The impact of passing air contaminated with a range of concentrations of toluene through an indoor air biofilter under a range of fluxes. The biofilter temperature set point was 25 °C and had an average thickness of 2.2 cm. The legend for all graphs is presented in (a).

FIGURE 3. The correlations of the individual toluene influent and effluent concentrations presented in Figure 2. The line of best linear fit (forced through ZERO) is also included in each plot. Similar graphs could be generated for each of the VOCs at each of the test temperatures through each of the biofilters but are not presented. VOC was released into the supply air of the room’s air handling system, dispersing the VOC throughout the room. VOC levels could not be actively lowered but occurred as a result of biofiltration or leakage. The user defined diurnal profile for each TEX started at 0 ppbv (parts per billion, volume) at 18:00, rose to 60 and 80 ppbv (a maximum of about 3 µmol m-3 or 300 µg m-3) at midnight, and back to 0 by 06:00 (see Figure 2). These influent concentrations were comparable to indoor levels (9) and mimicked the generation of the VOCs as a result of diurnal human activities. Data Analysis. The four fluxes were applied to the four biofilters on a diurnal basis (giving four replicates) at each of the three temperatures. Earlier work (8) and the work of others (12) indicated that once adapted to the VOCs, no additional acclimation period was required when the flux was changed. Although little change in the removal kinetics was seen after the first four diurnal cycles, the system was given 7 days to adjust to new temperatures. To evaluate the biofilters, the actual effluent values were compared to 242

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001

weighted averages based upon the two influent readings nearest in time to the effluent reading. The performance of the biofilters was quantified as the correlation of the effluent to the influent concentrations for each individual daily treatment cycle, while the intercept was forced through the origin (typical examples of the regressions are presented in Figure 3). The number of readings used to calculate the regressions ranged between 12 and 29. To determine differences in biological activity over the range of temperatures the first-order kinetic equation (15) was applied as follows

{ }

Ce -LKi ) exp Ci miUg

[1]

where Ce and Ci are the effluent and influent VOC concentrations, L is the direct path length through the biofilter, mi is the distribution coefficient for the VOC between the

FIGURE 4. The influence of three operating temperatures (a) 20 °C; (b) 25 °C; and (c) 30 °C and air flux on the ratio of the effluent and influent (i) toluene, (ii) o-xylene, and (iii) ethylbenzene concentrations. The individual data points were calculated by the regression of between 12 and 33 influent and effluent readings with the intercept forced through ZERO (as presented in Figure 3). Data from the four biofilters were pooled although due to the failure of equipment, three replicates were used in some treatments. Readings where either the influent or effluent levels were below the detection limit of the analysis system were not included. aqueous and gaseous phases (based on Henry’s Law), Ug is the influent flux, and Ki is the rate constant which was used as an estimate of microbial activity. Data analyses were conducted using SAS 6.12 (SAS Institute Inc. Cary, NC). Analytical procedures used the GLM Procedure.

Results and Discussion Figures 2 and 3 clearly indicate that the indoor air biofilter can remove substantial amounts of toluene present in very low concentrations from an indoor air mass. These plots are typical of the other VOCs and the different operating temperatures used in this study. The presented data also indicate that the system could accurately and repeatedly control the influent concentrations of the VOCs. Figure 3 indicates that despite being subjected to a relatively large range of concentrations, the effluent levels exhibited a good linear correlation with the influent level. Average r 2 for the three components over the range of temperatures and fluxes were 0.994 ( 0.005, 0.994 ( 0.007, and 0.995 ( 0.004 for ethylbenzene, o-xylene, and toluene, respectively.

The ratios of effluent to influent under the range of fluxes and operating temperatures for toluene, o-xylene, and ethylbenzene are presented in Figure 4. Due to equipment failure only three biofilters were used at 20 °C. All three compounds exhibited similar decreasing removal efficiency with increasing flux at all temperatures. This pattern could be made linear through the logarithmic transformation of flux (not presented). Decreasing removal efficiency with increasing air flow rates has been reported for other biofiltration systems loaded with similar VOCs at much higher levels (12, 16). Comparisons of the removal efficiencies under the different temperatures suggest improved removal at the lower fluxes with the higher temperatures. Other biofilters subjected to substantially higher influent concentration tend to show increase removal efficiency with increased temperatures (12, 16-18). It is important to note that the temperature of the biofilter was regulated via the influent water temperature. Figure 5 presents the effluent temperature under the range of fluxes and operating temperatures. Actual biofilter temperatures were modified from these starting temperatures due to mixing VOL. 35, NO. 1, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

243

TABLE 1. The Optimal Temperature of Operation and Its Corresponding Maximum Removal Rates by the Biofilters Subjected to TEX Influent Concentrations of 1.5 µmol m-3aira o-xylene

toluene influent air flux (m s-1) 0.025 0.050 0.100 0.200 a

optimum (C°)

maximum removal (µmol m-3 s-1)

25.90 24.62 21.52 19.13

1.42 1.56 1.93 2.11

ethylbenzene

optimum (C°)

maximum removal (µmol m-3 s-1)

optimum (C°)

maximum removal (µmol m-3 s-1)

25.99 24.73 21.46 18.30

1.26 1.59 2.31 2.48

26.41 24.43 21.86 19.80

1.301 1.78 2.35 2.43

The values were calculated from the polynomial lines of best fit presented in Figures 6-8.

FIGURE 5. The resulting biofilter effluent temperature associated with three different influent water temperatures under the range of air fluxes. The standard errors associated with each temperature and flux are included presented but may be smaller than the symbol. with “room temperature” air (ca. 20 °C) and due to the evaporative cooling. The strongest influence appears to be through evaporative cooling. The amount of cooling was proportional to the logarithm of the flux with little differences in the relative amount of cooling for the different operating temperatures. In all subsequent analysis the effluent temperatures were used. To determine the removal capacity of the biofilters, the amount of the VOCs removed per unit volume of the biofilter and per unit time when subjected to a theoretical influent VOC concentration of 1.5 µmol m-3air (the midpoint of the applied range) was calculated. This removal capacity is plotted as a function of temperature under the different fluxes for the different VOCs in Figures 6-8. At the slowest flux, a definite optimum temperature could be seen. And as the flux increased there was a shift to a lower optimum temperature. The data from each VOC were subjected to ANOVA using a polynomial model. Optimum removal temperatures and maximal removal capacities for each VOC across the range of fluxes were determined based upon the second derivative of these models and are presented in Table 1. The behavior across the range of fluxes were very similar for all three VOCs. For all three, maximum removal occurred at the most rapid fluxes, with only a relatively small difference between the removal rates at 0.100 and 0.200 m s-1 (between 5% and 10% greater removals at the more rapid flux). This may have been the result of more uniform VOC concentrations throughout the depth of the biofilter. Given the dimensions of the biofilters, the removals reported in Table 1 translate into removal of between roughly 0.5 gvoc m-3biofilter h-1 at the slowest flux to over 0.9 gvoc m-3biofilter h-1 at the most rapid flux. Table 1 clearly indicates the enhanced removal under the rapid influent fluxes was associated with a reduction in the optimum operating temperature. Each doubling of the flux through the filters was typically associated with a reduction of the optimum of 2 °C. Although industrial biofilters for styrene reported optimum removal within the range reported here (19), most temperature optima for other biofilters tend to be higher. Trickle bed biofilters subjected to BTEX loading rates of between 50 and 140 g m-3 h-1 and 244

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001

FIGURE 6. The influence of temperature on the removal capacity of toluene (µmol degraded per unit empty bed retention time). The individual data points were calculated from the data presented in Figure 4, normalized for the thicknesses (m) of the four different biofilters. The four fluxes, (a) 0.025; (b) 0.050; (c) 0.100; and (d) 0.200 m s-1, are presented separately along with their polynomial line of best fit. The r 2 are 0.50, 0.49, 0.04, and 0.35, respectively. a flux of approximately 0.002 m s -1 had optimum removal around 30 °C, irrespective of the VOC (11). Similarly, Sorial (12) found the removal of toluene by a peat based biofilter was still increasing at 32.2 °C. The ability of a biofilter to degrade VOCs is dependent on two factors: the partitioning of the VOC into the aqueous phase from the air and the rate at which the microbes can degrade the material. Either of these factors can be rate limiting and will be influenced by temperature differently. If microbial activity was limited, then removal would likely increase with increasing temperature (16, 17). However if the removal was diffusion limited, then the increased solubility of the VOCs under reduced temperature would lead to the greater removal due to a greater availability of substrate for the degraders. Calculated Henry’s law constant for the three VOCs (20) clearly indicates that, for a given amount the VOCs present in the air stream, the amount partitioned into the water column varied greatly over the range of temperatures examined (increasing with decreasing temperature). The rate constants (an estimate of biological activity of the biofilters) were calculated for each biofilter at each set temperature based upon eq 1, using the calculated partitioning coefficients for the VOCs. The rate constants (s-1) were subjected to ANOVA. Further analysis indicated no significant interactions between the biofilter and the influence of temperature on the rate constant for each VOC (all biofilters had similar responses to temperature for each VOC). The results are presented in Figure 9. For all three VOCs, there was a significant reduction in the rate constant at the cooler operating temperature (20 °C)

FIGURE 7. The influence of temperature on the removal capacity of o-xylene (µmol degraded per unit thickness of the biofilter per unit time). The individual data points were calculated from the data presented in Figure 4, normalized for the thicknesses (m) of the four different biofilters. The four fluxes, (a) 0.025; (b) 0.050; (c) 0.100; and (d) 0.200 m s-1, are presented separately along with their polynomial line of best fit. The r 2 are 0.55, 0.73, 0.33, and 0.48, respectively.

FIGURE 8. The influence of temperature on the removal capacity of ethylbenzene (µmol degraded per unit thickness of the biofilter per unit time). The individual data points were calculated from the data presented in Figure 4, normalized for the thicknesses (m) of the four different biofilters. The four fluxes, (a) 0.025; (b) 0.050; (c) 0.100; and (d) 0.200 m s-1, are presented separately along with their polynomial line of best fit. The r 2 are 0.75, 0.95, 0.24, and 0.27, respectively. with no significant difference between 23 and 26 °C. Examining the influence of temperatures between 15 and 50 °C on industrial trickle bed biofilter, Lu and co-workers (11) found the rate constant dropped quickly below 25 °C and that the rate of change in the rate constant increased with decreasing biofilter loading. Significant microbial population shifts have been reported at temperatures between 23 °C (19) and 30 °C (11). It is interesting to note that here, as in earlier studies (7, 8), there was no acclimation period associated with change in flux. Because of this, it is unlikely that the change in removal was due to changes in the potential number of degraders (changes in populations or activation of genetic pathways). However, changes in biofilter temperature required an acclimation period of up to 4 days. This supports the contention that there were changes in the degrader populations which could explain the observed differences in rate constants with temperature. Although the presented work was not designed to test the effectiveness of this biofilter relative to other systems, several observations can be made about the performance of a biofilter

FIGURE 9. The rate constant (s-1) (estimating biological activity of the biofilters) as a function of temperature (°C) for toluene (9); o-xylene (b); and ethylbenzene (1). The vertical bars represent the standard error of the mean (n ) 4). The presented temperatures are the average effluent temperature across the range of fluxes under the different operating temperatures. subject to very low influent concentration. First, results indicate that biofiltration can remove volatile organic compounds present at concentrations typical of the indoor environment. Thus the use of biofiltration to replace or supplement traditional ventilation techniques to maintain indoor air quality deserves further consideration. Obviously, whether the biofiltration of indoor air is an economic choice depends on both the relative efficiencies of the system and energy cost, both of which are beyond the scope of this study. Second, the effective removal of VOCs through a relatively thin biofilter could in part be due to the use of a living plant as a biofilter medium. Botanical systems are able to support a population of degrading microbes greater than could be expected given the VOC availability for a number of reasons. Green plants such as mosses are known to generate a number of volatile and semivolatile compounds as part of their normal metabolism which act as an alternative food source for VOC degrading microbes (21). Also mosses release compounds which alter the microbial composition of their microenvironment (22). The effect may also simply be physical. The small thin leaves of moss are typically only a single cell thick and therefore mosses possess very large surface area-to-mass ratios giving ample surface area for biofilm generation. Despite being resistant to microbial breakdown, the leaf tissue will senesce and decay. However being bioregenerative, the moss support media will constantly grow new surface area to replace the dead material. Also, the development of hypoxic regions in the biofilm can reduce microbial activity (23). The use of a photosynthetically active substrate could overcome this limitation. Third, biofilters subjected to these low VOC concentrations respond in a manner similar to industrial biofilters with the exception that under low influent concentrations the system is not limited by microbial activity (potentially a consequence of the use of living botanical as a filter medium) but rather is substrate limited. Thus the filter removes TEX most effectively at low temperatures. Operating the indoor air biofilter at cooler temperatures has benefits on indoor air quality beyond improved VOC removal. Because of passing through a water soaked support medium, the air exiting the biofilter will be at 100% relative humidity. If the biofilter is maintained at warm temperatures, then the humidity of the surrounding air mass will build up to the point of influencing IAQ and potentially damaging the building. However, if the biofilter is maintained several degrees cooler than the indoor space, then as the cooler saturated air exits the biofilter, it will be warmed and dried by the ambient air mass. This will reduce the negative impact of the water vapor. Also maintaining the biofilter at cool temperatures limits the probability of the pathogen Legionella. VOL. 35, NO. 1, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

245

The presented work clearly indicates that biofiltration of indoor air to remove aromatic volatile organic compounds is possible.

Acknowledgments This research was funded in part by Centre for Research in Earth and Space Technology (CRESTech) research contract and grants from the Ontario Ministry of the Environment and Energy (MOEE) and Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA). The authors also would like to acknowledge the generous support of Canada Life Assurance and their subsidiary Adason Properties and the participation of Genetron Systems whose patents (No. 4,754,571 (U.S.) and No. 1272026 (Canada)) are the basis of some of the systems studied and the help of Mr. J. Mallany in the preparation of this manuscript.

Literature Cited (1) Raaschou-Nielson, O.; Lohse, C.; Thomsen, B. L.; Skov, H.; Olsen, J. H. Environ. Res. 1997, 75, 149-159. (2) De Bortoli, M.; Kno¨ppel, H.; Peil, A.; Pecchio, E.; Schlitt, H.; De Wilde, H. In Indoor Air’90, Proceedings of the 5th International Conference on Indoor Air Quality and Climate, Toronto 1990; Canada Mortgage and Housing Corp: Ottawa, ON, pp 695700. (3) Jenkins, P. L.; Phillips, T. J.; Mulberg, E. J.; Hui, S. P. Atmos. Environ. 1992, 26A, 2141-2148. (4) Watanabe, M. E. Environ. Sci. Technol. 1997, 31, 182 A-186 A. (5) Shimp, J. F.; Tracy, J. C.; Davis, L. C.; Lee, E.; Huang, W.; Erickson, L. E.; Schnoor, J. L. Crit. Rev. Environ. Sci. Tec. 1993, 23, 41-47. (6) Darlington, A. B.; Dixon, M. A.; Arnold, K. A. SAE Technical Paper Ser. 1996, 961355. (7) Darlington, A.; Dixon, M. A.; Pilger, C. Life Support Biospheric Science 1998, 5, 63-69. (8) Darlington, A. B.; Dixon, M. A. SAE Technical Series Paper 1999, 1999-01-2069.

246

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001

(9) Kostiainen, R. Atmos. Environ. 1995, 29, 693-702. (10) Hwang, S.-J.; Tang, H.-M. J. Air Waste Manage. 1997, 47, 664673. (11) Lu, C.; Lin, M.-R.; Chu, C. J. Environ. Eng.-ASCE 1999, 125, 775779. (12) Sorial, G. A.; Smith, F. L.; Suidan, M. T.; Biswas, P.; Brenner, R. C. J. Hazard. Mater. 1997, 53, 19-33. (13) Darlington, A.; Chan, M.; Malloch, D.; Dixon, M. A. Indoor Air 2000, 10, 39-46. (14) Zilli, M.; Fabiano, B.; Ferraiolo, A.; Converti, A. Biotechnol. Bioeng. 1996, 49, 391-398. (15) Ottengraf, S. P. P. In Biotechnology: a comprehensive treatise in 8 volumes; Rehm, H. J., Reed, G., Eds.; VCH: Weinheim, Germany, 1986; Vol. 8, pp 425-452. (16) Kiared, K.; Fundenberger, B.; Brzezinski, R.; Viel, G.; Heitz, M. Ind. Eng. Chem. Res. 1997, 36, 4719-4725. (17) Hu, H.-Y.; Fujie, K.; Urano, K. J. Ferment. Bioeng. 1994, 78, 100104. (18) Zhang, W.; Bouwer, E.; Wilson, L.; Durant, N. Water Sci. Technol. 1995, 31, 1-14. (19) Arnold, M.; Reittu, A.; Wright, A. v.; Martikainen, P. J.; Suihko, M.-L. Appl. Microbiol. Biol. 1997, 48, 738-744. (20) Ashworth, R. A.; Howe, G. B.; Mullins, M. E.; Rogers, T. N. J. Hazard. Mater. 1988, 18, 25-36. (21) Pritchard, P. H.; Mueller, J. G.; Lantz, S. E.; Santavy, D. L. In Microbial Diversity and Ecological Function; Allsopp, D., Colwell, R. R., Hawksworth, D. L., Eds.; University Press: Cambridge, 1995; pp 161-184. (22) Asakawa, Y. In Bryophytes, their chemistry and chemical taxonomy; Zinsmeister, H. D., Mues, R., Eds.; Oxford, Clarendon Press: New York, Oxford University Press: New York, 1990; pp 369-410. (23) Hwang, S.-J.; Tang, H.-M.; Wen-Chang Environ. Prog. 1997, 16, 187-192.

Received for review February 28, 2000. Revised manuscript received October 13, 2000. Accepted October 16, 2000. ES0010507