BTOFTLTRATION FOR CONTROL OF VOLATILE ORGANIC COMPOUNDS CVOCS)
Dolloff F. Bishop
Risk Reduction Engineering Laboratory
U.S.Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
Department of Chemical Engineering
University of Cincinnati
Cincinnati, Ohio 45221
Air biofiltration is a promising technology for control of air emissions of biodegradable volatile organic
compounds (VOCs). In conjunction with vacuum extraction of soils or air stripping of ground water, it can be
used to mineralize VOCs removed from contaminated soil or groundwater. The literature (1) describes three
. major biological systems for treating contaminated air bioscrubbers, biotrickling filters and biofilters.
Bioscrubbers, which are not evaluated here, use counter current gas-liquid spray columns with microorganisms
freely suspended in the aqueous phase. Biofilters and biotrickling filters use microbial populations in biofilms
immobilized on support media to degrade or transform contaminants in air. Filter media can be classified as:
bioactive fine or irregular particulates, such as soil, peat, compost or mixtures of these materials; pelletized,
which are randomly packed in a bed; and structured, such as monoliths with defined or variable passage size
and geometry. The media can be made of sorbing and nonadsorbing materials. Nonbioactive pelletized and
structured media require recycled solutions of nutrients and buffer for efficient microbial activity and are thus
called biotrickling filters. Filters with bioactive fine or irregular particulates as media, referred to as biofilters,
usually do not recycle solutions of nutrients and buffers to prevent media compaction and gas channeling. All
filters humidify the contaminated air before biotreatment.
Soil biofilters are relatively large because soil pores are smaller and compounds have low permeability
in soil. They also have limited bed depths, required for maintaining humidity in soil and minimizing pressure
drop. Peat/compost biofilters, used commercially, are suitable for treating large volumes of air containing
biodegradable VOCs at low concentrations ( < 200 ppmv). However, both soil and peatkompost biofilters are
susceptible to channeling and maldistribution of air and require periodic media replacement.
Extensive work has been conducted to improve biofiltration by EPA’s Risk Reduction Engineering
Laboratory and the University of Cincinnati in biofilters using pelletized and structured media and improved
operational approaches. Representative VOCs in these studies included compounds with a range of aqueous
solubilities and octanol-water partition coefficients. The compounds include iso-pentane, toluene, methylene
chloride, trichloroethylene (TCE), ethyl benzene, chlorobenzene and perchloroethylene (PCE) and alpha (a-)
pinene. Comparative studies were conducted with peatlcompost bibfilters using isopentane and a-pinene.
Control studies were also conducted to investigate adsorption/desorption of contaminants on various media using
mercuric chloride solution to insure the absence of bioactivity,
MATERIALS AND METHODS
The typical experimental set-up of a biofilter bed (Figure 1) packed with support media was operated at
steady-state conditions to characterize process performance. Contaminated air, synthesized by injecting VOC
stock solutions into a controlled air stream, was passed through the biofilter bed. The inlet air was humidified
Figure 1. Typical experimental set-up of biofilter system.
before contamination. In the case of pelletized and structured media, solutions of nutrients and buffers were
introduced at the top of the bed and allowed to trickle countercurrent to the air flow through the media. The
solutions were collected at the bottom of the bed and recycled with appropriate additions for nitrogen
consumption and pH control. The water in the solutions insured effective wetting of the media.
Initially the pelletized and structured geometry (straight-passages) media filters were seeded with
acclimated biomass from an activated sludge process. Attachment or encapsulation of the seed on the various
media was successful. Garden variety peat humus (2204.6 gms) was thoroughly mixed with foam fluff (144.8
gms) providing a peat mixture with high void fraction. Similarly, compost (1358.5 gms) and foam fluff (145.5
gms) were mixed thoroughly for the compost column studies. Foam fluff was made by shredding polystyrene
foam into pieces with an average size of 2 mm. Control studies conducted with foam fluff showed that no
VOC adsorption occurred on the foam material. Contaminated air for all studies was prepared by injecting
appropriate VOC stock solutions into the humidified air flow controlled by a mass flow controller.
Gas samples to characterize biofilter performance were collected from the inlet and outlet ports of the
biofilter using gas tight syringes. Samples were collected in 250 mL glass sampling tubes by diverting the gas
flow through the tubes for 1 hour. The 250 microliter samples were injected into a gas chromatograph (Tracor
585 GC, with 703 PID 10.0 ev lamp, 1000 Hall ECD, Tekmar LSC 2000 purge and trap concentrator). Liquid
samples (5 mL) were collected in syringes with luer locks for purge and trap analysis.
In evaluating mineralization of the chlorinated VOCs, 100 mL samples of the effluent nutrient solution
were taken and mixed with 2 mL of 5 M sodium nitrate to adjust the ionic strength, and analyzed for chloride
ion with an Orion solid state combination electrode (9617 BN) in an Accumet 1003 pH/mV/ISE meter. The pH
of the effluent nutrient solutions was measured by the Accumet 1003 pH/mV/ISE meter. Carbon dioxide
concentrations in the inlet and outlet gas streams were determined in a Fisher 1200 gas partitioner.
The studies evaluated peat and compost biofilters, pelletized (activated carbon, ceramic and
encapsulated biomass) biofilters and structured geometry (ceramic and carbon coated ceramic straight-passages)
biofilters. Biofilter performances, expressed in terms of removal efficiencies of the compounds were calculated
from the amount of compound removed per unit time in the biofilter, expressed as a percentage of the amount
of that compound entering the biofilter per unit time.
Peat and compost biofilters:
Studies of peat and compost biofilters treating slightly soluble iso-pentane at high (360-960 ppmv)
concentrations in humidified air were completed using the indigenous microorganisms and nutrients of each
media to establish bioactivity. Initial abiotic control tests on peat and compost revealed poor adsorption of iso-
pentane. The effluent air content of iso-pentane became equal to the influent concentration (350 ppmv) in about
Pseudo-equilibrium removal efficiencies were established for iso-pentane by operating at selected air
retention times (2 minutes to 13 minutes) until the increasing removal efficiencies reached steady removal
values. Equilibrium removal values, achieved in one to three weeks of operation at each retention time,
increased with decreasing influent iso-pentane concentration and increasing gas residence time. The compost
biofilter exhibited substantially higher removal efficiencies at the 360 ppmv iso-pentane concentration than the
peat biofilter. At higher influent concentrations, the two materials exhibited roughly similar performances.
Water content in peat and compost effected removal efficiencies. Performance studies, with varying
water content in the media, revealed critical water contents of 0.48 gm of waterlgm of peat and 0.58 gm of
watedgm of compost, below which biofilter performance deteriorated substantially. The loss in removal
efficiencies below the critical water content were caused by irreversible shrinkage of the media permitting gas
by-passing. Subsequent addition of water did not expand the media nor eliminate the cracks and voids.
Removal of iso-pentane exhibited optimal removal efficiencies at 0.56 gm of waterlgm of peat and 0.67 gm of
water/gm of compost. As the water content increased above the optimal value, removal efficiencies gradually
decreased for both media. These decreasing efficiencies probably resulted from partial filling of bed void
fraction with free water and increased mass transfer difficulties for the slightly soluble iso-pentane. The losses
of efficiency, above and below the optimal water content, illustrated the need for maintaining appropriate media
water content through effective humidification of influent air entering peat and compost biofilters.
Studies of temperature effects on removal efficiency in fully humidified air revealed, for both media,
increasing efficiencies with increasing temperatures and maximum removal at 35°C. Below 25"C, the removal
efficiencies decreased nearly linearly with temperature. Thus for influent air temperatures below 25"C,
substantial improvement in biofilter performance could be achieved by heating the air to increase bed
A study using a-pinene was also conducted in a compost biofilter to assess the effects of pH on media
with limited buffering capacity. Monounsaturated a-pinene, a major contaminant in press vent and dryer gas, is
a slowly degradable complex bridged ring compound (2) and biodegrades through a variety of pathways,
producing organic acids and neutral compounds that accumulate as intermediate products. The study revealed
that a compost biofilter with an air retention time of 3.5 minutes supported nearly complete removals and
mineralization of the a-pinene at an inlet concentration of 25 ppmv. At inlet concentrations of 50 ppmv and
above, however, the removal efficiencies and mineralization sharply decreased. Tests on mixtures of the
compost and water revealed that organic acids were accumulating in the compost bed, limiting degradation and
indicating a need for additional buffer capacity in the compost for inlet a-pinene concentrations above 25 ppmv.
Biofilters with pelletimd media
Two biofilters, one packed with 3 mm activated carbon pellets (void fraction 0.35), the other with 6
mm porous ceramic (celite) pellets (void fraction 0.40) were studied to characterize performance of biofilters
randomly packed with pellets and using nutrient recycle. The research also revealed performance differences
and similarities between adsorptive and non-adsorptive media. The activated carbon biofilter, operated at 2
minutes of gas retention time and a nutrient solution loading of 1000 L/mzday, initially revealed high adsorptive
removals of the inlet contaminants; toluene (520 ppmv), methylene chloride (180 ppmv), and TCE (25 ppmv)
followed by a rapid decrease in contaminant removal as the wetted carbon became saturated with Contaminants.
In this first biofilter study, the initial buffered nutrient solution had insufficient ammonia which limited biomass
growth and substantially prolonged start-up time by delaying biomass growth. Indeed, the carbon media
fortuitously became saturated with adsorbed VOC contaminants, especially with toluene. Subsequent increase of
ammonia in the nutrient solution accelerated biofilm development. Removal efficiencies increased to nearly
100%for all three contaminants. The buffered nutrient also maintained the liquid phase pH above 6.2. The
activated carbon biofilter continued to effectively remove (>99%) of the contaminants for an additional 3
months before excessive pressure losses and flooding occurred.
The start-up of the ceramic celite biofilter with appropriate nutrients at a nutrient solution loading of
250L/m2day and a 2 minute air retention time revealed gradual increases in contaminant removals as the seeded
biomass grew on the celite pellets. After about 40 days, sufficient biofilm growth occurred to substantially
degrade ( > 9 5 % ) four of the five inlet contaminants; toluene (450 ppmv), methylene chloride (150 ppmv), ethyl
benzene (25 ppmv) and chlorobenzene (40 ppmv). In contrast to the performance of the activated carbon
biofilter, the fifth inlet contaminant, TCE (25 ppmv) was only partially removed by the bioactivity in the celite
Folsom et al. (3) showed that TCE is usually degraded only as a secondary substrate in the presence of
a primary metabolite, such as toluene or phenol. Examination of contaminant concentration profiles in the celite
biofilter revealed that toluene was rapidly removed in the bottom third of the biofilter. As a result toluene was
not available as a primary metabolite in the upper two-thirds of the filter. Thus the slowly degrading TCE was
degraded only in the bottom third of the column.
In the activated carbon biofilter, toluene should also rapidly degrade, probably in the bottom third of
the biofilter. Apparently the adsorbed toluene from the prolonged start-up of the activated carbon biofilter acted
as a toluene reservoir desorbing toluene to the attached biofilm on the carbon over the full height of the biofilter
bed. The desorbing toluene produced complete degradation of the TCE at the 2 minute retention time in
activated carbon bed but was unavailable over full height of the biofilm in the nonadsorbing celite biofilter, thus
limiting TCE degradation.
The celite biofilter with larger void space in the 6mm bed continued to substantially remove all of the
contaminants except TCE for about 210 days of operation after full acclimation. The biofilter then exhibited
increasing pressure losses and flooding similar to that which occurred in the activated carbon biofilter. The
biofilm plugging the biofilter, however, adhered less strongly to the celite pellets than to the carbon pellets and
was easier to remove. Studies on the degree of mineralization of the contaminants due to biodegradation were
also conducted on both pelletized media. Abiotic control columns, identical to the biofilter beds but without
biomass seeding, revealed breakthrough of the contaminants typical for adsorptive and non-adsorptive beds.
The amounts of contaminant absorbed in the small volume of nutrient solution were negligible compared to that
in the inlet air. There were no increases in carbon dioxide concentration in the air stream.
In the carbon and celite biofilters, carbon dioxide increases in the air stream during treatment of
contaminated air ultimately revealed nearly complete mineralization of the removed contaminants. Initially, a
portion of the contaminants removed by the biofilters were converted to biomass. With increasing biofilm
buildup and negligible releases of biomass, carbon dioxide increases in the effluent air approached increases
expected from full mineralization of the removed contaminants. This indicated that the biomass decay rates
began to approach the biomass growth rates in both biofilters. Chloride ion increases in the effluent nutrient
solution also revealed complete conversion of organic chlorine in the contaminants to chloride ion. These
results along with the absence of partially degraded VOC products in the effluent air, verified by gas
chromatography measurements, confirmed efficient mineralization of the removed VOCs.
A comparative study of a-pinene biodegradation was also conducted in a celite biofilter with trickling
nutrient and buffer solution. In contrast to the compost biofilter, the results revealed that nearly complete
removal and mineralization occurred at biofilter air retention times of 4 minutes for inlet concentrations of a-
pinene of 25, 50, and 75 ppmv. Reduction of the air retention time to 1.5 minutes at an inlet a-pinene
concentration of 75 ppmv did not significantly reduce celite biofilter performance.
PCE, a VOC recalcitrant to aerobic biodegradation, and TCE are easily dechlorinated under anaerobic
conditions. The partially dechlorinated products are also easily degraded under aerobic conditions. Thus a
nonbiodegradable hydrogel pellet (3 by 6mm cylinder), encapsulating biomass from an activated sludge process,
was developed to retard oxygen transport, producing an anaerobic core and an aerobic outer zone for
synchronous anaerobic/aerobic treatment of contaminated air. The pellet included an outer stainless steel mesh
for structural stability. Biofilter tests of the pellets with trickling nutrient and buffer solution revealed complete
mineralization of inlet TCE (21 ppmv) and PCE (20 ppmv) at air retention times of 1.5 minutes and 4 minute,
respectively. The hydrogel pellets exhibited long term stability.
Structured Geometry (Straight-Passages) Biofilters
Biofilter studies on straight-passage media with trickling nutrient and buffer solutions at typical loadings
of 150 L/m*/day included a celite plate biofilter and extruded cordierite and carbon coated cordierite biofilters.
The celite plate biofilter constructed of thick-walled plates with low surface area per unit volume (10 m2/m3),
required relatively high air retention times (15 minutes) for nearly complete biodegradation of four of the five
inlet contaminants; toluene (450 ppmv), methylene chloride (150 ppmv), ethylbenzene (25 ppmv), and
chlorobenzene (40 ppmv). As in the biofilter with celite pellets, the fifth contaminant, TCE (25 ppmv), was
only partially removed (35 %) because of lack of a cometabolite in the upper region of the biofilter. Increasing
the inlet toluene concentration in air to 850 ppmv and subsequently increasing the air flow rate in an attempt to
increase the availability of cometabolite produced little impact on TCE removal. Addition of phenol to the
nutrient solution, as an alternative TCE cometabolite available over the entire bed height, however, increased
the TCE removal, ultimately to nearly 100% removal.
Biomass release from straight-passages celite media was observed in the effluent nutrient solution. The
amount of released biomass depended upon the nutrient flow rate and biomass loading. In addition, during 240
days of operation, the straight-passage media in the bench-scale filter never exhibited pressure drop buildup.
These results suggested possible self-cleaning or easier cleaning for straight-passages media compared to the
extensive cleaning required for pelletized media.
Biofilter studies on isopentane using extruded (straight-passages) cordierite and carbon and activated
carbon coated (0.1 mm films) cordierite media, with high surface area per unit volume (80m2/m3)and high void
fraction (80%), were evaluated at isopentane concentrations varying from 360 to 960 ppmv in the inlet air to
compare their removal efficiencies with those of peat and compost biofilters. The removal efficiencies of
slightly soluble isopentane in the cordierite biofilter, as a function of contaminant concentration and air retention
time, were similar to those observed in peat and compost biofilters. The removal efficiencies in the cordierite
biofilter were limited by poor adhesion of the biofilm on the vertical walls. The removal efficiencies in the
carbon coated and activated carbon coated cordierite biofilter, however, were substantially better than observed
in peat, compost or uncoated cordierite media. The average biodegradation rates of isopentane in activated
carbon coated cordierite biofilters were also significantly greater than those in the other biofilters, substantially
reducing biofilter size.
Media surface effects on start-up and dynamic response to sudden contaminant concentration changes
were examined for straight-passages cordierite media without coatings and with coatings of resin, carbon (0.1
mm) and activated carbon (0.1 mm) as a function of time. Biofilter start-up with carbon coated cordierite was
faster than uncoated and resin coated cordierite media. Start-up for activated carbon coated media was fastest.
High adsorption capacity and strong adhesion of the activated carbon coating accelerated the start-up process.
Similarly, dynamic responses to sudden contaminant concentration changes were also fastest on the cordierite
media coated with activated carbon.
Kinetic experiments on aerobic degradation were conducted for various VOC on a range of media
types. Monod kinetics satisfactorily described the biodegradation of the tested VOCs and led to development of
design models for each type of biofilter media. The kinetic results on isopentane revealed that activated carbon
coated or activated carbon media provided best overall performance for slightly soluble VOCs. For soluble and
moderately soluble VOCs, non adsorbing pelletized and structured media with good biofilm adhesion provided
removals and kinetics equivalent to those obtained in biofilters with activated carbon or carbon coated media.
Biofilters using porous pellet or structured (straight-passages) media and recycled nutrient and buffer
solutions have potential advantages over conventional soil or peat/compost biofilters for removal of
biodegradable VOCs from contaminated air. These include:
0 improved distribution of air flow and moisture control in the media with lower operating pressure
losses and improved process Performance;
0 improved neutralization of acid degradation products (pH control);
e increased capacity for efficiently treating higher VOC loading (up to 800 ppmv); and
0 capability for removal of excess biomass from the biofilter, thus preventing clogging and eventual
Biofilters with porous pellet media, however, require development of effective cleaning methods for
field-scale applications. Biofilters with structured media (straight-passages) offer easier cleaning options,
including release of biomass (self-cleaning) into the recycling nutrient and buffer solution. Biofiltration with
activated-carbon-coated or activated carbon pelletized or structured media is an effective technology for control
of slightly soluble VOCs. Novel media supporting anaerobiclaerobic treatment extend the range o application
of biofiltration to VOCs recalcitrant to aerobic biodegradation.
1. Ottengraf, S.P.P. and R. Diks, "Biological Purification of Waste Gases," Chim Ogei 8 (5), 41 (1990).
2. Gibbon, G.H., Millis, N.F., and Pirt, S.J. "Degradation of @-pineneby Bacteria," Procs IV IFS:
Forment. Technol. Today, 609-612, (1972).
3. Folsom, B.R., Chapman, P.J., and Pritchard, P.H. "Phenol and Trichloroethylene Degradation by
Pseudomonas Cepacia G4: Kinetics and Interactions Between Substrates. " And. Environ. Microbiol.,
56; 1279, (1990).
FOR MORE INFORMATION:
Dolloff F. Bishop
Risk Reduction Engineering Laboratory
U.S.Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, OH 45268