Biological elimination of H2S and NH3 from wastegases by biofilter packed with immobilized heterotrophic bacteria

Biotreatment of various ratios of H₂S and NH₃ gas mixtures was studied using the biofilters, packed with co-immobilized cells (Arthrobacter oxydans CH8 for NH₃ and Pseudomonas putida CH11 for H₂S). Extensive tests to determine removal characteristics, removal efficiency, removal kinetics, and pressure drops of the biofilters were performed. To estimate the largest allowable inlet concentration, a prediction model was also employed. Greater than 95% and 90% removal efficiencies were observed for NH₃ and H₂S, respectively, irrespective of the ratios of H₂S and NH₃ gas mixtures. The results showed that H₂S removal of the biofilter was significantly affected by high inlet concentrations of H₂S and NH₃. As high H₂S concentration was an inhibitory substrate for the growth of heterotrophic sulfur-oxidizing bacteria, the activity of H₂S oxidation was thus inhibited. In the case of high NH₃ concentration, the poor H₂S removal efficiency might be attributed to the acidification of the biofilter. The phenomenon was caused by acidic metabolite accumulation of NH₃. Through kinetic analysis, the presence of NH₃ did not hinder the NH₃ removal, but a high H₂S concentration would result in low removal efficiency. Conversely, H₂S of adequate concentrations would favor the removal of incoming NH₃. The results also indicated that maximum inlet concentrations (model-estimated) agreed well with the experimental values for space velocities of 50–150 h⁻¹. Hence, the results would be used as the guideline for the design and operation of biofilters.     

Biotreatment of H2S- and NH3-containing waste gases by co-immobilized cells biofilter

Gas mixture of H2S and NH3 in this study has been the focus in the research area concerning gases generated from the animal husbandry and the anaerobic wastewater lagoons used for their treatment. A specific microflora (mixture of Thiobacillus thioparus CH11 for H2S and Nitrosomonas europaea for NH3) was immobilized with Ca-alginate and packed inside a glass column to decompose H2S and NH3. The biofilter packed with co-immobilized cells was continuously supplied with H2S and NH3 gas mixtures of various ratios, and the removal efficiency, removal kinetics, and pressure drop in the biofilter was monitored. The results showed that the efficiency remained above 95% regardless of the ratios of H2S and NH3 used. The NH3 concentration has little effect on H2S removal efficiency, however, both high NH3 and H2S concentrations significantly suppress the NH3 removal. Through product analysis, we found that controlling the inlet ratio of the H2S/NH3 could prevent the biofilter from acidification, and, therefore, enhance the operational stability. Conclusions from bioaerosol analysis and pressure drop in the biofilter suggest that the immobilized cell technique creates less environmental impact and improves pure culture operational stability. The criteria for the biofilter operation to meet the current H2S and NH3 emission standards were also established. To reach Taiwan's current ambient air standards of H2S and NH3 (0.1 and 1 ppm, respectively), the maximum inlet concentrations should not exceed 58 ppm for H2S and 164 ppm for NH3, and the residence time be kept at 72 s.     

A comparative study in treating two VOC mixtures in trickle bed air biofilters

Two independent parallel trickling bed air biofilters (TBABs) ("A" and "B") with two different typical VOC mixtures were investigated. Toluene, styrene, methyl ethyl ketone (MEK), and methyl isobutyl ketone (MIBK) were the target VOCs in the mixtures. Biofilter "A" was fed equal molar ratio of the VOCs and biofilter "B" was fed a mixture based on EPA 2003 emission report. Backwashing and substrate starvation operation were conducted as biomass control. Biofilter "A" and "B" maintained 99% overall removal efficiency for influent concentration up to 500 and 300 ppmv under backwashing operating condition, respectively. The starvation study indicated that it can be an effective biomass control for influent concentrations up to 250 ppmv for biofilter "A" and 300 ppmv for "B". Re-acclimation of biofilter performance was delayed with increase of influent concentration for both biofilters. Starvation operation helped the biofilter to recover at low concentrations and delayed re-acclimation at high concentrations. Furthermore, re-acclamation for biofilter "B" was delayed due to its high toluene content as compared to biofilter "A". The pseudo first-order removal rate constant decreased with increase of volumetric loading rate for both biofilters. MEK and MIBK were completely removed in the upper 3/8 media depth. While biofilter depth utilization for the removal of styrene and toluene increased with increase of influent concentrations for both biofilters. However, toluene removal utilized more biofilter depth for biofilter "B" as compared to biofilter "A".     

Long-term operation of biofilters for biological removal of ammonia

Biological removal of ammonia was investigated using two types of packing materials, compost and sludge in laboratory-scale biofilters (8l reactor volume). The aim of this study is to investigate the potential of unit systems packed with these supports in terms of ammonia emissions treatment. Experimental tests and measurements included analysis of removal efficiency, metabolic products, and results of long-term operation. The inlet concentration of ammonia applied was 20-200 mg m-3. The ammonia loading rates of 24.9-566 g NH3 m-3 d-1 to compost biofilter (BF3) and 24.9-472 g NH3 m-3 d-1 to sludge biofilter (BF4) were applied for 210 days, respectively. Removal efficiencies of the compost and sludge biofilters were in the range of 97-99% and 95-99%, respectively when the inlet concentration of ammonia was below 110 mg m-3, and the maximum elimination capacities were 288 and 243 g NH3m-3d-1, respectively. However, removal efficiency and elimination capacity of both biofilters significantly decreased as the inlet concentration increased to above 110 mg m-3. By using kinetic analysis, the maximum removal rate of ammonia, Vm, and the saturation constant, Ks, were determined for both packing materials and the value of Vm for compost was found to be larger. Periodic analysis of the biofilter packing materials showed the accumulation of the nitrification product NO3- in the operation. During the experiment, the pressure drops measured were very low. The use of both packing materials requires neither nutritive aqueous solution nor buffer solution.     

Removal of low concentrations of carbon tetrachloride in compost-based biofilters operated under methanogenic conditions.

Research was performed to demonstrate the removal of carbon tetrachloride (CT) using compost biofilters operated under methanogenic conditions. Biofilters were operated at an empty-bed residence time of 2.8 minutes using nitrogen as the atmosphere. Hydrogen and carbon dioxide were supplied as an electron donor and carbon source, respectively, during acclimation of the bed medium microbes. Once methanogenesis was demonstrated, CT flow to the biofilter was established. Biofilters were operated over a CT concentration range from 20 to 700 ppbv for 6 months. Bed medium microbes were able to remove up to 75% of the inlet CT. At excessively high CT concentrations (> 500 ppmv), methane production and hydrogen utilization by the bed medium microbes appeared to be inhibited. CT removal by the biofilter decreased when the hydrogen supply was removed from the biofilter inlet, indicating that hydrogen acted as the electron donor for reductive dechlorination. The removal efficiency and relatively low empty bed residence times demonstrated by these laboratory-scale biofilters indicate that anaerobic biofiltration of CT may be a feasible full-scale process.     

Backfitting Electrostatic Precipitators: Gas Velocity and Reliability Considerations

Abstract

The kinetic behavior of the toluene biofiltration process was investigated in this research. Toluene was used as a model compound for less water-soluble gas pollutants. The limiting factor in the overall toluene biofiltration process was determined by analyzing the effectiveness factor of the biofilm along the biofilter. Experiments were conducted in three laboratory-scale biofilters packed with mixtures of chaff/compost, D.E. (diatomaceous earth)/compost and GAC (granular activated carbon)/compost, respectively. A mathematical model previously proposed was verified in this study as being applicable to these biofilters packed with different filter materials. Both the experimental and theoretical results confirmed that the biodegradation rate along the biofilter followed the zero order, fractional order to first order kinetics as toluene concentration decreased. Moreover, at higher toluene concentration, biodegradation rate and mass flux of toluene were lower near the bottom of the biofilter due to substrate inhibition. Analysis of the effectiveness factor indicated that biofiltration of a less soluble compound such as toluene should not be operated at high gas flow rates (low gas residence times) due to the mass transfer limitation of such a system. At an approximate constant inlet toluene concentration of 0.9 g/m³, the toluene removal efficiency in these three biofilters would drop below 90% when the gas residence time decreased to 2.5, 2.5, and 2.0 min, respectively.     

Removal of Low Concentrations of Carbon Tetrachloride in Compost-Based Biofilters Operated under Methanogenic Conditions

ABSTRACT

Research was performed to demonstrate the removal of carbon tetrachloride (CT) using compost biofilters operated under methanogenic conditions. Biofilters were operated at an empty-bed residence time of 2.8 minutes using nitrogen as the atmosphere. Hydrogen and carbon dioxide were supplied as an electron donor and carbon source, respectively, during acclimation of the bed medium microbes. Once methanogenesis was demonstrated, CT flow to the biofilter was established. Biofilters were operated over a CT concentration range from 20 to 700 ppbv for 6 months. Bed medium microbes were able to remove up to 75% of the inlet CT. At excessively high CT concentrations (>500 ppmv), methane production and hydrogen utilization by the bed medium microbes appeared to be inhibited. CT removal by the biofilter decreased when the hydrogen supply was removed from the biofilter inlet, indicating that hydrogen acted as the electron donor for reductive dechlorination. The removal efficiency and relatively low empty bed residence times demonstrated by these laboratory-scale biofilters indicate that anaerobic biofiltration of CT may be a feasible full-scale process.     

Methane biofiltration in the presence of ethanol vapor under steady and transient state conditions: an experimental study

Methane (CH₄) removal in the presence of ethanol vapors was performed by a stone-based bed and a hybrid packing biofilter in parallel. In the absence of ethanol, a methane removal efficiency of 55 ± 1% was obtained for both biofilters under similar CH₄ inlet load (IL) of 13 ± 0.5 g_CH4 m^?3 h^?1 and an empty bed residence time (EBRT) of 6 min. The results proved the key role of the bottom section in both biofilters for simultaneous removal of CH₄ and ethanol. Ethanol vapor was completely eliminated in the bottom sections for an ethanol IL variation between 1 and 11 g_ethanol m^?3 h^?1. Ethanol absorption and accumulation in the biofilm phase as well as ethanol conversion to CO₂ contributed to ethanol removal efficiency of 100%. In the presence of ethanol vapor, CO₂ productions in the bottom section increased almost fourfold in both biofilters. The ethanol concentration in the leachate of the biofilter exceeding 2200 g_ethanol m^?3 _leachate in both biofilters demonstrated the excess accumulation of ethanol in the biofilm phase. The biofilters responded quickly to an ethanol shock load followed by a starvation with 20% decrease of their performance. The return to normal operations in both biofilters after the transient conditions took less than 5 days. Unlike the hybrid packing biofilter, excess pressure drop (up to 1.9 cmH₂O m^?1) was an important concern for the stone bed biofilter. The biomass accumulation in the bottom section of the stone bed biofilter contributed to 80% of the total pressure drop. However, the 14-day starvation reduced the pressure drop to 0.25 cmH₂O m^?1.     

Biofiltration of Cyclic Air Emissions of [H9251]-Pinene at Low and High Frequencies

Abstract

Biofiltration of periodically fluctuating concentrations of an α-pinene-laden waste gas was investigated to treat both high-frequency and low-frequency fluctuations. The effects of periodic concentration fluctuations on biofilter performance were measured. Controlled variables of periodic operation included cycle period and amplitude. The cycle period ranged from 10 min to 6 days, with the inlet α-pinene concentration fluctuating between 0 and 100 parts per million volume. At high-frequency concentration cycling (i.e., on the order of minutes), both cyclic and constant concentration biofilters maintained similar long-term performance with an average removal efficiency of 77% at an averaged loading rate of 29 g α-pinene/m³ bed/hr. A first approximation suggests kinetics that are time-independent, indicating that steady-state data can be used to predict transient behavior at this time scale. Cyclic biofilter operation with a cycle period of 24 hr (with equal on/off time) was achievable for biofilters without a significant loss in performance. At longer time scales, cyclic biofilter performance decreased at the restart of the ON cycle. The recovery time to previous levels of performance increased with increasing cycle period; the recovery time was less than 1 hr for a cycle period of 24 hr and between 6 and 8 hr for a cycle period of 6 days.     

Decomposition of chlorinated dioxins, odourous compounds and NOx from MSW incineraton plant by oxidizing catalyst

An experiment was carried out to decompose chlorinated dioxins (PCDDs, PCDFs) Chlorobenzenes, NOx and odourous compounds (H₂S, CH₄S, C₂H₆S₂, C₈H₈, C₂H₆S, C₂H₄O, NH₃) simultaneously using a catalyst in the MSW incineration plant. The experiments were conducted at temperatures from 200°C to 400°C and from 3000h⁻¹ to 6000h⁻¹ at space velocity. A catalyst containing V₂O₅ and WO₃ on the basis of TiO₂ is used, an oxidizing catalyst of the honeycomb type. The average decomposition efficiencis were 95%, 98%, 92% for PCDDs(48CDDs), PCDFs(48CDFs) and Chlorobenzenes(36CLBs) at a reaction temperature of 350°C and a space velocity of 3000h⁻¹, more than 90% for NOx at a reactiont temperature of 300°C and more than 80% for odourous compounds at the reaction temperature of 300°C and a space velocity of 6000h⁻¹. All those compounds were decomposed successfully with increasing contact time and surface. The rate-determing step was the chemical reaction of catalyst surface.     

管式生物过滤器去除乙苯废气

生物过滤由于其良好的成本效益和环境友好性已经成为控制挥发性有机化合物(VOCs)含量和气味气体排放的常规技术。营养物质的均匀分布、生物膜和介质床内的气体流是成就一个性能优良的生物过滤器至关重要的因素。而由本实验室开发的管式生物过滤器(TBFs)已被证明具备此优势。本实验的管式生物过滤器以聚氨酯海绵作为填料,研究在不同有机负荷、气体停留时间(EBCT)、进气量和表面活性剂等条件下乙苯废气的去除效率(RE)。实验同时记录了管式生物过滤器启动阶段的表现。初期使附着在填料上的微生物暴露在浓度为40 mg/m3的乙苯废气中40 d,此时的气体停留时间为15 s,使微生物慢慢适应并逐步降解乙苯废气;然后连续地控制管式生物过滤器的入口乙苯浓度为40、80、120和160 mg/m3,以使有机负荷逐步升高。结果表明,乙苯去除效率随着有机负荷的增大而逐步减小。当气体停留时间从15 s增加到30 s和60 s,而有机负荷控制在38.60 g/(m3·h)时,乙苯废气去除效率略微增加。此外,随着进气量的增大乙苯废气的最大平均去除效率有所下降而此时的降解容量增大,这个过程中乙苯进气浓度保持不变。结果还表明,在营养液中加入聚乙二醇辛基苯基醚这种表面活性剂可以提高乙苯废气的去除效率。     

Photochemical oxidation of reduced sulfur compounds in an urban location based on short time monitoring data

This study examines the oxidation of reduced sulfur compounds (RSCs) in urban ambient air. The photochemical conversions of RSC (such as DMS, CS₂, H₂S, DMDS, and CH₃SH) to a further oxidized form (e.g., SO₂, MSA, and H₂SO₄) were assessed using a photochemical box model. For our model simulation of RSC oxidation, measurements were taken at an urban monitoring station in Seoul, Korea (37.6° N, 127° E) during three separate time periods (e.g., Sept. 17–18, Oct. 23, and Oct. 27–28, 2003). The results indicate that DMS and H₂S were the dominant RSCs with concentrations of 370 ± 140 and 110 ± 60 pptv, respectively. The photochemical conversion of DMDS to SO₂ was found to occur more efficiently than other RSCs. The overall results of our study suggest that photochemical conversion of RSCs accounted for less than 15% of the observed SO₂ during the measurement period. The SO₂ production from DMS oxidation (mainly by the reaction with OH) was found to be affected primarily by the abstraction channel due to high NO_x levels during the experimental conditions.     

Biological removal of gaseous ammonia in biofilters: space travel and earth-based applications

Gaseous NH3 removal was studied in laboratory-scale biofilters (14-L reactor volume) containing perlite inoculated with a nitrifying enrichment culture. These biofilters received 6 L/min of airflow with inlet NH3 concentrations of 20 or 50 ppm, and removed more than 99.99% of the NH3 for the period of operation (101, 102 days). Comparison between an active reactor and an autoclaved control indicated that NH3 removal resulted from nitrification directly, as well as from enhanced absorption resulting from acidity produced by nitrification. Spatial distribution studies (20 ppm only) after 8 days of operation showed that nearly 95% of the NH3 could be accounted for in the lower 25% of the biofilter matrix, proximate to the port of entry. Periodic analysis of the biofilter material (20 and 50 ppm) showed accumulation of the nitrification product NO3- early in the operation, but later both NO2- and NO3- accumulated. Additionally, the N-mass balance accountability dropped from near 100% early in the experiments to approximately 95 and 75% for the 20- and 50-ppm biofilters, respectively. A partial contributing factor to this drop in mass balance accountability was the production of NO and N2O, which were detected in the biofilter exhaust.     

Biofiltration of cyclic air emissions of alpha-pinene at low and high frequencies

Biofiltration of periodically fluctuating concentrations of an alpha-pinene-laden waste gas was investigated to treat both high-frequency and low-frequency fluctuations. The effects of periodic concentration fluctuations on biofilter performance were measured. Controlled variables of periodic operation included cycle period and amplitude. The cycle period ranged from 10 min to 6 days, with the inlet alpha-pinene concentration fluctuating between 0 and 100 parts per million volume. At high-frequency concentration cycling (i.e., on the order of minutes), both cyclic and constant concentration biofilters maintained similar long-term performance with an average removal efficiency of 77% at an averaged loading rate of 29 g alpha-pinene/m3 bed/hr. A first approximation suggests kinetics that are time-independent, indicating that steady-state data can be used to predict transient behavior at this time scale. Cyclic biofilter operation with a cycle period of 24 hr (with equal on/off time) was achievable for biofilters without a significant loss in performance. At longer time scales, cyclic biofilter performance decreased at the restart of the ON cycle. The recovery time to previous levels of performance increased with increasing cycle period; the recovery time was less than 1 hr for a cycle period of 24 hr and between 6 and 8 hr for a cycle period of 6 days.     

Photocatalytic bacterial inactivation by polyoxometalates

The photocatalytic inactivation (PCI) of Escherichia coli (Gram-negative) and Bacillus subtilis (Gram-positive) was performed using polyoxometalate (POM) as a homogeneous photocatalyst and compared with that of heterogeneous TiO₂ photocatalyst. Aqueous suspensions of the microorganisms (10⁷–10⁸ cfu ml⁻¹) and POM (or TiO₂) were irradiated with black light lamps. The POM-PCI was faster than (or comparable to) TiO₂-PCI under the experimental conditions employed in this study. The relative efficiency of POM-PCI was species-dependent. Among three POMs (H₃PW₁₂O₄₀, H₃PMo₁₂O₄₀, and H₄SiW₁₂O₄₀) tested in this study, the inactivation of E. coli was fastest with H₄SiW₁₂O₄₀ while that of B. subtilis was the most efficient with H₃PW₁₂O₄₀. Although the biocidal action of TiO₂ photocatalyst has been commonly ascribed to the role of photogenerated reactive oxygen species such as hydroxyl radicals and superoxides, the cell death mechanism with POM seems to be different from TiO₂-PCI. While TiO₂ caused the cell membrane disruption, POM did not induce the cell lysis. When methanol was added to the POM solution, not only the PCI of E. coli was enhanced (contrary to the case of TiO₂-PCI) but also the dark inactivation was observed. This was ascribed to the in situ production of formaldehyde from the oxidation of methanol. The interesting biocidal property of POM photocatalyst might be utilized as a potential disinfectant technology.     

Sequential photochemical-biological degradation of chlorophenols

UV/TiO₂/H₂O₂, UV/TiO₂ and UV/H₂O₂ were compared as pre-treatment processes for the detoxification of mixtures of 4-chlorophenol (4CP), 2,4-dichlorophenol (DCP), 2,4,6-trichlorophenol (TCP) and pentachlorophenol (PCP) prior to their biological treatment. When each chlorophenol was initially supplied at 50 mg l⁻¹, UV/TiO₂/H₂O₂ treatment supported the highest pollutant removal, COD removal, and dechlorination efficiencies followed by UV/TiO₂ and UV/H₂O₂. The remaining toxicity to Lipedium sativum was similar after all pre-treatments. Chlorophenol photodegradation was always well described by a first order model kinetic (r² > 0.94) and the shortest 4CP, DCP, TCP and PCP half-lives of 8.7, 7.1, 4.5 and 3.3 h, respectively, were achieved during UV/TiO₂/H₂O₂ treatment. No pollutant removal was observed in the controls conducted with H₂O₂ or TiO₂ only. Inoculation of all the photochemically pre-treated mixtures with activated sludge microflora was followed by complete removal of the remaining pollutants. Combined UV/TiO₂/H₂O₂-biological supported the highest detoxification, dechlorination (99%) and COD removal (88%) efficiencies. Similar results were achieved when each chlorophenol was supplied at 100 mg l⁻¹. COD and Cl mass balances indicated UV, UV/H₂O₂, and UV/TiO₂ treatments lead to the formation of recalcitrant photoproducts, some of which were chlorinated.     

Effects of nitrogen and oxygen on biofilter performance

Three laboratory-scale biofilters packed with inert material were used to study the nitrogen and oxygen requirements for biofiltration of methanol. Mixtures of methanol with inorganic nitrogen (NH3 or NO3) at nitrogen-to-carbon (N:C) ratios ranging from 0.015 to 0.4 were employed to reveal nitrogen effects on biofiltration. In the oxygen study, mixtures of air and oxygen at different oxygen contents were used. At low nitrogen levels, the removal rate increased with increasing N:C ratio for both NH3 and NO3. However, at high concentrations, NH3 had an inhibitory effect on biodegradation while the removal rate reached a plateau at high NO3 concentrations. Biofiltration with 63% oxygen in the inlet gas stream increased the maximum removal rate from 120 to 145 g/m3/hr after 3 days in comparison with biofiltration with air. However, a further increase in oxygen content up to 80% did not lead to a further improvement in biofilter performance, suggesting that both oxygen and biofilm thickness can be the relevant factors limiting biofilter performance and creating the plateau in removal rates at high loadings.     

Design and performance characterization strategy using modeling for biofiltration control of odorous hydrogen sulfide

Biofilter, dynamic modeling software characterizing contaminant removal via biofiltration, was used in the preliminary design of a biofilter to treat odorous hydrogen sulfide (H2S). Steady-state model simulations were run to generate performance plots for various influent concentrations, loadings, residence times, media sizes, and temperatures. Although elimination capacity and removal efficiency frequently are used to characterize biofilter performance, effluent concentration can be used to characterize performance when treating to a target effluent concentration. Model simulations illustrate that, at a given temperature, a biofilter cannot reduce H2S emissions below a minimum value, no matter how large the biofilter or how long the residence time. However, a higher biofilter temperature results in lower effluent H2S concentrations. Because dynamic model simulations show that shock loading can significantly increase the effluent concentration above values predicted by the steady-state model simulations, it is recommended that, to consistently meet treatment objectives, dynamic feed conditions should be considered. This study illustrates that modeling can serve as a valuable tool in the design and performance optimization of biofilters.     

A combined plasma photolysis (CPP) method for removal of CS2 from gas streams at atmospheric pressure

A combined plasma photolysis (CPP) reactor that utilizes a dielectric barrier discharge (DBD) plasma and 207 nm UV radiation from discharge-driven KrBr^* excimers was designed and constructed. Gas streams containing CS₂ were treated with stand-alone DBD and CPP at atmospheric pressure. In comparison to DBD, CPP greatly enhanced the removal efficiency at the same applied voltage, waste gas concentration and gas residence time. Thus the applied voltage could be reduced to a certain extent in the plasma processing of industrial wastes. Influences of the KrBr^* radiation, inlet CS₂ concentration and gas residence time on CS₂ removal by CPP were also studied. In addition, the likely reaction mechanisms for the removal of CS₂ by CPP are suggested based on the byproducts analysis. The enhanced removal efficiency and reaction mechanisms implied that the CPP process would probably be suitable for the removal of a large number of gaseous pollutants.     

Two-stage biofilter for effective NH3 removal from waste gases containing high concentrations of H2S

A high H2S concentration inhibits nitrification when H2S and NH3 are simultaneously treated in a single biofilter. To improve NH3 removal from waste gases containing concentrated H2S, a two-stage biofilter was designed to solve the problem. In this study, the first biofilter, inoculated with Thiobacillus thioparus, was intended mainly to remove H2S and to reduce the effect of H2S concentration on nitrification in the second biofilter, and the second biofilter, inoculated with Nitrosomonas europaea, was to remove NH3. Extensive studies, which took into account the characteristics of gas removal, the engineering properties of the two biofilters, and biological parameters, were conducted in a 210-day operation. The results showed that an average 98% removal efficiency for H2S and a 100% removal efficiency for NH3 (empty bed retention time = 23-180 sec) were achieved after 70 days. The maximum degradation rate for NH3 was measured as 2.35 g N day(-1) kg of dry granular activated carbon(-1). Inhibition of nitrification was not found in the biofilter. This two-stage biofilter also exhibited good adaptability to shock loading and shutdown periods. Analysis of metabolic product and observation of the bacterial community revealed no obvious acidification or alkalinity phenomena. In addition, a lower moisture content (approximately 40%) for microbial survival and low pressure drop (average 24.39 mm H2O m(-1)) for system operation demonstrated that the two-stage biofilter was energy saving and economic. Thus, the two-stage biofilter is a feasible system to enhance NH3 removal in the concentrated coexistence of H2S.