Inlet NH3 Concentration Research Articles

Investigating the Adsorption Behavior of Polyaniline and Its Clay Nanocomposite towards Ammonia Gas.

Air pollution and control of gaseous air pollutants are global concerns. Exposure to these gaseous contaminants causes several health risks, especially exposure to irritant gases such as ammonia (NH3). Furthermore, the application of smart polymeric nanocomposites in environmental applications has gained significant interest in recent years. In this study, aniline was polymerized without and with clay using a carbon dioxide (CO2)-assisted polymerization technique, yielding PANI and PANC samples, respectively. The samples were characterized using different methods, such as Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), transmission electron microscopy (TEM), scanning electron microscope (SEM), and Brunner Emmett Teller (BET). The synthesized nanomaterials were utilized as gas adsorbents using a fixed bed reactor to investigate their adsorption behavior towards NH3. Three inlet NH3 concentrations were tested (35-150 ppm). The results revealed that the adsorption capacities of PANC nanocomposites were higher than nanostructured PANI for the studied concentrations. The adsorption capacities were 61.34 mgNH3/gm for PANC and 73.63 mgNH3/gm for PANI at the same inlet concentration (35 ppm). The highest NH3 adsorption capacity recorded was 582.4 mg NH3/gm, for PANC. This study showed the impressive adsorption behavior of the prepared PANI and PANC nanomaterials towards NH3 gas. Consequently, nanostructured PANI and PANC can be promising adsorbents that can be utilized to control different gaseous air pollutants.

Nitrogen Recovery from Different Livestock Slurries with an Innovative Stripping Process

Ammonia (NH3) emissions deriving from the management of livestock manure have a significant environmental impact, and therefore it is important to reduce them. Among the available options, the process of NH3 stripping is promising to remove NH3 from manures and digestates recovering it as a mineral fertilizer (e.g., ammonium sulfate) that is more widely adoptable on farms. The traditional stripping process takes place in batches; however, in this study, a continuous process was evaluated using a lab scale plant in which four reactors were used in series with different hydraulic retention times (HRTs) of 12 or 20 days. The NH3 recovery of each reactor was studied for the liquid fraction of pig slurry, dairy cattle slurry and digestate, applying simple headspace aeration. For 20 days of HRT, totals of 92%, 83% and 67% of NH3 were stripped from the digestate, pig slurry and dairy cattle slurry, respectively. For 12 days of HRT, total NH3 recoveries were 83%, 60% and 41% for the digestate, pig slurry and dairy cattle slurry, respectively. The inlet NH3 concentration and inlet total alkalinity had a positive and negative effect, respectively, on the specific NH3 removal rate for each reactor. Stripping NH3 on farm scale can abate NH3 emissions in response to the environmental concerns of European policies.

Numerical investigations on flow boiling heat transfer of ammonia water binary solution (NH3/H2O) in a horizontal microchannel

The flow boiling heat transfer characteristics of NH3/H2O mixture in a 2D single horizontal microchannel (0.4 mm width × 6 mm length) were investigated by Computational Fluid Dynamics (CFD) method. The multiphase VOF model and modified phase change Lee method were adopted to address the non-isothermal phase change process of the flowing zeotropic NH3/H2O mixture, while the variations of the binary mixture thermophysical properties were also taken into account. The effects of mass flux (46~552 kg/(m2∙K)), inlet NH3 concentration (0-35% by mole) and heating wall temperature (20.5~70 ∘C) on the overall and local flow boiling heat transfer performance have been comparatively evaluated under constant heating wall temperatures. According to the numerical results, the heat dissipation rate of NH3/H2O mixture flow boiling could reach up to 1.41 MW/m2 at a mass flux of 552 kg/(m2∙s), which was 2.05 times of water single-phase flow cooling under a same constant heating wall temperature of 50 ∘C. It was also revealed that, for NH3/H2O mixture flow boiling in the microchannel, there was a threshold of inlet NH3 concentration to maintain a certain level of heat dissipation rate at a given mass flow rate and further increasing the inlet NH3 concentration would no longer benefit the heat dissipation process. Furthermore, there were no local dry-outs found throughout the whole microchannel length under all the simulation conditions in this study, which could be attributed to the unique flow boiling behaviors of zeotropic NH3/H2O mixture. Therefore, it should be noticed that NH3/H2O mixture, under certain conditions, could be a good alternative coolant for preventing local dry-outs and maintaining a certain functional temperature of electronic components.

Interactions and microbial variations in a biotrickling filter treating low concentrations of hydrogen sulfide and ammonia

A lab-scale biotrickling filter (BTF) packed with porcelain Rasching ring and ceramsite was applied for co-treating of low concentrations of hydrogen sulfide (H2S) and ammonia (NH3), as major pollutants typically found in e.g., intensive livestock production facilities. In this study, the outlet gas concentrations of H2S and NH3 were used for indicators if the treated gas reached odor-free condition. Overall, excellent removal efficiencies were obtained for both H2S and NH3 in the BTF during Stage I (H2S alone) and Stage II (H2S and NH3). Specifically, the H2S outlet concentration was below the detection limit (∼3.6 ppbv) and the NH3 outlet concentration was less than 0.4 ppmv when the inlet concentrations of H2S and NH3 were around 1.8 ppmv and 35.3 ppmv, respectively. In this case, the running empty bed residence time was 10.2 s. During Stage II, the outlet H2S concentration was decreased significantly when the inlet NH3 concentration was increased, likely due to the influence by pH. Meanwhile, the outlet nitrous oxide (N2O) concentration was kept low (<2% NH3) during the experiment, suggesting a proper operation of the BTF. After the inlet gas shifted from H2S alone at Stage I to H2S and NH3 at Stage II, the main sulfur-oxidizing bacteria (SOB) species in the BTF switched from Acidithiobacillus to Thiobacillus.

Microbial and isotopomer analysis of N2O generation pathways in ammonia removal biofilters

Ammonia removal biofilters can be a potential source of nitrous oxide (N2O) production as a result of microbial nitrification and denitrification. In this study, these two N2O generation pathways was quantified using isotopic site preference values (SP, 33‰ for nitrification and 0‰ for denitrification) in a 204-d operation. Tests with two moisture conditions (45% and 55%) and three inlet NH3 concentrations (35, 18 and 0 ppmv) were performed. A 55+% NH3 removal efficiency was achieved in biofilters with 35 and 18 ppmv ammonia supply, but no significant difference (p > 0.05) was found between the moisture treatments. Results showed that biofilters were clearly net sources of N2O, and biofilters with higher moisture content generated significantly (p < 0.05) higher N2O concentration. The N2O generation did not stop even after the biofilters were terminated. The percentage of inlet NH3–N converted into N2O–N were 5.2%, 8.5% for biofilters with 45% moisture content, and 14.8%, 10.8% for those with 55% moisture content. Gene abundance of amoA and nosZ in packing materials (taken on days 64, 107, 140, 180 and 204) increased due to NH3 input reaching the highest on day 140 and then decreased in response to reduced NH3 supply on day 180 and 204. The changes of SP values suggested a shift between nitrification and denitrification with regard to N2O generation. Overall, the nitrification was the dominant pathway for N2O generation, but uncertainty exits as well. This study confirmed that NH3-loaded biofilters were net sources of N2O, and use of SP-N2O may be helpful in better understanding the processes responsible for such emissions.

Application of biological methods in the treatment of gaseous ammonia

In Vietnam, practical applications of biological methods in air pollution control are highly limited. This study evaluated and compared the ammonia removal performance in air of a cow-manure biofilter, commercial compost biofilter, and biotrickling filter with K3 biomedium cultured with attached microorganisms from activated sludge. The results indicated that with an inlet NH3 concentration of 65-80 mg/m3 (95-117 ppm), the treatment efficiency was highly promising with an output concentration in the range of 2-5 mg/m3 (3.0-7.5 ppm) and elimination capacity of 3-9 gNH3/m3.h. With an inlet concentration below 200 mg/m3, all three experimental models could remove ammonia to meet the emission standard (QCVN 19:2009/BTNMT) of 50 mg/m3. The study results indicated that the investigated biological technologies have potential for use in removing ammonia and other odorous gases in polluted air. This study investigated the removal of gaseous ammonia using cow manure, compost, and K3 material as biomedium in biofilters and biotrickling filters.

Experimental optimization of a spray tower for ammonia removal

Spray tower scrubbers ordinarily have low air resistance and gas removal efficiency. Although packed-bed wet scrubbers are efficient in gaseous contaminants treatment, a significant limitation of packed-bed wet scrubbers is that they have high pressure drop and primary costs. Appropriate features of the nozzle play an important role in system cost, efficiency, low operational costs, and optimization of spray towers. Operating pressure, nozzles size, and number of nozzles could increase mass transfer and removal efficiency and decrease investment and save operational costs. The objective of the present study was to develop a spray tower through optimization of the design and operating parameters for removal ammonia emissions from the air streams. Spray tower design parameters included nozzle type, number of stages of spray nozzle, and operating parameters such as operating pressure and inlet NH3 concentration. Among the studied parameters, only increasing ammonia concentration was in inverse proportion to the spray tower efficiency. The spray tower was optimized as equipped with an 8010SS spray nozzle with three stages working together, spraying 0.01% H2SO4 scrubbing liquid counter-current to the air stream with operating pressure of 12 bars and inlet NH3 concentration of 24.1 ppm. The highest removal efficiency was 97.92% at an 8010SS spray nozzle with three stages working together, H2SO4 solution, pressure 12 bars and inlet ammonia concentration of 24.1 ppm. The results of this study demonstrated that caustic spray tower could be a very effective technology for removal of NH3 from air stream.

Performance, limitations and microbial diversity of a biotrickling filter for the treatment of high loads of ammonia

In this work, limits and capacities of biotrickling filtration as a technique to compete with chemical scrubbing for the treatment of ammonia loads above 40gNm−3h−1 are investigated. Long-term operation of a lab-scale biotrickling filter under different inlet NH3 concentrations (100–1400ppmv) and gas contact times (2–46s) was performed to study the influence of operating parameters such as the make-up water flowrate and to assess biological and mass transfer limitations. The effect of the ammonia load over nitritation and nitratation rates was analyzed. At the same ammonia load, nitrification rates at gas velocities of 38.4mh−1 (large gas contact times) were 2-fold higher compared to those at gas velocities of 844mh−1 (short gas contact time) due to a better distribution of the NH3 along the packed bed, thus reducing the inhibitory effect of free ammonia (FA) and free nitrous acid (FNA) accumulation. Microbial diversity was compared between low and high-ammonia load conditions through tag-encoded FLX amplicon pyrosequencing showing that accumulation of FA and FNA promoted the growth of nitrifiers. Also, a large fraction of denitrifying species was found demonstrating that denitrification occurs also in aerobic biotrickling filters.

Semi-mechanistic modelling of ammonia absorption in an acid spray wet scrubber based on mass balance

A model to describe reactive absorption of ammonia (NH3) in an acid spray scrubber was developed as a function of the combined overall mass transfer coefficient Kyav. An experimental study of NH3 absorption using 1% dilute sulphuric acid was carried out under different operating conditions. An empirical correlation for Kyav with respect to droplet Sauter mean diameter, liquid flow rate, and inlet NH3 concentration was developed with an R2 = 97.12%. Air velocity positively correlated with Kyav at 30 ppmv, but did not exhibit an effect at higher concentrations (165–300 ppmv), while liquid flow rate showed the greatest effect on Kyav. The Kyav correlation was incorporated in the performance model to construct a semi-mechanistic model with high prediction accuracy (R2 = 97.60%, MSE = 0, RMSE = 0.03, MAPE = 5.24%). This generalised performance model can be used to predict NH3 removal efficiencies of our optimised acid spray scrubber under various operating conditions at animal facilities.

Statistical modelling of ammonia absorption in an acid spray scrubber

The use of acid spray wet scrubbers for recovering ammonia (NH3) emissions is promising due to its high NH3 removal efficiency, simplicity in design, and minimal pressure drop contribution on fans. An experimental study was conducted to evaluate the performance of a lab-optimised acid spray scrubber using simple modelling tools. Important parameters that significantly affect scrubber efficiency were identified as inlet NH3 concentration, air retention time, Sauter mean diameter of spray droplets, and liquid flow rate. Two statistical models (additive and multiplicative models) were developed from the experimental data using regression analysis of scrubber efficiency as a function of the significant operating parameters. The additive model had better performance accuracy with an R2 value of 0.93, MSE of 0, RMSE of 0.06, and MAPE of 8.89%. Both models have good predictive ability based on residual analysis, power analysis, and cross-evaluation. This study was able to develop simplified models to aid in predicting NH3 removal efficiency of an acid spray scrubber, which can be installed at animal facilities for NH3 abatement.

Improvement of biodegradation in compact co-current biotrickling filter by high recycle liquid flow rate: Performance and biodegradation kinetics of ammonia removal

As a microbial-environmental-control-type deodorizing system, we have developed a compact biotrickling filter system for small-scale livestock farms. The performance of the compact co-current biotrickling filter operated at high recycle liquid flow rates was systematically examined. In particular, we studied improvements in the nitrification ability of the system due to the resultant enhancement of absorption and dissolution of NH3 and absorption of O2 with the high flow rates of recycle liquid flowing downward co-currently with gas flow. At the empty bed residence time of 50s, almost complete removal of NH3 was obtained with recycle liquid flow rates of 103 and 205Lm−3day−1 for 20 days while the inlet NH3 concentration was increased from 200 to 500ppm. With a recycle liquid flow rate of 411Lm−3day−1 the removal efficiency remained above 95% for 57 days while the inlet NH3 concentration was increased from 200 to 700ppm. The biodegradation kinetics for NH3 removal was successfully analyzed using the Haldane substrate inhibition kinetics. The present data and kinetic analyses showed that the substrate inhibition was suppressed and the biodegradation of ammonia in the compact biotrickling filter could be improved by the high recycle liquid flow rate.

NH3 biofiltration of piggery air

An aboveground pilot-scale biofilter filled with wood chips was tested to treat ammonia emissions from a piggery located in Brittany (France). Two long-term tests (“summer” and “autumn” experiments) were carried out to improve biofilter applications for agriculture. The influence of climatic conditions on biofilter performance was taken into account.During summer 2012, the biofilter was operated for 74 days at different empty bed residence times (EBRTs) from 6 to 15 s. Inlet NH3 concentrations were relatively constant (around 15 mg m−3). Significant NH3 reductions were achieved at EBRT = 12 s (removal efficiencies, RE, ranged between 90 and 100% for loading rates, LR, of around 4 g m−3 h−1). At a lower EBRT (6 s), RE dropped to roughly 30–50%. This was due to the dramatic increase in the loading rate (LR up to 12 g m−3 h−1) but the results showed that the change in atmospheric conditions (temperature and relative humidity) also had a significant influence on biofilter performance. It was evidenced that the use of a humidifier upstream of the biofilter must be taken into account for large-scale biofilter design, but only for specific conditions (the spraying of the biofilter having to be carried out exceptionally).During autumn 2012, the biofilter was operated for 116 days at EBRT = 12 s. RE were around 80% for LR of around 3 g m−3 h−1. In such autumnal atmospheric conditions, a demister system should be installed upstream of the biofilter in order to avoid water accumulation in the bed material.Although biofiltration was suitable for NH3 treatment of piggery air, the need to control accurately the medium moisture content implies that biofilters would not be easily managed by a pig farmer.

A Prototype Acid Spray Scrubber for Absorbing Ammonia Emissions from Exhaust Fans of Animal Buildings

Mitigation of ammonia (NH3) emissions from animal production buildings has been a challenge because of the large volume of low NH3 concentration laden air being released. Among emission mitigation technologies for concentrated animal feeding operations, acid spray scrubbers have the greatest potential for adaptation to the existing large animal facilities because of their lower fan airflow reduction, ability to simultaneously remove particulate and gaseous pollutants, and viability for zero or less waste generation by recycling effluents as liquid fertilizer. A multi-stage wet scrubber prototype that can be operated with a maximum of three stages was developed and optimized for reducing NH3 emissions using simulated conditions typically encountered at an animal building exhaust. The parameters optimized for a single-stage wet scrubber include nozzle type, nozzle operating pressure, sulfuric acid concentration, spray coverage, and air retention time. The optimized single-stage wet scrubber settings can remove emissions from 60% ±1% at 5 ppmv inlet NH3 concentration (IAC) to 27% ±2% at 100 ppmv IAC at a normal exhaust superficial air velocity (SAV) of 6.6 m s -1 . A high concentration of droplets inside the contact chamber increased the rate of inter-collision between droplets, which led to high droplet coagulation and decreased surface area for gas-liquid contact. These phenomena were prevented by operating the nozzles in the higher stages co-current to the airflow and by using fewer nozzles in higher stage. The two-stage and three-stage wet scrubbers were therefore optimized by determining the least number of nozzles in each stage that provided the most effective NH3 removal. The optimized two-stage scrubber could remove NH3 emissions from 60% ±0% at 5 ppmv IAC and 35% ±1% at 100 ppmv IAC. The optimized three-stage scrubber could remove emissions from 63% ±3% at 5 ppmv IAC and 36% ±3% at 100 ppmv IAC. Airflow retention time was found to significantly affect NH3 absorption. Reducing the superficial air velocity to 3.3 m s -1 from 6.6 m s -1 , which increased the air retention time from 0.2 s to 0.4 s, improved NH3 removal efficiencies to 98% ±3% at 5 ppmv IAC and 46% ±2% at 100 ppmv IAC for the single-stage scrubber. Similarly, the performance of the two-stage scrubber at a SAV of 3.3 m s -1 improved to 77% ±0% at 20 ppmv IAC and 57% ±1% at 100 ppm IAC. Lastly, the performance of the three-stage scrubber at a SAV of 3.3 m s -1 improved to 70% ±1% at 30 ppmv IAC and 64% ±1% at 100 ppmv IAC. It was observed that the three-stage wet scrubber did not increase the overall wet scrubber performance, as predicted theoretically. Further studies are needed so that the application of these scrubber designs becomes feasible for treating air emissions from animal buildings. The wet scrubber caused an additional backpressure of 27.5 Pa, resulting in about 8% airflow reduction for a fan operating at 12.5 Pa.

AMMONIA ABSORPTION IN A VERTICAL SPRAYER AT LOW AMMONIA PARTIAL PRESSURES

In this article, ammonia (NH3) removal efficiency as affected by stripping factor, air mean retention time (0.8, 1.1,1.7, 2.8, and 8.1 s), and NH3 inlet concentration (5, 15, and 50 ppmv) was evaluated using a sprayer unit called thediffusion-coagulation-separation (DCS) system. The NH3 removal efficiency was calculated from the NH3 concentrationsmeasured from both air and water streams. The DCS system was found to effectively remove 84% of NH3 from the air streamat a stripping factor value of 0.5 or when the retention time of the air was 8.1 s. An empirical equation for NH3 removalefficiency as a function of stripping factor and inlet NH3 concentration was derived.

Control of H2S waste gas emissions with a biological activated carbon filter

AbstractThe removal of high concentrations of H2S from waste gases containing mixtures of H2S and NH3 was studied using the pilot‐scale biofilter. Granular activated carbon (GAC), selected as support material in this study, demonstrated its high adsorption capacity for H2S and good gas distribution. Extensive tests to determine removal characteristics, removal efficiency, and removal capacity of high H2S levels and coexisting NH3 in the system were performed. In seeking the appropriate operating conditions, the response surface methodology (RSM) was employed. H2S removal capacities were evaluated by the inoculated bacteria (biological conversion) and BDST (Bed Depth Service Time) methods (physical adsorption). An average 98% removal efficiency for 0.083–0.167 mg dm−3 of H2S and 0.004–0.021 mg dm−3 of NH3 gases was achieved during the operational period because of rapid physical adsorption by GAC and subsequently an effective biological regeneration of GAC by inoculated Pseudomonas putida CH11 and Arthrobacter oxydans CH8. The results showed that H2S removal efficiency for the system was not affected by inlet NH3 concentrations. In addition, no acidification was observed in the BAC biofilter. High buffer capacity and low moisture demand were also advantages of this system. The maximal inlet loading and critical loading for the system were 18.9 and 7.7 g‐H2S m−3 h−1, respectively. The results of this study could be used as a guide for the further design and operation of industrial‐scale systems. Copyright © 2004 Society of Chemical Industry

KINETICS OF AMMONIA REMOVAL IN A PILOT-SCALE BIOFILTER

A mobile, skid-mounted biofilter was designed and built to determine the kinetics of ammonia oxidation at a modern2400-sow farrow-to-wean unit. The biofilter system consisted of a variable-speed blower, packed-bed humidifier, andtwo reactors (packing volume of 0.35 m3 per reactor) configured in parallel. Prescreened, composted yard waste was usedsince compost contains a large number of active microorganisms, is relatively inexpensive, and is readily available. Ammoniaemissions (0 to 12 ppmv or 0 to 8.4 mg m-3) from the swine facility (and 0 to 25 ppmv or 0 to 17.4 mg m-3 in a simulated stream)were transported downward across the packing in the reactors, and spray nozzles at the top of each reactor were used to addmoisture to the packing. Ammonia removal ranged from 25% to 95% depending on the residence time and inlet NH3concentration. Using first-order kinetics, the measured ammonia degradation rate ranged from 0.06 to 0.80 mg NH3 m-3 s-1for volumetric loading rates ranging from 0.05 to 0.25 mg NH3 m-3 s-1. Residence time distribution (RTD) analysis was alsoperformed to determine the effect of scale-up on axial dispersion and deviation from plug flow. RTD analyses suggest thatnon-ideal reactor design equations may be required to predict reactor size for desired ammonia conversions. A reactor designmethod (i.e., sizing calculation) was developed based on the kinetics of ammonia oxidation. For example, assuming avolumetric flow rate of 13 m3 s-1, NH3 conversion of 95% (CNH3in = 25 ppmv), and a first-order rate constant of 0.08 s-1,a reactor volume of 487 m3 and residence time of 37.5 s is required. The size of the reactor will change depending on thecharacteristics of the swine facility. Pressure drops across the bed should be 50 Pa or less in order to utilize in-house fans.This will probably limit the height of the reactor to 1 m or less. The mass of compost required can be estimated from its bulkdensity. Moisture content must be maintained between 40% and 60% in the biofilter to maintain biological activity. Additionalresearch is required to develop inexpensive methods of emission humidification, online moisture analysis, and water addition.

Hydrogen sulfide effects on ammonia removal by a biofilter seeded with earthworm casts.

Ammonia (NH3) removal efficencies were evaluated when hydrogen sulfide (H2S) and NH3 in binary mixture gases were supplied to a ceramic biofilter seeded with earthworm (Lumbricus terrestris) casts. The effect of inlet H2S concentration and space velocity (SV) on the removal of NH3 was investigated after the acclimation of the biofilter with NH3 gas. When NH3 was singly supplied to the biofilter, NH3 removal was maintained at almost 100% until inlet NH3 concentration was increased up to 600 microL L(-1) and SV up to 330 h(-1), at which the elimination capacity of NH3 was 148 g N m(-3) h(-1). When H2S was supplied simultaneously, however, the accumulation of toxic sulfide ions showed dual effects on NH3 removal efficiencies. First, no effects were observed at inlet H2S loading below 60 g S m(-3) h(-1); however, inhibition by H2S at higher loading was observed above 60 g S m(-3) h(-1). The point at which loading achieved a maximum of more than 99% NH3 removal efficiency was 139 g N m(-3) h(-1), when inlet H2S concentration was held under 100 microL L(-1), but it dropped to 76 and 30 g N m(-3) h(-1) when the inlet H2S concentration increased to 220 and 460 microL L(-1), respectively. The critical points of inlet H2S loading that guaranteed over 99% NH3 removal were determined as 100, 100, 60, and 40 g S m(-3) h(-1) at inlet NH3 concentrations of 100, 200, 400, and 600 microL L(-1), respectively. Inlet NH3 loading had synergic effects of increasing the inhibition of inlet H2S loading on the NH3 removability of the biofilter.

Biological Removal of Gaseous Ammonia in Biofilters: Space Travel and Earth-Based Applications

ABSTRACT 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 ~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.