Greenhouse Gas Emissions From Biofilters for Composting Exhaust Ammonia Removal

Emissions of odorous compounds, such as ammonia (NH3), from composting have negative agronomic and environmental impacts. A biofilter is widely used for NH3 removal, with one of its potential detrimental by-products being nitrous oxide (N2O), which is a higher warming potential greenhouse gas (GHG). The aim of the study was to evaluate the effect of empty bed retention time (EBRT) on GHG emissions from biofilters for removing NH3 from composting. Composting experimental trials lasted 6 weeks, and composting materials were mixtures of dead pigs and manure. Three groups of biofilters with 1.2 m-height, 0.3 m-inner diameter, and 1.0 m media depth were conducted with EBRT of 30, 60, and 100s, respectively. Each treatment was performed in triplicate, and the gas was monitored using the dynamic emission vessel method. The Spearman’s correlation analysis showed a significantly positive correlation between inlet concentrations (ICs) of NH3 and increased N2O concentrations: ρ = 0.707, 0.762, and 0.607 with p ≤ 0.0001 for biofilters with EBRT of 30, 60, and 100s, respectively. The fraction of NH3-N denitrified into N2O-N in biofilters with EBRT of 60 and 100s was higher than that with EBRT of 30s. The total global warming potential (GWP) increased by 126%, 162%, and 144% for biofilters with EBRT of 30, 60, and 100s, respectively. These results indicated that biofilters with longer EBRT will lead to higher GWP production. Future research on odorous mitigation for composting with biofilters should focus more on greenhouse gas emissions.


INTRODUCTION
Composting has been applied worldwide as an environmentally friendly and cost-effective method for sanitation and recycling animal waste (Loyon, 2017;Zheng et al., 2020). However, one major complication during composting is the odor emission (Cheng and Hu, 2010;Huang et al., 2012;Paulot, et al., 2014), which not only poses a problem to the general public and environmental health but also causes adverse effects on vegetation surrounding the composting plants (Han et al., 2019). Ammonia (NH 3 ) is considered the major contributor to odor from composting (Zhang et al., 2016;Zhu et al., 2016), and the precursors of particulate matter can be produced by the reactions of NH 3 with sulfuric and nitric acid aerosols (Pinder, et al., 2007;Renner and Wolke, 2010;Kong et al., 2020).
Biofilters are widely used to reduce NH 3 emissions from composting (Turan et al., 2009;Mudliar et al., 2010;Janni et al., 2014). Many studies have focused on the physical, chemical, and biological parameters influencing the biofiltration process (Park, et al., 2002;Yuan et al., 2019). A number of studies have shown that nitrous oxide (N 2 O) generation in biofilters is often accompanied by NH 3 removal (Maia et al., 2012b;Yang et al., 2014a;Kong et al., 2020). Akdeniz and Janni (2012) observed that N 2 O generation ranged from −29.2% to 4.0% for a flat-bed biofilter with an empty bed retention time (EBRT) of 5s. Clemens and Cuhls (2003) observed that approximately 26% of NH 3 -N entered in the biofilters is converted into N 2 O-N. N 2 O emissions from biofilters can be affected by several factors, such as inlet NH 3 concentration, temperature, moisture content, and pH value (Maia et al., 2012b;Yang et al., 2014a;Yang et al., 2014b;Dumont et al., 2014). High moisture content can cause regional anaerobic zones and increase the microbial activity, which favors N 2 O emission (Yang et al., 2014a). Compared to a high pH value (8.0-9.5), a low pH value (4.5-6.0) of the media can inhibit the generation of N 2 O reductase and reduce the emission of N 2 O (Yang et al., 2014b). N 2 O from biofilters was correlated significantly with the NH 3 in the biofilters (Clemens and Cuhls, 2003). EBRT can affect microorganism absorption and the conversion process of NH 3 in the biofilters (Shang et al., 2020). Shorter EBRT means faster gas flow velocity, which can change the oxygen (O 2 ) gradients in the biofilter media and can lead to changes in denitrification (Maia, et al., 2012b). In addition, higher air flow can increase the emission rate of gas from composting. EBRT is one of the key parameters of biofilters for NH 3 removal (Liu et al., 2017); however, few studies have investigated the effects of EBRT on emissions of N 2 O. A better understanding of the effects of EBRT on the generation and emission of N 2 O needs to be elucidated. In this study, the effects of EBRT on greenhouse gas (N 2 O and methane, CH 4 ) emissions from pilot-scale biofilter systems were studied. The results can provide a promising tool for greenhouse gas reduction from full-scale biofilters.
CH 4 is another important greenhouse gas emitted during composting (Zhu-Barker, et al., 2017). Biofilters also can be used to reduce CH 4 emissions (Haubrichs and Widmann, 2006). Many studies performed on CH 4 biofiltration utilized an EBRT of at least 4 min (La et al., 2018), which posed obstacles to the biofilter application. Little information is available in the literature about CH 4 reduction capabilities during NH 3 biofiltration.
Global warming potential for N 2 O and CH 4 is 296 and 23 times higher than that of carbon dioxide (CO 2 ), respectively (Pachauri et al., 2014). It is very important to examine the generation of greenhouse gas due to odor treatment by biofilters. Thus, the objectives of this study were 1) to investigate the emissions of N 2 O from biofilters with different EBRT and 2) to assess the greenhouse gas (N 2 O and CH 4 ) emissions from biofilters for composting NH 3 removal.

Experiment Materials
This study was conducted in the Beijng Anding pig farm, located in Daxing District, Beijing, China (39°62′N, 116°50′E). The composting materials and composting process have been described by Shang et al. (2020).
The mature compost, composted for about 4 months, was used as the medium material. Before the experiment, the mature compost was inoculated with activated sludge from the aerobic fermentation plant for wastewater treatment in the pig farm, and the water content of packing materials at the beginning of composting was adjusted to 57.4 ± 2.5%, according to Akdeniz and Janni (2012). The mature compost used for biofilter media had total carbon (C) and nitrogen (N) contents of 30.8 ± 2.8% and 2.9 ± 0.7%, respectively, with a pH of 7.0.

Biofilter Design and Operations, Gas Sampling, and Analytical Method
The biofilters were constructed with circular unplasticized polyvinyl chloride (UPVC) pipes. The dimensions were 1.2 m (height) and 0.3 m (inner diameter), with a 1.0 m depth of media (corresponding to a bed material volume V = 0.07 m 3 ). Nine biofilters were divided into three treatments, and different EBRT (30s, 60s, and 100s) were set according to previous studies (Pagans et al., 2005). Each treatment was conducted with three replicates. EBRT = V/Q, where Q is the air flow rate in m 3 ·s −1 . The experimental arrangements are shown in Table 1. NH 3 , CH 4 , and N 2 O concentrations of air outside (1 sampling point), a gas inlet of biofilters (3 sampling points), and a gas outlet of biofilters (9 sampling points) were measured continuously and simultaneously using a photoacoustic multigas analyzer (model Innova 1412i, LumaSense Technologies, Ballerup, Denmark) every day. The deodorization system was constructed, as previously described (Shang et al., 2020), and the schematic is shown in Figure 1.
The experiment corresponding to the composting period was carried out for 42 days. During the experiment, no water was supplied to the media of biofilters.

Data Analyses
All data were analyzed by Microsoft Excel 2016 and SPSS 22. The variables were tested for significance using Spearman's correlation coefficient. The rations of gas concentrations in outlet and inlet biofilters were calculated (Dumont, 2018). A ration >1 indicates gas formation in the biofilter, and a ration <1 indicates gas depletion in the biofilters. The t-test (one side p < 0.05) was used to determine if there was a significant deviation from 1.

Global Warming Potential Calculations
Global warming potential (GWP) was quantified as CO 2 equivalent with a 100-year timescale: 1 kg CH 4 and N 2 O emitted are equivalent to 23 and 298 kg CO 2 , respectively (Pachauri et al., 2014). The CH 4 , N 2 O, and GHG emissions (in kg eqCO 2 ) were calculated by multiplying the aerobic rates, the concentration, and the GWP factor during the whole experiment which lasted for 42 days. In this study, the GWP of

Emission of Nitrous Oxide From Biofilters
NH 3 , N 2 O, and CH 4 concentrations in the air during the composting period were 5.6 ± 2.5, 1.0 ± 1.6, and 3.2 ± 1.6 mg m −3 , respectively. The NH 3 removal efficiencies have been described by Shang et al. (2020). The daily mean concentration of N 2 O from composting ranged between 1.5 and 11.5 mg m −3 , while the daily mean N 2 O concentrations from biofilters were about around 1.0 mg m −3 (Figure 2). The large variation in N 2 O production was due to the different concentrations of NH 3 and N 2 O from composting. The N 2 O concentrations at the biofilters' outlet were in the range of 20%-1250%, 23%-1652%, and 3-1352% higher than those at the biofilters' inlet for EBRT of 30, 60, and 100s. The outlet concentrations of N 2 O of biofilters with EBRT of 100 and 60s were higher than those with EBRT of 30s ( Figure 3). But there are no significant differences between the biofilters with different EBRT.
With biofiltration systems, about half of the inlet NH 3 is converted to nitrites or nitrates, and the other half is absorbed into the water as ammonium (Ottosen et al., 2011;Yasuda et al., 2017). Both ammonia oxidizer and denitrifier can produce N 2 O in a biofilter, but the majority of N 2 O was generated from denitrification (Kong et al., 2020). Maia et al. (2012a) pointed that higher EBRT may result in non-uniform oxygen distribution, which favors conditions for denitrification to generate N 2 O. In this study, the higher EBRT means a slower air flow rate, and the increasing N 2 O concentration from biofilters with increasing EBRT may result from more denitrification. In addition, when the air continuously goes through media in the biofilter, the higher EBRT implies that there is more content time of air and media, which can lead to high NH 3 -sorption, which favors the occurrence of nitrification-denitrification (Maia et al., 2012a). In this study, N 2 O concentrations were found to be higher, up to 16 times, than inlet concentrations, which may be due to the longer EBRT and higher inlet concentrations of NH 3 (Dumont et al., 2014;Liu et al., 2017).
The inlet concentrations, inlet loads, and elimination capacities of NH 3 may affect N 2 O generation. The Spearman's correlation analysis showed a strong positive correlation between inlet NH 3 concentrations and increased N 2 O concentrations: ρ = 0.707, 0.762, and 0.607 with (p ≤ 0.0001) for EBRT of 30, 60s and 100s, respectively. Inlet loads of NH 3 and increased N 2 O concentrations were positively correlated: ρ = 0.685, 0.750, and 0.579 (p ≤ 0.0001) for EBRT of 30, 60s, and 100s, respectively. Elimination capacities of NH 3 and increased N 2 O concentrations were also positively correlated: ρ = 0.706, 0.761, and 0.602 with p ≤ 0.0001 for EBRT 30, 60, and 100s, respectively. In this study, the inlet NH 3 concentrations were between 12 and 447 mg m −3 due to different composting processes, while the overall removal efficiencies were 85.4%, 88.7%, and 89.0% for EBRTs of 30, 60, and 100s, respectively. The elimination capacities of NH 3 have been discussed by Shang et al. (2020). NH 3 removal was considered the net source of N 2 O in biofilters (Maia et al., 2012a). N 2 O is considered normal at biofilters treating NH 3containing air, being a byproduct of nitrification and denitrification, and the pathways of N 2 O formation reported in the literature, however, are complex. Kong et al. (2020) found that the inlet NH 3 concentration can affect the nitrification process and the substrate availability for denitrification. In the present study, high inlet loads of NH 3 promoted N 2 O generation, which were consistent with the other study (Kong, et al., 2020); the denitrification is the main pathway for N 2 O formation.

Fraction of NH 3 -N Denitrified Into N 2 O-N
In the present study, there were significant differences in N 2 O emissions between biofilters with different EBRT due to different inlet loads of NH 3 resulting from different EBRT. The fractions of NH 3 -N converted into N 2 O-N were 4.6%, 5.6%, and 5.1% for biofilters with EBRT of 30, 60, and 100s, respectively. The results were consistent with other studies ( Table 2). Denitrification is considered the main source of N 2 O emission while nitrification is a trigger (Yang et al., 2014a;Dumont et al., 2014;Kong et al., 2020), and the presence of oxygen can inhibit the denitrification process. In this study, low EBRT means high gas flow rate, which results in more oxygen penetration into the media of the biofilter, further reducing the denitrification rates. Except for the EBRT, some other factors, such as the kind of media and moisture content may also affect N 2 O generation in biofilters. In Maia et al. (2012a), the compost of horse manure, cattle manure, chicken waste, woodchips, sawdust, and other materials were used as media for NH 3 removal, fraction of NH 3 -N denitrified into N 2 O-N was 14% and 19% for one biofilter and two other biofilters, respectively. The results of Yasuda et al. (2009) confirmed that NH 3 could be treated by biofilter with rockwool mixture without an extra increase of N 2 O. The percentage of N 2 O reduction efficiency ranging from 0.13 to 0.73% was found by Akdeniz (et al., 2011), who studied the removal of NH 3 and N 2 O using biofilters with lava rock as media at 5s EBRT. The part of NH 3 -N converted into N 2 O was estimated to range from 10% to 40% (Dumont et al., 2014). Yang et al. (2014a) found that there was a slight increase in N 2 O when the media moisture content increased from 35 to 55%, but further increasing the moisture content to 63% triggered N 2 O generation rapidly. Yasuda et al. (2009)    Frontiers in Bioengineering and Biotechnology | www.frontiersin.org June 2022 | Volume 10 | Article 918365 5 (2014b) showed that N 2 O concentrations ranged from 0.1 to 0.4 ppm with a pH of 8.0, and the acidified biofilters showed higher N 2 O concentrations than alkalized biofilters. In the present study, the values of the moisture content and pH of media change very little through the experiment. During the experiment, the pH of media was maintained at around seven for all biofilters, while the average moisture contents of media (mean ± SE, n = 12) were 48.3 ± 0.4, 48.5 ± 0.5, and 49.1 ± 0.4 for the biofilters with EBRT of 30, 60, and 100s, respectively.

Emissions of CH 4 From Biofilters
The daily mean concentrations of CH 4 from composting ranged between 5 and 149 mg m −3 , while the concentrations from biofilters ranged between 2 and 241 mg m −3 . The average daily REs of CH 4 were 6.5%, −10.5%, and −6.0% for EBRT 30, 60s, and 100s, respectively. Lim et al. (2012) reported the CH 4 reductions ranged from 0.1% to 1.9% in biofilters with 0.3s-0.6s EBRT. Fedrizzi et al. (2018) attained nearly 100% CH 4 100% removal efficiencies with 756s EBRT, so enough EBRT is needed for CH 4 removal. Akdeniz et al. (2011) achieved the removal efficiencies of 6.9%-25% for CH 4 by using the pilot-scale biofilters with lava rock media at the high moisture levels and low inlet CH 4 concentrations (90% moisture content and average of 31.0 ppm of inlet CH 4 concentration), while the EBRT was just 5s. Melse and van der Werf (2005) reported that the biological conversion of CH 4 in a biofilter is a slow process due to the low water solubility of methane (Henry's law constant = 1.5 × 10 -3 M atm −1 ). Biofilters with EBRT of 83-199s cannot reduce concentrations of CH 4 emitted from municipal solid waste composting, which has been reported by Clemens and Chuls (2003). Yasuda et al. (2009) also pointed out that several microbial CH 4 production and consumption reactions occurred, but a large portion of the CH 4 seemed to pass straightly through the biofilter with the EBRT of 100-200s 2014;Devinny, et al., 1999). In the present study, negligible CH 4 removal and no significant difference were found among three EBRTs, (Figure 4) and the result was similar to that of Akdeniz et al. (2011), Yasuda et al. (2009), and Lim et al. (2012. The possible reason is that the EBRT in this study is not enough for the transfer of CH 4 from the gas phase to the biofilm phase (Melse and Van der Werf, 2005). On the other hand, although the anaerobic methanogens were not tested in this study, the retention time for anaerobic suspended growth in  biofilters was about 50 days (Maia et al., 2012b); the slow acclimation of methanogens in the biofilter will lead to low CH 4 removal efficiency.

Global Warming Potential Emissions
Based on the measurements, the total global warming potential (GWP) emissions were calculated. Table 3 shows the GWP loading rate before and after passing through the biofilter for all the biofilters. The total GWP emission increased by 126%, 162%, and 144% for biofilters with EBRT values of 30, 60, and 100s, respectively. Significant differences were found in the total CO 2 -eq emissions from the biofilters. A lower percentage of GWP was emitted from the biofilter with 30s EBRT than in biofilters with 60 and 100s EBRT because the fraction of NH 3 -N denitrified into N 2 O-N is lower for the biofilter with 30s EBRT than that of 60 and 100s EBRT. These results showed that prolonged EBRT can improve RE of NH 3 but increase the GWP emissions accordingly. Melse and Hol (2017) evaluated three kinds of biofiltration of exhaust air from animal houses in which the total GWP emissions increased by 60%, 45%, and 0 in biofilters with EBRT of 1.4, 2.6, and 3.3s, respectively. In their experiment, the average inlet concentrations of NH 3 were 66, 10, and 15 ppm, respectively, and the removal efficiencies were 74%, 42%, and 38%, respectively. In the present study, the average inlet NH 3 concentrations were 124-163 mg m −3 , and the removal efficiencies were 82%-89%; both average inlet NH 3 concentrations and removal efficiencies of biofilters were higher than Melse and Hol (2017). With regard to GHG emissions, it can be concluded that the lower EBRT is more suitable for biofilter systems of composting, and the parameters of the control process, such as pH, temperature, dissolved oxygen, and water content, will be useful to prevent N 2 O formation and guarantee a good NH 3 removal efficiency.

CONCLUSION
The increased N 2 O concentrations from the biofilter were strongly and positively correlated with the elimination capacities of NH 3 . The total GWP emission increased by 54%, 62%, and 61% for biofilters with 30, 60, and 100s EBRT, respectively. The total GWP emission from biofilters increases by over 50% compared to a composting system without biofilters. More NH 3 converted into N 2 O due to higher EBRT suggested that lower EBRT is useful to prevent GHG from biofilters.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
BS: conceptualization, data curation, formal analysis, funding acquisition, writing-original draft, and writing-review and editing. TZ: conceptualization, data curation, formal analysis, and writing-review and editing. XT: formal analysis, investigation, methodology, and writing-review and editing. YC: investigation, validation, and writing-review and editing.