Release Reductions of Gaseous Ammonia and Nitrogen Oxides from Electrochemical Treatment of Swine Wastewater

Received: September 30, 2019
Revised: October 12, 2019
Accepted: October 16, 2019
Download Citation: RIS|BibTeX|https://doi.org/10.4209/aaqr.2019.09.0486  

Cite this article:

Huang, K.L., Liu, C.C., Lee, C.W., Ma, C.Y., Lin, T.C., Tsai, J.H. and Chen, S.J. (2019). Release Reductions of Gaseous Ammonia and Nitrogen Oxides from Electrochemical Treatment of Swine Wastewater. Aerosol Air Qual. Res. 19: 2490-2501. https://doi.org/10.4209/aaqr.2019.09.0486

HIGHLIGHTS

• Electrolysis can reduce the release of NH₃ from swine wastewater into the air.
• Electrolysis can reduce the release of NO from swine wastewater into the air.
• Anodes influence the release of NH₃ and NO_x from swine wastewater into the air.
• Initial NH₃-N concentration influences the release of NH₃ and NO_x in electrolysis.

INTRODUCTION

Pig farming including intensive or large-scale raising and breeding operations is still necessary to meet the increasing pork demand in several countries (Osada et al., 2017; Song et al., 2017). Despite the recent spread of African swine fever to new areas or countries, China, the United States and the European Union held the first, second, and third positions, respectively, for the global pork production in 2018 (Deblitz et al., 2018). This phenomenon causes the generation of considerable amounts of swine wastewater which needs to be well treated before discharge for the prevention of surface and underground water pollution, because several types of pollutants including high levels of organic and nitrogen compounds might be present in swine wastewater, which are regulated by water quality parameter limits (U.S. EPA, 2013; Ding et al., 2017; Osada et al., 2017; Song et al., 2017).

Another great environmental concern for pig farming is the emissions of ammonia (NH₃) and greenhouse gases (e.g., CH₄, CO₂, and N₂O) from the wastewater in purification facilities or the slurry in storages at pig farms prior to land application (Viguria et al., 2015; Osada et al., 2017). According to the European Environment Agency (EEA, 2019), ammonia and nitrogen oxides (NO_x) are two of four key air pollutants with emissions restricted by the EU legislation from 2010 onward, and emissions of ammonia from the agricultural sector continue to rise, posing a challenge for EU Member States in meeting EU air pollution limits. Guthrie et al. (2018) also indicates that the agricultural sector is the biggest contributor to ammonia pollution in the UK. Although the levels of other air pollutants have declined, ammonia is one of the main sources of nitrogen pollution, alongside nitrogen oxides, and ammonia emissions have been rising since 2013 in the UK, with significant implications for biodiversity and human health. NH₃ and NO_x are two of most important pollutants that need more intensive monitoring for the air pollution at Kathmandu Valley in Nepal (Kiros et al., 2016). These compounds are also two of N-containing acidifying pollutants which may deposit from the atmosphere to the ground (Wright et al., 2018) and impacts species composition through soil acidification (Guthrie et al., 2018). In addition to agricultural sources, NH₃ (reduced N) may also be produced from biomass burning, while NO_x (oxidized N) can be contributed from different sources, such as vehicle emissions, industry, and domestic combustion, and electricity generation (Wright et al., 2018). NH₃ and NO_x are regarded as precursor gases which may influence the chemical compositions of PM_2.5, visibility, and air quality (Li et al., 2018), so they are concerned for their spatio-temporal variations (Cai et al., 2018; Cheng et al., 2018) and long-range transport (Sadanaga et al., 2017).

Ammonia and hydrogen sulfide are the main odor gases related to manure swine houses (usually causing a nuisance for surrounding communities), although odor from inside swine production facilities is a complex mixture of gases (~330 different compounds), primarily due to manure decomposition under anaerobic conditions (Rahman and Borhan, 2012; Ubeda et al., 2013). Therefore, manure treatment systems are necessary for controlling the ammonia and odor emission from swine wastewater. Odor emission from liquid manure can be reduced by a three-step process: solid-liquid separation of manure, anaerobic digestion, and aerobic biological treatment (for the removal of organic and nitrogen compounds) (Rahman and Borhan, 2012; Montes et al., 2015). This process is also adopted in Taiwan for swine wastewater treatment; however, it has the disadvantages of time-consuming, low ammonia removal, and poor odor control. Moreover, ammonia is one of air toxic pollutants regulated by Air Pollution Control Act Enforcement Rules in Taiwan.

Electrochemical advanced oxidation process (EAOP) is a more efficient approach than the three-step one for swine wastewater treatment to alleviate the swine wastewater pollution along with its odor nuisance including ammonia emission. Generated in EAOP, hydroxyl radical (•OH) is a powerful oxidant which is highly reactive for degrading organic pollutants in wastewater. In the presence of chloride, active chlorine can be electrochemically generated in the EAOP and it is very efficient for removing ammonia and other nitrogen-containing compounds in wastewater (Li and Liu, 2009; Sirés et al., 2014; Radjenovic et al., 2015; Moreira et al., 2017). For this purpose and the effective production of •OH for chemical oxygen demand (COD) degradation, EAOP highly relies on the use of non-active anode materials (e.g., PbO₂ and boron-doped diamond [BDD]) (Li and Liu, 2009; Sirés et al., 2014; Radjenovic et al., 2015; Moreira et al., 2017). In our previous work, we explored the removal of COD and ammonia in swine wastewater at different operating parameters (Huang et al., 2018, 2019). However, little attention has been paid to release of NH₃ and nitrogen oxides (NO_x) from wastewater into the air in association with the variations in NH₃-N, NO₂^–-N, NO₃^–-N, and COD concentrations in the swine wastewater with/without electrolysis. Therefore, this study focuses on the investigation of gaseous ammonia and nitrogen oxides emitted from electrochemical treatment of swine wastewater using different anode materials.

MATERIALS AND METHODS

Electrolysis of Swine Wastewater

The electrolysis of swine wastewater with the addition of 0.05 M NaCl was performed in an undivided electrochemical cell at room temperature (25°C) and 0.25 A cm^–2 (Fig. 1). A Ti plate was used as the cathode, while the anode material was BDD-I (Nb/BDD from Neocoat, Germany), lab-prepared substance (Ti/PbO₂, graphite/PbO₂, or graphite/BDD), or graphite. The fabrication method for the lab-prepared PbO₂ was provided elsewhere (Chen et al., 2012), whereas the fabrication of graphite/BDD referred to an approach of hot filament chemical vapor deposition (HFCVD) (Pinheiro et al., 2012). A DC power supply (Twintex TP2H-20S, Taiwan,) was used to monitor the cell voltage and current over time in the electrolytic experiments.

Fig. 1. The scheme of experimental device (A: Ammonia analyzer, B: Sampling tube, C: Magnetic stirrer, D: DC power supply, E: thermostatic bath, F: Three-side rectangular box, +: Anode, and ‒: Cathode).

The degradation or removal efficiency of COD or NH₃-N was calculated using the following equation:

where C_t is the residual concentration of either COD or NH₃-N at a given electrolysis time t and C₀ is the initial concentration of COD or NH₃-N.

The specific energy consumption (E_SP) (kWh g-COD^–1 or kWh g-NH₃-N^–1) was determined using the following calculation (Martínez-Huitle and Ferro, 2006; Radjenovic et al., 2015):

where E_cv is the average cell voltage (V); I is the applied current (A); V_wv is the wastewater volume (L). 

Analysis of Ammonia and Nitrogen Oxides in the Air

An ammonia analyzer (Ecoech Serinus-44) was used to directly collect air samples and quantify the concentrations of NH₃, NO, NO₂, and NO_x in the air right above the electrolytic cell holding swine wastewater with or without electrolysis (Fig. 1). In experiments, the swine wastewater was well mixed using a magnetic stirrer (stir bar, length = 4 cm and diameter = 0.8 cm) operated at 320 rpm, which proceeded the thorough mixing of swine wastewater without vortex formation in the electrochemical reactor (diameter = 6 cm). The air entrance of sampling tube was placed at 4 cm right above the swine wastewater and the sampling space was three-side surrounded by a paper holder to reduce interference of air flow and provide temporary accumulation of species of interest for sampling. The method detection limits of these nitrogen compounds are all 0.4 ppb, while their upper limits of quantitation are the same (20 ppm). The reported concentrations of these four species in the air were all background (blank) corrected. These data were obtained in laboratories A and B, and the former had higher background air ammonia concentrations than the latter (~100 ppb).

Chemical Analysis of Swine Wastewater

The swine wastewater samples, obtained from two pig farms in southern Taiwan, were stored at 4°C before use. Prior to testing, each sample was filtered using a 0.45 µm filter to remove suspend solids. The methods set forth in the National Institute of Environmental Analysis (NIEA) (Taiwan) W517.52B, W203.51B, W448.51B, W418.53C, W419.51B, and W424.52A were used for the analyses of COD, conductivity (EC), ammonium nitrogen (NH₃-N), nitrite nitrogen (NO₂^–-N), nitrate nitrogen (NO₃^–-N), and pH respectively.

A COD analyzer (COD Reactor CR25, Rocker) coupled with a colorimeter (Hach-DR900) was used to analyze the COD concentrations of samples. The required digestion for the ammonia measurements was conducted by a Gerhardt VAP-200 Kjeldahl Nitrogen Distillation System. The NH₃-N in wastewater was spectrophotometrically measured at 640 nm (Hanna HI 96715) based on the reaction of ammonia with phenol and hypochlorite catalyzed with nitroprusside to form intensively blue indophenol in an alkaline medium (Berthelot reaction). Some ammonia measurements were also performed using a Hach-DR900 analyzer.

Sulfanilamide was used to diazotize the nitrite nitrogen of a sample followed by coupling with N-(1-naphthyl)-ethylenediamine dihydrochloride (NED), and the concentration of nitrite was then photometrically analyzed at 543 nm. Nitrate was also spectrophotometrically determined at 220 nm through deduction of the double absorbance at 275 nm for each sample. The chloride concentration was quantitatively determined using a redox potential titrator (Metrohm 702 SM Titrino), while the EC and pH were measured using an SC-170 conductivity meter (Suntex, Taiwan) and a TS-100 pH meter, respectively.

RESULTS AND DISCUSSION

Concentrations of NH₃ in the Air and NH₃-N in the Wastewater without Electrolysis

The air NH₃ samples were collected in two laboratories (A and B), so the background NH₃ concentrations in the air in the laboratories need to be measured before the tests of swine wastewater. In fact, the background NH₃ concentrations in the air in the laboratories varied with time. Figs. 2(a)–2(b) show two examples of the time-dependent concentration profiles of background air NH₃ collected in laboratory A. The air NH₃ concentrations fluctuated within the range of 262–404 ppb for a minute-based measurement (Fig. 2(a)), while those for an hour-based measurement ranged from 86–754 ppb (Fig. 2(b)). The latter had an increase in air NH₃ concentration during daytime, mainly due to the increase in worker activity during office hours in the laboratory, because the chemical composition of human perspiration includes urea and the ammonia nitrogen can be detected in human perspiration (Thurmon and Ottenstein, 1952).

Fig. 2. Air NH₃ concentration variations over time ((a) 2 and (b) 24 hrs) in laboratory A.

Figs. 3(a)–3(b) show the variation air NH₃ concentrations along with that of wastewater NH₃-N over time at similar initial wastewater NH₃-N concentrations (99.5 and 97.2 mg L^–1, respectively) in laboratory A. As can be seen from these two figures, the NH₃ concentrations in the air right above the electrolytic cell without operation were significantly higher that of background air. This phenomenon is associated with the release of NH₃ from the swine wastewater into the air, because ammonia is a volatile compound which may partition itself between the liquid and the gas above the liquid (Arogo et al., 1999). According to Arogo et al. (1999), the mass transfer rate of NH₃ from the swine wastewater into the air is proportional to the product of a mass transfer rate constant, the interfacial surface area, and a concentration driving force:

where m is the mass of ammonia, k is the overall mass transfer coefficient of ammonia, A_s is the interfacial surface area, C_w is the ammonia concentration in wastewater, C_air is the ammonia concentration in the air, and t is time.

Fig. 3. Air NH₃ and wastewater NH₃-N concentration variations over time at different initial wastewater NH₃-N concentrations or sampling durations ((a) 99.5 (8-hr), (b) 97.2 (8-hr), and (c) 69.1 (24-hr) mg L^–1) (laboratory A).

Moreover, the ammonia volatilization might be influenced by other factors such as air velocity, air temperature, liquid temperature, and system geometry. Although the fluctuation of NH₃-N concentrations in the swine wastewater was observed, it was almost negligible in comparison to that of NH₃ concentrations in the air. This trend was also found when increasing the hourly sampling from 8 to 24 hrs for the wastewater with a lower initial NH₃-N concentration (Fig. 3(c)). The gradual rise of ammonia concentration in the air above the swine wastewater with the increase of time is attributed to the continuing release of ammonia with an increasing time and its temporary accumulation in the sampling space.

Variations of NH₃ and Nitrogen Oxides Concentrations in the Air and of NH₃-N in the Wastewater with/without Electrolysis

Increasing over time like the trend shown in Figs. 3(a)–3(b), the air NH₃ concentrations (2186–12858 ppb) in Fig. 4(a) were significantly higher than those in Figs. 3(a)–3(b), because the swine wastewater (stirring [S] only) had a significantly higher initial NH₃-N concentration (661 mg L^–1). During the electrolysis (E) of this swine wastewater (S+E), the air NH₃ concentrations, also increasing over time at t ≤ 150 min, were significantly smaller (738–4985 ppb) than those without electrolysis; additionally, the decrease in air NH₃ concentration was observed at t > 150 min. This result is attributed to the decrease in driving force for ammonia volatilization (Eq. (3)) because of the decrease in wastewater NH₃-N concentration (from 661 mg L^–1 at t = 0 min to <0.01 mg L^–1 at t = 240 min) (pseudo first order rate constant of NH₃-N removal k_NH3-N = 1.33 × 10^–4 s^–1) (Table 1).

Fig. 4. (a) Air NH₃ and wastewater NH₃-N concentration variations and (b) air nitrogen oxides (NO, NO₂, and NO_x) concentration variations over time (E: electrolysis and S: stirring for wastewater) (laboratory B).

It is interesting to note that the NO concentration in the air was greater for the wastewater without electrolysis than for that with electrolysis, but an opposite tendency was observed for NO₂ in the air (Fig. 4(b)). However, their corresponding NO_x concentrations were similar at most of sampling intervals, although the NO concentrations were significantly higher at t = 90, 180, and 240 min for the wastewater with electrolysis than for that without electrolysis. This finding is related to the conversion of ammonia into other nitrogen compounds. Without electrolysis, it is known that two groups of bacteria, Nitrosomonas and Nitrobacter may cause the oxic conversion of ammonia to nitrite and nitrate, respectively. However, ammonia oxidizers are able to oxidize ammonia under anoxic conditions (Schmidt and Bock, 1997). According to Stuven and Bock (2001), significant NO and NO₂ production was observed during biological N-elimination from synthetic wastewater and sludge liquor from the municipal wastewater treatment plant wastewater with high ammonium concentration (about 700 mg N L^–1); moreover, Nitrosomonas was an efficient NO producer and significant NO₂ formed during simultaneous nitrification/denitrification. Nevertheless, various disinfecting oxidants such as chlorine, ozone, and hydroxyl radicals can be generated at BDD electrodes in NaCl solution (Yano et al., 1998). The free chlorine, active chlorine, or chloramines might involve in the inactivation of Nitrosomonas (europaea) (Chauret et al., 1998), which affected the production of NO and NO₂ from nitrification in the wastewater and thus the release of nitrogen oxides from wastewater into the air, leading to the variations of nitrogen oxides in the air.

During the electrolysis of swine wastewater containing 0.05 M NaCl in this study, the electrochemical production of active chlorine (Cl₂, HOCl, and OCl^–) could be achieved through the following two reactions (Kapałka et al., 2010).

Based on reactions 6–9, the HOCl from the above reaction could oxidize NH₃ or NH₄⁺ to produce nitrogen gas or nitrate (Li and Liu, 2009).

Or HOCl might react with NH₃ to form chloramines via reactions 10–12 (Kapałka et al., 2010; Pérez et al., 2012), while the formation of chloramines was indicated to be insignificant at pH > 8 (Pérez et al., 2012).

It was reported that ammonia can be electrochemically oxidized through several complicated steps to various nitrogen compounds such as N₂ and NO_x gases, although the former is more dominant than the latter (Wasmus et al., 1994; Kim et al., 2006). According to Kim et al. (2005), the ammonia was oxidized mainly to nitrogen gas through a direct electrolytic reaction in the basic solution, while ammonia could be partly decomposed to nitrogen by the hydroxyl radicals in the neutral and acid solution; besides, the ammonium ion could be oxidized to NO_x and nitrogen by electrochemically generated hypochlorite plus hydroxyl radicals in the neutral condition. This explanation is also appropriate for the finding of NO_x production from the electrochemical oxidation of ammonia in swine wastewater in this study. However, the evolution of oxygen on anode and that of hydrogen on cathode leading to an effect of ammonia stripping might also influence the rate of NO_x release from wastewater into the air.

Variations of Nitrite, Nitrate, and COD Concentrations over Time in Wastewater

The reactions on both anode and cathode might influence the production and decomposition of ammonia, nitrite, nitrate, and COD, which then resulted in the variations of NH₃ and NO_x concentration over time in the air. Fig. 5(a) shows the variations of nitrite and nitrate concentrations over time in wastewater. For comparison, the data of NH₃-N taken from Fig. 4(a) is also presented in Fig. 5(a). the concentration of NH₃-N decreased over time from 661 to ND, that of NO₂^–-N slightly increased initially and then decreased over time (0.14–2.36 mg L^–1), and that of NO₃^–-N significantly increased with an increasing time (88.5–259 mg L^–1) during the electrolysis of swine wastewater.

Fig. 5. Concentration variations in nitrogen compounds (NH₃-N, NO₂^–-N, and NO₃^–-N) (a) and COD and pH (b) over time in wastewater (Batch I, laboratory B).

As discussed in the previous section, the decrease in NH₃-N concentration over time was chiefly resulted from the reactions of free (or un-ionized) ammonia (NH₃) or ionized one (NH₄⁺) (pK_a of NH₄⁺ ↔ NH₃ = ~9.24 (Huang et al., 2008)) with electrochemically generated active chlorine, which led to the formation of NO₃^– or chloramines. Kapałka et al. (2010) concluded that the oxidation of ammonia at pH < 8 was mostly mediated by electro-generated in situ active free chlorine to enhance ammonia oxidation using BDD in the presence of chloride, because the dominant ammonia species was NH₄⁺ based on the pK_a = 7.50 for HOCl → OCl^– + H⁺ (Pérez et al., 2012). Nevertheless, they also indicated that direct (non-mediated) electrochemical oxidation of ammonia on BDD electrode proceeded mainly at pH > 8 via NH₃ oxidation.

In addition to N₂, NO₂^– and NO₃^– might also be produced on the anode via the further oxidation of NH₂OH or NO formed from electrochemical oxidation of ammonia (Kim et al., 2005) or by dissolution of nitrogen oxides, e.g., Eq. (15) in the aqueous phase (Bunce and Bejan, 2011).

In fact, these reactions are known to occur through more complicated steps. According to Bunce and Bejan (2011), the one-electron oxidation of water at the anode surface MO_n initially yields an adsorbed hydroxyl radical MO_nOH• (physisorbed active oxygen), which can be further oxidized to MO_n+1 (chemisorbed active oxygen) at noble metal oxides, such as Ti/IrO₂, Ti/RuO₂, and PtO_x. The MO_n+1 can initiate oxidation of a substrate by a two electron, one oxygen atom transfer mechanism. In the case of ammonia, this can form H₃N⁺–O^–, tautomeric to NH₂OH; however, at anodes where MO_nOH• cannot be oxidized further, such as PbO₂, Ti/SnO₂, and BDD, the anode acts as a source of hydroxyl species in its reactions with external substrates (including ammonia). Nevertheless, N₂, N₂O, and NO₂ may be formed from ammonia electrochemical oxidation on BDD in basic solution, and •OH might react with NH₃ to form NH₂•_ad (Eq. (16)), although the usual mechanism for oxidation of an extrinsic substrate at BDD is not yet clear (Bunce and Bejan, 2011).

The NH₄⁺ in the neutral or acid solution can also be partly decomposed to nitrogen by •OH (with a decomposition rate much lower than that in the basic condition), not by a direct electrode reaction (Kim et al., 2005).

The nitrite and nitrate could be reduced on cathode via the following reactions (Pérez et al., 2012), which should also participate in the variation in NH₃-N, NO₂^–-N, and NO₃^–-N concentrations shown in Fig. 5(a).

The small nitrite concentration detected in our experiments was related to its almost instantaneous oxidation by OCl^– in the presence of chloride during electrolysis (Pérez et al., 2012).

On the other hand, the concentration of COD decreased from 6266 to 397 mg L^–1 over time in wastewater electrolysis (Fig. 5(b)) (pseudo first order rate constant of COD removal k_COD = 1.21 × 10^-4 s^–1 in Table 1). The electro-generated •OH was principally responsible for the degradation and mineralization of COD (Li and Liu, 2009; Sirés et al., 2014; Radjenovic et al., 2015; Moreira et al., 2017). Moreover, the solution pH also decreased from 7.86 to 4.68 with time at t ≤ 180 min but it began to rise for the electrolysis after 180 min (Fig. 5(b)). The reactions 4–9 and 13–17 may cause the variation of pH in solution; furthermore, the electrochemical reactions for the evolution of oxygen on anode and that of hydrogen on cathode also involve in the pH change in solution. It was reported that the electrochemical production of •OH were not favored in solution at pH > 9 [25]. Cl^– may react with •OH to generate OCl^–, which can be further oxidized by •OH to from ClO₂^–, ClO₃^–, and ClO₄^– stepwise (Kapałka et al., 2010).

Effect of Anode Material on NH₃ and Nitrogen Oxides Concentration Variations in the Air

Fig. 6(a) shows the increase in air NH₃ concentration over time for the non-electrolyzed swine wastewater with an initial NH₃-N concentration of 107 mg L^–1, similar to the trend depicted in Fig. 4(a). However, the air NH₃ concentration was greater in Fig. 4(a) than in Fig. 6(a), because the initial NH₃-N concentration in wastewater was higher for the former than for the latter. The electrolysis of swine wastewater (initial NH₃-N concentration = 98.2 mg L^–1) using a Ti/PbO₂ anode exhibited an increase in air NH₃ concentration at t ≤ 60 min and then decreased at t > 60 min (Fig. 6(b)), while the NH₃-N concentration in wastewater decreased over time, also similar to the tendency shown in Fig. 4(a). This trend was also observed when replacing Ti/PbO₂ with BDD-I as the anode for the electrolysis of wine wastewater with an initial NH₃-N concentration of 82.0 mg L^–1) (Fig. 6(c)). However, the NH₃-N concentration reached ND at 60 min for using the BDD-I anode, but the NH₃-N concentration decreased smaller after 60 min and was only close to 10 mg L^–1 for using the Ti/PbO₂ one. The removal rate of NH₃-N was faster using the BDD-I than using the Ti/PbO₂ (k_NH3-N = 4.83 × 10^–4 and 4.63 × 10^–4 s^–1, respectively) (Table 1); hence, the increase in air NH₃ concentration was faster when using the BDD-I.

Fig. 6. Air NH₃ and wastewater NH₃-N concentration variations over time for the (a) wastewater without and (b) with electrolysis on Ti/PbO₂ (Batch II-01), and (c) BDD-I (Batch II-02) (laboratory A).

More anodes were tested to further explore the variations of air NH₃ and nitrogen oxides concentrations during the electrolysis of swine wastewater. When using BDD-I as the anode, the air NH₃ concentration increased with increase in electrolysis time, reached a maximum (4450 ppb) at 60 min, and then decreased over time (Fig. 7(a)). It was noticed that the air NH₃ concentrations for using different anodes in wastewater electrolysis followed the order BDD-I (Nb/BDD) >> graphite/BDD > graphite > graphite/PbO₂ ≈ Ti/PbO₂ (Fig. 7(a)). Therefore, using BDD anodes resulted in more release of NH₃ from wastewater into the air than using PbO₂ or graphite, regardless of the difference in electrode substrate material. Similarly, the BDD-I anode still exhibited a significantly greater release of NO (maximum = 8.1 ppb) from wastewater into the air than the other tested anodes (Fig. 7(b)). However, the air NH₃ concentrations for using graphite/BDD and graphite anodes were similar, and this tendency was also true for using graphite/PbO₂ ≈ Ti/PbO₂ ones. Like the trend of NO release from wastewater into the air, the air NO₂ concentrations for using these anodes were in the order BDD-I (Nb/BDD) (maximum = 216 ppb) >> graphite/BDD ≈ graphite > graphite/PbO₂ ≈ Ti/PbO₂ (Fig. 7(c)). Consequently, this trend was also true for the air NO_x concentrations (BDD-I, maximum = 222 ppb) when using these anodes. The significant deviation in air NO_x concentration resulted from using BDD-I (Nb/BDD) and graphite/BDD is probably attributed to the BDD crystalline structure was significantly better for the former than for the latter, and the graphite/BDD should significantly had more graphite structure than the BDD-I. Further study on electrode surface analysis is necessary to verify this point. Accordingly, NO_x formation was more favorable on BDD than on PbO₂ or graphite. More study is necessary to explore the reaction mechanisms related to this issue (see more discussion in the next section).

Fig. 7. Concentration variations in (a) NH₃, (b) NO, (c) NO₂, and (d) NO_x over time in the air (Batch II-1 to Batch II-5, laboratory A).

Effect of Anode Material on NH₃-N, NO₂^–-N, NO₃^–-N, and COD Concentration Variations in the Wastewater

The performance of NH₃-N degradation in wastewater for using the five tested anodes was in the order graphite/PbO₂ ≈ Ti/PbO₂ > BDD-I >> graphite/BDD > graphite (Fig. 8(a)), consistent with their pseudo first order rate constants of NH₃-N removal (k_NH3-N = 2.98 × 10^–4–6.35 × 10^–4 s^–1) (Table 1). This result is similar to another study which used swine wastewater with comparable NH₃-N and higher COD concentrations (Huang et al., 2019). The better removal rate of NH₃-N on PbO₂ than on BDD and graphite in wastewater was associated with the inherently different nature in electrode material and the mechanisms for ammonia removal. The difference in initial NH₃-N concentration (76.3–90.0 mg L^–1) was small (Table 1), so its influence on the NH₃-N removal should be minor in comparison to that of anode material.

Fig. 8. Concentration variations in (a) NH₃-N, (b) NO₂^–-N, (c) NO₃^–-N, and (d) COD_t/COD₀ over time in wastewater (Batch II-1 to Batch II-5, laboratory A).

As discussed earlier, the dominant ammonia species is NH₄⁺ in neutral or acidic aqueous solution, which is not favored for the direct (non-mediated) oxidation of ammonia on anode (Pérez et al., 2012), while NH₃ can be directly oxidized on BDD in basic solution (Kapałka et al., 2010), and both NH₄⁺ and NH₃ can be oxidized by •OH (Kim et al., 2005; Bunce and Bejan, 2011). The oxidation rate of ammonia was also slow by the •OH generated from using H₂O₂ plus UV irradiation (Huang et al., 2008). However, for the swine wastewater with the addition of 0.05 M NaCl in this study, the electro-generated active chlorine was more useful for mediated oxidation of ammonia than electro-produced •OH (Kapałka et al., 2010; Pérez et al., 2012). On the other hand, the BDD or graphite anode exhibited less formation of nitrite (Fig. 8(b)) or nitrate (Fig. 8(c)) than the PbO₂ one during the electrolysis of wastewater, also supporting that the electro-generated active chlorine dominated for the ammonia removal in wastewater, although the Ti cathode was also responsible for the variations of nitrite and nitrate concentrations.

Fig. 8(d) shows that the graphite/BDD and graphite anodes exhibited worse performance in COD removal (k_COD = 4.50 × 10^–5–1.55 × 10^–4 s^–1) than the other tested anodes (k_COD = 5.13 × 10^–5–1.19 × 10^–3 s^–1) (Table 1). This phenomenon is associated with the fact that the performance in electrochemical production of •OH in water solution is better on non-active electrodes (e.g., BDD and PbO₂) than on active electrodes (e.g., graphite and Pt); moreover, the electro-generated •OH is powerful for the degradation and mineralization of organic pollutants in solution (Comninellis, 1994; Sirés et al., 2014; Radjenovic et al., 2015; Moreira et al., 2017). According to Table 1, the initial COD concentration also affected the COD removal in wastewater. For example, the initial COD concentration in Batch I wastewater (6266 mg L^–1) was significantly higher than that in Batch II-1 (542 mg L^–1), so the former had a smaller k_COD (1.21 × 10^–4 s^–1) than the latter (5.13 × 10^–4 s^–1) when using the same BDD-I anode. In addition to anode material, the initial concentration of COD or NH₃-N also influenced its specific energy consumption in electrolytic removal (E_SP-C and E_SP-N, respectively). For comparison and consistency, the NH₃-N and COD reached ND at 0.5 (Fig. 8(a)) and 2 hrs (Fig. 8(d)), respectively, so the data of 0.5-hr E_SP-N and 2-hr E_SP-C are provided in Table 1. The E_SP-C (0.052–0.853 kWh g-COD^–1) or E_SP-N (0.316–6.703 kWh g-NH₃-N^–1) was lower as its initial concentration was higher. Therefore, the E_SP-C (0.052 kWh g-COD^–1) and E_SP-N (0.316 kWh g-NH₃-N^–1) of this study for BDD-I were lower than those in another our study (0.323 kWh g-COD^–1 and 1.398 kWh g-NH₃-N^–1, respectively) which had lower initial COD and NH₃-N concentrations (1025 and 86 mg L^–1), respectively (Huang et al., 2019). According to mass balance and using 1 L (10 cm × 10 cm ×10 cm) for the space volume above the wastewater, the total nitrogen mass (the mass of NH₃-N and NO_x-N in air plus that of NH₃-N, NO₂^–-N, and NO₃^–-N in wastewater) at 0-hr deducts that at 2-hr is denoted as ΔN_T  and the data of ΔN_T  for different runs are also given in Table 1. The ΔN_T  value of Batch I wastewater was 89.8 mg, significantly greater those of Batch II wastewater 

(9.97–17.4 mg), because the former had significantly higher ammonia concentration than the latter. TheΔN_T should be present as the other forms of nitrogen compounds (e.g., N₂ and N₂O), and the change in organic nitrogen concentration in the wastewater should also involve in the change of ΔN_T.

CONCLUSIONS

In this study, the NH₃ concentration in the air right above the swine wastewater without electrolysis was found to increase over time and this increase of air NH₃ concentration was greater when the wastewater NH₃-N concentration was higher. During the electrolysis of this swine wastewater using each of five tested anodes, the air NH₃ concentrations, increasing to a maximum and then decreasing over time, were smaller than those without electrolysis, mainly because the decrease in driving force for ammonia volatilization resulted from the decrease of wastewater NH₃-N concentration over time. The NO concentration in the air was greater for the wastewater without electrolysis than for that with electrolysis but an opposite tendency was observed for NO₂ in the air, while their corresponding NO_x concentrations were similar at most of sampling intervals.

Both anode material and initial pollutant concentration influenced the release of NH₃ and nitrogen oxides from wastewater into the air and the variations in NH₃-N, NO₂^–-N, NO₃^–-N, and COD concentrations in the wastewater. The air NH₃ concentrations for using different anodes in wastewater electrolysis followed the order BDD-I (Nb/BDD) >> graphite/ BDD > graphite > graphite/PbO₂ ≈ Ti/PbO₂, while for air NO, NO₂, or NO_x concentrations the order slightly shifted to: Nb/BDD >> graphite/BDD ≈ graphite > graphite/PbO₂ ≈ Ti/PbO₂. The graphite/BDD and graphite anodes exhibited worse performance in NH₃-N (k_NH3-N = 2.98 × 10^–4–6.35 × 10^–4 s^–1) or COD electro-degradation (k_COD = 4.50 × 10^–5–1.19 × 10^–3 s^–1) than the other tested ones, but they exhibited less formation of nitrite or nitrate than the PbO₂ anodes. The specific energy consumption of NH₃-N or COD was lower as its initial concentration was higher.

ACKNOWLEDGEMENT

The authors would like to thank the Ministry of Science and Technology, Taiwan for financially supporting this research under Grant No. MOST 106-2221-E-020-003-MY3. We are also grateful for the financial assistance of Prosperity Sci-Tech Co., Ltd. under a project (10700245) financially supported by the Kaohsiung Environmental Protection Bureau.