Environ Sci Pollut Res DOI 10.1007/s11356-015-4467-x

RESEARCH ARTICLE

Factors controlling nitrous oxide emissions from a full-scale activated sludge system in the tropics Ariane C. Brotto 1,3 & Débora C. Kligerman 2 & Samara A. Andrade 1 & Renato P. Ribeiro 1 & Jaime L. M. Oliveira 2 & Kartik Chandran 3 & William Z. de Mello 1

Received: 12 December 2014 / Accepted: 30 March 2015 # Springer-Verlag Berlin Heidelberg 2015

Abstract Despite interest in characterizing nitrous oxide (N 2 O) emissions from wastewater treatment plants (WWTPs) in several parts of the globe, there are few studies in tropical zones. This study focus on the contribution of the scientific knowledge of anthropogenic nitrogen greenhouse gas emissions to climate change in tropical countries, investigating factors controlling N2O emissions in a non-biological nitrogen removal municipal WWTP. In terms of operational parameters, dissolved oxygen (DO) concentrations displayed a biphasic impact on N2O production and emission, with the highest emission at DO of 2.0 mg O2 L−1. The low solids retention time of 3 days also played a significant role, leading to nitrite accumulation, which is an important trigger for N2O production during nitrification. Furthermore, other factor especially important for tropical countries, namely, temperature, also had a positive correlation with N2O production. Emission factors estimated for this study were 0.12 (0.02–0.31)% of the influent total nitrogen load and 8.1 (3–17) g N 2 O person−1 year−1, 2.5 times higher than currently proposed Responsible editor: Bingcai Pan Electronic supplementary material The online version of this article (doi:10.1007/s11356-015-4467-x) contains supplementary material, which is available to authorized users. * Débora C. Kligerman [email protected] 1

Departamento de Geoquímica, Instituto de Química, Universidade Federal Fluminense, Niterói, RJ 24020-141, Brazil

2

Departamento de Saneamento e Saúde Ambiental, Escola Nacional de Saúde Pública, Fundação Oswaldo Cruz, Rio de Janeiro, RJ 21041-210, Brazil

3

Department of Earth and Environmental Engineering, Columbia University, New York, NY 10027, USA

emission factors. Therefore, the highly variability and dependence on operational parameters reinforce the use of a single emission factor is inadequate, especially for developing countries with limited or variable extent of biological wastewater treatment and in regions of the world with widely varying climate patterns. Keywords Greenhouse gas . Nitrous oxide emission . Wastewater treatment . Activated sludge . Nitrite . Solids retention time

Introduction Emissions of nitrous oxide (N2O) from wastewater treatment processes are gaining increased recognition as the collective action on the reduction of anthropogenic greenhouse gas (GHG) emissions has been globally recognized (UNFCCC 2007). Nitrous oxide has a global warming potential of approximately 300 times that of CO2 at a 100-year lifespan and it is the main ozone-depleting substance in the twenty-first century (Intergovernmental Panel on Climate Change (IPCC) 2007; Ravishankara et al. 2009). Domestic wastewater treatment systems contribute with 1.6 % of the total worldwide N2O emissions in 2010, and over 20 years, those emissions increased by 1.6 Tg CO2 eq. (46 %) (USEPA 2012). The processes of N2O production in biological wastewater treatment results especially from biochemical transformations, called nitrification and denitrification, and are essentially those responsible for N2O production in soils, sediments, and water bodies as well (Wrage et al. 2001). Nitrification is a two-step biological process, first the oxidation of ammonia (NH3) or ammonium (NH4+) to nitrite (NO2−) carried out by ammonia-oxidizing bacteria (AOB), followed by the oxidation of NO2− to nitrate (NO3−) by nitrite-oxidizing bacteria

Environ Sci Pollut Res

(NOB) (Grady et al. 1999). Nitrous oxide is a by-product of the first step, termed nitritation, under both oxic and O2-limiting conditions (Chandran et al. 2011; Wrage et al. 2001). During denitrification, the sequential reduction of NO3− to nitrogen gas (N2) under anoxic conditions, N2O is a regular intermediate and is produced as a result of several conditions, such as low pH, organic carbon limitation, type of carbon source, dissolved oxygen (DO), and nitrite inhibition (Lu and Chandran 2010; Wrage et al. 2001). Denitrification in anoxic zones is associated as the dominant source of N2O in WWTPs (IPCC 2006); however, recent studies have indicated that nitrification, in aerobic zones, could also play a role in N 2 O generation and emission (Ahn et al. 2010a, b; Kampschreur et al. 2008). Despite of a wide variation range related to the fraction of total nitrogen (TN) conversion to N2O gas in full-scale surveys (0–25.3 %) (Ahn et al. 2010b; Foley et al. 2010; Rassamee et al. 2001), few studies have been conducted in both developing and tropical countries (Foley et al. 2010; Brotto et al. 2010; De Mello et al. 2013; Ribeiro et al. 2013). In Brazil, sewage from some of the major cities has been treated using centralized systems, such as conventional activated sludge, for biodegradable organic matter removal and, occasionally, also removal of nitrogen (N) but at the cost of producing substantial amounts of N2O, given that the process is not designed for N removal (Chandran et al. 2011). On the other side, in Europe and in the US, regulations concerning nutrient removal (e.g., N and phosphorus) has become increasingly restrictive and have spurred efforts to develop innovative biological nitrogen removal (BNR) systems (Ahn et al. 2010a). In addition, both water consumption and wastewater generation are influenced by climate factor. In tropical countries, a greater consumption of water is found, and consequently, domestic wastewater has distinct characteristics as of developed countries, including higher influent flow, lower organic matter, and TN load (von Sperling 2005). Furthermore, higher temperatures promote increase in the growth rate of microorganisms and, successively, favoring the degradation of organic matter and intensifying removal efficiencies (Rittman and McCarty 2001). Given that N2O emissions are strongly impacted by microbial activities, which in turn are controlled by temperature and wastewater loading, emission patterns from different parts of the globe could be rather distinct. Thus, the use of a single emission factor (EF) of 3.2 g N2O person−1 year−1 for non-biological nitrogen removal (non-BNR) (USEPA 2012; IPCC 2006) is perhaps inadequate for wastewater treatment systems in developing countries, where more than 80 % of the world population lives (Brazilian Institute of Geography and Statistics, acronym in Portuguese 2001), as well as for tropical countries, as temperature plays a decisive role on microbial transformations. It was hypothesized for this study that: (1) N2O emissions from WWTPs in the tropics would be higher than emissions in

temperate zones, given that microbial activity increases with increasing temperature and consequently intensifies conversion processes and production of N2O; (2) although N2O emissions from non-BNR systems are considered lower than from BNR, in developing countries the EF for non-BNR likely exceed the EF from BNR operations, assuming that WWTPs herein are either not adequately operated or are under-designed to accommodate full-nitrification and denitrification. Therefore, the principal motivation of conducting this detailed monitoring campaign was the necessity to expand the database for N2O emissions from WWTPs in developing countries and in tropical climates, and to gain a better understanding of the operating parameters responsible for the rise in N2O emissions in such systems. Based on these gaps, the specific objectives of this study were: (1) to determine the production and emission of N2O in non-BNR systems in a developing country; (2) to infer controlling factors that are related to N2O emissions in activated sludge WWTPs in the tropics, namely, nitrite, pH, DO, and temperature.

Experimental WWTP location and characteristics This study was carried out in a non-BNR activated sludge WWTP located in one of the cities that form the Metropolitan Region of Rio de Janeiro. The WWTP serves approximately 50,000 people and process nearly 14.7×103 m3 day−1 of domestic wastewater. This process is composed of preliminary and secondary treatment units, including screening and grit removal, aeration tank (AT), rectangular settling tank, and return activated sludge line (Fig. 1). The AT has surface area of 960 m2 and volume of 5, 760 m3. Aeration occurs continuously along the AT, with no anoxic zones for pre- or post-denitrification, at an air flow rate of 5.4×103 m3 h−1. The hydraulic residence time (HRT) and solids retention time (SRT) are 9 h and 3 days, respectively. Sampling and analysis Nitrous oxide measurement campaigns took place once a month from January to July. Nitrous oxide concentrations were measured at seven distinct points located along the activated sludge aerobic tank to capture spatial variability in emissions. The following tasks were carried out at each of the seven sampling points: (1) in situ measurements of the mixed-liquor temperature, DO, and pH with a Hanna HI988 multi-parameter probe; (2) collection of air released from the surface of the mixed-liquor, as a result from its aeration process, in four replicates for determination of N2O concentration; (3) collection of atmospheric air; (4) collection of mixedliquor sample for immediate extraction and later

Environ Sci Pollut Res

Fig. 1 Schematic diagram of the studied full-scale activated sludge WWTP (RAS: return activated sludge; WAS: waste activated sludge)

determination of dissolved N2O concentrations; and (5) collection of mixed-liquor grab sample for the determination of ammonium (NH4+), nitrite (NO2−), nitrate (NO3−), total nitrogen (TN), and chemical oxygen demand (COD), in triplicate, only in the last four campaigns (April–July). In order to determine N2O concentration in the air released from the surface of the mixed-liquor, an upturned funnel device was used (Brotto et al. 2010). The technique allows the capture of air bubbles released from the aerated surface by a polyethylene funnel of 30 cm diameter (area=0.071 m2) attached to a 3.5 cm (inner diameter) polyvinyl chloride (PVC) pipe, which is maintained semi-immersed in the mixed-liquor during sampling (Fig. 2). The other end of the PVC pipe is opened, allowing the air arising from the surface of the mixedliquor to pass through the pipe. The mass flow rate of air

entering the funnel, based on the AT area, the flow of air injected in the AT and the area of the funnel mouth, is 6.7 L min−1. The residence time of air, emitted from surface of the mixed-liquor, inside the funnel-PVC tube device is approximately 1 min. Inside the PVC pipe, there is a flexible plastic tube with 4 mm of inner diameter, which extends from the middle funnel neck to the other end of PVC pipe, through which the sample is withdrawn using a 20-mL polypropylene syringe. For gas sample collection, approximately 2/3 of the funnel is kept immersed in the mixed-liquor and sampling starts 3 min later, time required to homogenize air inside the funnel (funnel headspace). Four replicate samples were collected for analyses from each point to ensure stability of N2O concentration inside the funnel headspace. Before this procedure, inner air (ca. 60 mL) was withdrawn and discarded from the inner plastic tube using the same sampling syringe. During the exchange of sampling points within the AT, funnel was moved laterally without being removed from the mixed-liquor, preventing the entry of atmospheric air inside the funnel headspace. Nitrous oxide emission flux (F) was calculated based on the mass flow rate (Q) of injected air into the AT, the difference between N2O concentrations in the air released from the mixed-liquor and the atmospheric N2O concentration (ΔC) and the superficial area (A) of the AT, as follows: F¼

ðQ x ΔC Þ A

ð1Þ

To determine the concentration of N2O dissolved in the mixed-liquor, a 30-mL liquid sample was collected with a 60-mL syringe, followed immediately by 30 mL of ambient air (Brotto et al. 2010). The syringe was agitated vigorously for 2 min, and the air in the syringe headspace was transferred to a dry syringe interconnected via a three-way valve. Nitrous oxide concentration in the aqueous phase (ηmol L−1) was estimated using the following equation (Guimarães and de Mello 2008): Fig. 2 Upturned funnel device for the capture of air bubbles released from the aerated surface (adapted from Brotto et al. 2010)

     C aq ¼ K o xC hsð f Þ þ P RT x C hsð f Þ −C hsðiÞ

ð2Þ

Environ Sci Pollut Res

where Chs ( f ) e Chs (i) are, respectively, the final (after agitation) and initial (ambient air) N2O concentration in the syringe headspace (mol), Ko is the equilibrium constant expressing the N2O solubility (mol L−1 atm−1) (Weiss and Price 1980), P is the atmospheric pressure (1 atm), R is the gas constant (0.08205601 L atm K−1 mol−1), and T is the liquid temperature (K). Nitrous oxide concentrations were determined by gas chromatography, in a Shimadzu model GC-17A, with an electron capture detector (Ni(63)) operated at 340 °C, using argon with 5 % methane as carrier gas at a flow rate of 40 mL min−1. The air samples were injected in a 2 cm3 loop coupled to a stainless steel six-way valve. The separation column used was a packaged Porapak Q (80–100 mesh; 3 mm × 3.2 m). Two N2O analytical standards, with 356 and 840 ppb (White Martins Gases Industriais S.A.) were used for N2O quantification. All analyses were performed within 8 h of the sample collection. Analytical precision was of ±1 %, which is the coefficient of variation based on five replicates of those two standards. Influent, mixed-liquor, and effluent wastewater grab samples were collected in 500-mL polypropylene flasks, stored in an insulated container with artificial ice packs, transported to the laboratory, filtered through 0.22 μm diameter pore cellulose acetate membranes (Millipore®), and stored at −20 °C for subsequent determination of NH4+ (Indophenol Blue), NO2− (Griess diazotization reaction), and NO 3− (ion-selective electrode) all according to standard methods (Eaton et al. 2005). For TN and COD determination, unfiltered samples were acidified with concentrated H2SO4 at pH 2. Total N was determined spectrophotometrically following sample digestion with potassium persulfate solution at 105 °C using the Hach method 10072 (Hach, Loveland, CO). COD concentration was determined spectrophotometrically following sample acid digestion with H2SO4/K2Cr2O7 solution, with addition of H2SO4/Ag2SO4 solution to catalyze oxidation of certain organic compounds, according to Standard Methods (Eaton et al. 2005). All analyses were performed in triplicate. Statistical analysis Statistical analyses were performed using STATISTICA 7.0 for Windows software (StatSoft Inc., OK). All data were expressed as mean±standard deviation (SD) and median. Nitrous oxide emission rates, DO, NO2−, pH, and temperature were examined for statistical distribution to determine the appropriate form of statistical analysis. Those data were found to be not normally distributed, requiring the application of nonparametric statistics to the data, which enabled the Spearman’s correlation between them. The significance for all p values was 0.01.

Results and discussion Characteristics of the activated sludge and performance of the full-scale process The process performance of the full-scale wastewater treatment system during the study is summarized in Fig. 3. The TN and COD concentrations in the wastewater influent were 46±9.0 mg N L−1 and 460±117 mg L−1, respectively, characterizing it as a medium-strength wastewater (Jordão and Pessôa 2005). The average NH4+-N concentration in the wastewater influent was 31±7.4 mg N L−1. Moreover, concentrations of both NH4+-N and TN in the influent increased during the winter (June and July) due to the lower water usage on this period of the year, a typical pattern in tropical countries. The average (and median) removal efficiency in relation to the COD was 73 % (84 %) and, relative to the TN, was 52 % (51 %), which were both below the range considered for activated sludge wastewater treatment (von Sperling 2005; Jordão and Pessôa 2005). In addition, the removal efficiency in relation to NH4+-N was 52 % (40 %). Additional nitrogen mass balance can be found in Online Resource 1. The removal efficiency can be inferred in terms of process design and operation. In general, for medium-strength wastewater, it is required SRT ranging from 4 to 10 days due to slow growth rates of nitrifying bacteria and HRT from 6 to 8 h, for a process called conventional activated sludge (von Sperling 2012). However, the WWTP in the study operates at a limiting SRT varying from 1 to 3 days and greater HRT varying from 7 to of 13 h. It demonstrates that this plant is under-designed to operate and to achieve the removal efficiency of COD and TN required for the influent loading. Additionally, as N2O emissions are intrinsically linked with process design and the propensity for N2O emissions of any given WWTP configuration relies both on process operation and performance characteristics, this plant is anticipated to release significant amounts of N2O to the atmosphere for not achieving either complete nitrification or denitrification. In the AT, the average NH4+-N concentration was 18.0± 8.6 mg N L−1 while NO2−-N concentration in the range of 0.3 ±0.2 mg N L−1 and detected NO3−-N concentration of 0.9± 1.0 mg N L−1. These data suggest, once again, that the applied operational parameters conditions, such as SRT and HRT, are not in conformity to what is demanded in order to achieve required removal. Nitrous oxide emission factors Dissolved N2O concentration in the mixed-liquor varied from 45.4 to 425 ηmol L−1 and was over-saturated in relation to its concentration in the atmosphere (∼323 ppb), with saturation ratio ranging from 1311±30 % to 5,388±28 %. Nitrous oxide concentrations and emissions showed a significant temporal

Environ Sci Pollut Res Fig. 3 Variation of influent (black circle) and effluent (white circle) concentration and removal efficiency (gray triangle) of a COD (mg L−1), b TN (mg N L−1), and c NH4+ (mg N L−1) in the fullscale activated sludge system during the study period

variation during the period of the study (Table 1). Nitrous oxide production in the liquid-phase and emissions to the atmosphere (April–July) have a strong linear correlation in

activated sludge treatment system (p=0.76, n=47, p

Factors controlling nitrous oxide emissions from a full-scale activated sludge system in the tropics.

Despite interest in characterizing nitrous oxide (N2O) emissions from wastewater treatment plants (WWTPs) in several parts of the globe, there are few...
752KB Sizes 1 Downloads 7 Views