SWINE WASTEWATER NITROGEN REMOVAL AT DIFFERENT C/N RATIOS USING THE MODIFIED LUDZACK-ETTINGER PROCESS

2 Federal University of Technology Paraná, UTFPR/PPGEA/ Francisco Beltrão PR, Brazil. 3 Federal University of Technology Paraná, UTFPR/ Dois Vizinhos PR, Brazil. 4 University of the Contestado/ Concórdia SC, Brazil. 5 Embrapa Suínos e Aves/ Concórdia SC, Brazil. Received in: 6-20-2018 Accepted in: 9-14-2018 Engenharia Agrícola, Jaboticabal, v.38, n.6, p.968-977, nov./dec. 2018 Doi: http://dx.doi.org/10.1590/1809-4430-Eng.Agric.v38n6p968-977/2018


INTRODUCTION
Worldwide increasing demand for meat has caused the establishment of large concentrated animal feeding operations (CAFOs) for livestock production in order to reduce animal production costs and soil demanded for effluent disposal. Usually intensive breeding farms confine large numbers of animals in an area of land proportionally small, breaking the relationship between crops and animal production. Thus, large amounts of manure are applied to the soil without undergoing any kind of treatment or stabilization, and without considering agronomic and legal criteria (Kunz et al., 2005). According to Williams (2008), much of the environmental impact generated by swine farming is a result of the lack of adequate management of solid and liquid waste. The waste produced in CAFOs often exceeds the amount that can be used as biofertilizer due to land requirement limitations (Kinyua et al., 2014). Nutrient removal, mainly nitrogen and phosphorus from different wastewaters, has become the most important concern for the wastewater treatment plants in the past three decades. Nitrogen can be removed from wastewater by biological and physical chemical processes (Kunz & Mukhtar, 2016).
The biological Nitrification/Denitrification (NDF) process (Equation 1 and 2) is worldwide used for effluent nitrogen removal when an available carbon source is not a problem. In this process, during the aerobic and autotrophic step (nitrification) NH4 + is oxidized to NO3followed by the anoxic and heterotrophic step (denitrification) where NO3is converted to gaseous Engenharia Agrícola, Jaboticabal, v.38, n.6, p.968-977, nov./dec. 2018 nitrogen (N2) (Lan et al., 2011). Nitrogen removal via nitrite Nitritation/denitritation (NDT) (Eq. 3 and 4) was reported to be technically feasible and economically favorable, especially when treating wastewater with high nitrogen concentrations and low C/N ratios (Yang et al., 2007). According to Peng & Zhu (2006), the methods used to accumulate nitrite are based on the appropriate regulation of temperature, pH, dissolved oxygen, cell retention time and initial nitrogen concentration.
Nitrogen removal via NDT can be achieved by Nitrite Oxidizing Bacteria (NOB) inhibition (e.g.. DO restriction, sludge wash out), due to different physiological characteristics and responses to operating conditions in order to restrict nitrite oxidation to nitrate (Ge et al., 2015). For anoxic phase (denitrification or denitritation) the availability of biodegradable organic carbon becomes critical in wastewater with high ammonia concentration and low C/N ratio. An alternative to solve this problem is the use of NDT process saving carbon during denitritation step. For this, a possible alternative is the DO reduction, for NOB activity restriction. This process has attracted attention in recent years, especially for treatment of wastewater with low C/N ratio (Gujer, 2010;Ge et al., 2015). In this way, nitritation can be obtained and maintained at low DO concentration.
The present study aims the NDT establishment by DO restriction evaluating the influence of different C/N ratios on TOC (Total Organic Carbon) consumption and N removal efficiency comparing NDF and NDT for swine wastewater nitrogen removal.

Wastewater Sampling and Characterization
The wastewater samples were collected from a swine manure treatment system (SMTS) at Embrapa Swine and Poultry, located in Concordia, Santa Catarina, Brazil. For the experiments, were used the primary settling tank effluent (STE), and the UASB reactor effluent (URE), in order to have samples at different C/N ratios (Table1).

Experimental design
A schematic representation of the experimental MLE (modified Ludzak-Etinger) process configuration, that was under operation at laboratory scale in Embrapa Swine and Poultry, is presented in Figure 1. The aeration was provided using an air pump (Big Air, A420) and controlled by a dissolved oxygen controller (Hach, SC200). The anoxic reactor was continuously fed (Qin = 1.5 mL min -1 ) for a nitrogen loading rate of 0.35±0.07 kg m -3 d -1 of N), the recirculation rate was equivalent to 5.5*Qin and the sludge recirculation rate was 1*Qin using peristaltic pumps (Masterflex, 7518-60). The anoxic and oxic reactors were maintained under continuous stirring (IKA-RW90) at 15 rpm and 200 rpm, respectively. The study was conducted in four different phases: Phase I (presence of AOB and NOB at high DO concentration, NDF process), between days 1 -37, DO concentration in the oxic reactor between 2.0 -3.0 mg L -1 of O2. The system influent was composed by the STE presenting a TOC and NH3-N concentration of 1,100 mg L -1 of N and 719 mg L -1 of N respectively, C/N ratio of 1.5.
Phase II (DO concentration reduction favoring inhibition of NOB, NDT process) was conducted between days 38-101. At the first 20 days (38-53) the DO concentration in oxic tank was kept between 0.5 to 0.6 mg L -1 of O2, from day 54 to 67, the DO concentrations ranged from 0.4 to 0.5 mg L -1 of O2, and between days 68-101 with DO concentration between 0.6 to 0.7 mg L -1 of O2. The system influent was fed with SMTS STE with TOC and NH3-N concentration of 1,097 mg L -1 and 734 mg L -1 of N respectively, resulting in a C/N ratio of 1.5.

Oxygen Uptake Rates (OUR) on NDF and NDT Activity
The respirometer, consisted of a 400 mL glass conical flask with three exits at the top for an oxygen probe insertion (Hanna, HI 98186), a pH probe (Hanna, HI 98183) and the injection of the ammonium solution (NH4Cl) (De Prá et al., 2016). All nitrifying (Phase I: NDF and Phase II: NDT) respirometric tests were carried out at the same biomass concentration (0.47 g L -1 of VSS) with the biomass provided from the oxic reactor. A nutrient solution without total ammonia nitrogen (TAN), described by Campos et al. (1999), was prepared to carry out the washing of biomass between tests through suspension, centrifugation and discarding the supernatant. Once the mixed liquor reached DO saturation (9.0±0.5 mg L -1 of O2), ammonia substrate (NH4Cl) Where: SOUR: Specific Oxygen Uptake Rate (mgO2. gSSV -1 . min -1 ); OUR: Oxygen Uptake Rate (mg. min -1 of O2); X : biomass concentration (g of SSV), ER: endogenous respiration (mgO2. gSSV -1 . min -1 ).

Analytical Methods
Alkalinity, TOC, NO2 --N, NO3 --N and NH3-N was determined according to APHA, AWWA, WEF (2012). TOC analyzes were performed using a TOC analyzer (Analytik Jena, Multi C/N 2100). While NO2 --N, NO3 --N and NH3-N were determined based on a colorimetric method using a flow injection analysis system (FIAlab -2500). Alkalinity was determined using the titrimetric method . DO concentrations in oxic reactor were measured and controlled using a DO controller (Hach, SC200).

DO restriction effect on NDF process
At the beginning of phase I the main objective was to keep the conventional NDF process. Throughout this stage, between day 1 -37, the system presented an N removal efficiency of 82.3%, with 70% of TOC consumption ( Figure 2). During this phase NDF process has been successfully established and kept during the 37 days.
However, even after severe reduction of DO concentration, N removal efficiency via nitrate presented a low decrease when compared to phase I ( Table 2). The literature reports that complete nitrification is kept with higher values of DO concentration of 1.5 mg L -1 of O2 (Yang et al., 2012), while at lower DO levels 0.3-0.7 mg L -1 of O2 nitrite is accumulated by the nitritation process prevalence (Ma et al., 2009;Zeng et al., 2013). As nitritation was not stablished in Phase II-a, DO concentration was restricted again in the oxic reactor between days 54-67 (Phase II-b, Figures 2 and 3). The DO concentration was decreased from 0.55 ± 0.05 mg L -1 of O2to 0.45 ± 0.05 mg L -1 of O2. At this condition, it was observed a NH3-N accumulation (220.02 ± 117.4 mg L -1 of N) in oxic reactor, caused by the reduction of AOB (Ammonia oxidizing bacteria) activity.
On phase II-c, DO concentration was again increased in the oxic reactor from 0.45 ± 0.05 mg L -1 of O2 to 0.65 ± 0.05 mg L -1 of O2 (between days 68-101) in order to recover the AOB efficiency to completely remove NH3-N from the MLE process. During this phase (Figures 2 and  3) NH3-N removal was increased about 61%, compared to phase II-b with a slight improvement in N removal process ( Table 2).

Effect of DO concentration and different C/N ratios in N removal process via NDF and NDT
Initially the system was operated at C/N ratio of 1.5 (phase III-a, Figures 4 and 5) and DO 0.65 ± 0.05 mg L -1 of O2, under the same operating conditions of phase II-c, reaching a similar N removal efficiency that was also quite similar to phase I when NDF was established (Table 2).
At this phase III-b (between days 116 -135) NH3-N removal was around 90 % (Table 2). However, even with the reduction of C/N ratio at this stage to 0.9, TOC consumption was reduced 23% when compared to Phase III-a, without significant impact on the N removal efficiency (Table 2). According to Yang & Yang (2011), a reduction in the consumption of organic matter switching NDF to NDT can reach in 40% of carbon economy for heterotrophic process.
Thereafter a new reduction of C/N ratio from 0.9 to 0.75 (Phase III-c, Figures 4 and 5) between days 136-149 was performed. This strategy affected the N removal efficiency in almost 30 % comparing phases III-c with IIIb (Table 2). However, with the reduction of C/N ratio, there were accumulation of NO2 --N (60.11 ± 21.9 mg L -1 of N) and NO3 --N (138.2 ± 123.5 mg L -1 of N). The accumulation of NOX-N (NO2 --N + NO3 --N) occurs when there is absence of TOC in the denitrifying reactor because the TOC is the donor source of electrons, and nitrite and nitrate, are the electrons acceptors. Associated to NOX-N accumulation, and alkalinity consumption, pH will decrease favoring the generation of Free Nitrous Acid (FNA) that can also cause inhibition on nitrification process (Mohan et al., 2016;Hou et al., 2014).  Initially at phase IV, DO concentration in the oxic reactor was increased from 0.65 ± 0.05 mg L -1 of O2 to 2.5 ± 0.5 mg L -1 of O2, and C/N ratio adjusted to 1.5. At these conditions the N and TOC removal reached 88.45% and 86.45% respectively. Comparing Phase IV-a with Phase III-a (C/N 1.5 and DO 0.65 ± 0.05 mg L -1 of O2) ( Table 2) was observed a similar N removal and TOC consumption in both phases, around 80% and 87% respectively.
In order to prove that oxygen was not the limiting agent in the process, C/N ratio was successively reduced as performed in phase III. At phase IV-b, C/N ratio was reduced from 1.5 to 0.9 (Figures 4 and 5 Comparing with Phase III-b (Table 2), it is observed that besides of O2 economy (about 74 %). Meng et al. (2015) obtained N removal of 87% using nitritation/denitritation in an upflow microaerobic sludge reactor (UMSR), operating with swine influent at C/N ratio of 0.84.
In phase IV-c, C/N ratio was reduced from 0.9 to 0.75 (Figures 4 and 5) between days 197-211. Compared to Phase IV-b was observed a decrease in the N removal reaching 40.11% (Table 2).    Although SOURs were the highest for test 3 (Figure   7), the oxygen consumption reduction between NDF and NDT was not the highest found during the respirometric tests. Figure 8 shows the performance between SOURs and S ((NH4)2SO4) concentrations for the two studied processes. Whilst for 50 mg L -1 of NH3-N 11.5% oxygen saving is reached for a S concentration of 800 mg L -1 of NH3-N was achieved around 37% in oxygen saving ( Figure 8), that is higher than that is reported in the literature (Turk & Mavinic, 1986;Yang & Yang, 2011;Zhu et al. 2008;Fu & Zhao, 2015).       Figure 8 shows a logarithmic tendency curve indicating a possible stability at substrate concentrations above 800 mg L -1 of NH3-N. However, a great advantage to operate the NDT process compared to NDF process, maintaining efficiency next to 80% in N removal (Table 2 and Figure 5) and parallel consuming less 36.8% of carbon for denitrification ( Figure 6).

CONCLUSIONS
NDT was successfully stablished for swine wastewater nitrogen removal at DO of 0.6 -0.7 mg L -1 of O2, reaching N removal of 75%. At C/N ratio of 0.9 the nitrogen removal rate for NDT reached 0.31 kg m -3 d -1 of N that is 133% higher than the obtained for NDF at the same conditions with a reduction of TOC consumption of 27 %. The results show that it is possible to obtain efficient nitrogen removal efficiency for MLE configuration process for swine wastewater treatment operating the process at low DO and C/N ratios creating the possibility of application and operating with influent low carbon and high nitrogen, such as digestate from anaerobic processes.