Effect of cycle time and airflow in biological nitrogen removal from poultry slaughterhouse wastewater using sequencing batch reactor

This study aimed to evaluate the influence of airflow (0.25, 0.50 and 0.75 L.L.min) and cycle time (10.45 h, 14.25 h and 17.35 h) on a sequencing batch reactor (SBR) performance in promoting nitrification and denitrification of poultry slaughterhouse wastewater. The operational stages included feeding, aerobic and anoxic reactions, sedimentation and discharge. SBR was operated in a laboratory scale with a working volume of 4 L, keeping 25% of biomass retained inside the reactor as inoculum for the next batch. In the anoxic stage, C: N ratio was maintained between 5 and 6 by adding cassava starch wastewater. A factorial design (2) with five repetitions was designed at the central point to evaluate the influence of cycle time and airflow on total inorganic nitrogen removal (N-NH4 +N-NO2 +N-NO3 ) and in the whole process (nitrification and denitrification). The highest total inorganic nitrogen removal (93.3%) was observed for airflow of 0.25 L.L.min and a cycle time of 14.25 h. At the end of the experiment, the sludge inside the reactor was characterized by fluorescent in situ hybridization (FISH), indicating the presence of ammonia and nitrite oxidizing bacteria.


INTRODUCTION
Poultry slaughterhouses in Brazil present great economic importance, occupying the third position on world chicken meat's production with 12.23 millions of tons in 2010(ABEF, 2010)).However, the sector is characterized by generation of large volumes of wastewater, the equivalent to 18 L.chicken -1 (KIST et al., 2009), composed of high organic matter loads and nutrients, such as nitrogen (DEL POZO & DIEZ, 2005).
The removal of nitrogen compounds is an important aspect of wastewater treatment, once these nutrients cause several inconveniences, with subsequent damages to human health and to the environment (METCALF & EDDY, 2003;WU et al., 2008;SADEQ et al., 2008;DE NARDI et al., 2011).
A biological nitrogen removal by composed systems, such as sequencing batch reactors (SBR), contemplates aerobic and anoxic stages with sequencing operations throughout the time.The nitrification occurs, in aerobic conditions, under action of chemoautotrophic bacteria that oxidize ammoniacal nitrogen to nitrite and nitrate, sequentially.During denitrification, heterotrophic bacteria reduce nitrogen-oxidized forms to gaseous nitrogen, using an organic carbon source as electron donor, in anoxic condition (METCALF & EDDY, 2003).
Research in laboratory scale has applied SBR as post-treatment of industrial wastewater of animal raw material processing.Such studies inspected information about operational parameters that influence biological nitrogen removal, such as initial concentration of ammoniacal nitrogen, the ratio carbon: nitrogen, airflow and cycle time.ZENATTI et al. (2009), ANDRADE et al. (2010) and KUMMER et al. (2011) evaluated SBR with immobilized biomass, for nitrogen removal of wastewater from fish slaughterhouse.DALLAGO et al. (2012) andMEES et al. (2011) applied SBR in the wastewater removal from poultry slaughterhouse and refrigerated warehouse, contributing with noteworthy database about dimensioning parameters of this treatment unity.
Accordingly, the aim of this study was to assess the influence of airflow (0.25; 0.5 and 0.75 L. L. -1 .min - ) and cycle time (10h45; 14h25 and 17h35) on biological removal of total inorganic nitrogen of poultry slaughterhouse wastewater in SBR, applying response surface methodology, which enables investigation of factors' influence as well as the interaction between them (RODRIGUES & IEMMA, 2009).
Moreover, phylogenetic characterization of nitrifying microbial community found in biological sludge, using the fluorescent in situ hybridization technique.The knowledge about microbial flora established in reactor is of utmost importance for wastewater treatment systems, as it can provide relevant information about reactor performance (HIRASAWA et al., 2008).

Wastewater
The wastewater used as substrate derived from poultry slaughterhouse located in Western region of Paraná, and was collected at the exit of anaerobic ponds.The substrate was homogenized in a single recipient and, subsequently conditioned in polyethylene packages of 2L, stored at temperature of -5 ºC, until the moment of its use in reactor.
The wastewater from processing to obtain starch used as carbon source, in anoxic stage, was collected in a starch industry lo cated in Western region of Paraná.

Characterization of treatment system and ope rational conditions
The sequencing batch reactor (SBR), in laboratory scale, was manufactured in polyethylene with cylindrical format (185 mm of diameter and 205 mm of height) and 4 L working volume.The reactor was operated following the stages of feeding, aerobic reaction, anoxic reaction, sedimentation and discharge.
In order to start the reactor, 1 L of inoculum derived from another SBR was added, used by MEES et al. (2011), containing 12,888 mg.L -1 of VSS and 3 L of poultry wastewater, obtaining concentration of VSS in liquid mass (SSVML) of 3,299.85mg.L -1 .
The aeration system consisted of an aerator coupled to a flow meter, from which the airflow was manually adjusted.Air diffusion in liquid mass was conducted with the use of two porous stones.In this stage, the alkalinity was adjusted with sodium bicarbonate solution 1 M (GAPES & KELLER, 2009), in order to maintain the ratio alkalinity/ammoniacal nitrogen near the recommended by JANG et al. (2004), which is of 0.61 mg of NaHCO 3 per mg of oxidized N-NH 4 + .
In the anoxic stage, a mechanic mixer was kept with speed between 25 to 30 rpm (DALLAGO et al., 2012), to promote biomass and substrate contact, by incorporating air in the liquid environment.In the beginning of this stage, an external source of inorganic carbon was added (wastewater of cassava starch), adjusting the ratio DQO: (N-NO 3 -+ N-NO 2 -) between 5 and 6 (MEES et al., 2011).
After sedimentation stage, the removal of clarified wastewater was promoted (supernatant) of reactor by sonication.At each batch, 25% of total volume was retained (1 L) (sedimented biomass) for the next batch, therefore the volume of discharged wastewater and of re-feeding was of 3 L.
In order to control biomass concentration in reactor, at the end of each batch, volatile suspended solids concentration (VSS) was determined in sedimented sludge by measuring optical density in spectrophotometer at 700 nm (DAMASCENO et al., 2003).

Analytical monitoring
Samples of reactor were collected at the beginning and end of each operational stage (aerobic and anoxic), centrifuged at 5.000 rpm during 10 minutes (FONTENOT et al., 2007) and analyzed according to APHA (1998), regarding ammoniacal nitrogen, nitrite, nitrate, chemical oxygen demand, pH, total alkalinity, dissolved oxygen and temperature parameters.To analyze data obtained through experiments via statistics, the software STATÍSTICA for Windows version 7.0 was employed (STATSOFT TM , 2004).

Evaluation of linear effect of cycle time with constant airflow
Sequentially, due to results obtained with experimental planning, in which only the cycle time factor was significant, linear regression with plateau response was elected, in which a linear equation of response to increment regarding cycle time is undertaken, up to a point where reactor no longer presents increasing response.The point between linear equation and plateau, presents the great level of time.
The linear-plateau regression model (LPR) enables comparison among alternative combinations of straight lines and plateaus and then chooses, as the best adjusting option, the combination that presents the lowest sum of variance squares.
In this stage, six cycle times were evaluated, with reactor sub mitted to, in aerobic stage, constant airflow of 0.25 L.L -1 .min - .Eighteen tests of whole cycle were conducted, distributed in 6 treatments (levels of CT), with 3 repetitions each (Table 3).The efficiency results of total inorganic nitrogen removal were submitted to variance analysis at 5% of probability and, in the case of significant difference evidence, the Linear-Plateau regression analysis was applied, with 5% of probability.The software employed to conduct statistical analysis was the SAEG -System for Statistical Analysis (UFV, 2007).

Phylogenetic characte rization of microbial community through fluorescent in situ hybridization (FISH)
The microbial characterization of reactor biological sludge was conducted through the fluorescent in situ hybridization technique (FISH), using specific oligonucleotides probes for the members of Bacteria Dominium (EUB338), ammonia oxidizing bacteria (NSO190), Nitrospira spp.(NSR1156) and nitrite oxidizing bacteria Nitrobacter spp.(NIT3).
The FISH technique involved the following stages: cellular detachment, washing and fixation of cells, filtration, hybridization, fluoresce microscopy and microorganisms quantification, according to methodology described by HIRASAWA et al. (2008).
The phylogenetic probes labeled with the fluorescent dye Cyanine (Cy3) were synthesized by Eurofins MWG Operon (Ebersberg, Germany).The list of probes used and their characteristics are presented on Table 5.The samples stained with DAPI were excited with UV light and the fluorescence was recovered with the aid of the filter U-MWU2, while probes labeled with Cy3 were excited with green light and the fluorescence recovered with the filter U-MSWG2 using an Olympus BX 60 Pro-Plus 6 microscope.
Ten counts were carried out in random fields, with the percentage of hybridized cells in each field calculated in relation to total number of cells stained with DAPI.The counting was conducted with the aid of the "freeware" program Image Tool-UTHSCA (University of Texas Health Science Center, San Antonio -Texas/USA).

Effects of cycle time and airflow factors on the total inorganic nitrogen re moval
On Table 6, the experimental design matrix is displayed, with levels of each factor and responses obtained throughout the conducted tests.The parameters that presented p-value lower than 0.05 are significant in the confidence interval of 95%.
It can be interfered that, at 5% level of significance, only the cycle time factor was significant.Cycle time effect presented noteworthy and positive value, indicating that the process of total inorganic nitrogen removal was favored with the greatest applied cycle times; therefore, the removal efficiency tends to increase with the raise of reaction time.Regarding effects of airflow and interaction between variables, they do not presented significant value for tested conditions (p-value > 0.05), indicating the possibility of applying lower airflows, with the same removal efficiency, causing reduction of energy consumption for this purpose.ZENATTI et al. (2009), when evaluating nitrification in SBR for treatment of tilapia slaughterhouse wastewater, obtained average efficiency of 81.90% on ammoniacal nitrogen removal when the reactor was operated with 12 h of reaction, airflow of 6 L.min -1 and working volume of 2.5 L; however, it is observed that the conversion is only significantly affected by reaction time.
During the first experimental stage, tests 1 and 2 presented the lowest responses for total inorganic nitrogen removal; therefore, in these cases, there was no significant alkalinity consumption in aerobic stage, neither its generation in anoxic stage, probably due to cycle time.The remaining tests demonstrated alkalinity consumption during the oxidation of N-NH 4 + with average value of 82%, and increase of alkalinity value at the end o f denitrification stage, with recovery of approximately 69%.
Among the levels of evaluated factors, the best result regarding efficiency of total inorganic nitrogen removal, in whole process, 85.75%, was obtained with airflow of Effect of cycle time and airflow in biological nitrogen removal from poultry slaughterhouse wastewater Eng.Agríc., Jaboticabal,v.35,n.3,,maio/jun.2015 573 0.5 L.L -1 .min-1 and cycle time of 14h25, validating the results from BROWN et al., (2011) andDALLAGO et al. (2012).2011) also observed positive effect of cycle time factor, when evaluating the ratio carbon: nitrogen (3; 6 and 9) and cycle time (8h, 12h and 16h) on nitrogen removal of poultry slaughterhouse wastewater, in SBR (working volume of 5 L).The authors obtained optimized values of total inorganic nitrogen removal (84.32%), when the SBR was operated with cycle time of 16h (10h15, aerobic stage and 4h35, anoxic stage).Concerning the C: N ratio, authors observed a value of 6, which is similar to this study, with cycle time of 14h25 (10h15, aerobic stage and 3h, anoxic stage), ratio C:N of 5.28 and airflow of 0.5 L.L -1 .min-1 , which resulted in concentration of 3.1 mg.L -1 of dissolved oxygen.

MEES et al. (
The pH during the nitrification process, in all conducted tests, varied from 6.78 to 6.82.WEF et al. (2005) suggested great pH nitrification ranging from 6.5 to 8.0.During denitrification, pH varied from 5.73 to 6.72, with the industrial wastewater of cassava starch contributing to pH decrease reaching critical values, which was reestablished during the process.
The observed average temperatures were of 19.5 °C in the beginning of the nitrification stage of and of 22.4 °C at the end of denitrification stage.
The most efficient tests at removing total inorganic nitrogen, with cycle time of 14h25 and airflow of 0.5 L.L -1 .min - , presented average concentration of ammoniacal nitrogen in reactor's wastewater of 13.69 mg.L -1 .This result meet the maximum limit allowed to wastewater release, which is 20 mg.L -1 , defined on Conama Resolution n o 430 (BRASIL, 2011).

Evaluation of plateau effect on total inorganic nitrogen removal
The airflow factor was conclusive in the previous experiment stage, demonstrating that, for studied conditions, and of 95% de confidence, the airflow of 0.25 L.L -1 .min - was sufficient to promote aeration of nitrification stage.Therefore, the addition of greater concentrations would represent unnecessary energy expenditure.
In sequence, a planning referring to regression was chosen, setting the airflow at 0.25 L.L -1 .min - and increasing cycle times, progressively, until the constant value of nitrogen removal efficiency is observed.
In this experimental stage, the airflow of 0.25 L.L -1 .min - maintained the dissolved oxygen concentration, in average, 2.64 mg.L -1 in aerobic stage and 0.51 mg.L -1 at the end of anoxic stage.The average temperature was of 23.6±1.19o C, and the average concentration of VSS of sedimented biomass (sludge) was of 9,886.37±103.48mg.L -1 .The pH during nitrification was kept around 6.7±0.35, and in denitrification, it was of 6.4±0.38.
Through regression analysis with plateau response, the maximum point of inflexion of generated curve corresponds to cycle time of 14h25, a level recommended for obtaining maximum total inorganic nitrogen removal, verified in 93.3%.Eng.Agríc., Jaboticabal, v.35, n.3, p.567-577, maio/jun. 2015 574 FIGURE 1. Linear-plateau response between total inorganic nitrogen removal and cycle time.
In Table 8, the results of linear plateau regression analysis are presented.The equation of estimated regression is represented by: Ŷ= -92.3+12.9*X,for cycle time of 10.4≤14.25 (1) Y= 93.3, for 14.25≤X≤32.1 These values corroborate the results obtained during factorial design and indicate that the application of greater cycle times did not promote increase of nitrogen removal.

Phylogenetic characterization of the microbial community by FISH
The photomicrographs of microbial cells present in sludge samples, subjected to the FISH technique, are shown in Figure 2. MISSAGIA et al. ( 2007) evaluated microbial flora in biofilms in a percolator biological filter installed after UASB reactor, with 90 days of operation.Based on probe specificity, the authors could detect, at least, four different populations of ammonia oxidizing bacteria, such as: (i) two different populations of Nitrosomonas oligotropha, (ii) N. europaea and (iii) N. communis.The probe resulted in a positive sign for all samples, differing in the relative abundance.The community's relative abundance was verified by the ratio between all bacteria stained with DAPI and total bacteria labeled with NSO190 probe, with the highest value of 18.3% found in support of plastic rings.
In a study of HASEBORG et al. (2010), nitrifying bacteria and archaebacteria were fed in fixed-bed biofilm reactors, with different concentrations of ammonia and nitrite in synthetic and real wastewaters, and the bacteria were detected by FISH.In high concentrations of nitrite (5 to 10 mg.L -1 ), the abundance of Nitrobacter spp.species was of 88%.When the ammonia concentration was of 60-80 mg.L -1 , the obtained value of Nitrosomonas spp. was of 31%.
In the present study, the ammonia concentration was 81.9 mg.L -1 .The quantification of ammonia oxidizing bacteria, belonging to β-Proteobacteria group (detected by NOS and NSR 1156 probes) and of nitrite oxidizing bacteria (detected by NIT 3 probe) resulted, respectively, in 32.8% and 24.3% of abundance relative to DAPI, which corroborates the previously cited studies.

CONCLUSIONS
According to obtained results, it is possible to conclude that, in conditions established during factorial design, the effect of cycle time presented significant and positive value, while effects of airflow and interaction between variables did not present significant values for tested conditions.However, throughout linear-plateau regression tests, the best result regarding efficiency of total inorganic nitrogen removal was of 93.3%, obtained with airflow of 0.25 L.L -1 .min - and cycle time of 14h25, demonstrating applicability of SBR for biological nitrogen removal.
This study enabled the identification of microbial flora composition that participates in treatment processes, demonstrating that the FISH tool can be very useful in evaluating the bioreactors performance, in which the stability and efficiency are highly dependent of microbial communities interactions.

Figure 1
Figure 1 presents plateau determination between cycle time and total inorganic nitrogen removal.

FIGURE 2 .
FIGURE 2. In situ identification of Bacteria Dominium after hybridization with Cy3 labeled probes (a).Identical microscopic field was observed for cells stained with DAPI (b).

Table 1
displays a summary of average values obtained in physical-chemical characterization of poultry slaughterhouse wastewater and cassava starch processing wastewater.

2015 570 Effect evaluation of cycle time and airflow
Carla L. Lopes, Juliana B. R. M ees, Luciane Sene, et al.In order to verify the influence of cycle time and airflow factors on total inorganic nitrogen removal (N-NH 4 a factorial design was conducted 2 2 with 5 repetitions at the central point, resulting in 9 tests.The factors with coded and real values are presented on Table2.

TABLE 2 .
Factorial design matrix (22 ) with coded and real factors and values.

TABLE 4 .
Cycle time distribution over the operational stages during the experiment.

TABLE 5 .
Sequences of the phylogenetic probes used.

TABLE 6 .
Carla L. Lopes, Juliana B. R. M ees, Luciane Sene, et al.Factorial design ma trix (2 2 ) with independent variables and their responses on nitro gen removal.

Table 7
presents estimated effects for response variable of total inorganic nitrogen removal.

TABLE 7 .
Estimated effects of total inorganic nitrogen removal efficiency.

TABLE 8 .
Linear-plateau regression analysis results to predict total inorganic nitrogen removal.