Banafsheh Azimi, Elham Abdollahzadeh-Sharghi*, Babak Bonakdarpour,*

1 Chemical Engineering Department, Amirkabir University of Technology, Tehran, Iran

2 Environmental Group, Energy Department, Materials and Energy Research Center, Alborz, Iran

Keywords:Wastewater Anaerobic Aerobic Anaerobic moving bed sequencing batch biofilm reactor Anaerobic sequencing batch reactor Aerobic membrane bioreactor

A B S T R A C T In this study,the effect of number of stages and bioreactor type on the removal performance of a sequential anaerobic-aerobic process employing activated sludge for the treatment of a simulated textile dyeing wastewater containing three commercial reactive azo dyes was considered. Two stage processes performed better than one stage ones, both in terms of overall organic and color removal, as well as the higher contribution of anaerobic stage to the overall removal performance,thereby making them a more energy efficient option. The employment of a moving bed sequencing batch biofilm reactor, which uses both suspended and attached biomass,for the implementation of the anaerobic stage of the process,was compared with a sequencing batch reactor that only employs suspended biomass. The results showed that,although there was no meaningful difference in biomass concentration between the two bioreactors,the latter reactor had better performance in terms of chemical oxygen demand(COD)removal efficiency and rate and color removal rate.Further exploratory tests revealed a difference between the roles of suspended and attached bacterial populations,with the former yielding better color removal whilst the latter had better COD removal performance. The sequential anaerobic-aerobic process, employing an aerobic membrane bioreactor in the aerobic stage resulted in COD and color removal of 77.1 ± 7.9%and 79.9 ± 1.5%, respectively. The incomplete COD and color removal was attributed to the presence of soluble microbial products in the effluent and the autoxidation of dye reduction metabolites, respectively. Also, aerobic partial mineralization of the dye reduction metabolites, was experimentally observed.

1. Introduction

Large amount of dyes are produced annually, 50% of which are used in the textile industry. Large amount of water is used during the wet process in textile industry [1,2]. Dyeing process is one of the important wet processes in the textile industry and generates a high chemical oxygen demand (COD) wastewater with a very deep color [1]. It is estimated that about 15% of dyes used in the dyeing process are discharged in the wastewater[3].Among different class of dyes,azo dyes are the most common and consist of 70%of dyes used in textile wastewater [2]. Azo dye consists of one or more azo bonds (-N＝N-) connecting aromatic amines [4].

Different chemical and physical methods are used for the treatment of dyeing wastewater like filtration, adsorption, ozonation and coagulation/flocculation. The main drawbacks of these methods are high costs and the generation of large amounts of sludge.On the other hand, biological treatments are less expensive and also eco-friendly[4-6].The most commonly studied biological process for the treatment of azo dye containing wastewaters is the combined anaerobic-aerobic process. In this process, the azo dye is anaerobically reduced resulting in the formation of more toxic and harmful aromatic amines which cannot be degraded under anaerobic condition but are potentially aerobically biodegradable[5].

Although the applicability of combined anaerobic-aerobic processes for treatment of textile dyeing wastewaters has been demonstrated in numerous studies, some questions are still left unanswered. For example, these processes can be carried out as a one or a two stage process.In the former case,the mixed bacterial population in activated or anaerobic sludge is exposed to alternating anaerobic/aerobic environment,whereas in the latter case,the bacterial population is exposed to only one environment.Most previously reported studies have been carried out as either a one stage or a two stage process but there is very limited information on the comparison of these two process strategies.Bonakdarpouret al.[7]and Vyrideset al. [8]showed that the COD and color removal performance of combined anaerobic-aerobic processes is strongly influenced by the number of stages employed in the process,although the exact effect was found to be dependent on the dye and NaCl concentrations. However, the synthetic media employed by these authors contained only one reactive azo dye and the composition was not necessarily representative of real dyeing wastewaters. Also, these authors did not consider two stage processes in which activated sludge was employed in both stages.

In terms of reactor types employed in the anaerobic stage of the combined anaerobic-aerobic process, in previous studies suspended, attached and mixture of suspended and attached growth bioreactor systems have been employed. Examples of anaerobic suspended growth bioreactor configurations include membrane bioreactors (MBR) [9,10], baffled reactor [11], sequencing batch reactor (SBR) employing flocculated sludge [12,13] or granular sludge [14], and upflow anaerobic sludge blanket reactor [15,16].Examples of the attached growth bioreactor configurations employed for treatment of dye-containing wastewaters include the integrated biofilm reactor [17] and immobilized-cell SBR[18]. The results of these studies show that suspended and attached growth reactor types each have their advantages and disadvantages.A potentially better bioreactor type is one which combines the relative advantages of suspended and attached growth types. These advantage include higher process stability and biomass concentration [19], the lack of need for sludge recycle [20],and lower biomass wash out [21]. One such design is the moving bed biofilm reactor (MBBR). In previous studies, anaerobic batch MBBR performed better in term of COD and color removal compared to suspended activated sludge system for treatment of real textile wastewater[22,23].However,more study is needed regarding the comparative performance of MBBR with suspended or attached growth only cultivation systems as bioreactor types to be employed in the anaerobic stage of a combined anaerobic-aerobic wastewater treatment process,especially regarding the mechanisms responsible for the difference in performance.

Various reactor configurations have also been employed in the aerobic stage or phase of the combined anaerobic-aerobic wastewater treatment process. These have included aerobic SBR [12-14],continuous stirred tank reactor[15]and aerobic MBBR[22].A particularly promising reactor configuration employed in the aerobic stage of a combined anaerobic-aerobic treatment used for dye containing wastewater treatment has been the aerobic MBR which combines biological treatment with membrane filtration[9,10,24]. This combination has advantages compared to conventional activated sludge system such as higher quality effluent[24,25], independent sludge retention time control [10] as well as low maintenance [24]. de Jageret al. [26] investigated real textile wastewater treatment in a pilot scale anaerobic-anoxic-aerobic process and reported that discharge standards could only be met when a UF-membrane module was added after the aerobic stage.Yanet al. [11] used anoxic baffled reactor followed by an MBR for treatment of real dyeing wastewater and achieved 97% and 88% COD and color removal, respectively. A combined anaerobicaerobic MBR has been reportedly employed for treatment of synthetic wastewater containing an azo dye with the aerobic stage removing COD with high efficiency but only 30%-40% of the color[2]. In a study performed by You and Teng [27] aerobic MBR was coupled to an anaerobic SBR and reportedly achieved efficient degradation of the residual COD and color in the anaerobic effluent.Membrane bioreactors are, therefore, suitable choice of bioreactor type to be employed for the aerobic stage of the combined anaerobic-aerobic process for treatment of dyeing wastewaters.

The aim of the present study was therefore twofold: The first was to study the effect of the number of stages on the COD and color removal performance of a combined anaerobic-aerobic process used for treatment of a simulated textile dyeing wastewater containing mixture of three commercial important reactive azo dyes. For this purpose, the performance of a one stage process was compared with a two stage process. For ease of comparison,activated sludge, which is almost exclusively used in one stage processes, was employed in both processes. The second aim was concerned with the comparison of the performance of two bioreactor types for carrying out the anaerobic stage of the process,namely a moving bed sequencing batch biofilm reactor (MBSBBR),which consists of both suspend and attached growth cultivation systems,with a bioreactor type employing suspended only cultivation system (SBR). Finally, the performance of an aerobic MBR for the treatment of the effluent of the anaerobic bioreactor was evaluated.

2. Materials and Methods

2.1. Simulated wastewater

Since real textile wastewaters can significantly vary in composition from one batch to the next, in order to be able to meaningfully interpret the results obtained under varying test conditions a simulated textile dyeing wastewater was used throughout this study.The simulated wastewater was prepared taking into account the composition of a local textile dyeing plant (Takmiliran textile dyeing factory, Alborz province, Iran). Based on the analysis of the wastewater carried out at different seasons during the year,the average COD,color,and total dissolved solids(TDS)were found to be (900 ± 325) mg·L-1, (300 ± 90) Space Units (SU) and(9500±1500)mg·L-1,respectively.Because of the usage of sodium hydroxide, acetic acid, sodium carbonate and sodium chloride in the dyeing process, sodium acetate (produced by the reaction of acetic acid with sodium carbonate)and sodium chloride were used as COD and TDS sources in the simulated wastewater,respectively.The simulated wastewater contained sodium acetate(1500 mg·L-1)(COD = 900 mg·L-1), NH4Cl (191.1 mg·L-1), K2HPO4(37.6 mg·L-1),KH2PO4(14.5 mg·L-1),NaCl(10000 mg·L-1),trace element solution(1 ml·L-1) (Table S1), and three dyes, each at the same concentration, at a total concentration of 80.0-100.0 mg·L-1(250-350 SU)and total COD of 70 mg·L-1. Three commercial reactive azo dyes(Colorant ltd., India), namely, Black B (C.I. number: Black 5; CAS number: 17095-24-8; appearance: dark blue powder; molecular formula: C26H21N5Na4O19S6; molecular weight: 991.82 g·mol-1),Black WNN (C. I. number: Mix; CAS number: NA; appearance:black powder; molecular weight: NA), and Red 3BS (C. I. number:Red 195; CAS number: 93050-79-4; appearance: red powder;molecular formula: C31H19ClN7Na5O19S6; molecular weight:1136.32 g·mol-1) were obtained from a local textile dyeing plant(Takmiliran textile dyeing factory, Alborz province, Iran). Black B[Fig.1(a)][7,28]was a bifunctional dyes containing two azo bonds and two sulfatoethylsulfone groups, whereas Red 3BS [Fig. 1(b)]contains mono azo bond and triazine as a rective group [30]. The structure of Black WNN is not known. For obtaining the dominant form of dyes present in dye bath effluents,these dyes without further purification were hydrolyzed by the procedure described by Louren?oet al[31].

Fig. 1. Structure of (a) Black B [28] and (b) Red 3BS [29].

2.2. Activated sludge inoculum

Settled activated sludge from the downstream sedimentation unit of an aeration tank of a full-scale biological wastewater treatment unit treating poultry slaughterhouse wastewater (Iran Boorchin poultry processing plant,Tehran,Iran)was used as inoculum in both anaerobic and aerobic phases of batch and reactor experiments. In order to improve the sludge condition, the activated sludge was cultivated in a shaker incubator at 170 r·min-1and 30 °C every 48 h where it was fed for 10 days with simulated wastewater in which the ratio of COD:N:P was adjusted to 100:5:1.At the end of 10 days the concentration of mixed liquor suspended solids (MLSS) reached 4500 mg·L-1.

2.3. Abiotic test

In order to quantify the contribution of the dye adsorption to biomass to the decolorization during the anaerobic process, batch studies were conducted under anaerobic conditions using 250 ml Erlenmeyer flasks containing 150 ml of a mixture of autoclaved activated sludge and simulated wastewater. The media employed in these experiments was prepared according to the composition reported in Section 2.1 but excluding sodium chloride and additionally containing one dye.Adsorption of each dye at three different concentration(32.3 mg·L-1, 64.5 mg·L-1and 96.8 mg·L-1) was examined at 170 r·min-1and 30 °C for 48 h.

2.4. Batch experiments

The anaerobic phase of the batch experiments were conducted using 100 ml serum bottles which contained 80 ml of a mixture of activated sludge and simulated wastewater. The serum bottles were sealed with rubber stoppers and aluminum crimps and flushed with nitrogen gas for attaining anaerobic conditions. The aerobic phase of the batch experiments were performed as shake flask cultivation in 250 ml Erlenmeyer flasks.Both the serum bottle and shake flask experiments were conducted in a shaker incubator at 170 r·min-1and 30 °C. The media employed in these experiments was prepared according to the composition reported in Section 2.1 but excluding sodium chloride and additionally containing 80 mg·L-1of either single or a mixture of the three dyes.In the experiments aimed at studying the effect of the number of stages of the anaerobic-aerobic process only a mixture of the three dyes was employed.

In batch experiments simulating one stage anaerobic-aerobic process, at the end of the anaerobic phase both sludge and the decolorized solution was transferred to Erlenmeyer flasks. In the case of two stage process runs,after the end of the anaerobic stage the sludge was allowed to settle and the solution above the settled sludge was subsequently transferred to the Erlenmeyer flasks containing their own activated sludge.In both one stage and two stage process runs the duration of the anaerobic and aerobic phase were 48 h and 72 h,respectively.These times,which were chosen based on preliminary batch experiments, were necessary to achieve steady state color and COD removals in the anaerobic and aerobic phase/stage respectively (Fig. S1). The anaerobic-aerobic cycles were repeated until steady state condition in terms of color and COD removal were achieved in the anaerobic and aerobic phases,respectively. All the experiments were performed in triplicate.

2.5. Reactor experiments

2.5.1. Anaerobic reactors

Both anaerobic reactors (MBSBBR and SBR) had cylindrical shape (height = 28 cm, diameter = 24 cm) with total and working volume of 10 L and 6 L,respectively.In both cases the reactor cycle(anaerobic phase), which was controlled by a digital timer, consisted of: 0.25 h fill -23 h react -0.5 h settle -0.25 h decant. No activated sludge was discarded as waste during the operational period.At the beginning of react phase,reactors were flushed with nitrogen gas for 15 min.Hydraulic retention time(HRT),cycle time and volumetric exchange rate (VER) were 48 h, 24 h and 0.5,respectively. At the beginning of each cycle, 3 L of new simulated wastewater prepared according to the composition presented in Section 2.1 and additionally containing 80 mg·L-1of a mixture of the three azo dyes,was fed to each anaerobic reactor.A low speed gear motor (160 r·min-1) driving three paddle-shaped impellers was used during the anaerobic reaction phase to provide adequate mixing conditions (Fig. 2). Both reactors were kept at 30 °C in a water bath.

About 30% of the working volume of MBSBBR was filled with polyurethane(PU)cubic foam,with average size of 1 cm,on which biofilm had pre-formed.Use of PU foam as a support matrix for the immobilization of anaerobic biomass has been successfully reported in previous studies[22,23].The biofilm formation process was carried out under anaerobic conditions using shake flask cultivation in 1000 ml Erlenmeyer flasks containing activated sludge,the simulated wastewater (without dye and sodium chloride),and PU foam (70% V/V) at 120 r·min-1and 30 °C at an HRT of 48 h for a period of 40 days.Biofilm formation process was carried out in the absence of dye since previous studies have shown that dyes and their metabolites can have an adverse effect on the formation of the biofilm[32].In the case of the runs with SBR the initial MLSS in SBR was set at 3500 mg·L-1.

2.5.2. Aerobic reactor

The effluent of SBR was stored in Erlenmeyer flasks and fed to the MBR via a peristaltic pump(Fig.2).MBR was made of Plexiglas with total and working volume of 12 L and of 9 L,respectively.Flat sheet microfiltration chlorinated polyethylene membrane(KUBOTA, Japan) with 0.4 μm pore size and surface area of 0.1 m2was immersed in MBR. Aeration at a rate of 6-7 L·min-1was used to supply dissolved oxygen(DO)concentration of at least 3 mg·L-1to aerobic microorganisms and for membrane scouring.Details of aeration has been described elsewhere[33].The temperature of the mixed liquor inside the bioreactor was controlled by a heater at around 30°C.Effluent flowed out through the membrane by a peristaltic pump. An intermittent filtration cycle,i.e. 8 min suction followed by 2 min relaxation (non-suction), was adopted to retard membrane fouling. The MBR was equipped with level sensor and sampling ports. HRT and flux of the reactor were 72 h and 1.3 LMH, respectively and no sludge was discharged during its operation. Initial MLSS in the MBR was around 5300 mg·L-1.The operation time of both MBSBBR and combined SBR-MBR process was 60 days.

Fig. 2. (a) Schematic diagram, and (b) actual photograph of the SBR-MBR setup.

2.6. Examination of the role of attached and suspended biomass

To determine the role of attached and suspended biomass in MBSBBR, a shake flask experiment was carried out at the end of the reactor operation. The experiment was carried out in three 500 ml Erlenmeyer flasks under anaerobic conditions. Each Erlenmeyer was filled with 500 ml of a mixture of simulated wastewater(as reported in Section 2.1 but containing a mixture of the three dyes with total concentration of 80 mg·L-1) and either suspended sludge(with same concentration as that in the reactor),suspended sludge plus immobilized PU foams or only the immobilized PU foams. The PU foams were taken from the MBSBBR at the end of its operation and their volume inside the Erlenmeyer flask were 150 ml. Incubation was carried out in a shaker incubator at 30 °C and 130 r·min-1for 48 h.

2.7. Analytical methods

For COD, color, nitrate and ammonium analysis, the samples were initially centrifuged at 4000 r·min-1for 15 min.COD concentration was measured according to Standard Methods 5220 D[34].Ammonium test kit (Merck Spectroquant, Cat. No. 1.14752.001)and nitrate test kit (Merck Spectroquant, Cat. No. 1.09713.002)were utilized to determine concentration of ammonium and nitrate, respectively.

To prevent autoxidation of the reduced azo dyes, all samples used for measuring color were diluted immediately with a solution of phosphate buffer (10860 mg·L-1NaH2PO4·2H2O; 5380 mg·L-1Na2HPO4·H2O) containing ascorbic acid (200 mg·L-1) as described previously [35]. The concentration of each dye in samples from runs employing a single dye was determined by reading the absorbance at the dye maximum absorption wavelength (λmax) of 590,585 and 540 nm for Black B,Black WNN and Red 3BS,respectively,using a UV-Vis spectrophotometer (T80 + UV-Vis spectrophotometer PG instrument. Ltd., England). The optical density (OD)readings were converted to concentration values using a calibration curve. Due to fact that the purity of these three commercial dyes was unknown, the data on color concentration presented in the paper are based on the weight of the used powder dyes and do not necessarily represent actual color concentrations.

For the measurement of the dye concentration in samples containing a mixture of dyes, two methods were employed. In one,similar to the measurement in samples containing a single dye,the OD at λmaxfor mixture of colors, which was about 558 nm,was determined. In the other method, the visible spectra from 400 nm to 700 nm for the samples was scanned and the areas beneath the absorption-wavelength curve was calculated and expressed as SU [14]. For qualitative determination of the formation and fate of anaerobic dye reduction metabolites (aromatic amines), samples were scanned through ultra violet range (200-400 nm) [36,37].

MLSS, mixed liquor volatile suspended solids (MLVSS) and specific oxygen uptake rate (SOUR) were measured according to Standard Methods [34]. Soluble microbial products (SMP) extraction and measurement of protein (SMPprotein), carbohydrate(SMPcarbohydrate) and humic (SMPhumic) content was performed according to methods described in a previous study [38]. For extraction of SMP, the samples were initially centrifuged(30 min, 3200 min-1) to separate the liquid and the biomass and then the supernatant was filtered through 0.45 μm membrane filter.The SMPproteinand the SMPhumicwere measured using the Folin phenol reagent method, whereas the SMPcarbohydratewas determined by the phenol-sulfuric acid method. The SMPtotalwas estimated as the sum of these three components. Standards of bovine serum albumin, humic substances and glucose were used to determine protein,humic acid and carbohydrate concentrations.All analysis was repeated three times.

For determination of biofilm concentration, 10 immobilized PU foams were taken from the MBSBBR and the biofilm measured according to previous work [39]. The carrier elements were separated from the water and dried overnight in an oven at 105°C until constant weight was obtained.The dried samples were weighed in order to determine the total mass (Mtot), composed of carrier element mass (Mcarrier) and the fixed biomass. The biomass was then washed off, the clean carriers were weighed, and the amount of biofilm solids attached to the 10 carrier elements(BS10)calculated[Eq. (1)]:

The biofilm concentration (mg·L-1) concentration (mg) in the reactor could then be determined from the value of the filling grades (FG) and the number of carrier elements at 100% filling grade (BS10) using Eq. (2):

In the present study, the filling grade was set at 30% and the number of carrier elements at 100% filling grade were 1000 L-1.

To prepare the samples for scanning electron microscopy(SEM),samples were rinsed with phosphate buffer (pH 8), fixed with 3%glutaraldehyde for 2 h, rinsed with trishydrochloric acid buffer and deionized water and gradually dehydrated after successive immersions in increasingly concentrated ethanol solutions (30%,60%, 80%, 96% and 100%). Each rinsing and dehydrating cycle took 15 min. The particles were then coated with gold powder and attached to supports with silver glue. SEM analyses were performed using a VEGAII digital scanning microscope (VEGAII LMU,TESCAN Co., Czech Republic).

2.8. Statistical analysis

Statistical analyses of the experimental data,including one-way ANOVA and linear correlation analysis, were performed using Minitab version 17 (Minitab Inc., State College PA, USA). Differences between the data obtained under different test conditions were considered statistically significant whenP＜0.05 [40].

3. Results and Discussion

3.1. Batch experiments

3.1.1. Dye removal during the anaerobic-aerobic process

Preliminary anaerobic batch experiments simulating the one stage anaerobic-aerobic process employing single dyes (Fig. S1)indicated that the rate of biological decolorization for Black WNN and Red 3BS was similar and reached a final removal efficiencies of about 94.2% and 91.4%, respectively. However, in the case of the run with Black B, a lower steady state dye removal efficiency(85.7%) was achieved at a later period which means that Black B is less amenable to anaerobic decolorization compared to the other two dyes. The same batch anaerobic test was subsequently performed on the mixture of the three dyes. The results (Fig. S1)showed that the corresponding rate of anaerobic decolorization for the mixture of three dyes approached those for Black WNN and Red 3BS, while the color removal efficiency was about 78%.

The results of abiotic tests showed that adsorption of Black B and Red 3BS to the sludge is insignificant, whereas in the case of Black WNN, at dye concentration of 96.8 mg·L-1, a maximum of 19.4% ± 0.6% of the dye adsorbed to the sludge. Therefore, it was concluded that the major mechanism of azo dye removal during the anaerobic phase or stage of anaerobic-aerobic process runs was biologically mediated anaerobic reduction.

Different observations, depending on the dye type, was made when the effluent of the abovementioned experiments was exposed to air. In the case of experiments with Black B and Black WNN,the effluent immediately turned into blue and green respectively,whereas for Red 3BS no color change occurred.Two mechanisms of color formation in the aerobic stage or phase of anaerobic-aerobic process has been put forward for Reactive Black B in previous studies.The first mechanism postulates that there is only a partial biologically-mediated reduction of Reactive Black B during the anaerobic phase or stage leading to the formation of hydrazo compounds (-HN-NH-), which, being air sensitive, oxidize back to the azo group when exposed to oxygen [28]. The second postulated mechanism for aerobic color reformation in the case of Reactive Black 5 attributes the aerobic color formation to the autoxidation of 1,2,7-triamino-8-hydroxynapthalene-3,6-disul fonic acid which is one of the reduction metabolites during anaerobic decolorization of Reactive Black B [7,41,42]. This results in a compound which has a λmaxwhich is very similar to that of the original dye solution. This results in the aerobic reformation of the original color of the dye. None of these mechanisms, however,can be responsible for color reformation in the case of Black WNN,since with this dye the original color did not reappear in the presence of oxygen [28]. The elucidation of the mechanism of aerobic color reformation with this dye requires further study.

Fig. 3 shows the result of experiments in which the mixture of dyes was exposed to two consecutive one stage anaerobic-aerobic cycles. The results reveal that aerobic color formation also occurs with mixture of dyes but that around 20.2 ± 3.3% removal of the formed color occurs during the aerobic phase. Since most azo dye and oxidized aromatic amines cannot be degraded under aerobic condition [5,41], the removal is probably through the adsorption of the oxidized aromatic amine to the sludge. However, as shown in Fig. 3, the re-establishment of anaerobic conditions leads to complete removal of the aerobically formed color although the color quickly returns upon re-exposure to aerobic conditions.

Fig. 3. Color profile of mixture of three dyes in anaerobic phase and during exposure of the anaerobically reduced dyes mixture to a sequential aerobic/anaerobic/aerobic environment (two cycles of one stage anaerobic-aerobic process).

3.1.2.Effect of number of stages on the performance of the anaerobicaerobic process

Two different combined anaerobic-aerobic processes, namely one stage and two stage processes, were compared for treatment of the simulated textile dyeing wastewater containing mixture of the three commercial reactive azo dyes. In the two stage process,activated sludge was exposed to anaerobic or aerobic condition whereas in one stage process, activated sludge was exposed to alternating anaerobic-aerobic environment.

The results presented in Fig. 4(a) show that in both processes,anaerobic and aerobic COD removal increased in repeated batches,and the overall COD removal reaching to (84.8 ± 1.2)% and (89.3 ±1.5)%,respectively for one stage and two stage processes.This suggests that in both one stage and two stage processes the exposure of the bacterial population in activated sludge to the simulated textile dyeing wastewater and anaerobic conditions has led to the adaptation of this population, probably as a result shift in the bacterial community structure which results in enrichment of facultative bacterial populations. Bonakdarpouret al. [7] have reported that in one stage processes, exposing anaerobic sludge to alternating anaerobic-aerobic environment decreases anaerobic COD removal,whereas exposing activated sludge to the same environment has the opposite effect.

Fig.4. (a)COD,and(b)color removal at the end of the anaerobic and aerobic phases of 1st and 6st batches of one stage and two stage combined anaerobic-aerobic processes.

The contribution of anaerobic COD removal was also significantly higher in two stage (70.1%) compared to one stage (around 50%) process. Higher contribution of the anaerobic stage to COD removal in two stage process runs has also been reported in previous studies in which anaerobic sludge or activated sludge have been employed in the anaerobic stage [7,27,41,43,44]. This is despite the fact that the mechanism of anaerobic COD removal in the case of anaerobic and aerobic sludge are very different to each other. In previous work with one stage process runs employing activated sludge, it has been reported that the majority of COD is removed in the aerobic phase of the anaerobic aerobic process[7,14,45]. The findings of previous work, together with the results obtained in the present work,suggests that,irrespective of the type of sludge used in the anaerobic stage,two stage process run can be considered more energy efficient than one stage process, since in the former the same amount of COD can be removed through the use of less energy.

The results presented in Fig. 4(b) show that two stage process performed better in terms of anaerobic and overall color removal after acclimation, compared to one stage processes. However,when anaerobic sludge was employed in the first stage of an anaerobic-aerobic process for treatment of media containing only reactive Black B, Bonakdarpouret al. [7] reported better anaerobic decolorization at concentration of 100 mg·L-1for one stage processes, whereas in the presence of sodium chloride, Vyrideset al.[8]reported higher anaerobic decolorization efficiency in two stage processes.

Also as Fig. 4(b) illustrates, although the majority of color was removed in the anaerobic phase/stage in both processes, the relative contribution of anaerobic color removal in the two stage process (98.9%) was higher compared to one stage process (74.1%).Similar results were reported by Bonakdarpouret al. [7] and Vyrideset al. [8] when comparing one stage processes that employed anaerobic or activated sludge with two stage processes that employed anaerobic sludge in the first stage.

The poorer color and COD removal in one stage anaerobic-aerobic process runs at the end of the repeated batches can be partly related to the higher decrease in MLSS during the one stage process runs. MLSS was about (2500 ± 250) mg·L-1in first batch for both one and two processes which,at the end of the sixth batch,dropped to(150±20)mg·L-1for the one stage runs whereas it only dropped to(370±35)mg·L-1and(1400±89)mg·L-1at the end of the anaerobic and aerobic periods of two stage process runs,respectively.It was previously reported that biomass concentration can influence azo dye decolorization, with lower concentrations leading to decrease in decolorization rates [18,31,46]. Another factor that can explain the lower color and COD removal obtained in one stage compared to two stage anaerobic-aerobic processes is the fact that in the former the same bacterial population is exposed to alternating anaerobic/aerobic environments. The negative effect of alternating anaerobic-aerobic process during one stage processes employing either activated or anaerobic sludge has been reported in some previous studies [7,31], although in one study this phenomenon resulted in better mineralization of anaerobic decolorization metabolites[47].

3.2. Reactor experiments

Based on the data obtained in batch experiments, the reactor runs were performed as a two stage process.

3.2.1. Anaerobic reactors

SBR and MBSBBR was operated for 60 days during which MLSS,color, COD, nitrate and ammonia were monitored.

3.2.1.1. Biomass concentration.Fig. 5(a) illustrates the trend of change of MLSS in the first stage of the anaerobic-aerobic process which was carried out in a SBR.There was a continuous decrease in MLSS in the initial 27 days operation of the reactor after which MLSS attained a steady state value of (309 ± 63) mg·L-1.

Fig. 6. Variation of (a) color concentration, and (b) COD concentration during one cycle in MBSBBR and SBR.

In MBSBBR, MLSS of the suspended biomass gradually decreased in the first 20 days of the reactor operation and fluctuated between (172 ± 48) mg·L-1and (759 ± 70) mg·L-1during the rest of the operation(results not presented).On the other hand,the attached biomass concentration rose to a value of (16.7 ± 0.5)mg·L-1at the end of the operation. Statistical analysis revealed no meaningful difference between the total steady state biomass concentration in the MBSBBR and SBR. It was found that 96% of biomass was suspended and only 4%was attached to the PU foam.In some previously reported studies with MBBR,the majority of the biomass was also reported to be suspended[48],whereas in some others the opposite result was observed [39]. The distribution of the biomass between suspended and attached in a MBBR is expected to affect its pollutant removal performance.

Fig.5. (a)MLSS concentration in SBR during 60 days operation, and (b) UV-visible spectra (200-700 nm) for influent sample, MBSBBR and SBR effluent samples.

3.2.1.2. Color removal.Color removal reached steady state condition after 15 and 25 days in SBR and MBSBBR, respectively.Average steady state color removal was (89.2 ± 0.8)% in SBR and(87.3 ± 1.1)% in MBSBBR. The UV-Vis spectrum of the simulated textile dyeing wastewater feed and the samples withdrawn from the reactors is presented in Fig.5(b).It can be seen that the absorbance peak at 558 nm was almost completely absent in the sample from the effluent of both reactors; however, due to the formation of aromatic amines the UV-region absorbance area increased after the anaerobic stage [7,28,36,37], and this increase was slightly higher in the MBSBBR.The latter suggests a slightly higher production of anaerobic dye reduction metabolites in the MBSBR but statistical analysis showed that there was no meaningful difference between color of the effluent of MBSBBR and SBR under steady state condition.

Trend of change in the concentration of color during one cycle is presented in Fig.6(a).The results show that color concentration in the MBSBBR was almost steady after 6 h while the color inside the SBR reached a steady value equal to that of MBSBBR only after 20 h.This implies that the anaerobic color removal rate was higher in the MBSBBR.Higher color removal rate was also observed in MBBR in other studies when suspended sludge system and MBBR were compared for real textile wastewater treatment [22,23].

3.2.1.3. COD removal.Under steady state conditions, COD removal in MBSBBR and SBR was(70.5±5)%and(51.0±6.1)%,respectively.There was virtually no COD removal during the first 30 days of SBR operation after which there was a significant increase in the COD removal efficiency. The reason for this difference was explored in further experiments reported in Section 3.3. Higher COD removal rate in MBBR compared to suspended sludge system has also been reported previously [22,23].

Measurement of COD concentration during one cycle during the end of the operation of the two reactors [Fig. 6(b)] showed that during the initial period of operation in both SBR and MBSBBR,there was no pronounced change in the COD concentration whereas afterwards it gradually decreased and achieved steady concentrations during the last ten hours in both reactors. This means that both COD removal efficiency and rate were higher in MBSBBR compared to SBR.

3.2.2. Aerobic reactor

SBR effluent was fed to the MBR to evaluate the MBR performance for aerobic treatment of anaerobically treated the simulated textile dyeing wastewater.Average COD removal efficiency of MBR was about(71.6±13.6)%.Incomplete COD removal was also previously reported by García-Martínezet al.[49]who also used sodium acetate as the carbon source for decolorization of acid orange 7 in a continuous up flow stirred packed-bed reactor followed by aerobic MBR.Average total COD removal in anaerobic-aerobic process was about(77.1± 7.9)%.Since the carbon source used in the simulated textile dyeing wastewater (sodium acetate) is biodegradable, in order to elucidate the source of the residual COD in the effluent of the aerobic stage,the trend of change of SMPhumic,SMPcarbohydrateand SMPproteinconcentration in MBR effluent were monitored during its operation [Fig. 7(a)]. The results show an increase in the concentration of SMPhumicand SMPproteinduring MBR operation.Also SMPhumicconcentration was much higher than both SMPproteinand SMPcarbohydrate.One of the reasons for the increase in SMPhumicconcentration could be the autoxidation of some aromatic amines to humic like oligomer and polymeric structures in the aerobic bioreactor [5].

Furthermore, at the beginning of the operation, SMPprotein/SMPcarbohydrateratio was 0.35 but gradually increased to 2 at the end of the operation. The increase in this ratio can be explained by the fact that the carbohydrates are more easily degraded by bacteria than proteins [50]. These results, together with the fact that sodium acetate is biodegradable, indicate that the majority of the residual COD in the effluent of the MBR was contributed by SMP,especially SMPhumicand SMPprotein. Abdollahzadeh Sharghiet al.[33] treating real vegetable oil refinery wastewater using submerged MBR also observed that the effluent COD was mainly composed of humic substances followed by proteins. Baetaet al. [51]during azo dye (Remazol Golden Yellow RNL) removal in a combined anaerobic/aerobic reactor showed that all the residual soluble COD effluent from the aerobic phase was due to SMPs produced in anaerobe and aerobic reactors.

It has been reported that the production of SMP can be induced by the presence of toxic compound[33,52].The presence of SMP in the MBR effluent during its operation can therefore,at least partly,be attributed to the presence of anaerobic dye reduction metabolites (usually aromatic amines). According to previous studies,these compounds can be more toxic to the bacterial population compared to the parent azo dye, and are only partially degraded aerobically [32,51,53]. It has been reported that the toxicity of dyes, especially sulfonated azo dyes [54], and aromatic amine,leads to increased cell lysis and EPS production and hydrolysis in the anaerobic stage of an anaerobic/aerobic process, resulting in increase in SMP concentration [51].

Trend of change of MLSS and MLVSS concentration during operation of MBR is presented in Fig. 7(b). MLSS was steady at about(5800 ± 377) mg·L-1during the first 30 days, but with decrease in COD concentration of SBR effluent, the MLSS dropped to(3780 ± 120) mg·L-1. Based on statistical analysis, MLSS reached steady state conditions after 45 days. According to linear correlation analysis,MLSS of MBR was dependent on COD of the SBR effluent(rp= 0.873,p-value= 0.00). The trend of change of SOUR as an indication of the aerobic metabolic activity of microorganisms during MBR operation changed from an average value of(4.5±0.2)mg O2· (g MLVSS)-1·h-1during the initial 30 days to average value of(2.8 ± 0.2) mg O2· (gMLVSS)-1·h-1during the second half of the operation of the MBR. The decrease in SOUR might be explained by the decrease in MLSS values [40], the toxic effect of the anaerobic dye reduction metabolites (aromatic amines) as well as increase in the concentration of SMP compounds especially the refractory humic substances during MBR operation. MLVSS/MLSS ratio was almost constant and its average throughout MBR operation was about (0.5 ± 0.05).

In terms of color removal,MBR reached steady state conditions after 15 days and the average steady state color of the effluent was(54.6±4.1)SU[Fig.7(c)].Average color removal in the anaerobicaerobic process was(79.9±1.5)%.Despite this relatively high color removal in SBR-MBR reactors,due to autoxidation of aromatic amines formed in the anaerobic reactor, the color of the MBR effluent was always blue [Fig. 7(c)]. The aerobic re-formation of color was also observed during batch runs as reported in Section 3.1.1.

According to the data presented in Fig.8(a),the peak at 260 nm,which corresponds to the aromatic amine produced in the anaerobic reactor, decreased in the MBR effluent. This indicates that the aromatic amines partially degraded under aerobic conditions.Aerobic aromatic amines degradation in conjunction with reduced peak in UV region of UV-Vis spectra was observed in other studies[7,27,28,35,45,55].Another reason for decreasing aromatic amines in the MBR could be covalent bonding to the SMP, primarily SMPhumic. Bonding can occur through nucleophilic addition of the amine group with the carbonyl functionality of the humic substances and/or oxidative mechanisms [56-58].

Further confirmation of aromatic amine degradation was obtained through measurement of nitrate in the MBR effluent.Fig. 8(b) shows that the nitrate concentration in the MBR effluent varied between (118.1 ± 4.4) mg·L-1and (278 ± 5) mg·L-1during its operation. This is significantly higher than the (5.06 ± 2.02)mg·L-1value present in the SBR effluent. Increase in nitrate concentration resulted in increase in the peak at 220 nm of the UV spectra [Fig. 8(a)]. Since the ammonia concentration in the SBR effluent varied between (7.12 ± 0.17) mg·L-1and (17.46 ± 0.33)mg·L-1and fell to less than 1 mg·L-1in the MBR effluent, the nitrate formation in the MBR can be mainly attributed to the oxidation of the aromatic amices formed during the anaerobic dye reduction process. Increase in nitrate concentration due to aromatic amines degradation has been reported before. de Jageret al. [26] demonstrated that degradation of aromatic amines in an aerobic tank caused increase in nitrate level to 3 mg·L-1when real textile was treated. Increase in nitrate concentration from 16 mg·L-1to 83 mg·L-1was observed during decolorization of Direct Black 38 in anaerobic/aerobic sequential reactors [59].

3.3. The role of attached biomass and suspended biomass

Fig.9(a)shows the SEM image of the surface of a PU foam taken from the MBSBBR. The biofilm found on the surface of the media was seen with a cluster of cells connected by an irregular arrangement of interconnected extracellular polymeric substances projections.Comparison of this image with the corresponding image for a raw PU foam [Fig. 9(b)] confirms the attachment of the bacterial population in activated sludge during the biofilm formation process.

Fig. 7. (a) Variation of effluent SMPprotein, SMPcarbohydrate and SMPhumic in MBR, (b) MLSS and MLVSS concentration of MBR, and (c) variation of MBR effluent color and total color removal of SBR-MBR during 60 days operation.

In order to explain the difference in the performance of SBR and MBSBBR, and to better understand the role of attached and suspended biomass in COD and color removal, separate batch experiments were performed with suspended biomass, PU foams containing attached biomass, and the mixture of the two using samples taken from the MBSBBR at the end of its operation. As the data presented in Fig. 10 show, different color and COD removal pattern was observed during runs in which either attached or suspended biomass were used on their own. In the former case, (37.1 ± 0.3)% color removal was achieved during 48 h incubation whereas in the latter case and in the runs with mixture of suspended and attached biomass the color removal was (77.3 ± 0.2)% and (78.1 ± 0.1)%, respectively. COD concentration was almost unchanged during 30 h of incubation in runs with suspended biomass, whereas in the runs with either attached biomass or a mixture of attached and suspended biomass,COD concentration gradually decreased. COD removal in the runs with attached biomass, suspended biomass and combination of suspended and attached biomass was (76.3 ± 6.9)%, (37.6 ± 2.3)%and (88.5 ± 0.3)%, respectively. According to Mahendranet al.[48],due to difference in physiochemical and microbial properties of the biofilm and floc, mechanisms of contaminant removal and sorption by their surfaces are different.

Fig. 8. (a) UV-visible spectra (200-700 nm) of influent sample, SBR and MBR effluent samples, and (b) trend of change of NO3 concentration of MBR effluent during its operation.

Fig. 10. The effect of attached biomass, suspended biomass and mixture of suspended and attached biomass on color and COD concentration during 48 h incubation; (a) Color concentration profile, and (b) COD concentration profile.

The above results points to an indirect correlation between COD and color removal, with attached biomass exhibiting a high COD removal and a poor color removal whereas exactly the opposite trend was found with suspended biomass. Previous studies have shown that the nature of the electron donating compound can have a marked effect on the rate and extent of anaerobic decolorization, with acetate being generally considered to be a poor electron donor [60]. Other compound like ethanol and glucose can reduce the azo dye faster than acetate. It is not clear what the electron source was in the runs with suspended biomass.Some studies have reported that the sludge organics can be used as an electron donor for azo dye decolorization [60,61].

Fig.9. SEM images of microorganisms attached to the PU foam surface;(a)At the end of the operation(×10 k,5 μm),and(b)before attachment of microorganisms(×10 k,5 μm).

Distinctive microbial community in biofilm and suspended sludge has been observed in hybrid suspended growth - biofilm wastewater treatment system [48]. According to García-Martínezet al.[49]and Silvaet al.[62]characteristic of bed materials affects the bacterial community structure that develops in biofilms formed on supports. Therefore the proper choice of support used in MBSBBR can, through change in structure and population of the bacterial community on the support surface, enhance its performance.

4. Conclusions

In the present study, a simulated textile dyeing wastewater containing three commercial-grade reactive azo dyes was subjected to combined anaerobic-aerobic process employing activated sludge and the following major results obtained:

(1) Preliminary experiments showed that,in sequential anaerobic-aerobic process, depending on the type of commercial dye, the anaerobically reduced color can aerobically reform during the initial periods of the aerobic phase or stage although the reformed color is partially removed by the end of the aerobic period and totally removed during the subsequent anaerobic period of the process.

(2) Comparison of one stage and two stage anaerobic-aerobic processes that employ activated sludge showed both a better overall COD and color removal performance, as well a higher contribution of the anaerobic phase/stage to the overall COD and color removal, for two stage processes. The latter suggests that, when activated sludge is employed in the process, two stage process are more energy efficient than one stage processes since anaerobic phase/stage requires less energy to operate compared to the corresponding aerobic one.The difference in performance of one and two stage processes is related to the fact that during one stage processes activated sludge is exposed to alternating anaerobic-aerobic conditions whereas in two stages processes the bacterial population in activated sludge is subjected to an unvarying anaerobic or aerobic environment.

(3) Comparison of two reactor types as potential candidates for the implementation of the anaerobic stage of the two stage process showed a better COD removal efficiency and a higher color removal rate in the case of the type (MBSBBR)which employs both suspended and attached biomass compared to the type(SBR)in which only suspended biomass is used. However, similar color removal efficiencies was obtained with both reactor types.

(4) In order to better understand the underlying reason for the difference in the performance of MBSBBR and SBR, the role of suspended and attached biomass in the removal of COD and color was further explored. The results revealed that suspended bacterial populations had a better color removal performance whereas the bacterial community present in the biofilm had a superior COD removal performance. However,the comparatively low percentage of attached biomass in the MBSBBR (around 4%) implied that further work on carrier type and material could lead to improved removal performance for this reactor type.Also,for better elucidation of the difference in contaminant removal mechanisms of attached and suspended biomass, microbial community analysis should be performed.

(5) The employment of a MBR in the aerobic stage of the anaerobic-aerobic process resulted in (77.1 ± 7.9)% and(79.9 ± 1.5)% total COD and color removal respectively. The incomplete color removal was partly the result of the autoxidation of anaerobic reduction metabolites whereas the incomplete COD removal was mainly attributed to the production of SMP (especially SMPhumicand SMPprotein),which in turn seems to have been induced by the toxicity of the azo dye reduction metabolites produced during the anaerobic stage. The occurrence of partial mineralization of the anaerobic dye reduction metabolites during MBR operation,probably as a result of both aerobic biodegradation and covalent bonding to SMPhumic, was experimentally demonstrated.

Acknowledgements

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The present study was supported by Takmiliran textile dyeing factory (272219601); and Materials and Energy Research Center(MERC) (99392003).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.10.006.