Journal of Hazardous Materials 286 (2015) 416424

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journa l homepage: www.e lsev ier .com/ locate / jhazmat

Organics and nitrogen removal from textile auxiliaries wastewaterwith A2O-MBR in a pilot-scale

Faqian Suna, Bin Suna,b, Jian Hua, Yangyang Hea, Weixiang Wua,

a Institute of Environmental Science and Technology, Zhejiang University, Hangzhou 310058, Chinab Shanghai Electric Group Co. Ltd. Central Academe, Shanghai 200070, China

h i g h l i g h t s

A pilot-scale A2O-MBR system treating textile auxiliaries wastewater was assessed. Organic matter and recycle ratio strongly affected the performance of the system. GC/MS analysis found some refractory organics in the MBR permeate. Combination of organic foulants and inorganic compounds caused membrane fouling.

a r t i c l e i n f o

Article history:Received 12 October 2014Received in revised form 8 December 2014Accepted 10 January 2015Available online 13 January 2015

Keywords:Textile auxiliaries wastewaterAnaerobicanoxicaerobicMembrane bioreactorRecycle ratioRefractory organics

a b s t r a c t

The removal of organic compounds and nitrogen in an anaerobicanoxicaerobic membrane bioreactorprocess (A2O-MBR) for treatment of textile auxiliaries (TA) wastewater was investigated. The resultsshow that the average efuent concentrations of chemical oxygen demand (COD), ammonium nitro-gen (NH4+N) and total nitrogen (TN) were about 119, 3 and 48mg/L under an internal recycle ratio of1.5. The average removal efciency of COD, NH4+N and TN were 87%, 96% and 55%, respectively. Gaschromatographmass spectrometer analysis indicated that, although as much as 121 different types oforganic compounds were present in the TA wastewater, only 20 kinds of refractory organic compoundswere found in theMBR efuent, which could be used as indicators of efuents from this kind of industrialwastewater. Scanning electronmicroscopy analysis revealed that bacterial foulantswere signicant con-tributors to membrane fouling. An examination of foulants components by wavelength dispersive X-rayuorescence showed that the combination of organic foulants and inorganic compounds enhanced theformation of gel layer and thus caused membrane fouling. The results will provide valuable informationfor optimizing the design and operation of wastewater treatment system in the textile industry.

2015 Elsevier B.V. All rights reserved.

1. Introduction

As one of the largest industries in the world, the textile indus-try consumes large quantities of textile auxiliaries (TA), whichincludemore than 100 kinds of specialty chemicals, such as soften-ing agent, phosphates, polyamide resins, acrylic chelating agents,polyurethane coating agents and stiffening agents. In the TA pro-duction, considerableamountsofTAwastewater aregenerated. TheTA wastewater is very chemical-intensive and known to containhigh concentrations of organic matter, non-biodegradable matter,toxic substances and ammonia [13]. These compounds producelong-term environmental impacts, it is therefore important to

Corresponding author. Tel.: +86 571 88982020; fax: +86 571 88902020.E-mail address: weixiang@zju.edu.cn (W. Wu).

remove these organics and ammonia from the TA wastewater forreducing their harm to the environment.

TA wastewater could be treated by physicochemicaland biological methods or suitable combinations of them.Physicochemical methods involve adsorption, ion exchange,coagulationocculation, as well as advanced oxidation processes(such as Fenton oxidation, ozonation, photocatalytic oxidationand electrochemical oxidation) [46]. However, methods suchas coagulation, ion exchange, and adsorption only transfer theorganic pollutants from one phase to another. Advanced oxidationprocesses, such as Fenton oxidation, ozonation, photocatalyticoxidation and electrochemical oxidation, are very efcient andthe fastest way for destruction of organic compounds, but theyare expensive and could not be adopted commercially. Comparedwith physicochemical methods, biological treatment is often themost economical alternative for pollutant removal [7].

http://dx.doi.org/10.1016/j.jhazmat.2015.01.0310304-3894/ 2015 Elsevier B.V. All rights reserved.

F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424 417

Due to the low biodegradability and high toxicity of manytextile chemicals, a conventional activated sludge system is inad-equate in removing high concentrations of refractory organics andammonia. In recentyears, themembranebioreactor (MBR), becauseof complete biomass retention, has been successfully integratedwith an anaerobicanoxicaerobic (A2/O) system for industrialwastewater treatment [8]. The combined system was efcientand cost-effective in removing refractory pollutants and ammo-nia, especially at high and varying loading rates [9]. The use ofanaerobic process as a pretreatment process to partially convertrefractory organics to intermediates that are more readily degrad-able, could be attractive for refractory wastewater treatment. Theanoxicaerobic process is a good option for achieving biologicalnitrogen removal via pre-denitrication and aerobic nitrication.MBR can keep a long sludge retention time (SRT), which allows thesystem to keep a sufcient amount of slow-growing bacteria, suchas ammonia oxidizing bacteria and those specializing in degradingrefractory compounds [10]. Therefore, the A2O-MBR system is veryattractive for chemical-intensive industrial wastewater treatment.

The efciencies of organic compounds and nitrogen removal intheA2O-MBR systemare inuencedbymany factors, such as chem-ical characteristics of wastewater, internal recycle ratio, hydraulicretention time (HRT)andsubstrate loading rate. [1113]. Thechem-ical characteristics of TA wastewater determined the biodegradingcapacity of the system. Internal recycle ratio was found to have animportant role in nitrogen removal performance of the system. Inaddition, membrane fouling was inevitable in MBR systems. In theprevious studies reported in the literature, therewas little informa-tion on the nitrogen removal and behaviors of organic compoundsin a A2O-MBR system for TA wastewater treatment particularly atpilot-scale operation.

The objectives of the present research were: (1) to investigatethe pollutant removal performance in a A2O-MBR system, (2) tocharacterize in detail organic chemical composition in the rawwastewater and the efuents, (3) to compare the performance oftwo kinds ofmembranemodules and examine themembrane foul-ing behaviors.

2. Materials and methods

2.1. Experimental set-up and operating conditions

Thepilot-scale experimentswere carriedout in a sequential sys-tem (in Fig. 1) of an anaerobic reactor (A1), an anoxic reactor (A2),followed by an aerobic membrane bioreactor (O-MBR). The work-ing volumes of the three reactors were 3.8m3, 7.5m3 and 5.6m3,respectively. Reactor A1 was packed with bamboo carbon, whichwasgeneratedat600 Cundera slowpyrolysisprocesswithadiam-eter of 35mm. In the reactor A2, amechanical stirrerwas installedto agitate its content. The O-MBR could divide into buffer zoneand reaction zone. An internal pump was installed in the bufferzone andmixed liquor from the bottomwas continuously pumpedto the reaction zone. Two different kinds of membrane moduleswere immersed and symmetrically placed in the reaction zone. Onewas a hollow ber (HF)membranemodulemade of polyvinylideneuoride (PVDF)with anominal pore sizeof 0.1manda totalmem-brane surface area of 12.5m2 (SMM-1013, Memstar TechnologyCo. Ltd., Singapore). The other one was a at-sheet (FS) membrane

module with a nominal pore size of 0.1m and a total membranesurface area of 10.4m2 (DF80, Jiangsu Dafu Membrane Technol-ogy Co. Ltd., China). Air diffusers were installed underneath themembrane module to provide dissolved oxygen (DO) as well asto control membrane fouling and clogging. DO was maintained at35mg/L in the MBR during the experiments. Membrane ltrationthrough a suction pump was carried out in an intermittent suc-tionmodewith 9min of suction followed by a 2min release. In situmembrane chemical cleaning was performed to reducemembranefouling by 0.5% NaClO for 1h when the trans-membrane pressure(TMP) increased to 20kPa.

The sequential system was operated in a pre-denitricationmode, which consisted of three different steps. Firstly, TAwastewater was continuously pumped into reactor A1 for hydrol-ysis/acidication. Secondly, the water passed to reactor A2, wheredenitrication took place and a fraction of organic matter wasdegraded by heterotrophic bacteria. Thirdly, the water passedto the O-MBR where nitrication and degradation of remainingorganic matter occurred, and the permeate water was separatedby membrane modules while nitrate was recycled to A2 by par-tial recirculation. The inoculating sludge was drawn from theanaerobic sludge in a conventional TA wastewater treatment plant(Hangzhou, China). During the study, different internal mixedliquor recycle ratios from O-MBR to A2 were employed to assesschemical oxygen demand (COD) and nitrogen compounds removalof the system. Recycle ratios ranged from 0.5 to 2.5 of the inu-ent ow rate. At the start-up period (Phase I), the inuent owrate was kept at 250 L/h, and the recycle ratio was controlled atabout 0.5. The system was operated stably for 20 days. Afterwards,the recycle ratio was changed to about 1.5 and the system wasthen operated in this condition for 74 days (day 2194, Phase II).Fromday 95 onward, the recycle ratiowasmaintained at 2.5 (PhaseIII) and methanol at a equivalent concentration of about 240mg/LCOD was supplemented as an organic source in the A1 efuent.Because the recycle ratio was high at all times, mixed liquor sus-pended solids (MLSS) concentrations in the reactors were similar.During the whole period of the study, no sludge was removedfrom the plant except for some incidents with accidental sludgeloss. MLSS concentration in the O-MBR and the A2 uctuated at35005000mg/L. Theexperimentwas conductedunder anambienttemperature of 2025 C. Furthermore, parameters such as tem-perature, level of the tanks, TMP, ow rates of partial recirculationand the permeate inside the O-MBR were measured automaticallyand registered continuously in a database with the aid of Realinfosoftware (Realinfo., China).

2.2. Characteristics of TA wastewater

TA wastewater was collected from Transfar Chemicals Co. Ltd.,one of the largest TA manufacturing factories located in the south-east of China. Itwasrstlypumped into an intermediate tankbeforebeing continuously pumped into the bioreactor. A summary of theinuent characteristics at different experimental phases are shownin Table 1. The raw wastewater had a highly variable composi-tion. It can be seen that the average COD concentrations in thethree phaseswere 657, 944 and 828mg/L, while total nitrogen (TN)concentrations were 115, 106 and 121mg/L, respectively.

Table 1Characteristics of TA wastewater observed during the experiments.

Phase pH COD (mg/L) NH4+N (mg/L) TN (mg/L) COD/TN Number of data points

I 7.299.37 657 127 87 5 115 7 6.9 1.8 10II 7.428.71 944 163 90 19 106 19 8.0 2.0 35III 7.508.74 828 216 94 21 121 31 6.0 1.3 23

418 F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424

Fig. 1. Schematic diagram of the pilot-scale A2O-MBR system.

2.3. Analytical methods

Ammonium nitrogen (NH4+N), nitrite nitrogen (NO2N) andnitrate nitrogen (NO3N) were analyzed according to the stan-dardmethods for the examination ofwater andwastewater (APHA,1998). COD and TN were measured by the Hach COD kits and HachTN kits, respectively (Hach, USA). The pH was measured with a pHmeter (SG2,Mettler-Toledo, Greifensee, Switzerland) and DO by anoxygen probe (YSI 550A, YSI, Ohio, USA). MLSS was determined byvacuum ltration of 10mL of sludge and then dried at 105 C for2h.

Gas chromatographmass spectrometer (GCMS) was used toanalyze of the samples with liquidliquid extraction pretreatmentusing CH2Cl2 (chromatogrampure grade, Tianjin Chemical Factory,Tianjin, China). Subsequently a 200mL sample was extracted using10mL of CH2Cl2 (chromatogram pure grade, Fisher, USA) three

times at neutral, alkaline and acidic conditions, respectively [14].Then, a 1mL pretreated sample was analyzed by 6890N/5975BGCMS system (Agilent Corporation, USA). A DB-5MS capillary col-umn with an inner diameter of 0.25mm and a length of 30m wasused in the separation system. The GC column was operated in atemperature programmed mode by maintaining the temperatureat 40 C for 4min, then increasing to 300 C with an increment of8 C/min, and total run timewas46min. Theelectron impact energywas set at 70 eV. Organic compounds analysiswas undertakenwithreference to the instrument library (NIST 05. L) database.

At the end of ltration, a piece of membrane was cut fromthe middle of the membrane modules. The sample was xed with3.0% glutaraldehyde in 0.1M phosphate buffer at pH 7.2. The sam-ple, dehydrated with ethanol and silver-coated by a sputter, wasobserved using the scanning electronmicroscopy (SEM) (JEOL JSM-5600LV, Tokyo, Japan). Inorganic elementary analysis was carried

Fig. 2. Variation of COD concentrations in the pilot-scale A2O-MBR system during the 3 phases of operation.

F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424 419

out on an Elementary Vario El (Germany) system. Wavelength dis-persive X-ray uorescence (WD-XRF) data were obtained from aBruker S4 Explorer spectrometer.

3. Results and discussion

3.1. COD removal

The COD removal of the pilot-scale A2O-MBR during the wholeoperation period is presented in Fig. 2. It shows that the A2O-MBRperformed well in organic matter removal. COD concentration inthe TA wastewater was in the range of 3721291mg/L with anaverage value of 879mg/L. In the reactor A1, about 100mg/L CODwas steadily removed. The relatively low COD removal efciencyimplied organic compounds with macromolecular structure andlarge molecules in the raw wastewater were partially convertedinto small-molecule substrates through hydrolysis-acidicationprocesses in an anaerobic environment. A drop in mean pH of 0.7unit appears to support this (datanot shown). In the reactorA2,CODconcentration was diluted by mixed liquor recirculation. The aver-age COD concentration of A2 efuent was 320mg/L. In the O-MBR,most of the COD was removed. Despite the uctuations of inuentCOD concentrations, the average efuent COD concentration was131mg/L, with an average total COD removal of 85%. Moreover, itwas also observed that efuent CODconcentrationwas not affectedwhen the recycle ratio was increased from about 0.5 to 2.5. Theseresults indicated the system could maintain a consistently goodperformance of COD removal. The efuent could meet the nationaldischarge standard ofwater pollutants for industrial wastewater inChina (COD

420 F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424

Fig. 3. Proles of internal recycle ratio, ammonia and TN concentration in the A2O-MBR system.

not promote the increase innitrogen removal. Thesedata show thatthe limited denitrication capacity is attributable to the oxidationof organicmatter in the anoxic reactorwith the oxygen of the inter-nal recycle at high ratios. High DO concentration (above 5mg O2/L)present in the recycle mixed liquor deteriorated TN removal ef-ciency [15,22]. Therefore, it could be concluded that a recycle ratioaround 1.5 should be recommended to provide a proper comprisebetween removal efciency and cost under normal conditions.

3.3. Behavior of organic compounds

In order to gainmore insight into the organic composition of TAwastewater and the removalmechanisms for these organics duringthe treatment process, GCMS analysis was carried out. The chro-matograms of the raw wastewater and the efuent from each unitof A2O-MBR are shown in Fig. S1. It could be seen that, at least 121types of organic components were present in the TA wastewater,which contained 92 kinds of alkanes and 29 other kinds of organ-ics belonging to chemical additives and their derivatives. Aftertreatment by anaerobic reactor, the categories of organic com-pounds decreased to 90. These organics were more readily

degraded in the subsequent anoxic stage with the categories fur-ther decreasing to 55. At last, only 20 different kinds of organiccompounds were found in the permeate from MBR.

The details of organic compounds are summarized in Table 3 .It appears that the alkanes and butylated hydroxytoluene (BHT),corresponding to 47.6% and 26.6% of the total integration area,respectively, dominated the organic matter in the TA wastewater.BHT, a commonantioxidant,was likely to come from the rawchem-icals used for TA production [23]. Certain compounds, such as pyri-dine, tetramethylbutanedinitrile (TMSN), butoxytrimethylSilane,p-xylene and 1,3-dimethylbenzene, which each covered morethan1%of integration area,were the sub-dominant organic species,whereas other compounds were the minor species. These organiccompounds in the inuent are generally composed of synthesis-related chemicals and their derivatives. After the treatment byanaerobic tank, most of organic compounds have decreasedslightly, such as alkanes, BHT, pyridine and butoxytrimethylSi-lane. Compared to the alkanes with higher molecular weights,alkanes with low molecular weights were removed with rela-tively high efciency. Meanwhile, some new intermediates wereproduced. For instance, several alkaneswere probably converted to

F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424 421

Table 3Main organic compounds present in the inuent and the efuent by GCMS analysis.

No Organic compounds Relative mass percentage (%) Total removal efciencya(%)

TA wastewater A1 efuent A2 efuent MBR permeate

1 Alkanes 47.63 40.23 44.17 8.44 982 Butylated hydroxytoluene 26.57 16.73 8.8 6.05 983 Pyridine 4.81 0.77 1004 Tetramethylbutanedinitrile 4.61 5.1 15.59 34.02 195 Silane, butoxytrimethyl- 2.95 0.52 1006 1,2-Benzenedicarboxylic acid, mono(2-ethylhexyl) ester 1.72 1.52 3.18 2.06 877 p-Xylene 1.45 2.06 0.43 1.87 868 Benzene, 1,3-dimethyl- 1.11 1.69 1009 Acetamide, 2,2,2-triuoro-N-(trimethylsilyl)- 0.94 100

10 Ethylbenzene 0.88 1.9 0.68 0.57 9311 Acetic acid, (trimethylsilyl)- 0.83 0.28 10012 Methane, dimethoxy- 0.77 1.05 10013 Sulfur 0.64 0.47 10014 Benzene, 1,2,4,5-tetramethyl- 0.51 0.33 0.36 10015 Naphthalene, decahydro-2-methyl- 0.44 10016 Styrene 0.41 0.79 1 3.65 217 Cyclotetrasiloxane, octamethyl- 0.4 10018 Bis(trimethylsilyl) triuoroacetamide 0.39 10019 Phosphorous acid, tris(decyl) este 0.37 10020 2H-Benzotriazole, 2-ethyl- 0.36 0.43 0.61 1.22 6321 Hexa(methoxymethyl) melamine 0.32 0.43 0.92 1.94 3422 Silanamine, N-methoxy-1,1,1-trimethyl-N-(trimethylsilyl)- 0.3 10023 Naphthalene, decahydro-2,6-dimethyl- 0.3 10024 Morpholine, 4-methyl- 0.26 0.72 10025 3-Methyl-2-(2-oxopropyl) furan 0.25 10026 Formamide, N,N-dimethyl- 0.21 0.51 10027 Toluene 0.2 0.21 0.24 0.87 5228 Propanoic acid, 2-methyl-, trimethylsilyl ester 0.19 10029 1,4-Dioxane 0.18 0.15 0.95 1.63 130 d-Limonene 7.16 31 Bicyclo[410]hept-2-ene, 3,7,7-trimethyl- 5.57 32 Hexanoic acid, 2-ethyl- 3.28 33 Hydrazine, 1-methyl-1-(2-propenyl) 2.09 10.1 17.3 34 Benzene, 1-methyl-2-(1-methylethyl)- 2.05 35 Benzene, 1-ethyl-2,4-dimethyl- 0.81 36 Cyclohexene, 1-methyl-4-(1-methylethylidene)- 0.64 37 6-Tetradecanesulfonic acid, butyl ester 0.6 1.07 38 Tridecanol, 2-ethyl-2-methyl- 0.56 39 2-Ethylhexanoic acid, trimethylsilyl ester 0.3 40 Phenol, 3-(1,1-dimethylethyl)-4-methoxy- 1.86 41 Dodecane, 4,6-dimethyl- 0.65 42 Decane, 3,7-dimethyl- 0.56 43 Pentadecane 0.5 44 3-tert-Butyl-4-hydroxyanisole 1.86 3.46 45 2,4-Dimethoxy-N,N-dimethylbenzylamine 5.28 14.6 46 Others 1.05 1.09 2.32

Total 100 100 100 100

a Total removal efciency was calculated according to the peak area under the same condition.

2-ethyl-hexanoic acid, and 6-tetradecanesulfonic acid, butyl ester,and consistent with the general decrease of pH in the anaerobictank due to hydrolysis/acidication.

After the treatment in theanoxic tank, it canbe seen fromTable3that a considerable amount of benzene series organics and alka-nes, which were hardly biodegraded at the anaerobic stage, weredegraded. Pyridine, butoxytrimethylSilane, 1,3-dimethyl benzene,were completely removed. Furthermore, most of new interme-diates produced from the anaerobic tank were consumed atthe anoxic stage, probably leading to a satisfactory denitri-cation performance. By contrast, some new compounds, suchas 3-(1,1-dimethylethyl)-4-methoxy-phenol, 2,4-dimethoxy-N,N-dimethylbenzylamine etc. were produced. Therefore, in order totreat TA wastewater or other refractory wastewater, it is desir-able to have an anaerobic unit as a pretreatment method for betterdenitrication performance [24].

Most of the residual compounds in the anoxic efuents werefurther degraded at the aerobic stage. In order to make aquantitative comparison of the relative removal of organiccompounds from the TA wastewater, the areas of the peaks under

the chromatograms were determined under the same conditions.It can be observed that, alkanes, the main components in TAwastewater, were reduced by 98%, and only 6 types of alkanescould be detected in the MBR efuent. Similarly, BHT, the sub-dominant component, was reduced by 98%, which is better thanthat reported in municipal wastewater treatment using a com-bined systemofUASB and two constructedwetlands (1030%) [25].Although the majority of the organic pollutants were removed,there were still a few refractory contaminants present in the nalefuent. It seems that some compounds, such as TMSN, styrene andhexa-methoxymethylmelamine, were quite stable in the A2O-MBRsystem and often used as organic indicators of industrial efuentsfrom certain chemical production industries [26]. In addition, 1,4-Dioxane, a well-known hazardous pollutant, was hardly removedin the system. Despite the fact that 1,4-Dioxane biodegradationwas achieved through bioaugmentation and the use of enrichedmicrobial pure cultures [27], this compound could not be oxi-dized effectively by conventional biological process due to its highresistance to biotransformation, and could only be removed effec-tively by the advanced oxidation processes [28,29]. Suh and

422 F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424

Fig. 4. TMP evolution during 88-d continuous ltration operation in the MBR.

Mohseni [29] found that O3/H2O2 could be used to eliminate1,4-Dioxane, and the oxidation rate was rst order at the diox-ane concentrations lower than 50mg/L. Therefore, much attentionshould be paid attention to some refractory contaminants remain-ing in the nal efuent.

3.4. Membrane fouling behavior

Themembrane foulingof twodifferentkindsofmembranemod-ules (HF and FS) were observed bymonitoring TMP throughout theoperation of MBR. The variations of permeation ux and TMP are

Fig. 5. SEM micrographs showing the surfaces of clean membranes and fouled membranes. (a) New FS membrane surface, (b) fouled FS membrane surface, (c) new HFmembrane surface, (b) fouled HF membrane surface (FS =at sheet, HF=hollow ber).

F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424 423

Table 4Components of membrane foulants measured by WD-XRF.

Element Atomic ratio (%)

C F Si S Cl Ca Ti P N Na Mg Al K FeClean FS 61.17 38.74 0.01 0.01 0.03 0.01 0.03 Fouled FS 73.41 5.01 0.46 0.19 0.06 0.12 0.04 0.52 19.28 0.10 0.12 0.62 0.02 0.05Clean HF 44.21 41.48 0.01 0.01 8.24 5.01 1.02 0.02 Fouled HF 49.60 27.61 0.07 0.10 0.02 0.12 0.34 21.65 0.21 0.04 0.12 0.01 0.11

presented in Fig. 4. During the operation, the average permeationux of HF and FS in the MBR system were 8.52 and 9.74 L/(m2 h),respectively. As shown in Fig. 4, noticeable membrane fouling ofHF was rst observed on day 56 when the fouling was removedby on-line chemical cleaning. Subsequently, themembrane foulingof HF was observed on days 87 and 102. The trend of membranefouling of FS was very similar to that of HF. However, the TMP ofHFwas alwaysmuch higher than that of FS. On day 102, the TMP ofFS decreased almost to its initial level after the chemical cleaning,implying that the fouling in the FSwas smaller andmore reversiblethan fouling observed in the HF.

The cake layerwas obviously observed on FSmembrane surface,while itwasnotevident forHF.Afterushing thecake layerwith tapwater, the gel layer was seen on both membranes. SEM images ofclean and fouled membranes were taken to determine the depositmorphology of the gel surface. The membrane pores were clearlyseen on the new membranes (Fig. 5a and c), whereas, gel layerswere observed blocking the membrane pores on the fouled mem-branes and seemed to be more foulants accumulation (Fig. 5b andd).

WD-XRF was also used to measure the chemical componentson the clean and fouled membrane surface. Table 4 shows that therelative contents of C and N on the fouled membrane surfaces (HFand FS) weremuch higher than those on cleanmembrane surfaces,indicating that membrane fouling wasmainly governed by organicfouling. Moreover, several inorganic elements, such as Al, P, S, Si,Ca and Mg, were found on the fouled membrane. It is well knownthat metal ions may bridge the deposited cells and biopolymers,resulting in ahighly fouling layer and substantial TMP increase [30].Although the relative contents of these elements were small, theyplayed an important role in fouled layer formation [31]. The com-parison of chemical components between fouledHF and FS impliedthat FS exhibited more foulants accumulation than HF on mem-brane surface. In fact, SEM of longitudinal direction of membraneshowed that the gel layer of FS and HFwere about 60mand 5min thickness (Fig. S2). It could be deduced that gel layer and cakelayerwere easily formed on FSmembrane, and the cake layer couldprotect the gel layer from being removed by aeration from below.Fan and Huang [32] reported that the gel layer was indispensableand played a key role in the dynamic MBR rejection capacity ofthe ne particles. It was observed that small colloids and macro-molecules may adsorb on the thin gel layer of the membrane andresult in severely clogged pores [33]. These might be the reason ahigher ux under lower TMP could be obtained in the FSmembranemodule. Therefore, the chemical cleaning method using a combi-nation of acids and oxidizing agents is recommended for fouledmembrane caused by organic foulants and inorganic compounds.

4. Conclusions

Results suggest that the A2O-MBR system is very efcient insimultaneousorganicmatter andnitrogen removal in thechemical-intensive TA wastewater treatment. An increase of internal recycleratio from 0.5 to 1.5 had a greater effect than an increase from1.5 to 2.5 in the nitrogen removal performance. Some refrac-tory compounds present in the MBR permeate, such as TMSN,

styrene, hexa-methoxymethylmelamine and 1,4-Dioxane, shouldbe discharged into municipal sewage treatment plant for furthertreatment. The bridging between organic foulants and inorganiccompounds was a signicant contributor to membrane fouling,and the combination of acids and oxidizing agents could be a goodoption for chemical cleaning of fouled membrane.

Acknowledgements

Thisworkwas nancially supported by Fuji Electric Co. Ltd., andby Open Project of Zhejiang Province Key Laboratory of Environ-mental Pollution Control Technology. We thank Ms. Lijuan Maothe technician of 985-Institute of Agrobiology and Environmen-tal Sciences of Zhejiang University, for the assistance in using6890N/5975B GCMS system. The authors would like to thankemeritus professor Goen Ho of Murdoch University for his contri-butions to improve the manuscript.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at http://dx.doi.org/10.1016/j.jhazmat.2015.01.031.

References

[1] B. Smith, Future Pollution Prevention Opportunities and Needs in the TextileIndustry, 1994.

[2] A.F. Bertea, R. Butnaru, Polyamide dyeing wastewater recycling afterFenton-like oxidative treatment, Ind. Textila 63 (2012) 322326.

[3] S.F.A. Shahf, A.K. Shahf, A. Mehdi, A.A. Memon, K. Harijan, Z.M. Ali, Analysisand treatment of industrial wastewater through chemicalcoagulationadsorption process a case study of Clariant Pakistan Limited,AIP Conf. Proc. 1453 (2012) 353358.

[4] Y.L. Pang, A.Z. Abdullah, Current status of textile industry wastewatermanagement and research progress in Malaysia: a review, Clean-Soil AirWater 41 (2013) 751764.

[5] A. Chiavola, Textiles, Water Environ. Res. 81 (2009) 16961730.[6] R. Aplin, T.D. Waite, Comparison of three advanced oxidation processes for

degradation of textile dyes, Water Sci. Technol. 42 (2000) 345354.[7] A. Demirbas, Agricultural based activated carbons for the removal of dyes

from aqueous solutions: a review, J. Hazard. Mater. 167 (2009) 19.[8] W.T. Zhao, X. Huang, D.J. Lee, X.H. Wang, Y.X. Shen, Use of submerged

anaerobicanoxicoxic membrane bioreactor to treat highly toxic cokewastewater with complete sludge retention, J. Membr. Sci. 330 (2009) 5764.

[9] W.T. Zhao, Y.X. Shen, K. Xiao, X. Huang, Fouling characteristics in a membranebioreactor coupled with anaerobicanoxicoxic process for coke wastewatertreatment, Bioresour. Technol. 101 (2010) 38763883.

[10] T. Yu, R. Qi, D. Li, Y. Zhang, M. Yang, Nitrier characteristics in submergedmembrane bioreactors under different sludge retention times, Water Res. 44(2010) 28232830.

[11] X. Zhou, Y. Li, Y. Zhao, X. Yue, Pilot-scale anaerobic/anoxic/oxic/oxic biolmprocess treating coking wastewater, J. Chem. Technol. Biotechnol. 88 (2012)305310.

[12] Y. Chen, C. Peng, J. Wang, L. Ye, L. Zhang, Y. Peng, Effect of nitrate recyclingratio on simultaneous biological nutrient removal in a novelanaerobic/anoxic/oxic (A2/O)-biological aerated lter (BAF) system,Bioresour. Technol. 102 (2011) 57225727.

[13] K.G. Song, J. Cho, K.H. Ahn, Effects of internal recycling time mode andhydraulic retention time on biological nitrogen and phosphorus removal in asequencing anoxic/anaerobic membrane bioreactor process, BioprocessBiosyst. Eng. 32 (2009) 135142.

[14] Y.M. Li, G.W. Gu, J.F. Zhao, H.Q. Yu, Y.L. Qiu, Y.Z. Peng, Treatment of coke-plantwastewater by biolm systems for removal of organic compounds andnitrogen, Chemosphere 52 (2003) 9971005.

424 F. Sun et al. / Journal of Hazardous Materials 286 (2015) 416424

[15] T.W. Tan, H.Y. Ng, Inuence of mixed liquor recycle ratio and dissolvedoxygen on performance of pre-denitrication submerged membranebioreactors, Water Res. 42 (2008) 11221132.

[16] J.A. Baeza, D. Gabriel, J. Lafuente, Effect of internal recycle on the nitrogenremoval efciency of an anaerobic/anoxic/oxic (A2/O) wastewater treatmentplant (WWTP), Process Biochem. 39 (2004) 16151624.

[17] Y.M. Kim, D. Park, C.O. Jeon, D.S. Lee, J.M. Park, Effect of HRT on the biologicalpre-denitrication process for the simultaneous removal of toxic pollutantsfrom cokes wastewater, Bioresour. Technol. 99 (2008) 88248832.

[18] C.C. Tseng, T.G. Potter, B. Koopman, Effect of inuent chemical oxygendemand to nitrogen ratio on a partial nitrication/complete denitricationprocess, Water Res. 32 (1998) 165173.

[19] F.Q. Sun, S.W. Wu, J.J. Liu, B. Li, Y.X. Chen, W.X. Wu, Denitrication capacity ofa landlled refuse in response to the variations of COD/NO3N in theinjected leachate, Bioresour. Technol. 103 (2012) 109115.

[20] M. Henze, G.H. Kristensen, R. Strube, Rate-capacity characterization ofwastewater for nutrient removal processes, Water Sci. Technol. 29 (1994)101107.

[21] S.P. Sun, C.P.I. Nacher, B. Merkey, Q. Zhou, S.Q. Xia, D.H. Yang, J.H. Sun, B.F.Smets, Effective biological nitrogen removal treatment processes for domesticwastewaters with low C/N ratios: a review, Environ. Eng. Sci. 27 (2010)111126.

[22] J. Carrera, T. Vicent, J. Lafuente, Effect of inuent COD/N ratio on biologicalnitrogen removal (BNR) from high-strength ammonium industrialwastewater, Process Biochem. 39 (2004) 20352041.

[23] I. Holme, Dyes and nishes for synthetic bres, Int. Dyer 189 (2004) 911.[24] J. Wang, Z.J. Zhang, L.N. Chi, X.L. Qiao, H.X. Zhu, M.C. Long, Z.F. Zhang,

Performance of anaerobic process on toxicity reduction during treatingprinting and dyeing wastewater, Bull. Environ. Contam. Toxicol. 78 (2007)531534.

[25] C. Reyes-Contreras, V. Matamoros, I. Ruiz, M. Soto, J.M. Bayona, Evaluation ofPPCPs removal in a combined anaerobic digester-constructed wetland pilotplant treating urban wastewater, Chemosphere 84 (2011)12001207.

[26] O. Botalova, J. Schwarzbauer, N.a Sandouk, Identication and chemicalcharacterization of specic organic indicators in the efuents from chemicalproduction sites, Water Res. 45 (2011) 36533664.

[27] H.L. Huang, D.S. Shen, N. Li, D. Shan, J.L. Shentu, Y.Y. Zhou, Biodegradation of1,4-Dioxane by a novel strain and its biodegradation pathway, Water Air SoilPollut. 225 (2014) 2135.

[28] J.Y. Choi, Y.J. Lee, J. Shin, J.W. Yang, Anodic oxidation of 1,4-dioxane onboron-doped diamond electrodes for wastewater treatment, J. Hazard. Mater.179 (2010) 762768.

[29] J.H. Suh, M. Mohseni, A study on the relationship between biodegradabilityenhancement and oxidation of 1,4-dioxane using ozone and hydrogenperoxide, Water Res. 38 (2004) 25962604.

[30] Q. Li, M. Elimelech, Organic fouling and chemical cleaning of nanoltrationmembranes: measurements and mechanisms, Environ. Sci. Technol. 38(2004) 46834693.

[31] F. Meng, H. Zhang, F. Yang, L. Liu, Characterization of cake layer in submergedmembrane bioreactor, Environ. Sci. Technol. 41 (2007)40654070.

[32] B. Fan, X. Huang, Characteristics of a self-forming dynamic membranecoupled with a bioreactor for municipal wastewater treatment, Environ. Sci.Technol. 36 (2002) 52455251.

[33] I. Chang, S. Kim, Wastewater treatment using membrane ltrationeffect ofbiosolids concentration on cake resistance, Process Biochem. 40 (2005)13071314.

Organics and nitrogen removal from textile auxiliaries wastewater with A2O-MBR in a pilot-scale1 Introduction2 Materials and methods2.1 Experimental set-up and operating conditions2.2 Characteristics of TA wastewater2.3 Analytical methods

3 Results and discussion3.1 COD removal3.2 Nitrogen removal3.3 Behavior of organic compounds3.4 Membrane fouling behavior

4 ConclusionsAcknowledgementsAppendix A Supplementary dataReferences