Residual chemical oxygen demand (COD) fractionation in bio-treated coking wastewater integrating solution property characterization

Journal of Environmental Management 246 (2019) 324–333

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Research article

Residual chemical oxygen demand (COD) fractionation in bio-treated coking wastewater integrating solution property characterization

T

Cong Weia, Hengping Wua, Qiaoping Konga, Jingyue Weia,b, Chunhua Fenga, Guanglei Qiua, Chaohai Weia,∗, Fusheng Lia,b,∗∗ a b

School of Environment and Energy, South China University of Technology, Guangzhou, 510006, PR China River Basin Research Center, Gifu University, 1-1 Yanagido, Gifu, 501-1193, Japan

ARTICLE INFO

ABSTRACT

Keywords: Chemical oxygen demand (COD) Bio-treated coking wastewater Solution properties Correlation analysis Water quality assessment

The refractory nature of residual COD in bio-treated coking wastewater (BTCW) creates barriers for its further treatment and reclamation. It is necessary to fractionate the residual COD in BTCW associated with characterization of solution properties. In this paper, a stepwise process composed of membrane filtration, coagulation, adsorption and ozonation was proposed to fractionate residual COD in the BTCW, in which the COD was stepwise reduced to near zero. In addition, the correlation between COD and water quality indexes as well as solution properties were discussed together with a safety assessment of the water quality. Results showed that the residual COD fractionation percentage contributed by suspended solids, colloids, dissolved organics and reductive inorganic substances in the BTCW was 43.7%, 22.1%, 26.2% and 4.9%, respectively. By stepwise fractionating of these substances, the residual COD was reduced from 168.8 to 5.2 mg L−1, and the UV254 value decreased from 1.90 to 0.15 cm−1. In addition, the particle size of the dominant substances contributing to the residual COD was smaller than 450 nm. Among these substances, the hydrophobic fraction accounted for 78.66% (in the term of TOC). Three-dimensional excitation-emission matrix (3D-EEM) analysis showed that hydrophobic neutral substances (HON) were the main fluorescence constituent in the BTCW, which was highly removable by adsorption. The residual COD after adsorption was mainly composed of reductive inorganic substances. Apart from pursuit of high COD removal rates, more emphasis should be given to the removal of toxic COD. Correlations were observed between the residual COD and water quality indicators as well as solution properties, providing a guideline for optimized removal of residual COD in the BTCW. In summary, these results gave a referential information about the nature of residual COD in the BTCW for the selection of advanced treatment technologies and the management of water quality safety.

1. Introduction Industrial wastewater is one of by-products during industrial development process and its generation in China is increased rapidly with the booming development of China's society and economy. As a typical industrial wastewater, coking wastewater is associated with high concentration of organics (with COD concentrations about 3500–8000 mg L−1), together with high amounts of inorganic compounds such as NH3, NH4+, CN−, SCN−, F−, Cl−, S2− and NO2− (Jiang et al., 2016). It is generated from coal coking, coal gas purification and byproduct recovery processes. Sufficient treatment is needed before its discharge into the environment. In 2017, the amount

of coking wastewater generated in China was around 3.12×108 m3, accounting for 1.59% of the total industrial wastewater discharge (EPA of China, 2017). To date, the treatment of coking wastewater mainly relays on biological processes which could remove 95-98% COD from the raw coking wastewater. Residual COD (2-5% of the COD in raw coking wastewater) is commonly present in bio-treated coking wastewater (BTCW) because of biological reaction kinetic limits. Anaerobicanoxic-oxic (AAO) process is widely employed as the main biological process in coking wastewater treatment plants. Moreover, bio-treated effluent is still a failure to meet water reclamation standard even though lower COD of effluent in the oxic-hydrolytic-oxic (O–H–O) system was achieved comparing to the AAO process (Zhang et al., 2017;

Corresponding author. School of Environment and Energy, South China University of Technology, 382 Waihuan East Road, Guangzhou Higher Education Mega Center, Guangzhou, 510006, China. ∗∗ Corresponding author. River Basin Research Center, Gifu University, 1-1 Yanagido, Gifu, 501-1193, Japan. E-mail addresses: [email protected] (C. Wei), [email protected] (F. Li). ∗

https://doi.org/10.1016/j.jenvman.2019.06.001 Received 13 January 2019; Received in revised form 23 May 2019; Accepted 1 June 2019 0301-4797/ © 2019 Elsevier Ltd. All rights reserved.

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

Pan et al., 2019). Generally, the residual COD in BTCW ranged from 84.6 to 235.3 mg L−1 (Zhang et al., 2014; Li et al., 2017; Yang et al., 2017, 2018). The observed differences in the residual COD in BTCW among different reports are due to: (i) the raw coking wastewater was diluted consciously or unconsciously by some steel enterprises; (ii) the sample of BTCW was collected from the secondary sedimentation tank or the coagulation tank located after the secondary sedimentation tank; and (iii) the process configurations, operation parameters and level of management differed greatly with wastewater treatment plants. The COD discharge limit for coking wastewater has been reduced from 100 to 80 mg L−1 in China. In areas with high land development density and fragile ecological environment COD less than 40 mg L−1 for the final effluent was set as the discharge limit (EPA of China, 2012). In addition, there are increasing demands for the treated wastewater reuse due to water shortage in some regions, which require a nearly complete COD removal during the coking wastewater treatment. Commonly, advanced treatment of BTCW is typically necessary as BTCW has residual COD exceeding the discharge standard due to the presence of refractory contaminants so as to affect the efficiency of the subsequent desalination process. For example, a recyclable aminated hyper-crosslinked polymeric adsorbent was synthesized by Yang et al., which achieved 75-93% COD removal in the BTCW (Yang et al., 2013). Around 85.3% COD removal was achieved through optimizing operating conditions (regarding coagulation dosage or adsorbent dosage) of coagulation and adsorption processes (Li et al., 2018). Although a number of studies have been carried out on the removal of residual COD in BTCW though optimizing operating parameters, knowledge gaps still exist on the relationships between residual COD and the characteristics of the BTCW, with a limited number of researches being performed to date. The lack of in-depth understanding of the nature of the compounds contributing to the residual COD often results in impertinent process selection and low efficiency for its removal. Solution properties can be used as a tool to help understand the nature of residual COD, which is a set of chemical/physical/biological indicators, including hydrophilicity/hydrophobicity (Revilla et al., 2016), particle size distribution (Karahan et al., 2008), adsorbability (Kaur et al., 2016), fluorescence (Hudson et al., 2008), ultraviolet absorbance (Tran et al., 2015), and toxicity (Valitalo et al., 2017). However, the residual COD in BTCW is a mixture of different species so that COD fractionation is necessary before solution properties characterization. The inspiration of residual COD fractionation stemmed from particle size classification principle in which contaminant components in wastewater were generally divided into suspended solids (particle size > 450 nm), colloids (1 nm < particle size < 100 nm) and dissolve substances (particle size < 2 nm). Similarly, residual COD in the BTCW can be fractionated by particle size. To achieve the fractionation, a stepwise process was used to remove each particle-size fraction from the BTCW. Membrane filtration process (with 0.45 μm membranes) was a preferable method for eliminating suspended solids in wastewater (Touati et al., 2018). Coagulation process could achieve a reduction of colloids by increasing the coagulant dose and optimizing reaction pH, in which large organic molecules with hydrophobic property were removed preferentially (Zularisam et al., 2009). Adsorption process could achieve the removal of soluble organics in which the adsorbates were accumulated in adsorbents' outer or inner surface though Vander Waals force, electrostatic force and/or hydrogen bond. The adsorption efficiency depends on the properties of adsorbate and adsorbent, absorbent dose, pH and temperature in wastewater (Li et al., 2018). Inorganic components in industrial wastewater could be oxidized completely by ozone (Gounden et al., 2018). Each of these advanced technologies mentioned above targeted pollutants with different particle size. Therefore, a stepwise use of these advanced technologies could effectively fractionate BTCW's residual COD. Herein, the objectives of this research were to propose a method for COD fractionation in industrial wastewater, and to characterize the properties of different COD fractions. A stepwise process composed of

membrane filtration, coagulation, adsorption and ozonation was established to fractionate residual COD in the BTCW into suspended solids, colloids, dissolved organics and reductive inorganic substances. Specially, the particle size of colloids was defined as in a range of 100–450 nm while the particle size of dissolved organics was considered to be less than 2 nm in this study. Colloid and dissolved organic fractions in residual COD along the stepwise process were characterized by the changes in solution properties, including particle size distribution (PSD), hydrophilicity/hydrophobicity, ultraviolet absorbance and fluorescence. The Nemerow index method was used to evaluate the water quality safety and toxicity. The correlation between COD and other water quality indexes as well as solution properties was also discussed. It is anticipated that this paper will give an in-depth understanding of the nature of the residual COD in BTCW, providing rationality and references for technology selection of residual COD removal, thus benefiting an improved management of BTCW treatment. 2. Material and methods 2.1. Wastewater samples The wastewater sampling point was located in Songshan coking wastewater treatment plant in Shaoguan Steel Company, Guangdong Province, China. The WWTP has an average treatment capacity of 1680 ± 100 m3 d−1 and has been running for 14 years. The raw influent was first air-floated to remove oil stick at the degreasing tanks. Then, anaerobic-oxic-hydrolytic-oxic (A-O-H-O) system coupled with biological fluidized-bed was applied for biological treatment of the coking wastewater. The hydraulic retention time (HRT) of the anaerobic tank, the first oxic tank, the hydrolytic tank and the second oxic tank was about 28 h, 12 h, 20 h and 16 h, respectively. The designed COD loading of the anaerobic tank, the first oxic tank, the hydrolytic tank and the second oxic tank were about 1.6, 2.4, 0.5 and 0.4 kg·(m3·d)−1, respectively. The BTCW then went through secondary clarifiers, where activated sludge and remaining particles were coagulated and removed from the mixed liquor. Finally, the secondary effluent was discharged. The BTCW used in this work was collected from the second oxic biological fluidized-bed effluent, with brown-yellow color and offensive odor. The BTCW was collected every 3 days and the sampling campaign lasted two months. The samples were stored in a refrigerator (4 °C) for further analysis within five days. 2.2. A stepwise process for fractionating the residual COD in BTCW As shown in Fig. 1, the stepwise process including membrane filtration, coagulation, adsorption and ozonation was proposed to fractionate the residual COD in the BTCW. The investigated BTCW was firstly filtered through a 0.45 μm cellulose membrane to remove suspended solids. The filtrate was refrigerated at 4 °C. The selection of coagulant and the setting of coagulation conditions in the coagulation process were performed according to Cui (Cui et al., 2015). The laboratory reagent poly ferric sulphate (PFS’ dosage at 800 ± 50 mg L−1 with solution pH being controlled at 7.0) was purchased from Aladdin Reagents Co., Ltd., China. A sixbeaker jar-test apparatus (ZR4-6, Shenzhen Zhongrun, China) was used to mix the effluent from filtration process with PFS evenly. Each beaker contained 300 mL filtrated effluent. The coagulation procedure involved 10 min of rapid mixing at 200 rpm, followed by 20 min slow mixing at 25 °C (rotating speed was 50 rpm). Water samples from each beaker were vacuum filtered through 0.45 μm cellulose membrane to accomplish the separation of floc and water. Response surface method with Box-Behnken design (RSM-BBD) was employed to investigate the separation efficiency of colloids and low molecular mass in the coagulation process (design details were given in the supplementary material). For adsorption, the selection of adsorbent and the setting of adsorption conditions in adsorption process were performed according 325

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

Fig. 1. Flowchart of the stepwise process.

to our previous publication (Zhou et al., 2018). Powder activated carbon (PAC, Xinhua Activated Carbon Company, Shanxi, China) was added into 100 mL of the effluent from coagulation process, stirred at 120 rpm for 4 h at 25 °C, with a dosage of 4 g L−1. Adsorption process was conducted by optimizing the dosages of absorbent and adsorption pH from 2 to 6 coupled with response surface method with BoxBehnken design (COD removal of filtration effluent after adsorption process under different pH conditions using the same dosage of power activated carbon were investigated and design details in the supplementary material). Therefore, adsorption pH = 2.0 was chosen. Then, water samples were vacuum filtered through 0.45 μm cellulose membrane to accomplish the separation of PAC and water samples. Ozonation treatment was applied to the effluent after adsorption process for inorganic components oxidation. Gaseous ozone was generated using pure oxygen by an ozone generator (CH-ZTW, Ozone Electric Appliance Co, Ltd, China). The ozone-containing gas with inlet ozone concentration of 31 ± 3 mg L−1 at a flow rate of 1.0 L min−1 was delivered into a gas-liquid contacting column containing 100 mL of the water after adsorption treatment with initial pH of 10.0 for 60 min. Excess ozone in the off-gas was trapped using 20% KI solution for safety precaution.

measured according to the method described in our previous publication (Pan et al., 2018). Cl− was determined using ion chromatographymass spectrometer (ICP-MS, Elan 6000). 2.3.2. Characterization of solution properties 2.3.2.1. Particle size fractionation. The particle size fractionation of the BTCW and the effluent generated after filtration, coagulation, adsorption was assessed using sequential ultrafiltration (the particle size fractionation of the ozonation effluent was not measured since the concentration of organics in ozone effluent was below detection limit). The water samples were fractionated using six types of cellulose membranes (Millipore Corp) with nominal particle size cutoff of 13 nm, 8 nm, 5 nm, 4 nm, 3 nm, and 2 nm, respectively. The effective surface area of the membrane was 31.75 cm2. Prior to filtration, the membranes was first rinsed using ultrapure water to remove any possible organics until the TOC values in permeate reached below 0.1 mg L−1. High purity nitrogen (99.999%) was used to pressurize the filtration process (0.6 atm for 13 nm, 1.2 atm for 8 nm and 5 nm, 2.4 atm for 4 nm, 3 nm and 2 nm, respectively). A total amount of 200 mL for each sample was filtered with the aid of magnetic mixing. Afterwards, COD and TOC of each filtrate was determined.

2.3. Analytical methods

2.3.2.2. Hydrophilic/hydrophobic components fractionation. The hydrophilic/hydrophobic components fractionation of the BTCW and the effluent obtained after filtration, coagulation, adsorption was performed according to the reference (the ozonation effluent was not measured because the concentration of organics in ozone effluent was below detection limit) (Wang et al., 2007). Amberlite (Sigma-Aldrich) acrylic ester macro porous resin XAD-8 was used to separate the watersoluble organic substances into four groups: hydrophobic acid (HOA), hydrophobic base (HOB), hydrophobic neutral (HON), hydrophilic substances (HIS). A total volume of 100 mL for each water sample was pump through the XAD-8-resin-packed column. The eluted water was collected for further separation. The HOB fraction was obtained by back-flushing the resin column with 100 mL of 0.10 M HCl. The eluted water collected was acidified to pH at 2.0 and recirculated through the resin column to obtain the HIS fraction. The HOA fraction was obtained by backflushing the resin column with 100 mL of 0.10 M NaOH. The HON fraction was extracted by Soxhlet-extracted with anhydrous methanol. Excess methanol was removed by vacuum rotary evaporation at 45 °C. This fractionation procedure was illustrated in Fig. 2. After fractionation, the HOB fraction in 0.1 M HCl, HOA fraction in 0.1 M NaOH and HIS fraction in the effluent from the XAD-8 column were quantified by a total organic carbon (TOC) analyzer. Deionized water was added to

According to the stepwise process described above, five types of water samples (the BTCW, the filtration effluent, the coagulation effluent, the adsorption effluent and the ozonation effluent) were analyzed for water quality and solution properties. The pH values of all the water samples were adjusted to 7.0 ± 0.2 and then diluted to required concentrations by ultrapure water before further analysis. The pH values were measured using a pH meter (pHS-3C, Shanghai Scientific Instrument Co., Ltd., China). Water quality parameters analyzed include COD, total organic carbon (TOC), biochemical oxygen demand (BOD), phenols, lipids, suspended solids (SS), total nitrogen (TN), S2−, SCN−, CN−, Cl−, NO3−, and NH4+-N. Solution properties characterization includes particle size fractionation, hydrophilic/hydrophobic components fractionation, UV absorbance and three-dimensional fluorescence contour spectra (3D-EEMs) profiling. Each water sample was analyzed three times to produce an average value with a coefficient of variation below 5%. 2.3.1. Water quality analysis A total organic carbon analyzer (TOC-VCPH, Shimadzu, Japan) was used to determine the TOC values. COD, BOD, NH4+-N, phenols, lipids, SS, TN were analyzed in accordance with the National Standard Methods in China (EPA of China, 2002), CN−, SCN−, and S2− were 326

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

Table 1 Integrated water quality index from pollution level classification. Pollution Level

Severe

Heavy

Medium

Slightly

I

I≥6

3≤I < 6

1≤I < 3

I<1

I=

Pi2, max + Pi2, ave 2

(2)

where I is integrated water quality index, the classification of pollution levels according to integrated water quality index was shown in Table 1. P2i,max is the maximum water quality's single factor index of class i pollutants. P2i,ave is the average water quality's single factor index of class i pollutants. Fig. 2. Schematic diagram of the procedure for fractionation procedure. HOB, hydrophobic bases; HOA, hydrophobic acids; HON, hydrophobic neutrals; HIS, hydrophilic substances.

2.5. The correlation analyses between COD and water quality indicators Along with the evolution in the water quality indicators and solution properties in the stepwise process, factors affecting COD were investigated through analyzing the correlations between COD and water quality indicators as well as solution properties. Correlation analyses were performed using SPSS 24. The significances of the correlations in the statistics was evaluated using p-values (p < 0.05 or p < 0.01).

maintain the HON fraction volume to 100 ml then the HON fraction was also quantified by a TOC analyzer. 2.3.2.3. UV absorbance measurements. The ultraviolet-visible adsorption spectrometry of each water sample was conducted using a UV-visible spectrophotometer (Evolution 300, Thermo Fisher Scientific, United States) from 200 to 700 nm at 1 nm intervals (the UV absorbance measurement of the ozonation effluent was not performed since the concentration of organics in ozone effluent had reached below detection limit). UV254 represents the absorbance at 254 nm. The specific UV254 absorbance (SUVA), which was employed as a surrogate of dissolved organic substances aromaticity, was calculated as the UV254 absorbance divided by TOC of the water sample.

3. Results and discussion 3.1. Key water quality in the BTCW during the stepwise process Membrane filtration, coagulation, adsorption and ozonation were performed to fractionate the residual COD of BTCW. The COD and other key water quality parameters of each process effluent were shown in Table 2. The average residual COD of the BTCW was 168.8 mg L−1 and the BOD5 to COD (B/C) ratio of the BTCW was around 0.08 (Table 2), indicating that biological treatment is infeasible to further remove the residual organics in the BTCW. The average COD values of the effluent of filtration, coagulation, adsorption and ozonation were 95.1, 57.9, 13.6 and 5.2 mg L−1, respectively (Table 2). To be specific (Fig. 3), Suspended solids contributed 43.7% of the residual COD by comparing the COD values of the BTCW and the filtration effluent. The suspended solids were consisted of sludge flocs and residual micelle (Prasse et al., 2015). Colloids (with particle size between 2 and 450 nm), dissolved organics (with particle size less than 2 nm) and reductive inorganic substances constituted 22.1%, 26.2% and 4.9% of the total residual COD, respectively, based on the changes in effluent COD in the stepwise process. The contribution of inorganic reducing compounds to COD in wastewater with low organic concentration may not be ignored. According to the principle of COD determination and redox potential, potassium dichromate as oxidant could oxidize all substance whose reduction potential is less than 1.33 V, hence both the organics and the inorganic components in wastewater could be oxidized and counted as COD values. The presence of the inorganic components could lead to positive deviation of COD value. It was found that several inorganic components which could potentially contribute to the COD value existed in coking wastewater, for instance, Cl−, NO2−, SCN−, CN−, S2−, SO32− and Fe2+. The theoretical oxygen demand (ThOD) of some common inorganic substances were calculated according to the law about electron donor-acceptor balance between oxidant and reductant in Table 3, which was used as a reference for the correction test. Much attention has to be paid on the interference of Cl− on COD test since Cl− commonly present in BTCW. Two reasons of Cl− contributing to COD in analysis process were listed as follows:

2.3.2.4. Three-dimensional fluorescence contour spectra (3D-EEMs) detection. A HITACHI F7000 fluorescence spectrometer was applied for three-dimensional fluorescence contour spectra (3D-EEMs) detection. The detection parameters used were listed as follows: 150 W Xenon light; 400 V PMT voltage; 5 nm slit widths for both excitation and emission; Auto response model for response time; 2400 nm min−1 scan rate. The EEM spectra were collected with scanning excitation spectra from 200 nm to 450 nm at 5 nm increments by varying the emission wavelength from 200 nm to 550 nm at 5 nm intervals. The position (excitation and emission wavelength) and maximum fluorescence intensity in fluorescence units (AFU) of each peak was recorded. An ultrapure water blank EEM was subtracted to eliminate water Raman scatter peaks, and the datum of Raleigh scattering area (Em ≤ Ex+5 nm and Em ≥ Ex +300 nm) were set to zero to eliminate the interference from Raleigh scattering. 2.4. Water quality safety assessment model—the nemerow index method As a single factor evaluation method, the Nemerow index method was used to evaluate the degree of water pollution. The expression of water quality's single factor index Pi was described as follow:

Pi =

Ci Si

(1)

where Ci represents the measured concentrations of class i contaminant, and Si is class i contaminant's emission limit as according to GB161712012 in China (EPA of China, 2012). The water pollution level was evaluated comprehensively by using the water quality index parameters and referring to the pollution classification standard of the Nemerow index method (Chen et al., 2016). The mathematical expression of the Nemerow index method was stated as follow:

K2Cr2O7+4NaCl+6H2SO4→2KHSO4+4NaHSO4+2CrO2Cl2+3H2O

327

(3)

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

Table 2 Key quality parameters of the BTCW during the sampling period (The unit of all the water quality index is mg·L−1, each value represents the average value of triplicated measurements). Parameter

COD

BOD

TOC

Phenols

Lipid

SS

TN

NH4+-N

S2-

SCN−

CN−

Cl−

BTCW Filtration effluent Coagulation effluent Absorption effluent Ozonation effluent

168.8 95.1 57.9 13.6 5.2

14.52 12.34 9.76 1.22 0.54

58.20 32.82 20.24 2.95 1.96

0.20 0.15 0.11 0.05 0.01

1.94 1.52 1.16 0.34 0.12

52.18 5.20 3.36 2.12 0.42

56.4 50.3 43.6 36.1 27.5

3.44 3.12 2.54 1.35 0.32

1.56 1.38 1.18 0.55 0.02

1.33 1.24 0.85 0.39 0.03

1.12 0.86 0.66 0.42 0.05

780.3 767.2 760.5 751.2 100.1

Abbreviations Note. COD-Chemical oxygen demand; BOD-Biochemical Oxygen Demand; TOC-Total organic carbon; SS- Suspended solids.

3.2. Characterization of solution properties during the stepwise process Suspended solids are removed effectively though membrane filtration and reductive inorganic substances were oxidized by ozonation to eliminate the interference of residual COD. However, colloids and dissolved organics were composed of diverse organics. To understand the nature of these organic compounds, various techniques were used to characterize their properties. 3.2.1. Characterization of particle size distribution (PSD) Sequential ultra-filtration was applied to investigate the contribution of different particle size fractions to COD and TOC after filtration, coagulation and adsorption. The obtained results (Fig. 4) suggested that the organic fraction with particle size < 2 nm accounted for the largest portion of COD and TOC (57.35% and 55.78%, respectively. The COD and TOC values of each fraction during the stepwise process were shown in Table S5) among all compounds with particle size < 450 nm. The total COD and TOC of organics in the particle size range of 2–450 nm was 38.19 mg L−1, 11.91 mg L−1, respectively. After coagulation process, the COD of pollutants with particle size < 2 nm was 52.19 mg L−1, accounting for 90.62% of the total COD of coagulation effluent. It was indicated that the rest of COD and TOC after coagulation was not removed though the mechanism of electrical neutralization, trap effect of net-catching and complexation -precipitation. As for the effluent of the adsorption process, COD and TOC of organics with particle size less than 2 nm was 12.46 mg L−1 and 2.67 mg L−1, respectively. It showed that the majority of dissolved organics was removed via adsorption. The organics in the BTCW with particle size < 2 nm (the corresponding molecular weight of < 1000 Da) represented a majority of dissolved COD remaining after biological treatment. Yang et al. found that the molecular weight distribution of refractory organic compounds, such as aromatic compounds, ranged from 500 Da to 1000 Da in BTCW (Yang et al., 2018). For instance, 2-chloro-4-nitroaniline (molecular weight of 156.5 Da) is a kind of dissolved organic component in coking wastewater (Zhang et al., 2012). Wen et al. studied 2chloro-4-nitroaniline degradation kinetics under different hydraulic retention time in a simulated aerobic biodegradation experiment and found that the degradation rate of 2-chloro-4-nitroaniline was less than 20% even with a HRT of 24 h (Wen et al., 2016). The reason for the low degradation rate is that the existence of –NO2 and -Cl groups in 2chloro-4-nitroaniline significantly reduces the electron cloud density of the benzene ring, leading to the weakening of electrophilic aggressiveness of microorganisms and related enzymes. Thus, the biological removal of 2-chloro-4-nitroaniline was significantly hindered. Other

Fig. 3. The proportion of COD contributed by suspended solids, colloids, dissolved organic substances and reductive inorganic substances in total COD of the BTCW.

Chloride is not oxidized by potassium dichromate in aqueous solution as the standard redox potential of dichromate is close to that of Cl−. Sulphuric acid results in a higher conditional redox potential of potassium dichromate than that of Cl− according to Nernst equation and, thus, chromyl chloride (CrO2Cl2) is formed.

Cr2O72 + 14H+ + 6Cl

(4)

2Cr3 + + 7H2 O+ 3Cl2

Cl− could be oxidized quantitatively to chlorine by potassium dichromate (K2Cr2O7) during the 2 h of dichromate oxidation of organic compounds at 148 °C. The chloride interference was reduced as the organic matter concentration increased (Vyrides and Stuckey, 2009). But wastewater pollution degree may be misled by Cl− and chemical reagents and energy may be wasted for the pursuit of lowing the effluent COD (caused by high Cl−). On the other hand, chloride is the trace element needed by all living things. It is a component which does not need to be removed at low concentration for the sake of environment protection. However, it may contribute to the residual COD. Special attention needs to be paid on Cl− when evaluating the residual COD. The average COD of the ozonation effluent was 5.2 mg L−1, which means a majority of the residual COD in the BTCW were contributed by suspended solids, colloids, dissolved organics and reductive inorganic substances. The corresponding relationship between residual COD and different fractions also suggested that the stepwise process is relevant for COD fractionation. Further characterization of solution properties of different particle size fractions may promote a rational selection of advanced treatment techniques for improved management of BTCW treatment. Table 3 The theoretical oxygen demand of reductive inorganic substances. Reductive Substance

Cl−

Br−

H2O2

SCN−

NO2−

S2-

S2O32-

Fe2+

ThOD(mg·mg−1) Reductive Substance ThOD(mg·mg−1)

0.226 CN− 3.070

0.100 F− 0.421

0.471 Mn2+ 0.291

1.170 Sn2+ 0.136

0.348 SO320.200

2.000 NH4+ 1.778

0.571 I− 0.063

0.143 ClO− 0.901

328

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

Fig. 4. The particle size distribution of COD and TOC during the stepwise process (F-filtration process, C-coagulation process, A-adsorption process).

solution property was considered necessary to characterize dissolved organic compounds where the adsorption process played a key role in their removal.

Hydrophobic fraction accounted for 78.66% of the total TOC in the BTCW. HOA and HON components could easily be removed by coagulation, agreeing with the relatively high removal rate of HOA (66.25%) and HON (41.56%) components during coagulation process. In addition, adsorption process seemed to be non-selective and all fractions were removed efficiently. The TOC was only 2.87 mg L−1 after PAC adsorption. From this point of view, the majority of colloids and residual organics seem to show a hydrophobic nature.

3.2.2. Characterization of hydrophilicity/hydrophobicity The colloids and dissolved organic compounds in the BTCW were categorized into four major groups: HOA, HOB, HON and HIS (He et al., 2011). Characterization of hydrophilicity/hydrophobicity during the stepwise process was shown in Fig. 5, HOA and HON accounted for 35.11% and 30.03% of the total TOC in the BTCW. While, HIS contributed to 19.28% of the total TOC. HOB accounted for the lowest portion (15.58%). He et al. found that HIS and HOA fractions were the major TOC components in the BTCW, which were resulted from the coke quenching process (He et al., 2011).

3.2.3. Characterization of ultraviolet absorbance UV254 and SUVA reflected the unsaturation and aromaticity of the organics in the wastewater. SUVA might provide a quantitative index of aromaticity of organic matter. High SUVA value (> 4 L mg−1·m−1) indicates a comparatively high content of hydrophobic, aromatic and high molecular weight (MW) compounds, while low SUVA value (< 3 L mg−1·m−1) suggests a relatively high content of hydrophilic and low MW compounds in the sample (Kaewsuk and Seo, 2011). The UV254, and SUVA patterns of colloids and dissolved organics in BTCW after membrane filtration, coagulation and adsorption process were shown in Fig. 6. The UV254 of raw coking wastewater was 13.35 cm−1 and was reduced to 2.51 cm−1 after biological treatment. UV254 decreased gradually from 1.90 to 0.15 cm−1 during the stepwise process. The dominated fraction of UV254 was HON followed by HOA in the colloids and dissolved organic compounds. After coagulation, the UV254 of these two components were highly reduced, although they were still the major COD contributors. After adsorption, all the components were reduced to low UV254 levels (< 0.10 cm−1). The results indicated that aromatic organics were mainly concentrated in the HON and HOA component, and the HIS and HOB fractions presented low aromaticity and thus low UV254. The order of SUVA was consistent with that of UV254, in which the concentration of HON was the highest. HON fraction was the only component whose SUVA excessed the filtrated raw effluent, indicating that the organic substances with aromaticity was high in HON. Similar results were observed in the coagulation and adsorption groups where HON's SUVA excessed the unfractionated sample, suggesting that aromatic organic substances were concentrated in HON.

Fig. 5. Characterization of hydrophilicity/hydrophobicity during the stepwise process (RAW-the raw bio-treated coking wastewater, HOA-hydrophobic acid, HOB-hydrophobic base, HON-hydrophobic neutral, HIS-hydrophilic substances).

3.2.4. Characterization of fluorescence 3D-EEM technology is also used to characterize the structural 329

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

Fig. 6. Characterization of UV254 and SUVA for dissolved organic components during the stepwise process (F-filtration process, C-coagulation process, A-adsorption process, RAW-the raw bio-treated coking wastewater, HOA-hydrophobic acid, HOB-hydrophobic base, HON-hydrophobic neutral, HIS-hydrophilic substances).

properties of the colloid and dissolved organic substances. According to Quaranta's study, four notable peaks can be identified: aromatic-protein-like substances (peak T1), soluble microbial byproduct-like material (peak T2), fulvic-acid-like substances (peak A) and humic-acid-like substances (peak C) (Quaranta et al., 2012). As showed in Fig. 7, peak A and peak C were distinctly reduced after coagulation, and all fluorescence peaks were almost disappeared after adsorption. HON was the main fluorescence constituent with the blueshift of peak A being observed (Fig. S3). The fluorescence peak shift provided spectral information on the chemical composition changes. A red shift was related to the increase of polar group in fluorophores (Chen et al., 2003). Hydroxyl, carboxyl and alkoxy could give rise to red-shift in excitation or emission wavelength. While, blue-shift was bound up with lower aromatic degree or the break-up of large molecules into fragments (Hong et al., 2016). 3D-EEMs spectra were divided into five regions, which represented the first kind of aromatic protein like substances (Region 1), the second kind of aromatic protein like substances (Region 2), fulvic acid like substances (Region 3), soluble microbial byproduct-like material (Region 4), and humic acid like substances (Region 5), respectively. The changes in the proportion of five Ex/Em regions of the dissolved

organic fractions after filtration, coagulation and adsorption process were shown in Fig. 8. In the filtration effluent, Region 5 occupied the highest portion (25.29%), followed tightly by the Region 3 (23.80%) and the Region 2 (23.49%), respectively. After coagulation, the proportion of each region remained largely unchanged. After adsorption, Region 1 and Region 2 (representing aromatic proteins) was dominated with a total proportion of 66.14%. The proportion of Region 4 and Region 5 decreased dramatically and the proportion of Region 3 was roughly unchanged. During filtration, coagulation and adsorption, the proportion of HIS remained largely the same. As for HOA, HOB and HON, the sum of the proportion of Region 1 and Region 2 increased gradually during the stepwise process, suggesting that aromatic compounds remained as the main residual constituent in the adsorption effluent even though fluorescence peaks were barely detectable, showing their low concentrations. 3.3. Water quality safety assessment For the sake of water quality assessment, the ozonation effluent quality was evaluated by using the Nemerow index method with TN, S2−, SCN−, CN−, COD, BOD, NH4+-N, Phenols, SS and Lipid as

Fig. 7. Characterization of 3D-EEMs in the BTCW during the stepwise process. 330

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

Fig. 8. FRI distribution for filtration, coagulation and absorption effluent with different hydrophilicity/hydrophobicity (F-filtration process, Ccoagulation process, A-adsorption process, RAW-the raw bio-treated coking wastewater, HOA-hydrophobic acid, HOB-hydrophobic base, HON-hydrophobic neutral, HIS-hydrophilic substances, Pi stands for the proportion of region i, i = 1, 2, 3, 4, 5).

Table 4 Water quality evaluation results obtained using the Nemerow index method. Single factor index(Pi) Sample Number 1 2 3 4 5 Average Integrated water quality index (I) Pollution Level

TN 2.82 2.52 2.18 1.81 1.38 2.14 2.50 Medium

S23.00 2.60 2.20 1.00 0.20 1.80 2.47 Medium

SCN− 6.50 6.00 4.00 1.50 0.15 3.63 5.26 Heavy

CN− 5.50 4.00 3.00 2.00 0.25 2.95 4.41 Heavy

pollution indexes (Table 4). Pollution levels fell into three categories: heavy pollution, medium pollution and slightly pollution. The indexes in heavy pollution category including SCN− and CN−, which was the main pollutants in the final effluent. The medium pollution indexes included COD, S2−, and TN, and the slightly pollution indexes included BOD, NH4+-N, phenols, SS and lipid. It was indicated that (i) COD value did not accurately represent the real pollution level of the effluent, namely COD reduction being out of the step of pollution level reduction; and (ii) COD does not well-correlated with the pollution degree of heavy metals, micro-pollutants, pathogenic microorganisms and eutrophic elements like nitrogen or phosphorus in wastewater. Toxicity index is more indicative of the real pollution degree than COD when considering the discharge of the final effluent into the receiving aquatic environment. A multivariate analysis was also conducted to assess the potential relationships between the toxicity and coking wastewater factors (Zhao et al., 2014). Five test species belonging to different trophic levels were used to evaluate the toxicity of wastewater from different treatment stages in the same coking wastewater treatment plant. The results showed that the toxic unit of the bio-treatment effluent was reduced to 2.26, a level that still has adverse effects for the green alga. For acute toxicity protection, the criteria maximum concentration should be below 0.3 toxic units for the green alga. As a result, the effluent after biological treatment needed to be diluted for 8 times before its safe

COD 2.11 1.19 0.72 0.17 0.07 0.85 1.61 Medium

BOD 0.73 0.62 0.49 0.06 0.03 0.38 0.58 Slightly

NH4+-N 0.34 0.31 0.25 0.13 0.03 0.21 0.28 Slightly

Phenols 0.67 0.50 0.37 0.17 0.03 0.35 0.53 Slightly

SS 1.04 0.10 0.07 0.04 0.01 0.25 0.76 Slightly

Lipid 0.76 0.60 0.44 0.12 0.04 0.39 0.60 Slightly

discharge. In addition, the change in hazardous materials in coking wastewater at different treatment stages and the effects of them on the development of maize embryos and the activity of amylase and protease in maize seeds were also investigated (Wei et al., 2012). Results showed that biodegradable and the refractory organic were the most toxic compounds in coking wastewater. Unfortunately, the residual refractory organics in the BTCW still showed a significant inhibiting effect on maize embryo development. Photo-degradation processing can increase the toxicity on maize embryo development due to the increase in phenolic compounds transformed from other species of organics. However, ozone oxidation can reduce the content of refractory organics significantly in coking wastewater, so the toxicity of the wastewater on maize embryo development is reduced (Prasse et al., 2015). The changes in the toxicity of the effluent is relevant to the changes in ultraviolet absorbance and fluorescence which are an indicator of refractory organics and phenolic compounds properties. To our best knowledge, some previous studies suggested that advanced treatment processes (flocculation, tertiary clarification, dual-media filtration, chlorine disinfection and dechlorination) contribute markedly to the enhanced reduction of estrogenic potency following biological treatment but information regarding the mechanism behind toxicity reduction by these techniques is limited (Valitalo et al., 2017). Based on the literature results and the findings of this study, it is clear that highly toxic organics should be removed in a higher priority when pursuing 331

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al.

high COD removal rates, which is an important guideline to be followed for the design of advanced treatment technology for the tertiary treatment of BTCW.

size > 0.45 μm). However, the relatively high removal rate of HOA (66.25%) and HON (41.56%) components suggested that the efficiency of coagulation depended greatly on the hydrophobicity of colloids. (3) The value of UV254 was 0.76 cm−1 in the coagulation effluent and a few fluorescence peaks such as peak-T1 (aromatic protein like substances), and peak-C (humic acid like substances) were still remained as reflected by the 3D-EEM. It was suggested that the residual dissolved organic matter and reducing inorganic substances needed to be further separated by physical and chemical methods. Subsequently, the UV254 value was reduced to only 0.15 cm−1 in the adsorption effluent and all fluorescence peaks were barely detectable, indicating that the adsorption process can effectively remove unsaturated and fluorescence-generating dissolved organic matter. Ozone oxidation is effective for the removal of residual reductive inorganic substances. (4) Safety assessment and correlation analysis suggested that BTCW posed potential ecological risk, mainly due to the presence of residual toxic compounds. In view of this, the removal of highly toxic organics should be a primary consideration, rather than simply pursuing high COD removal rates for the design and selection of advanced treatment technology and for the development of water safety management solutions.

3.4. Correlations analyses The correlation between residual COD and various water quality indicators as well as solution properties was investigated. The residual COD showed strong correlation (R2 > 0.8, p < 0.05) with BOD, TOC, A (representing the fulvic-acid-like substances), C (representing the humic-acid-like substances), T1 (representing the aromatic-protein-like substances), T2 (representing the soluble microbial byproduct-like material), phenols, lipid and substances reflected by UV254 (Table S6). BOD, TOC, phenols and lipid are water quality parameters which represent organic concentration from different aspects. By establishing a correlation between residual COD and these parameters, it is helpful to consider more comprehensively about the solution of further treatment to the BTCW. For example, the ratio of BOD to COD ratio shows the biodegradability of the substances in BTCW. The priority removal order of organic COD and inorganic COD in the BTCW may be based on the ratio of TOC to COD as a theoretical reference. The degree of unsaturation in organics in the residual COD of the BTCW is reflected by combining evaluation of the UV254, phenols and COD values, thereby guiding the matching of target pollutants with advanced treatment techniques. The correlation coefficients of fluorescence EEM peaks (T1, T2, C, A) with residual COD all exceeded 0.8, thus the total fluorescence intensity of the peaks might be used as indicators to characterize the comprehensive content of dissolved organic matter in the BTCW. The organics reflected by peak-T1 organics can be combined with those reflected by peak-A, peak-C and peak-T2 to coexist in the BTCW, which was confirmed that π-π interactions play important roles in this process (Wang et al., 2018). In addition, the correlation coefficients of Cl−, SCN−, CN−, S2− and residual COD did not exceed 0.8. S2− was almost completely removed by bio-oxidation so that its influence on COD can be neglected in the BTCW. SCN− and CN− were converted into NO3− or NO2−. Although residual COD caused by NO2− could be further oxidized into NO3− by advanced oxidation, it associates with another problem that the rising NO3− concentration causes TN of the effluent not to meet the standard discharge limit as NO3− is generally removed by denitrification or anaerobic ammonium oxidation (Ferraz and Yuan, 2017). Therefore, a rational solution to the removal of residual COD based on different actual situation is needed for consideration by combining the correlation analysis of COD with water quality indicators as well as solution properties.

Further work will be required for establishing a wastewater solution property-treatment technology-water quality management-ecological risk nexus. Acknowledgements This work was supported by the State Program of the National Natural Science Foundation of China (No. 51878290, No. 51808297), and the Program for Science and Technology of Guangdong Province, China (No. 2015B020235005, No. 2015A020215008 and No. 2016A020221037). Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jenvman.2019.06.001. References Chen, R.H., Li, F.P., Zhang, H.P., Jiang, Y., Mao, L.C., Wu, L.L., Chen, L., 2016. Comparative analysis of water quality and toxicity assessment methods for urban highway runoff. Sci. Total Environ. 553, 519–523. Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003. Fluorescence excitationemission matrix regional integration to quantify spectra for dissolved organic matter. Environ. Sci. Technol. 37, 5701–5710. Cui, J., Jing, C., Che, D., Zhang, J., Duan, S., 2015. Groundwater arsenic removal by coagulation using ferric (III) sulfate and polyferric sulfate: a comparative and mechanistic study. J. Environ. Sci. (China) 32, 42–53. EPA of China, 2002. Water and Wastewater Monitoring Analysis Method, fourth ed. Publisher, Chinese Environment Science, Beijing. EPA of China, 2012. Emission Standard of Pollutants for Coking Chemical Industry (GB 16171-2012). State Environmental Protection Agency of China, Beijing. EPA of China, 2017. Annual Statistic Report on Environment in China. Chinese Environment Science Publisher, Beijing. Ferraz, F.M., Yuan, Q., 2017. Nitrite Interference with soluble COD measurements from aerobically treated wastewater. Water Environ. Res. 89, 549–554. Gounden, A.N., Singh, S., Jonnalagadda, S.B., 2018. Simultaneous removal of 2,4,6-tribromophenol from water and bromate ion minimization by ozonation. J. Hazard Mater. 357, 415–423. He, X., Xi, B., Wei, Z., Jiang, Y., Yang, Y., An, D., Cao, J., Liu, H., 2011. Fluorescence excitation-emission matrix spectroscopy with regional integration analysis for characterizing composition and transformation of dissolved organic matter in landfill leachates. J. Hazard Mater. 190, 293–299. Hong, T., Dang, Y., Zhou, D., Hu, Y., 2016. Study on the oxidative characteristics of organics in bio-treated textile wastewater by VUV/US/O3 process. Chem. Eng. J. 306, 560–567. Hudson, N., Baker, A., Ward, D., Reynolds, D.M., Brunsdon, C., Carliell-Marquet, C., Browning, S., 2008. Can fluorescence spectrometry be used as a surrogate for the

4. Conclusions A stepwise process composing of membrane filtration, coagulation, adsorption and ozonation was proposed to fractionate the residual COD. Changes in the solution properties during the stepwise process was characterized. Key findings were described as follow. (1) Residual COD in the BTCW can be fractionated into suspended solids, colloids, dissolved organics and reductive inorganic substances, with their contributions to the total COD being 43.7%, 22.1%, 26.2% and 4.9%, respectively. Residual COD could be reduced to near zero (from 168.8 to 5.2 mg L−1) when suspended solids, colloids, dissolved organics and reductive inorganic substances were removed from the BTCW, thus meeting the requirements for water reuse. (2) Suspended solids, colloids, dissolved organics and reductive inorganic substances may overlap each other during the stepwise process but it didn't affect the property characterization of these different COD fractions with particle sizes. Suspended solids and colloids accounted for 65.8% of the residual COD in the BTCW. Membrane filtration effectively removed suspended solids (particle 332

Journal of Environmental Management 246 (2019) 324–333

C. Wei, et al. biochemical oxygen demand (BOD) test in water quality assessment? An example from South West England. Sci. Total Environ. 391, 149–158. Jiang, B., Tan, L., Ning, S., Shi, S., 2016. A novel integration system of magnetically immobilized cells and a pair of graphite plate-stainless iron mesh electrodes for the bioremediation of coking wastewater. Bioresour. Technol. 216, 684–690. Kaewsuk, J., Seo, G.T., 2011. Verification of NOM removal in MIEX-NF system for advanced water treatment. Separ. Purif. Technol. 80, 11–19. Karahan, O., Dogruel, S., Dulekgurgen, E., Orhon, D., 2008. COD fractionation of tannery wastewaters-particle size distribution, biodegradability and modeling. Water Res. 42, 1083–1092. Kaur, K., Mor, S., Ravindra, K., 2016. Removal of chemical oxygen demand from landfill leachate using cow-dung ash as a low-cost adsorbent. J. Colloid Interface Sci. 469, 338–343. Li, J., Liu, X., Cheng, F., 2017. Bio-refractory organics removal and floc characteristics of poly-silicic-cation coagulants in tertiary-treatment of coking wastewater. Chem. Eng. J. 324, 10–18. Li, J., Yuan, X., Zhao, H., Li, F., Lei, Z., Zhang, Z., 2018. Highly efficient one-step advanced treatment of biologically pretreated coking wastewater by an integration of coagulation and adsorption process. Bioresour. Technol. 247, 1206–1209. Pan, J., Ma, J., Wu, H., Chen, B., He, M., Liao, C., Wei, C., 2019. Application of metabolic division of labor in simultaneous removal of nitrogen and thiocyanate from wastewater. Water Res. 150, 216–224. Pan, J., Wei, C., Fu, B., Ma, J., Preis, S., Wu, H., Zhu, S., 2018. Simultaneous nitrite and ammonium production in an autotrophic partial denitrification and ammonification of wastewaters containing thiocyanate. Bioresour. Technol. 252, 20–27. Prasse, C., Stalter, D., Schulte-Oehlmann, U., Oehlmann, J., Ternes, T.A., 2015. Spoilt for choice: a critical review on the chemical and biological assessment of current wastewater treatment technologies. Water Res. 87, 237–270. Quaranta, M.L., Mendes, M.D., MacKay, A.A., 2012. Similarities in effluent organic matter characteristics from Connecticut wastewater treatment plants. Water Res. 46, 284–294. Revilla, M., Galan, B., Viguri, J.R., 2016. An integrated mathematical model for chemical oxygen demand (COD) removal in moving bed biofilm reactors (MBBR) including predation and hydrolysis. Water Res. 98, 84–97. Touati, K., Gzara, L., Mahfoudhi, S., Bourezgui, S., Hafiane, A., Elfil, H., 2018. Treatment of coastal well water using ultrafiltration-nanofiltration-reverse osmosis to produce isotonic solutions and drinking water: fouling behavior and energy efficiency. J. Clean. Prod. 200, 1053–1064. Tran, N.H., Ngo, H.H., Urase, T., Gin, K.Y., 2015. A critical review on characterization strategies of organic matter for wastewater and water treatment processes. Bioresour. Technol. 193, 523–533. Valitalo, P., Massei, R., Heiskanen, I., Behnisch, P., Brack, W., Tindall, A.J., Du Pasquier, D., Kuster, E., Mikola, A., Schulze, T., Sillanpaa, M., 2017. Effect-based assessment of

toxicity removal during wastewater treatment. Water Res. 126, 153–163. Vyrides, I., Stuckey, D.C., 2009. A modified method for the determination of chemical oxygen demand (COD) for samples with high salinity and low organics. Bioresour. Technol. 100, 979–982. Wang, L., Hu, H., Wang, C., 2007. Effect of ammonia nitrogen and dissolved organic matter fractions on the genotoxicity of wastewater effluent during chlorine disinfection. Environ. Sci. Technol. 41, 160–165. Wang, L., Li, Y., Zhang, P., Zhang, S., Li, P., Wang, P., Wang, C., 2018. Sorption removal of phthalate esters and bisphenols to biofilms from urban river: from macroscopic to microcosmic investigation. Water Res. 150, 261–270. Wei, X.X., Zhang, Z.Y., Fan, Q.L., Yuan, X.Y., Guo, D.S., 2012. The effect of treatment stages on the coking wastewater hazardous compounds and their toxicity. J. Hazard Mater. 239–240, 135–141. Wen, G., Linjun, Z., Jining, L., Guosong, C., Lili, S., Yanhua, X., 2016. Degradation characteristics of three aniline compounds in simulated aerobic sewage treat system. Environ. Sci. 37, 240–246. Yang, W., He, C., Wang, X., Zhang, Y., Cheng, Z., Dai, B., Zhang, L., 2017. Dissolved organic matter (DOM) removal from bio-treated coking wastewater using a new polymeric adsorbent modified with dimethylamino groups. Bioresour. Technol. 241, 82–87. Yang, W., Li, X., Pan, B., Lv, L., Zhang, W., 2013. Effective removal of effluent organic matter (EfOM) from bio-treated coking wastewater by a recyclable aminated hypercross-linked polymer. Water Res. 47, 4730–4738. Yang, W., Wang, J., Hua, M., Zhang, Y., Shi, X., 2018. Characterization of effluent organic matter from different coking wastewater treatment plants. Chemosphere 203, 68–75. Zhang, S., Zheng, J., Chen, Z., 2014. Combination of ozonation and biological aerated filter (BAF) for bio-treated coking wastewater. Separ. Purif. Technol. 132, 610–615. Zhang, T., Wei, C., Ren, Y., Feng, C., Wu, H., 2017. Advances in airlift reactors: modified design and optimization of operation conditions. Rev. Chem. Eng. 33. Zhang, W., Wei, C., Feng, C., Yan, B., Li, N., Peng, P., Fu, J., 2012. Coking wastewater treatment plant as a source of polycyclic aromatic hydrocarbons (PAHs) to the atmosphere and health-risk assessment for workers. Sci. Total Environ. 432, 396–403. Zhao, J., Jiang, Y., Yan, B., Wei, C., Zhang, L., Ying, G., 2014. Multispecies acute toxicity evaluation of wastewaters from different treatment stages in a coking wastewatertreatment plant. Environ. Toxicol. Chem. 33, 1967–1975. Zhou, H., Wei, C., Zhang, F., Liao, J., Hu, Y., Wu, H., 2018. Energy-saving optimization of coking wastewater treated by aerobic bio-treatment integrating two-stage activated carbon adsorption. J. Clean. Prod. 175, 467–476. Zularisam, A.W., Ismail, A.F., Salim, M.R., Sakinah, M., Matsuura, T., 2009. Application of coagulation-ultrafiltration hybrid process for drinking water treatment: optimization of operating conditions using experimental design. Separ. Purif. Technol. 65, 193–210.

333