Journal of Cleaner Production 200 (2018) 568e577 Contents lists available at ScienceDirect Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro Treatment of centrifugal mother liquid of polyvinyl chloride by internal circulation aerobic bioﬁlm reactor: Lab to plant scale system udia G. Silva d, Salma Tabassum a, b, *, Qinhong Ji c, Chunjie Li b, Lina Chi b, Cla Amin F.A. Ajlouni a, Chunfeng Chu b, Rua Alnoman a, Zhenjia Zhang b, ** a Chemistry Department, Faculty of Science, Taibah University, Yanbu Branch, 46423, Yanbu, Saudi Arabia School of Environmental Science and Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China New Energy Research Center, CNOOC Research Institute Ltd., Beijing, 100028, China d Laboratory of Separation and Reaction Engineering - Laboratory of Catalysis and Materials (LSRE-LCM), Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias s/n, 4200-465, Porto, Portugal b c a r t i c l e i n f o a b s t r a c t Article history: Received 21 September 2017 Received in revised form 21 July 2018 Accepted 27 July 2018 Available online 30 July 2018 Centrifugal mother liquid (CML) is a typical poorly biodegradable organic wastewater generated during the production of polyvinyl chloride (PVC). A method of treating PVC centrifugal mother liquid by a high efﬁciency internal circulation aerobic bioﬁlm reactor (four reactors lined up together) is presented in this study. The optimized technique was tested and evaluated in a lab and in a plant scale system. The effect of chemical oxygen demand (COD) removal efﬁciency was studied under different hydraulic retention time (HRT): 40, 20, 17, 14 and 12 h, respectively. Considering the efﬁciency and removal rate, the best HRT was 14 h. At HRT of 14 h, the average volume loading was 0.326 kg COD/(m3∙d). During the product shift from TL-1000 to TL-800 (two kinds of PVC products with different polymerization degree), the treatment efﬁciency of the system was not affected. Compared with the plant scale system, the results showed that in the lab setup the removal rates of centrifugal mother liquid of TL-1000 (total organic carbon (TOC) ¼ 70e75 mg/L) and TL-800 (TOC ¼ 20 mg/L) were 90%e95% and 80%e90%, respectively. While in the plant scale system, the removal rates of centrifugal mother liquid of TL-1000 and TL-800 were 80% e90% and 65%e80% (efﬂuent TOC 7.5e12 mg/L), respectively. Bioﬁlm formation by the microbial reproduction as well as intensive ﬁlamentous bacteria settlements were observed. © 2018 Elsevier Ltd. All rights reserved. Keywords: Centrifugal mother liquid Polyvinyl alcohol Membrane bioreactor Wastewater Filler 1. Introduction Worldwide production of Polyvinyl chloride (PVC) has risen sharply over recent years (Gravenfors et al., 2015; Wang et al., 2016). Regardless of the claims that PVC production has many negative effects on the natural environment and human health (Bidoki and Wittlinger, 2010), it is still extensively used (Jia et al., 2016). China has become the world's leading PVC manufacturer and consumer in recent years with an output of PVC reaching 16.10 million tons (Bing, 2012). Centrifugal mother liquid (CML) is one of the main sources of wastewater which comes from the production of PVC resin by suspension polymerization process in PVC * Corresponding author. ** Corresponding author. E-mail addresses: firstname.lastname@example.org, email@example.com (S. Tabassum), firstname.lastname@example.org (Q. Ji), email@example.com (Z. Zhang). https://doi.org/10.1016/j.jclepro.2018.07.276 0959-6526/© 2018 Elsevier Ltd. All rights reserved. production industry. Three to ﬁve tons of deionized water is required for every 1 ton of PVC production (Li and Zhang, 2001), and the CML accounts for about 70% of the total PVC production wastewater. An annual output of 100 thousand tons of PVC factory discharges more than 1000 m3 centrifugal mother liquid per day (Zhao et al., 2011). Various treatment methods have been used for Polyvinyl Alcohol (PVA) wastewater treatment, such as adsorption (Zhang et al., 2012), biological treatment (Yu et al., 1996), chemical oxidation (Kim et al., 2003) and membrane ﬁltration (Blanco et al., 2015). However, the conventional biological system is not able to degrade PVA efﬁciently due to the low biological oxygen demand (BOD) and chemical oxygen demand (COD) ratio i.e. BOD5/COD (0.01) of PVA (Wang et al., 2014). Moreover, due to foam generation, low treatment efﬁciency and difﬁcult operation were observed during PVA wastewater treated by biological reactors (Zhang and Yu, 2004). Clogging of membrane was also observed by PVA, as PVA could pass through the biotreatment processes during membrane-based S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 wastewater treatment techniques (Blanco et al., 2015). CML contains relatively low concentrations of refractory organic pollutants. CML has low biodegradability, low hardness, low salinity, and high turbidity (Zhao et al., 2011). Its COD value usually ranges from about 100 to 300 mg/L, while its BOD5/COD value ranges from about 0.2 to 0.3 (Huang, 2007). Due to the low biodegradability, the biotreatment efﬁciency of CML is usually low, and this can cause problems, such as the bulking of activated sludge and difﬁculties in microbe culture, and low treatment efﬁciency (Phugare et al., 2011; Zhao et al., 2011). It also contains the dissolved polymers (mainly PVA), which are used as dispersing agents in the polymerization process. CML may also contain highly toxic residual vinyl chloride monomer (VCM) (Labunska et al., 2008). Various conventional CML treatment methods have been used such as coagulation/ﬂocculation (AlMubaddal et al., 2009), ﬁltration (Zhang et al., 2003), biotreatment (Song and Li, 2009) and membrane technologies (Yang et al., 2005), etc. Coagulation/ﬂocculation or ﬁltration can effectively remove suspended solids but cannot remove dissolved pollutants effectively (Huang, 2007). Other disadvantages of ﬁltration include low ﬁltration speed, high electricity consumption, etc. (Zhao et al., 2011). These techniques have certain advantages such as simple operation and high treatment efﬁciency, but the membrane can be easily clogged by residual PVA, and therefore they need more periodic reverse cleaning and replacement, which lead to additional cost (Huang, 2007; Yang et al., 2005). Recently (Ye et al., 2017), the environmental impact generated from the PVC industry was reported. Due to the strong surface activity of PVA, it poses a large threat to the environment by forming high amounts of foam. Because of the water-solubility of PVA, its impact on environment is not so obvious, unlike other polymers, such as solid plastic. Yet, with the increasingly amounts of industrial wastewater containing PVA being discharged, environmental consequences have to be taken into account, as PVA usually takes about 900 days to be decomposed completely in aquatic environment (Zhang and Zhou, 2003). This creates a serious threat to the sources of fresh water supply and environment (Avio et al., 2017; Halima, 2016). Therefore, it is necessary to develop efﬁcient and improved alternative processes for treating this type of waste (Quartey et al., 2015). According to many research works, PVA is degradable both by aerobic and anaerobic methods. Currently, the treatment of wastewater containing PVA mainly uses techniques combining biochemical technology with AOPs (advanced oxidation processes) or physicochemical methods (Punzi, 2015). Combining biochemical technology with physicochemical technology is the most popular way and also a growing trend, because of its low cost and good treatment result. Moreover, it is necessary for wastewater treatment systems to be designed by the engineers on the natural ways of cleaning wastewater through the use of microorganisms. Compared to suspended growth systems, wastewater treatment with bioﬁlm systems have several advantages such as low space requirement, operational ﬂexibility, reduced hydraulic retention time, increased biomass residence time, resilience to changes in the environment, high active biomass concentration, enhanced ability to degrade recalcitrant compounds as well as lower sludge production (Sehar and Naz, 2016). Over time, the bioﬁlms in the suitable environment will grow by feeding off the organic matter and nutrients in the wastewater that ﬂow over them and become strongly attached to the surface they lez, 2012). The surface area and geometry of supgrow on (Gonza port materials affect the hydrodynamic conditions in the reactor and thus affect bioﬁlms formation, which in turn affects wastewater treatment (Matos et al., 2011; Yu et al., 2008). Nowadays, 569 different synthetic and natural materials have been proposed by researchers as bio-ﬁlter media in ﬁxed bioﬁlm reactors for wastewater treatment, such as polystyrene (Naz et al., 2013), polypropylene (Khatoon et al., 2014), tire-derived rubber (Naz et al., 2014) and pebbles (Khan et al., 2015). Tianjin LG Dagu Chemical Co., Ltd. (LG-DAGU), located in Tanggu District in Tianjin, China, is a Sino-Korean joint venture, which is set up by LG Chemical (the largest chemical company in South Korea) and Tianjin Dagu Chemical Co., Ltd. LG-DAGU is one of the largest PVC manufacturer in China, which uses chlorethylene to produce two kinds of PVC (TL-800 and TL1000) by suspension polymerization process. With the increasing production, water usage has also increased, which stands at around 2 million tons per year. For the further development, LG-DAGU needs sufﬁcient water, which is a difﬁcult task, because of water shortage in Tianjin and strict municipal government control over industrial water usage. Project collaboration was setup between LG-DAGU and Shanghai Jiao Tong University to ﬁgure out ways to reduce water consumption and improve water efﬁciency. In the current study a laboratory was setup in the waste water treatment plant that utilizes the BioAX water treatment technology patented by our group (a novel environmental biotechnological aerobic process with internal circulation or internal circulation aerobic bioﬁlm reactor) (Zhang, 26 December 2002e9 November 2005). A great amount of work is needed for implementation. But our research group has effectively and practically worked on achieving the best result from BioAX in last 4 years based on the research performance not only in laboratories but also employing successfully the installation and operation in the pilot scale. For instance, we have successfully implemented this reactor (Li et al., 2014) in the treatment of one of the most difﬁcult wastewater around the world i.e. coal gasiﬁcation wastewater (CGW). Also, BioAX exhibited high efﬁciency during a key in-depth technological research (Ji et al., 2015) for treatment of highly contaminated and toxic CGW. Later, Tabassum et al., 2015a used BioAX along with other reactors for advanced treatment of coal gasiﬁcation wastewater. Moreover, the same reactor has been used in the successful running of 5 m3/h Palm Oil Mills Efﬂuent (POME) zero discharge Pilot plant, built at MPOB (Malaysian Palm Oil Board) Experimental Palm Oil Mill, Labu, Malaysia (Tabassum et al., 2015b). The research team conducted an experiment on the treatment of PVC-CML by internal circulation aerobic bioﬁlm reactor or BioAX. The main objectives of this study are (i) the inoculation and start-up of BioAX reactor system; (ii) the continuous operation performance of BioAX reactor system when there is a sudden change in CML; (iii) wastewater pollutant degradation and bioﬁlm condition in different reaction phases; (iv) comparing the performances from lab to plant scale condition. 2. Materials and methods Two kinds of PVC (TL-800 and TL-1000) were produced by LGDAGU, which have different degrees of polymerization. Therefore, the discharged CMLs have different characteristics (Table 1). This experimental study used the CML samples taken from LG-DAGU. At the beginning of the experiment, the plant produced two kinds of PVC periodically, leading to periodical changes of CML in the apparatus both in the lab and in the plant scale system. Due to the change of the water-supply pipeline, the lab scale system processed only TL-1000 CML, while the plant scale system processed the mixture CML from the two kinds of PVC. As shown in Table 1, the conductivity of CML is just a third of that of tap water, because it comes from deionized water. CML has little amount of organics, among which two are major ones, extremely small PVC particle and PVA. The PVA is used as 570 S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 Table 1 Characteristics of CML based on two kinds of PVC. TL-800 TL-1000 Temperature ( C) pH COD (mg/L) TOC (mg/L) SS (mg/L) Turbidity (NTU) Conductivity (m S/cm) 50e55 45e50 8.5e9.0 8.5e9.5 60 200 20 65 20 50 40 180 120 180 dispersant by LG-DAGU, which has a degree of polymerization (DP) as high as 2400, making it as a major pollutant. 2.1. Centrifugal mother liquid treatment process system setup This experiment used an original internal circulation aerobic bioﬁlm reactor or BioAX, which is constituted by four reactors connected in series with roughly the same size and volume, as shown in Fig. 1a. The four reactors are made of plexiglass column, with effective volumes of 46.3 L, 46.3 L, 46.2 L and 46.4 L, respectively. Each reactor is 140 cm tall with 22 cm inner diameter. Each reactor has a wastewater inlet on its lower part, a wastewater outlet on its upper part and a sludge outlet at its bottom. The vertical cylindrical tubes are used for aeration, ﬁxed in the bottom of the reactors. The mechanism of internal circulation in the aerobic bioﬁlm reactor can be observed in Fig. 1b, which is also well described in our previous research articles about coal gasiﬁcation wastewater treatment (Ji et al., 2015; Li et al., 2014). The resilient plastic ﬁlling is ﬁxed outside the vertical cylinder tube for bioﬁlm formation. The resilient plastic ﬁlling we used was like brush and the material was polypropylene (PP), relatively strong and better for bioﬁlm formation (Qureshi et al., 2005). PP is a popular synthetic ﬁlament, which is used in a wide variety of applications. Polypropylene has excellent stiffness when wet, is nonbrittle, is inert to most solvents, such as oils, acids and chemicals, and is fungus resistant (Maddah, 2016). The polypropylene, which we used, is natural in colour and 0.5 mm in diameter. It is elastic and not easy to sag when it is impacted by the downward ﬂow of wastewater. Therefore, compared with other biological membrane ﬁllers, this kind of ﬁller is more suitable for the reactor we developed. The system operates in internal circulation and down-ﬂow type mode. The picture in Fig. 1c shows the experimental laboratory device. CML wastewater from the factory ﬂows through a pipeline to the distributing gutter in the laboratory, which is then pumped into the experimental system from the lower part of reactor No.1. After treated by the experimental system, the CML wastewater ﬂows out from the last reactor (Fig. 1a). Organic matters in the CML wastewater were degraded by the microorganism which grew on the biological ﬁllers. 2.2. Reactor start-up Sludge collected from LG-DAGU's domestic sewage disposal system was blended before it was put into the reactors. The mixed liquor suspended solids (MLSS) and mixed liquor volatile suspended solids (MLVSS) value of the internal circulation aerobic bioﬁlm reactors were 3000 mg/L and 1680 mg/L, respectively, and the value of Volatile suspended solids (VSS) to total suspended solids (SS) ratio was 0.56. Each reactor was ﬁlled with CML and Fig. 1. (a) Schematic diagram of the experimental apparatus of plant scale internal circulation aerobic bioﬁlm reactor process; (b) an insight of ﬂow direction inside the reactor; and (c) Photograph of the laboratory setup of the test system. S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 loaded with 10 g of glucose. DO (dissolved oxygen) was kept at the level of 4e6 mg/L by the DO analyzer with its detector going below the bafﬂe (in the gap of the ﬁlling). Then the reactors were maintained in this way while being exposed to air. After two days, a thin layer of grey bioﬁlms appeared on the surface of the ﬁlling, which suggested the completion of ﬁrst stage inoculation and bioﬁlm formation. At the beginning of the bioﬁlm forming process, the sludge is mixed evenly. Later, it is gradually divided into layers, with relatively clear water in the upper layer and the sludge that failed to form bioﬁlms, sinks to the bottom. The inﬂuent was kept at a low ﬂow rate. After two days, a small amount of sludge was found in the wastewater, mostly inorganic, and the wastewater turned clear after settlement. The bioﬁlm formation just took two days, as the temperature was kept at around 25 C, which was favourable for the growth and rapid reproduction of the bacteria. Also, there were abundant microorganisms in domestic sewage, which grew rapidly with sufﬁcient dissolved oxygen. Moreover, the surface of resilient ﬁlling was favourable for the growth of microorganisms. After two days of inoculation, wastewater began to be pumped into the reactor. The condition of CML in the reactor was monitored periodically. 2.3. Analytical methods Certain indicators of the CML and the efﬂuent were monitored over a long period of time. The condition of aerobic bioﬁlms was analyzed from time to time. COD, total organic carbon (TOC), pH, total suspended solids (TSS), volatile suspended solids (VSS), wastewater temperature, dissolved oxygen (DO), turbidity and electro-conductibility analyses were carried out by the standard methods for the examination of water and wastewater (Eaton et al., 2005). 3. Results and discussion There is a large room for error using potassium dichromate method for measurement of low-density liquid (Davies, 2005). In this study, we paid more attention to the degradation of all carboncontaining organic matters in wastewater, not just PVA. Therefore, based on the conventional practice in the ﬁeld of wastewater treatment, we used TOC index to evaluate the concentration of all carbon-containing organic compounds in wastewater. Therefore, this study measured TOC instead of COD to analyse the amount of organisms in the liquid (Kadlec and Wallace, 2008; Mara and Horan, 2003). Therefore, the speciﬁc components of wastewater carbon-containing organic compounds were not given. 3.1. Performance of system under different hydraulic retention time (HRT) HRT is a key factor for the treatment result of the reactor, because it has a direct impact on volume loading. Ideally, HRT should be as short as possible while maintaining the quality of the efﬂuent. TL-800 did not have a signiﬁcative contribution to the sewage treatment system because of the small amount of its production. Besides, the COD of the TL-800 CML was just around 60 mg/L. The main contribution was the CML of TL-1000. During the experiments, the CML of TL-1000 (COD around 200 mg/L) was continuously pumped into the reactor. Temperature was kept at 25e35 C and DO 4e6 mg/L. Nutrients ((in mg/L): FeSO4∙7H2O 15, MgSO4∙7H2O 50, MnCl2∙4H2O 0.5, ZnCl2 0.5, CuCl2 0.5, NaBO2∙10H2O 0.3, AlCl3 0.5, CoCl2∙2H2O 0.5 and NiCl2∙2H2O 0.5) were added to the reactor with a COD, N and P proportion of 100:5:1. Different HRTs (HRT ¼ 40,20,17,14 and 12 h) were set to 571 observe the respective COD removal by measuring COD concentration in the inﬂuent and efﬂuent of the reactor. When HRT was 40 h (Fig. 2), the average volume load of the system was 0.102 kg COD/(m3∙d). The COD concentration of the efﬂuent in reactor No.2 was not so much different from that in reactor No.4, which indicated that pollutant degradation just took about 20 h to complete. Wastewater in reactor No. 1 was turbid, having rapid bioﬁlms growth. PVC particles were found around the inside of the reactor and also on the bioﬁlms after about two days. Wastewater in the other three reactors was clear, with bioﬁlms turning beige. Bioﬁlms in reactors No. 3 and No. 4 tended to form very slowly, which was attributed to long HRT and the lack of organisms. When HRT was 20 h (Fig. 2), the average volume loading was 0.233 kg COD/(m3∙d). Efﬂuent was clear without impurity or unpleasant smell. COD removal rate ranged between 75% and 90%, with COD concentration in the system efﬂuent mostly below 50 mg/L. The reactors’ adaptability improved after the initial stage. COD removal rate was also on a rise because of longer time in this period. The results showed that the removal rate of reactor No.1 directly inﬂuenced the total removal rate of the entire system. At the beginning of the experiment, the removal rate of reactor No.1 was relatively low, just 17% on the 4th day and the total system removal rate was 61% only. This was probably caused by the inﬂuence of the CML on reactor No.1. After some time, the removal rate gradually picked up, which suggested the increasing adaptability of reactor No. 1. Bioﬁlms in the ﬁrst two reactors tended to form more quickly than in the other two reactors. In reactor No.1, PVC particles were found on the bioﬁlms which turned beige and jelly-like, sticking to the ﬁlling in the later period of this experiment. In some areas of the reactor, air ﬂow rate had a clear decrease due to limited space in the reactor, which might be one of the reasons for the ﬂuctuation of removal rate. In reactor No. 2, threadlike bioﬁlms were found on the ﬁlling. The presence of bioﬁlms in the last two reactors were more when HRT was 20 h. Reactor No. 4 did very little contribution to the total pollutant removal, with the COD concentration of efﬂuent below 10 mg/L, which meant that HRT can be further reduced. When HRT was 17 h (Fig. 2), the average volume loading was 0.246 kg COD/(m3∙d), which did not increase much because of relatively low COD in the inﬂuent in this period. But due to the more stable COD concentration of the inﬂuent in comparison with when the system HRT was 20 h, the removal results were also more stable. Efﬂuent condition of the whole system was steady, with clear water without impurity or unpleasant smell. The COD removal efﬁciency ranged between 70% and 80%, which might further increase experiment time, considering that both the system stability and bioﬁlm adaptability were enhanced with time. When HRT was 14 h (Fig. 2), the average volume loading was 0.326 kg COD/(m3∙d). Efﬂuent condition of the whole system was steady, with clear wastewater without impurity or unpleasant smell. The COD removal efﬁciency was between 73% and 80%, respectively. Suspended substances were sometimes observed in the efﬂuent, which were caused by the bioﬁlms falling off during the process of its metabolism. With the lower HRT, the ﬂow rate was higher, making it easier for the fallen bioﬁlms to be ﬂushed out. The COD concentration of efﬂuent was 12.5 mg/L, which was reduced to 3.3 mg/L after the coagulant was added. CML in the ﬁrst two reactors was more turbid, while CML in the other two reactors was clearer. In reactor No. 4, green algae were found near the outlet. The HRT was further decreased to 12 h (Fig. 2), the average volume loading was 0.402 kg COD/(m3∙d). Efﬂuent of the whole system was steady, with clear wastewater without impurity or unpleasant smell. Fallen bioﬁlms were often seen in the efﬂuent due to the high water ﬂow rate. Total removal rate was between 60% and 70%. The 572 S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 Fig. 2. Comparison of COD degradation curve and removal efﬁciency when HRT was 40 h, 20 h, 17 h, 14 h, and 12 h. COD concentration of efﬂuent was between 55 mg/L and 70 mg/L. After inoculation and adaptation, bioﬁlms could effectively degrade organic pollutants in the CML. Fig. 2 illustrated the relationship between COD removal results and HRT. Under the same HRT, the removal rate picked up with time as the system got adapted to the CML. As HRT decrease, COD removal rate also decreased. With the lowest HRT of 12 h, COD concentration in the efﬂuent no longer met the COD requirement (COD<60 mg/L) of reused water. Taking both efﬁciency and removal rate into account, the most appropriate HRT was 14 h (2/3 of 20 h), with the removal rate being slightly lower in comparison to the HRT of 20 h. Each reactor made a different contribution to the total removal rate. Reactor No. 1 was the greatest contributor under different volume loading conditions, with the removal rate as high as 40%. Also, the ﬂuctuation of removal rate in reactor No. 1 directly inﬂuenced the total removal result. A large amount of PVC particles was found on the bioﬁlms in reactor No. 1, which led to the appearance of beige jelly-like substances on the ﬁlling in the later experimental period. This showed that reactor No. 1 removed COD through degradation as well as adsorption of large amount of PVC particles and other macromolecular organic pollutants, making this reactor to maintain a high removal rate. The COD concentration of CML of polyvinyl chloride was low, only 100e200 mg/L, while the HRT of CML of treatment process system was relatively long. Moreover, the MLSS concentration of the system was about 5000e6000 mg/L, so the biomass-based organic loading/sludge organic loading was very low, which leaded to a sludge yield of only 0.01e0.05 kg MLVSS/kg COD. This meant that sludge yield of internal circulation aerobic bioﬁlm reactor was only 5%e10% of the sludge yield of activated sludge process. 3.2. Impact of product shifts on treatment According to the production plan of LG-DAGU, the two products TL-800 and TL-1000 were produced alternately. There was a shift of product every 30 days, with TL-1000 being the main product. Product shift exerted great inﬂuence on the treatment system due to the change of CML. In order to analyse the impact of product shift, temperature was kept at 25e35 C, DO 4e6 mg/L and HRT 14 h. Nutrients were added to the reactor with COD, N and P proportion (100: 5:1). Then, TOC concentration of the inﬂuent and the fourth reactor efﬂuent were continuously measured. Removal rate curve during product shift is shown in Fig. 3a (TOC curve of each outlet) and Fig. 3b (TOC removal rate curve of each reactor). During product shift, removal rate and outlet wastewater changed due to different adaptation of microorganisms to different CMLs. The CMLs of the two products had different types and amount of organic pollutants, which required different microorganisms for the reactor to adapt to them. Therefore, every time the product was shifted, S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 573 Fig. 3. (a) TOC curve of each outlet; and (b) TOC removal rate of each reactor. CML changed, which required microorganisms in the reactor to change. Hence, the treatment results were inﬂuenced during this process. When production was shifted from TL-1000 to TL-800, the treatment result was basically not inﬂuenced, because TL-800 CML had a very low TOC level. In addition, 14 h of HRT was sufﬁcient for organisms in the CML to degrade. During the production of TL-800, reactor No. 1 maintained a high removal rate, while TOC level in No. 2 outlet was no more than 2 mg/L higher than that in reactor No. 4, which suggested that degradation was almost completed in the ﬁrst two reactors. A small amount of suspended substances was found in reactor No. 4 in the ﬁrst two days after the shift from TL-800 to TL-1000, probably due to the fallen bioﬁlms during metabolism. Moreover, removal rate in reactor No. 1 decreased sharply, because it was not prepared for a sudden increase of TOC level in the CML. It took the reactor a week to adapt to the new CML. TOC concentration of another reactor efﬂuent was also inﬂuenced by the shift. But removal rate in these three reactors picked up after a small ﬂuctuation, which contributed to decrease the system TOC to a normal level in about four days. The reason was probably that microorganisms in the latter three reactors were hungry for organisms due to the low TOC concentration of the CML of TL-800, which allowed them to grow extremely fast once provided with abundant organics during the shift to TL-1000. But as for reactor No.1, it already developed a complete microorganism system adapted to the CML of TL-800, which made it more difﬁcult to adapt to the shift. According to the observation, bioﬁlms in reactor No.1 turned cotton-like, with outlet wastewater turning turbid in the ﬁrst two days after the shift. The process of the CML shift inﬂuencing treatment results was not frequently monitored due to limited conditions. In addition, research was not carried out on the best HRT for TL-800 CML as well as each reactor's contribution to pollutant removal, due to the short production period of TL-800. According to the data, removal rate in reactor No. 2 increased two days after the shift from TL-1000 to TL-800, probably because the reactor had formed a microorganism system adapted to the CML. Nevertheless, more research is needed to conﬁrm it. 574 S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 3.3. Comparing treatment results of lab and plant scale system In LG-DAGU, there is also an internal circulation system for CML treatment. It has four reactors connected in a line, just like the lab scale. The treatment results of the two systems were compared. As can be seen in Fig. 4, the HRT of lab scale treatment system was set at 14 h, just 2/3 of the HRT of the plant scale system which was 20 h. The experiment took place in November and December. Temperature in the lab could be maintained at 25e35 C by heating devices. But for the plant scale system, temperature was lower. Although, while entering the plant scale system, the PVC-CML was relatively warm (30 C) but the CML temperature dropped sharply due to exposure to air. The temperature drop was more rapid when inﬂuent was blocked for maintenance during product shift. According to some studies, when the temperature of sewage is below 13 C, bioactivity will be reduced, which leads to the decrease of organics removal rate. The low temperature also caused the decrease of the removal rate of the plant scale system. Comparing the treatment results of the two systems, it was clear that the lab scale setup outperformed the plant scale system when treating two kinds of CML. When dealing with the CML of TL-1000 (TOC ¼ 70e75 mg/L), the removal rate of the lab scale system was 90%e95%, and the TOC concentration of efﬂuent was 4e6 mg/L. As for the CML of TL-800 (TOC ¼ 20 mg/L), the removal rate of the lab scale system was 80%e90%, and the TOC concentration of efﬂuent was 2e4 mg/L. The removal rates of plant scale system for CML of the two products were 80%e90% and 65%e80% respectively, and the TOC concentration of efﬂuent was 7.5e12 mg/L. The performance of the plant scale system was also less stable. Firstly, the reasons for the different pollutant removal rates is the accurate temperature control, added nutrients and aeration for the lab scale system, which could help to increase the pollutant removal rate. Secondly, the bioﬁlms of the lab scale system tend to form more readily, because a smaller system is better at selecting and inoculating microorganisms. Bioﬁlms in the lab scale system were evenly attached to the ﬁllers. The amount of microorganisms was higher due to the larger density of the ﬁlling, but there was a smaller amount of microorganisms in the plant scale system. In addition, on the ﬁlling of the latter two reactors there were very few bioﬁlms which fell off easily. Although both systems used internal circulation, they had slightly different ways of aeration and sewage circulation due to different scales. The lab scale system was more efﬁcient in oxygen usage and transmission because of better control over wastewater circulation. But in the later period of the experiment, bioﬁlms growth and PVC particles attached on the ﬁlling prevented the CML from fully contacting with the bioﬁlms in some areas of reactor No.1. In comparison, in the plant scale system the CML was more evenly distributed, but with less efﬁciency in oxygen usage. Thirdly, the lab scale system was more vulnerable to the change of CML during the shift from TL-800 to TL-1000. From day 9e11, TOC removal efﬁciency of both systems dropped due to sudden increase of TOC, with 7.6% reduction of the removal rate of the lab scale system. But in this period, the removal rate of the plant scale system just decreased by 4.5%. On the 31st day, when there was another product shift, the removal rate of the lab scale system even dropped below the plant scale system level to 65%. The reason is that the plant scale system is larger and with more bio-variety than the lab scale system, which makes it more resilient to the sudden increase of volume loading. But microorganisms in the lab scale system are stronger in adaptation and reproduction, which can bring up the removal rate to its normal level within two or three days under the favourable condition. The lab scale system was more vulnerable to the change of CML during the shift from TL-1000 to TL-800. According to our observation, the removal rate of the plant scale system dropped to 50% following a sudden decrease of TOC, with the removal rate of the lab scale system falling to around 80%. The following are the possible reasons. Due to TL-1000 has a longer production period, both systems will struggle to readapt and reconstruct after the production shifted to TL-800. But the lab scale system was less vulnerable to the shift because of its more favourable conditions for the optimal growth of microorganisms, including temperature, nutrients and aeration. In the plant scale system, more bioﬁlms fell off than that in the lab scale system, leading to more suspended substances in outlet water, which increased the TOC concentration of this stream. Besides, during the product shift, more bioﬁlms fell off. In addition, TOC concentration in CML of TL-800 was relatively low in the ﬁrst place. Therefore, the removal rate was even lower. 3.4. Observation of bioﬁlms growth This experiment used resilient ﬁlling was made of PVC ﬁber and Fig. 4. TOC removal curves of laboratory and plant scale system. S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 braided into a bundle (diameter 150 mm). The ﬁlling was ﬁxed on the internal circulation tube with thick iron wires as long as 1500 mm coiling around the tube. The resilient ﬁlling has many advantages (Ji et al., 2015; Li et al., 2014) as it is light-weighted with stable physical and chemical forms. The ﬁber bundles have a threedimensional structure in the ﬂowing water, which increases the surface area of the ﬁber bundles. The bioﬁlms can easily attach to the ﬁlling (Li et al., 2018), making good contact with CML. The ﬁber bundles can sway with the water, which can prevent blockage caused by bioﬁlms. This design is useful for treatment of sewage with high organics concentration (Li et al., 2016). There were no signiﬁcant differences between the growth conditions of bioﬁlms in one reactor under different HRT (less than 20 h) due to the low system volume loading. As the TOC concentration gradually decreased, bioﬁlms in each of the reactors started to show a different growth pattern. Bioﬁlms in the former two reactors tend to form faster than that in the other two reactors. After a month of steady-state operation (HRT 14 h), sludge was taken from all the four reactors. Sticky clumps of bioﬁlm were observed in reactor No. 1 (Fig. 5a). The bioﬁlms were jelly-like, difﬁcult to be peeled off from the ﬁlling. PVC particles were attached to the bioﬁlms. The bioﬁlms in reactor No. 1 had a very different look from those in other reactors. Fig. 5c shows a picture of the ﬁlling before it was placed into the reactor. Bioﬁlms in reactors No. 2 (Fig. 5d) and No. 3 (Fig. 5b and e) looked quite similar. Bioﬁlms were clearly seen attached on the PVC ﬁber. Bioﬁlms in these two reactors were easier to peel off. Reactor No. 4 (Fig. 5f) had loose bioﬁlm architecture, making it easier to peel off from the ﬁlling. 3.5. Microscopic examination of bioﬁlms Analysing the growth of the bacteria is the best way to evaluate the process and result of sewage treatment. Yet, it takes a long time for observation and identiﬁcation of the bacteria, which makes difﬁcult to use the analysis results to guide and forecast the future production. Protozoa and metazoan are interdependent from bacteria (Sapp, 1994). They are also easier to be observed due to the larger sizes and easier to reﬂect environmental changes because they are more sensitive to environmental changes than bacteria (Madoni, 2011). Therefore, analysis of the types, amount, growth and changes of protozoa and metazoan can also reﬂect the growth and changes of bacteria, which can help in learning about the 575 treatment process results. The information gathered can facilitate in setting future production plans. During the experiment, bioﬁlms were examined by the microscope to observe their growth as well as protozoa. As the inoculation sludge was collected from domestic sewage, organisms on the bioﬁlms were diversiﬁed, with lots of Vorticella, Nematode, Paramecia, Litonotus, Rotifera and sometimes Colpidium and Amoeba. After ﬁve days of operation, Zoogloea was observed including mainly Vorticella and Nematode, and sometimes Litonotus. Moreover, when the operation went stable, Zoogloea was observed with Vorticella, Nematode and lots of Potifera. The latter two reactors hosted fewer Protozoons, probably due to fewer bacteria. After product shift, some cotton-like sludge was found in reactor No. 1, which might be the fallen bioﬁlms and also small amount of sludge was found in the outlet of reactor No. 4. Internal circulation aerobic bioﬁlm reactor has the advantage of low sludge production, it is explained in detail in our previous research article about treating coal gasiﬁcation wastewater (Ji et al., 2015). According to microscopic examination, Zoogloea fell apart, which indicated the renewal and reconstruction of the bioﬁlms. Similar to the lab scale system, bioﬁlms in the former two reactors of the plant scale system tended to form better than those in the latter two, but were also easier to fall off than those in the lab scale system due to the large inﬂuence from air ﬂow, water ﬂow and gravity. Bioﬁlms in the latter two reactors of the plant scale system fell off easily, leading to lots of suspended sludge that were found in the efﬂuent. More Vorticella and Rotifera were found in the ﬁrst two reactors. At the beginning of sludge inoculation, treatment result was very unstable. Later, bioﬁlms tended to form better. Microscopic examination revealed the presence of Vorticella, Rotifera and Nematode, which indicated a good treatment result. It was also found that product shift had some impact on the bioﬁlms (Fig. 6aed). 3.6. Operational problems and countermeasures Several problems must be acknowledged as a part of this study along with the remediation's are listed below. 3.6.1. Blocked area in Reactor No.1 In the lab scale system, blocked areas were often formed in reactor No. 1, by PVC particles (CML) getting attached to the Fig. 5. (a) #1 MBR effective volume: before packing 45.7 L, after packing 46.3 L; (b) #3 MBR effective volume: before packing 46.7 L, after packing 46.2 L; (c) ﬁller; and (d)e(f) photographs of bioﬁlm on the surface of ﬁller (lush breeding microbial colonies). 576 S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 Fig. 6. Microorganisms under the Microscope (a) Rotifera; (b) Nematode; (c) Vorticella; (d) nematode and Vorticella; (e) Chironomus larvae growing stages and (f) Chironomus (adult). bioﬁlms. The blockage often got in the way of air ﬂow and wastewater ﬂow. In order to tackle this problem, some measures can be taken to prevent PVC attachment on the bioﬁlms, including stirring and more aeration. In practice engineering design, we periodically wash off the excess bioﬁlms on the PVC ﬁber ﬁlling in the reactors. It is also used to remove the PVC particles that are trapped in the bioﬁlms due to bioﬁlms ﬁltration and ensure the normal operation of the system. After two months of operation, sludge can be removed by the above measures for the reactor to be inoculated again. In the plant scale system, there were no blocked areas due to different ways of ﬁlling and wastewater ﬂow. Sometimes, PVC was found attached to the bioﬁlms in the plant scale system, but not as much as in the lab scale system. So, in the plant scale system, sludge can be removed less frequently. 3.6.2. Chironomus larvae boom After two months of stable operation, red bugs were found in reactor No. 2. And later, same bugs were also seen in the third and the fourth reactors, which were red and 1e2 cm long, sticking to the bioﬁlms and reactor walls, mostly on the middle and upper parts. These bugs were conﬁrmed to be chironomus larvae. A week later, chironomus were found on top of the reactors (Fig. 6e and f). In recent years, the research focused on adult chironomid taxonomy and benthic macroinvertebrate community help in evaluating water quality (Nicacio and Juen, 2015). In the sewage treatment process, chironomid larvae appearance is an indication of stable and effective sewage treatment process. Chironomid larvae's initial appearance in the early stage can be utilised to accelerate the bioﬁlms detachment and renewal (Babu, 2011). When they reach a large number, they might harm the bioﬁlms. Precautions should be taken at this stage by increasing the amount of inﬂuent water in the reactor which can lead to the death of the larvae. Then the reactors can be re-inoculated. This whole process takes two days. 4. Conclusion PVC centrifugal mother liquid was treated by a high efﬁciency internal circulation aerobic bioﬁlm reactor (four reactors lined up together) in this present study from lab to plant scale system. The treatment in the ﬁrst reactor has a great inﬂuence on the efﬁciency of the whole system, and the removal efﬁciency accounts for about 40% of the total removal efﬁciency. The change in the product shift greatly inﬂuenced the treatment efﬁciency of the system. Each change needs 3e4 days for the system to adapt, therefore, uniform water quality can promote the stable operation of the system. According to observation, there was usually a large amount of foam ﬂoating in reactor No. 1. In reactor No. 2, a small amount of foam sometimes appeared. No foam was found in the other two reactors. From this observation, we can conclude that PVA was successfully degraded in the ﬁrst two reactors. Considering the high PVA concentration in the CML and from the above results, it can be deduced that almost all of the PVA molecules were destroyed or transformed. Therefore, it was reasonable to suppose that the subsequent biological treatment would remove most of the remaining COD or TOC from the CML which gone through the internal circulation aerobic bioﬁlm reactor treatment system. Under optimized treatment conditions, most of the COD was removed from the CML and the biodegradability of the CML increased signiﬁcantly. Though the sludge production was less in the system, the PVC particles caused the reduced bioﬁlm activity and chironomid larvae growth caused the serious damage to the bioﬁlms. Future study will be carried out by our research group on the adsorption of PVC particles and the growth regularity of chironomid larvae. This lab to plant scale study is expected to account to a good foundation for a successful implementation effort for the treatment of poorly biodegradable organic wastewater produced during the production of polyvinyl chloride. Acknowledgement The authors express their gratitude to the School of Environmental Science and Engineering, Shanghai Jiao Tong University for providing the research facilities. Dr. Tabassum was ﬁnancially supported by Talented Young Scientist Program supported by the Ministry of Science and Technology, P. R. China as an Assistant Researcher (IND-15-003) at School of Environmental Science and S. Tabassum et al. / Journal of Cleaner Production 200 (2018) 568e577 Engineering, Shanghai Jiao Tong University, China. Dr. Tabassum, also thankful to the Department of Chemistry, Taibah University, Saudi Arabia. C.G. Silva acknowledges the FCT Investigator Programme (IF/00514/2014) with ﬁnancing from the European Social Fund and the Human Potential Operational Programme and ﬁnancing though Project POCI-01-0145-FEDER-006984 e Associate Laboratory LSRE-LCM funded by FEDER through COMPETE2020 ~o Programa Operacional Competitividade e Internacionalizaça (POCI) e and by national funds through FCT - Fundaç~ ao para a ^ncia e a Tecnologia. Cie References AlMubaddal, F., AlRumaihi, K., Ajbar, A., 2009. Performance optimization of coagulation/ﬂocculation in the treatment of wastewater from a polyvinyl chloride plant. J. Hazard Mater. 161, 431e438. Avio, C.G., Gorbi, S., Regoli, F., 2017. Plastics and microplastics in the oceans: from emerging pollutants to emerged threat. Mar. Environ. Res. 128, 2e11. Babu, M., 2011. . Effect of Algal Bioﬁlm and Operational Conditions on Nitrogen Removal in Wastewater Stabilization Ponds. Wageningen University Institute for Environment and Climate Research (WIMEK). Wageningen University, Netherlands, CRC Press/Balkema, pp. 1e142. Bidoki, S.M., Wittlinger, R., 2010. Environmental and economical acceptance of polyvinyl chloride (PVC) coating agents. J. Clean. Prod. 18, 219e225. Bing, J., 2012. Analysis on situations of China's PVC industry in the year 2011. Polyvinyl Chloride 40, 1e8 (in Chinese). Swinnen, N., Prieto, D., Negro, C., 2015. Blanco, L., Hermosilla, D., Blanco, A., Assessment of the performance of membrane bioreactors applied to the treatment of industrial efﬂuents containing poly(vinyl alcohol). Ind. Eng. Chem. Res. 54, 5442e5449. Davies, P.S., 2005. The Biological Basis of Wastewater Treatment. Strathkelvin Instruments Ltd, pp. 1e20. Eaton, A.D., Clescen, L.S., Rice, E.W., Greenberg, A.E., 2005. APHA. Standard Methods for the Examination of Water & Wastewater. American Public Health Association, American Water Works Association, Water Environment Federation. lez, J.K.I.M., 2012. The Study of Biological Wastewater Treatment through Gonza Bioﬁlm Development on Synthetic Material Vs. Membranes. Worcester Polytechnic Institute. Gravenfors, E., Wiberg, K., Alkell, K., Eskhult, G., Fischer, S., Freij, L., Nylander, A., €m, F., Westerholm, E., 2015. Phthalates Which Are Toxic for Olsson, F., Ramstro Reproduction and Endocrine-disrupting e Proposals for a Phase-out in Sweden,Report from a Government Assignment. Report 4/15, Arkitektkopia, p. 160. Stockholm. Halima, N.B., 2016. Poly (vinyl alcohol): review of its promising applications and insights into biodegradation. RSC Adv. 6, 39823e39832. Huang, P., 2007. Study on Treatment of Thermophilic PVC Wastewater in Aerobic Moving Bed Bioﬁlm Reactor. Dissertation, Harbin Institute of Technology, Harbin (in Chinese). Ji, Q., Tabassum, S., Yu, G., Chu, C., Zhang, Z., 2015. A high efﬁciency biological system for treatment of coal gasiﬁcation wastewater - a key in-depth technological research. RSC Adv. 5, 40402e40413. Jia, X., Xin, G., Qian, Y., Qian, Y., 2016. Sectoral co-control of air pollutants: case of a chlor-alkali/polyvinyl chloride sector in China. J. Clean. Prod. 112, 1667e1675. Kadlec, R.H., Wallace, S., 2008. Treatment Wetlands, second ed. CRC Press. Khan, Z.U., Naz, I., Rehman, A., Raﬁq, M., Ali, N., Ahmed, S., 2015. Performance efﬁciency of an integrated stone media ﬁxed bioﬁlm reactor and sand ﬁlter for sewage treatment. Desalin. Water Treat 54, 2638e2647. Khatoon, N., Naz, I., Ali, M.I., Ali, N., Jamal, A., Hameed, A., Ahmed, S., 2014. Bacterial succession and degradative changes by bioﬁlm on plastic medium for wastewater treatment. J. Basic Microbiol. 54, 739e749. Kim, S., Kim, T.-H., Park, C., Shin, E.-B., 2003. Electrochemical oxidation of polyvinyl alcohol using a RuO2/Ti anode. Desalination 155, 49e57. Labunska, I., Brigden, K., Santillo, D., Johnston, P., Walters, A., 2008. Characterisation of Wastewater Discharges from Chloralkali Plants and Associated Chlorinated Chemical Production Facilities in Spain, Greenpeace Research Laboratories Technical Note (12/2008). University of Exeter, Exeter. Li, C., Tabassum, S., Zhang, Z., 2014. A novel environmental biotechnological aerobic process (BioAX) for the treatment of coal gasiﬁcation wastewater. RSC Adv. 4, 35156e35162. Li, M., Zhang, L., 2001. Studies on the treatment and reuse of PVC wastewater. Qilu Petrochem. Technol. 29, 211e214 (in Chinese). Li, Y., Tabassum, S., Chu, C., Zhang, Z., 2018. Inhibitory effect of high phenol concentration in treating coal gasiﬁcation wastewater in anaerobic bioﬁlter. J. Environ. Sci. 64, 207e215. 577 Li, Y., Tabassum, S., Zhang, Z., 2016. An advanced anaerobic bioﬁlter with efﬂuent recirculation for phenol removal and methane production in treatment of coal gasiﬁcation wastewater. J. Environ. Sci. 47, 23e33. Maddah, H.A., 2016. Polypropylene as a promising plastic: a review. Am. J. Polym. Sci. 6, 1e11. Madoni, P., 2011. Protozoa in wastewater treatment processes: a minireview. Ital. J. Zool. 78, 3e11. Mara, D., Horan, N.J., 2003. Handbook of Water and Wastewater Microbiology, ﬁrst ed. Academic Press, London (Elsevier Science). Matos, M., Alves, C., Campos, J.L., Brito, A.G., Nogueira, R., 2011. Sequencing batch bioﬁlm reactor: from support design to reactor operation. Environ. Technol. 32, 1121e1129. Naz, I., Batool, S.A.-U., Ali, N., Khatoon, N., Atiq, N., Hameed, A., Ahmed, S., 2013. Monitoring of growth and physiological activities of bioﬁlm during succession on polystyrene from activated sludge under aerobic and anaerobic conditions. Environ. Monit. Assess. 185, 6881e6892. Naz, I., Khatoon, N., Ali, M.I., Saroj, D.P., Batool, S.A.U., Ali, N., Ahmed, S., 2014. Appraisal of the tire derived rubber (TDR) medium for wastewater treatment under aerobic and anaerobic conditions. J. Chem. Technol. Biotechnol. 89, 587e596. Nicacio, G., Juen, L., 2015. Chironomids as indicators in freshwater ecosystems: an assessment of the literature. Insect Conserv. Diver 8, 393e403. Phugare, S.S., Kalyani, D.C., Surwase, S.N., Jadhav, J.P., 2011. Ecofriendly degradation, decolorization and detoxiﬁcation of textile efﬂuent by a developed bacterial consortium. Ecotoxicol. Environ. Saf. 74, 1288e1296. Punzi, M., 2015. Treatment of Textile Wastewater by Combining Biological Processes and Advanced Oxidation, Biotechnology. Lund University, Sweden, p. 160. Quartey, E.T., Tosefa, H., Danquah, K.A.B., Obrsalova, I., 2015. Theoretical framework for plastic waste management in Ghana through extended producer responsibility: case of sachet water waste. Int. J. Environ. Res. Publ. Health 12, 9907e9919. Qureshi, N., Annous, B.A., Ezeji, T.C., Karcher, P., Maddox, I.S., 2005. Bioﬁlm reactors for industrial bioconversion processes: employing potential of enhanced reaction rates. Microb. Cell Factories 4, 24. Sapp, J., 1994. Evolution by Association: a History of Symbiosis. Oxford University Press. Sehar, S., Naz, I., 2016. Role of the bioﬁlms in wastewater treatment. In: Dhanasekaran, D., Thajuddin, N. (Eds.), Microbial Bioﬁlms - Importance and Applications., InTech, pp. 121e144. Song, Y., Li, Y., 2009. Reutilization of mother liquid of PVC and biological treatment technology. China Chlor-Alkal 1, 36e39 (in Chinese). Tabassum, S., Wang, Y., Zhang, X., Zhang, Z., 2015a. Novel Mass Bio System (MBS) and its potential application in advanced treatment of coal gasiﬁcation wastewater. RSC Adv. 5, 88692e88702. Tabassum, S., Zhang, Y., Zhang, Z., 2015b. An integrated method for palm oil mill efﬂuent (POME) treatment for achieving zero liquid discharge e a pilot study. J. Clean. Prod. 95, 148e155. Wang, C., Wang, H., Liu, Y., Huang, L., 2016. Optimization of surface treatment for ﬂotation separation of polyvinyl chloride and polyethylene terephthalate waste plastics using response surface methodology. J. Clean. Prod. 139, 866e872. Wang, T., Ma, T., Qu, G., Liang, D., Hu, S., 2014. Performance evaluation of hybrid GaseLiquid pulse discharge Plasma-Induced degradation of polyvinyl alcoholcontaining wastewater. Plasma Chem. Plasma Process. 34, 1115e1127. Yang, Z., Xu, Z., Gao, W., Huang, S., Song, Y., 2005. Study on the treatment of polyvinyl chloride industrial wastewater by membrane bioreactor. Water Purify Technol 24, 21e23 (in Chinese). Ye, L., Qi, C., Hong, J., Ma, X., 2017. Life cycle assessment of polyvinyl chloride production and its recyclability in China. J. Clean. Prod. 142, 2965e2972. Yu, H., Gu, G., Song, L., 1996. Degradation of polyvinyl alcohol in sequencing batch reactors. Environ. Technol. 17, 1261e1267. Yu, Y., Feng, Y., Qiu, L., Han, W., Guan, L., 2008. Effect of grain-slag media for the treatment of wastewater in a biological aerated ﬁlter. Bioresour. Technol. 99, 4120e4123. Zhang, B., Zhou, Y., 2003. The sizing pollution and environmental protection. Cotton Text Technol 31, 17e20 (in Chinese). Zhang, S.J., Yu, H.Q., 2004. Radiation-induced degradation of polyvinyl alcohol in aqueous solutions. Water Res. 38, 309e316. Zhang, X., Lv, Y., Jin, Q., 2003. Testing on treatment of emulsiﬁed wastewater from PVC production by conventional coagulationeﬁltration method. Ind. Water Wastes 34, 47e49 (in Chinese). Zhang, Y., Chen, W., Lv, G., Lv, F., Chu, P.K., Guo, W., Cui, B., Zhang, R., Wang, H., 2012. Adsorption of polyvinyl alcohol from wastewater by sintered porous red mud. Water Sci. Technol. 65, 2055e2060. Zhang, Z., vol. 26 December 2002e9 November 2005. Aerobic Bioﬁllm Device for Wastewater Treatment, Pat., ZL 02157780.3, China. Zhao, D., Wei, Y., Wang, Y., 2011. Treatment and use of PVC mother liquid. China Chlor-Alkal 5, 11e13 (in Chinese).