Application of integrated ozone biological aerated filters and membrane filtration in water reuse of textile effluents

Bioresource Technology 133 (2013) 150–157

Contents lists available at SciVerse ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Case Study

Application of integrated ozone biological aerated filters and membrane filtration in water reuse of textile effluents Yaozhong He a, Xiaojun Wang a,⇑, Jinling Xu a, Jinli Yan a, Qilong Ge a, Xiaoyang Gu b, Lei Jian b a b

College of Environment And Energy, South China University of Technology, Guangzhou 510006, China HuaLu Environmental Technology Co., Ltd., Guangzhou, China

h i g h l i g h t s " Integrated ozone biological aerated filters and membrane filtration were first applied for water reuse of textile effluents. " Manganese ore grains were first used as ozonation catalyst in the reactor. " The reverse osmosis concentrates could be discharged directly without further disposal. " 100% polyvinyl alcohol removal rate was obtained. " Economic benefits could be obtained by producing high quality reclaimed water.

a r t i c l e

i n f o

Article history: Received 31 July 2012 Received in revised form 11 January 2013 Accepted 13 January 2013 Available online 31 January 2013 Keywords: Textile effluents Water reuse Ozone Biological aerated filter Membrane filtration

a b s t r a c t A combined process including integrated ozone-BAFs (ozone biological aerated filters) and membrane filtration was first applied for recycling textile effluents in a cotton textile mill with capacity of 5000 m3/d. Influent COD (chemical oxygen demand) in the range of 82–120 mg/L, BOD5 (5-day biochemical oxygen demand) of 12.6–23.1 mg/L, suspended solids (SSs) of 38–52 mg/L and color of 32–64° were observed during operation. Outflows with COD 6 45 mg/L, BOD5 6 7.6 mg/L, SS 6 15 mg/L, color 6 8° were obtained after being decontaminated by ozone-BAF with ozone dosage of 20–25 mg/L. Besides, the average removal rates of PVA (polyvinyl alcohol) and UV254 were 100% and 73.4% respectively. Permeate water produced by RO (reverse osmosis) could be reused in dyeing and finishing processes, while the RO concentrates could be discharged directly under local regulations with COD 6 100 mg/L, BOD5 6 21 mg/L, SS 6 52 mg/L, color 6 32°. Results showed that the combined process could guarantee water reuse with high quality, and solve the problem of RO concentrate disposal. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction 1.1. Background Textile industries can cause serious environmental problems due to the high water consumption, as well as a wide variety of biorefractory dyes and recalcitrant chemicals used in dyeing and finishing processes (Babu et al., 2007). Some dye groups, mainly azo dyes which are extensively used, can be transformed into mutagenic and carcinogenic aromatic amines (Akceylan et al., 2009) in vivo through cleavage of the azo groups (AN@NA), so that

⇑ Corresponding author. Address: Room 301, College of Environment and Energy, South China University of Technology, Guangzhou Higher Education Mega Center, Guangzhou 510006, China. Tel.: +86 20 13802767806; fax: +86 20 85640936. E-mail addresses: [email protected], [email protected] (X. Wang). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.01.074

textile effluents are harmful to environment if the contaminants are not eliminated. In recent years, various kinds of advanced treatment processes for textile effluents were investigated, such as catalytic ozonation (Legube and Karpel Vel Leitner, 1999), Fenton/photo-Fenton oxidation (Pérez et al., 2002), electrochemical treatment (Chatzisymeon et al., 2006), photocatalytic oxidation involving UV/H2O2 (Garcia et al., 2007) and ultrasonic-assisted ozone oxidation treatment (Zhou et al., 2012). However, most of these technologies were conducted at laboratory scale and not suitable for large-scale applications. As one of the advanced treatment processes in common use, ozonation is known to oxidize organic molecules through a combination of molecular and radical reactions. The molecular reactions are selective and they proceed through electrophilic attack in most cases, while the radical reactions involving hydroxyl radicals are generally fast and non-selective (E0 = 2.80 V). Ozone radical reactions can be promoted at high pH, by addition of H2O2, UV

Y. He et al. / Bioresource Technology 133 (2013) 150–157

irradiation or catalysts (Glaze et al., 1987), and such treatment processes involving hydroxyl radicals are called Advanced Oxidation Processes (AOPs), which are widely used to promote the degradation rates of recalcitrant organic matter (Tizaoui and Grima, 2011). However, AOPs are commonly used to achieve partial oxidation of non-biodegradable organic matter with high molecular weight or recalcitrant organic matter because of the large consumptions of energy and chemicals for complete oxidation and mineralization (Bijan and Mohseni, 2005). Otherwise, a certain research found that AOPs were unable to convert certain organic matter into carbon dioxide at even high oxidant dosage (Assalin et al., 2004). Hence, AOPs are not economically and technically feasible techniques if used alone, but can help produce biodegradable intermediates which are then decontaminated biologically. On the other hand, it is quite important and necessary to reclaim textile effluents because of the worsening of water shortage all around the world. Among various of water reuse technologies, membrane filtration technologies such as UF (ultrafiltration) and RO have been proven to be doable alternatives for water reuse of biologically treated textile effluents (Liu et al., 2011), and they are becoming more and more commonly used for the reason that they are one of the recommended treatment methods for recycling textile effluents in the BAT (Best Available Techniques) reference document (Capar et al., 2008). Among these membrane technologies, RO is widely applied to produce purified water for reuse. Operating with water recovery rates from 35% to 85%, RO plants generate great amount of concentrates which are commonly discharged into water bodies nearby, and pose potential threat to aquatic environment (Pérez-González et al., 2010). To deal with the problem of concentrate disposal, several methods such as evaporation ponds, deep well injection, electrodialysis, or even zero liquid discharge system have been developed, but they are either too restricted in application or too expensive in operation (Zhou et al., 2011; Greenlee et al., 2009). This work focused on the advanced treatment and water reclaimation of the chemical–biologically treated effluents from a cotton textile mill (Esquel Textile Co., Ltd.) in Guangdong Province, China. The project was carried out basing on the pilot study conducted by Qi et al. (2011). In this case, the two-stage integrated ozone-BAFs were first applied as pretreatment of membrane filtration in actual project for the purpose of providing high quality feed stream for membrane filtration units and solving the problem of RO concentrate disposal as well. 1.2. Ozone-BAF It is commonly recognized that ozonation is not an economically feasible technique if used alone, but can help produce biodegradable intermediates which are then treated biologically. In the past few years, various of combined processes including ozonation and biological treatment were investigated for the purpose of providing feasible alternatives to remove biorefractory compounds and recalcitrant organic matter from wastewaters effectively and economically (Chu et al., 2012; Battimelli et al., 2010; Sangave et al., 2007; Bijan and Mohseni, 2005), but most of them were not carried out in engineering applications. Moreover, all of these combined processes involved reactions taking place in separated devices, so that the capital cost would be expensive if implemented in industry application. The newly developed integrated ozone-BAF which has been registered at State Intellectual Property Office of PR China (patent number: ZL200710028632.9) was a successful attempt carried out by Xiaojun Wang’s research group, for the intention of cutting down the capital cost (about 20–30% off than separated devices) of a petrochemical wastewater advanced treatment project in Guangdong, China. However, ozone is well known as a bactericide, and it

151

is impossible to add ozone into bioreactor simply. During the development period, a certain kind of heterogeneous catalytic hardcore bed was considered to be essential to make full use of ozone and enable the growth of microorganisms in a single reactor. The ideal catalyst grains should meet several requirements described as follows: Firstly, they must be long lasting and effective, so that most of the ozone could react with organic matter in water and then decompose into oxygen without significant harm to the microbes living in the upper area of BAF. Secondly, they should be inexpensive and easy to obtain, or else the cost will increase sharply. Moreover, they should be strong enough to support the weight of packing layer above, and should be heavy enough to prevent backmixing during the backwash operation. Based on these assumptions, several kinds of mineral grains such as hematite grains, rutile grains, manganese ore grains, granite grains and dolomite grains were tested for potential application. Finally, the manganese ore grains turned out to be a suitable kind of catalytic material, which was capable of integrating ozonation and biodegradation in a single reactor. The ozone-BAF can be divided into catalytic ozonation layer and biodegradation layer from bottom to top as shown in Fig. 1, with a volume ratio between 1:5 and 1:8. The two layers are filled with manganese ore grains and ceramic grains respectively. Manganese ore is a kind of mineral rich in manganese oxide, which has been recognized as a high performance ozonation catalyst for organic matter such as aromatic compounds and alkanes by transforming ozone to active oxygen species that can oxidize these compounds (Einaga and Ogata, 2009) or by the generation of highly oxidative intermediate species (Legube and Karpel Vel Leitner, 1999). With the huge activated surface provided by catalytic grains, the organic pollutants such as aromatic compounds can be partially oxidized into biodegradable organic matter including phthalic acid, maleic acid or oxalic acid by controlling the ozone dosage. The formation of those intermediate products enables easier biodegradation in the upper area of the reactor. In some cases, the organic pollutants could also be destroyed due to the hydroxyl radicals generated by ozone decomposition in aqueous solution at rather high pH (Baig and Liechti, 2001). It is worth mentioning that coagulation sedimentation or sand filtration is required to keep the influent low of SS. It is necessary to aerate continuously to inhibit the obstruction of catalytic grains and to remove partially oxidized products on the catalyst surface. Besides, aeration is also essential to sufficient biodegradation of organic pollutants. The ozone-BAF can guarantee the growth and reproduction of microorganisms without noticeable toxicity on biomass owing to the fully ozone consumption of catalytic ozonation layer. Hence, the reactor is free of ozone destroyer, so as to lower down the equipment cost. Additionally, it has been proven that integration of ozone oxidation and biological treatment in a single device could allow better performance than that of coupled process in some cases (Di Iaconi, 2012; Qi et al., 2011). So far, the ozone-BAF has been successfully applied in several projects of wastewater treatment in China, such as liquid chemical wharf wastewater, petrochemical wastewater and tannery wastewater advanced treatment, showing its flexibility and enhanced degradation ability of recalcitrant and/or toxic organic pollutants (influent COD varies in the range of 100–160 mg/L).

2. Methods 2.1. Design influent and effluent The feed streams for the water recycling plant were the secondary effluents treated by coagulation sedimentation, anaerobic–aerobic activated sludge treatment, secondary sedimentation and sand leach in sequence. Design permeate water production

152

Y. He et al. / Bioresource Technology 133 (2013) 150–157

Fig. 1. Sketch of integrated ozone-BAF: ozone can be sucked into the influent by a venturi tube. After that, the influent goes through catalytic ozonation layer and biodegradation layer under continuous aeration, and flows out through screen mesh.

Table 1 Influent and effluent quality designed. Index

Influent

O3-BAF effluent

Permeate water

COD BOD5 SS Color pH

6110 mg/L 630 mg/L 670 mg/L 664° 7–8

640 mg/L 610 mg/L 620 mg/L 64° 7–8

610 mg/L 61 mg/L 61 mg/L 0 7–8

The influent was treated by ozone-BAF and membrane filtration (UF & RO) in sequence.

employed. In the membrane filtration plant, water produced by SF (sand filtration), UF (ultrafiltration) and RO units was first stored in specified reservoir for the next step, respectively. Besides, security filters, scale inhibitor and disinfectant were also adopted to protect the UF and RO membrane units. Since the influent of membrane filtration was rather clean, the concentrates from RO units could be discharged under local regulations, instead of additional disposal. 2.3. Main process parameters

3

capacity was 3000 m /d in this case, with feed stream rate of 5000 m3/d (1/3 of the total effluents discharged by the textile mill). The combined process consisted of two steps shown as follow. Design water quality indexes of influent and effluent for the water reclaimation plant are listed in Table 1. 2.2. Technological process Based on the results of pilot study (Qi et al., 2011), the combined process was consisted of ozone-BAF and membrane filtration in sequence, as demonstrated in Fig. 2. Ozone-BAF was used as pretreatment of membrane filtration for the purpose of alleviating the load of membrane units to allow a longer time before membranes need to be washed, and for the sake of allowing direct RO concentrate discharge instead of highly expensive post-treatment. The pretreatment units composed of eight primary ozone-BAFs with individual size of 5 m  5 m  5.5 m (L  W  H) and five secondary ozone-BAFs with individual size of 5 m  5 m  4.5 m (L  W  H). After being treated by the two-stage ozone-BAFs, the effluents gathered into clean–water reservoir for membrane filtration and backwash water supply. The backwash concentrates from ozone-BAFs were recycled into influent buffer reservoir after sand filtration. Pretreated water could not be reused directly due to its high salinity, hence membrane units including UF and RO were

The ozone-BAF units ran at an influent flow rate of about 208 m3/h continuously. There were eight primary ozone-BAFs and five secondary ozone-BAFs working in sequence. The total volume of catalytic grains and ceramic pellets was 90 m3 for each primary ozone-BAF, and 75 m3 for each secondary ozone-BAF. The catalytic grains used were 20–30 mm in size, with manganese oxide content of about 30%, while the ceramic pellets used were 3–5 mm in size, with porosity of 40% approximately. For a single primary ozone-BAF, the effective volume, HRT (hydraulic retention time) and air to water flow ratio were about 87.5 m3, 3.3 h and 6 respectively. For a single secondary ozone-BAF, the effective volume, HRT and air to water flow ratio were about 75 m3, 1.8 h and 3, respectively. The two ozone generators working with liquid oxygen were customized products (GUOLIN Co. Ltd., Qingdao, China), which worked alternately. The maximum ozone productivity for a single generator was 6 kg/h with electric power of 60 kW. Besides, all the pumps and blowers were standard products. For the membrane filtration system, there were four sand filters, two UF units and two RO units working in sequence. The operating pressures of SF, UF and RO units were about 0.04 MPa, 0.06 MPa and 0.95 MPa, respectively. Both of the UF and RO units worked alternately. A total of 6160 m2 PVDF (polyvinylidene fluoride) membranes with permeation flux of 35–120 L/m2 h and 10,500 m2 polyamide membranes with permeation flux of 10– 15 L/m2 h were used in the UF and RO units, respectively. For the

Y. He et al. / Bioresource Technology 133 (2013) 150–157

153

Fig. 2. Flowchart of treatment process. (1) Influent buffer reservoir; (2) primary ozone-BAFs; (3) secondary ozone-BAFs; (4) clean water reservoir; (5) back-wash concentrate reservoir; (6) Back-wash concentrate sand filter; (7) sand filters; (8) sand filtration reservoir; (9) UF units; (10) UF filtrate reservoir; (11) RO units; (12) RO filtrate reservoir; and (13) security filters.

RO units, the ratio between permeate flow and concentrate flow was 3:2. Besides, alkaline detergents (30% caustic soda and 10% sodium hypochlorite solution) and acid detergents (20% citric acid and 30% hydrochloric acid) were used to wash the membrane units. 2.4. Analytical methods Textile effluents are very complicated due to the addition of chemicals including various of dyes, carriers, biocides, bleaching agents, complexion agents, ionic and non-ionic surfactants, sizing agents, and so on (Selcuk, 2005). Consequently, it is hard to explain the overall degradation of the organic matter in textile effluents individually. Thus, several parameters including COD, BOD5, SS, color, UV254 and concentrations of PVA (a sizing agent extensively used in the textile mill) are employed to monitor the degradation of organic matter by ozone-BAF treatment. During this study, water samples were taken weekly for analysis: COD was measured by the potassium dichromate method (Neelavannan et al., 2007), BOD5 by a BOD5 measurement equipment (CY-II, Taihong medical equipment Co., Ltd., China), color by dilution multiple method (values were expressed as x°, when 0° indicated the sample was as clear as pure water), SS by weighing method with a precise electronic balance (XB224, CANY Precision Instruments Co., Ltd., China), pH by a pH meter (PHSJ-4A, INESA Scientific Instrument Co., Ltd., China). UV254 was detected by an UV–vis spectrophotometer (UV-765, CANY Precision Instruments Co., Ltd., China). The PVA measurement was performed at a kmax of 690 nm using the same UV–vis spectrophotometer, basing on the blue color produced by reaction of PVA with iodine in the presence of boric acid (Finley, 1961). Besides, all the reagents used were analytical grade and purchased from Tianjin Kemiou Chemical Reagent Co., Ltd., China. 3. Results and discussion 3.1. Optimal ozone dosage In order to bring the whole system into work, the ozone dosage was increased weekly to determine the optimal dosage according to the 7-day average removal rates of COD, color and SS as shown in Fig. 3. It can be seen that the removal rates of COD, SS and color increase with the ozone dosage, especially for the color removal rates. Generally, the decolorization by molecular ozone occurs rapidly due to the rapid destruction of the conjugated chains of the dye molecules which are responsible for color (Tizaoui and Grima, 2011). Meanwhile, the conjugated chromophores of dyes can be broken into molecules with relatively simple structures, so that higher biodegradability of textile effluents is available. Under fluctuant operating conditions that influent COD varied between

Fig. 3. Effects of different O3 dosage on COD, color and SS removal rates: during the optimization procedure, ozone productivity increased from 0 kg/h to 6 kg/h step by step, and the influent rate was 5000 m3/d for the ozone-BAF system.

83 mg/L and 110 mg/L, the optimal ozone dosage taking account of running cost and treatment efficiency was found between 20 mg/L and 25 mg/L. According to estimation, the total ozone consumption was about 0.4 kgO3/kgCODremoved most of the time, and the ratio of ozone dosage between primary ozone-BAFs and secondary ozone-BAFs was kept constant at 3:1 by controlling the valves. It has been proven that ozone is the key factor determining the function of ozone-BAF. During the running period, it was found that both of the effluent COD and color took a turn for the worse after forced outage of ozone generators by accidents. Without the addition of ozone, most of the biorefractory dyes and chemicals could not be transformed into organic compounds that were capable of biodegradation. However, the purpose of domestication was to achieve survival of the fittest microbes that could adapt to the trace amount of residual ozone in water and to strengthen the biodegradation ability of ozone-BAF by promoting the growth of microbes. But so far, the microbial population and its dispersed characteristics in ozone-BAF still remain unknown. In order to fill this blank, an experimental study is in progress to investigate the relationship between microbial community succession and residual ozone concentration in water along the ozone-BAF. 3.2. Running results and discussion Apart from the 1-month trial running period, the whole system has come into service for 3 months. Monitoring data including

154

Y. He et al. / Bioresource Technology 133 (2013) 150–157

color, COD, BOD5, SS, UV254 and PVA concentrations are shown in Fig. 4a–e below. The pH values are absent for the reason that the pH of all water samples maintained in the range of 7–8 without significant fluctuation. Fig. 4a shows that the ozone-BAF pretreatment is competent for decolorization with color removal rates of 93.75% for most of the time (color 6 4°). It is reasonable to suppose that, in the catalytic ozonation layer, the electrophilic attack by ozone tends to occur on sites possessing high negative charge density, including multiply bonded species such as AC@CA or AN@NA and atoms such

as N, P, O, and S (Tizaoui and Grima, 2011). The first stage of decolorization is supposed to occur by ozone molecular reaction with the double bond azo chromophoric groups AN@NA or with the double bond AC@CA connecting aromatic rings that are responsible for color (Lopez et al., 2004), resulting in rapid decolorization of wastewater. However, the more organic matter the influents contain, the worse decolorization rate is obtained. When the COD of influents were over 110 mg/L on May 23 and June 20 respectively, the color of effluents increased to 8°, color removal rates dropped to 87.5%. With constant ozone dosage, higher organic matter

Fig. 4. Removal of contaminants by two-stage ozone-BAFs: (a) color, (b) COD, (c) BOD5, (d) PVA and UV254, and (e) SS: the two-stage ozone-BAFs ran at an influent flow rate of 5000 m3/d with ozone dosage of 20–25 mg/L.

Y. He et al. / Bioresource Technology 133 (2013) 150–157

content tends to impair the decolorization of residual dyes and other chromophore groups by ozone competition, so that the color removal rate declines. Ozone-BAF is effective in removing COD and BOD5 from textile effluents as presented in Fig. 4b and c. During the monitoring period, the influent COD and BOD5 distributed in the range of 82– 105 mg/L and 12.5–18.4 mg/L respectively, with an average BOD5/COD ratio of about 0.16 which indicated the difficulty in degrading the organic pollutants with conventional biological methods (a BOD5/COD ratio higher than 0.3 is preferred). Nevertheless, effluents with COD of 30–40 mg/L and BOD5 of 4.7– 6.5 mg/L were obtained, and the average removal rates of COD and BOD5 were 63.7% and 62.0% respectively. When the influent COD increased beyond 110 mg/L, such as 120 mg/L on May 23 and 117 mg/L on June 20, the effluent COD increased to 45 mg/L and 44 mg/L respectively. Besides, effluents with BOD5 less than 8 mg/L were obtained when influent BOD5 came to 23.1 mg/L and 19.5 mg/L respectively. The BOD5/COD ratio rose to around 0.26 after primary ozone-BAF treatment, with average COD and BOD5 of 45.2 mg/L and 11.9 mg/L respectively. Both of the ozonation and biodegradation were supposed to be responsible for pollutant removal. As the mass ratios between ozone dosage and influent COD were in the range of 0.17–0.25 in this case, ozone could only be capable of partial oxidation other than complete oxidation of the soluble organic matter at such low dosage, resulting in the formation of more biodegradable molecules but no significant decrease of COD (Selcuk, 2005). Hence, organic pollutants especially the recalcitrant chemicals were supposed to be partially oxidized by ozone first, then transform into organic molecules with simple structures that could be adsorbed and degraded by microbes living in the upper biodegradation layer. Based on the fact that the concentrations of biorefractory aromatic compounds correlated well with the values of UV254, and the fact that PVA was extensively used as sizing agent in this mill which was responsible for COD as well, UV254 and PVA concentrations were analyzed for the purpose of confirming the assumption above. Fig. 4d reveals that ozone-BAF was capable of degrading aromatic substances and PVA with remarkable efficiencies. The removal rates of UV254 distributed in the range of 72% and 80% under design running conditions, but dropped to 66% and 70% when the influent UV254 increased to 1.72 and 1.75 on May 23 and June 20, respectively. Since the UV254 removal rates were rather stable with constant ozone dosage, most of the aromatic compounds were considered to be degraded rather than being absorbed by microorganisms. Besides, 100% PVA removal rates were obtained when influent PVA concentrations distributed in the range of 21.5– 25.8 mg/L. As a kind of superpolymer, the BOD5/COD ratio of PVA is merely about 0.064 (Lin and Lo, 1997), so that it is hard to be degraded by traditional biological treatment processes. However, it has been reported that PVA could be removed by ozone oxidation, and the removal of PVA by ozonation was not due to complete mineralization but to the breakdown of macromolecules to micromolecules (Shin et al., 1999), which could be further biodegraded. Accordingly, the PVA was supposed to be partially oxidized into biodegradable micromolecules in the catalytic ozonation layer first, then to be adsorbed and biodegraded by the microbes in the upper biodegradation layer. Fig. 4e indicates that the effluents produced by the two-stage ozone-BAF process were rather limpid with low content of suspended particles (SS 6 15 mg/L). As a kind of biological aerated filters, ozone-BAF provides enormous surface area for microbe inhabitance and enables strengthened filtration of the suspended particles by filling with granular media. Furthermore, it has been reported that some microorganisms could produce extracellular exopolymers during organism growth, which could be used as biological flocculants to enhance the performance of water treatment

155

(Cloete and Oosthuizen, 2001). Since the biofilms attaching to the grains were composed of microorganisms, it was reasonable to believe that the secretions produced during their metabolism acted as biological flocculants that could adsorb and flocculate the suspended particles to achieve high SS removal and improve biological oxidation (Qiu et al., 2007). However, to keep the high performance of ozone-BAFs, they required to be backwashed weekly to rush away the accumulated suspended solids and the excess biomass produced. The sludge containing flow with SS varying in the range of 350–630 mg/L was then recycled into the influent buffer reservoir after sand filtration. The sand filter also required to be backwashed every 2 days, and the sewage produced was discharged into the initial buffer tank of the mill to circulate along the whole treatment process. In addition to providing stable feed streams for membrane filtration process with high quality, the ozone-BAFs could also allow direct emission of RO concentrates in accordance with local discharge limits of water pollutants (standard limits: COD 6 100 mg/L, BOD5 6 30 mg/L, SS 6 70 mg/L, color 6 50°). It can be seen from Fig. 5 that, under design running conditions (seen in Table 1), the COD, BOD5, SS and color of the concentrates emitted are 80–96 mg/L, 13.8–17.5 mg/L, 40–51 mg/L and 8–16°, respectively. Nevertheless, the qualities of concentrates were rather poor on May 23 and June 20 due to the deterioration of influents with COD of 108 mg/L and 112 mg/L, BOD5 of 20.3 mg/L and 21 mg/L, respectively. Based on operating experience, COD was the control parameter for the emission of RO concentrates, so a COD online monitor (HACH CODmax plus sc) was employed to control the flow direction of concentrates automatically. The concentrates could be discharged directly most of the time, but also could be refluxed into the influent buffer tank while COD P 100 mg/L without noticeable influence on the system. In general, as long as the influents met the design requirements, it was unnecessary to deal with the problem of RO concentrate disposal under local regulations, allowing more environmental friendly use of membrane filtration technologies. However, even though the impact of salt emission of RO concentrates on water bodies could not be neglected, the salinity problem was not concerned about in this case because no economically feasible desalinization technology for RO concentrates was available all over the world (Greenlee et al., 2009). For the safety of ecology environment, the Chinese government advocates enterprises to carry out cleaner production to relieve salinity pollution, instead of prohibiting the salt emissions with strict discharge standards. After being purified by RO procedure, permeate flows with pH = 6.5–7.0, COD < 5 mg/L, BOD5 < 1 mg/L, SS < 1 mg/L, color = 0°,

Fig. 5. Characteristics of RO concentrates: The permeate flow to concentrate flow ratio was 3:2, with concentrate flow rate of 2000 m3/d.

156

Y. He et al. / Bioresource Technology 133 (2013) 150–157

electrical conductivity = 50–100 lS/cm, total hardness < 2 mg/L and alkalinity = 20–40 mg/L (provided by research and development department of Esquel Textile) were obtained during the 3month running period, which allowed water reuse in dyeing and finishing processes with excellent quality. Nevertheless, to maintain the high performance of membrane filtration plant, the sand filters, UF and RO units required to be washed regularly, and their optimal cleaning frequencies were 10 days, 40 min and 3 months, respectively. For each washing operation, 100 m3 sludge containing concentrates were discharged from each sand filter, 18 m3 from an UF unit, and 105 m3 from a RO unit, respectively. The average SS values of these concentrates distributed in the range of 1210– 1380 mg/L, 175–220 mg/L and 23–42 mg/L respectively (water samples were taken at every 3 min during each washing operation). In general, the sludge (dry weight) produced by washing operation was estimated to be less than 13 kg/d for a sand filter, less than 75 kg/d for a UF unit, and less than 0.04 kg/d for a RO unit, respectively. For disposal, the sludge containing concentrates were recycled into the initial influent buffer tank of the mill, so that most of the sludge can be separated by coagulation sedimentation, secondary sedimentation and sand leach operation from wastewater. The separated sludge is then transported to a brick factory for brick making after dewatered by plate-and-frame filter presses.

and maintenance of other materials were not included. Based on the present price of high quality industrial water (4 Yuan/m3), it was estimated that each cubic meter of permeate water could save 1.65 Yuan for the enterprise, hence the payback period of the total investment should be about 8.6 years with permeate water capacity of 3000 m3/d.

3.3. Economical analysis

This work was supported by the National Natural Science Foundation (NSFC) of China, Grant No. 5107 8149, and the Scientific Management of Technological Planning Projects of Guangzhou, China, Grant No. 2009Z1-E751. We also appreciated the staffs of Esquel Textile for their feedbacks and assistance to us, especially for some valuable monitoring data.

To deal with the problem of sludge caused by water treatment, there was a set of sludge dewatering facility in the mill, with original dewatered sludge production of 85–90 t/d. After being dewatered by plate-and-frame filter presses, the sludge moisture content decreased from 98–99.4% to 70–72% (provided by research and development department of Esquel Textile), the dewatered sludge was then transported to a brickyard for brick making. The average sludge production was estimated about 0.21 kgSS/ kgCODremoved for primary ozone-BAFs and about 0.25 kgSS/ kgCODremoved for secondary ozone-BAFs, less than the values of 0.4–0.6 kgSS/kgCODremoved that normally reported for conventional activated sludge process (Lotito et al., 2012). Besides, the sludge production by membrane filtration plant was less than 200 kg/d (dry weight) according to the estimation above. As the total dewatered sludge production maintained at around 85–90 t/d without significant growth, the running cost of newly increased sludge disposal caused by ozone-BAF backwash and membrane washing operations was not concerned about in this case. The total investment of this project was about RMB 15.49 million Yuan. Table 2 lists the main items in detail. The general engineering contained underground drainage pipeline, power station, electrical control room, compressor room and other facilities. Most of the devices were automatically controlled to allow convenient and efficient management, and make the whole process free of extra personnel expense. The running cost including electricity consumption (0.75 Yuan/kW), liquid oxygen cost and deterioration of the UF and RO membranes were 0.74 Yuan/m3 for ozone-BAF and 1.61 Yuan/m3 for membrane process, respectively. The production cost of permeate water was 2.35 Yuan/m3, while the depreciation

Table 2 Investment and running cost. Item

Ozone-BAF

Membrane

Foundation construction Equipment General engineering Running cost

2.85 Million Yuan 3.06 Million Yuan 1.58 Million Yuan 0.74 Yuan/m3

1.62 Million Yuan 6.38 Million Yuan 1.61 Yuan/m3

The general engineering included underground drainage pipeline, power station, electrical control room, compressor room and other facilities.

4. Conclusions Integrated ozone-BAFs were first applied as pretreatment of membrane filtration. Design influents were characterized by COD 6 110 mg/L, BOD5 6 30 mg/L, SS 6 70 mg/L and color 6 64°, with flow rate of 5000 m3/d. Effluents with COD 6 40 mg/L, BOD5 6 6.5 mg/L, SS 6 15 mg/L, color 6 4° were obtained by ozone-BAFs with ozone dosage of 20–25 mg/L. Owing to the pretreatment, RO concentrates could be discharged directly in accordance with local discharge limits. Results showed that the combined process including ozone-BAFs and membrane filtration was a promising solution for RO concentrate disposal which derived from water reuse in textile industry. Acknowledgements

References Akceylan, E., Bahadir, M., Yilmaz, M., 2009. Removal efficiency of a calix[4]arenebased polymer for water–soluble carcinogenic direct azo dyes and aromatic amines. J. Hazard. Mater. 162, 960–966. Assalin, M.R., Rosa, M.A., Duran, N., 2004. Remediation of kraft effluent by ozonation: effect of applied ozone concentration and initial pH. Ozone-Sci. Eng. 26, 317–322. Battimelli, A., Loisel, D., Garcia-Bernet, D., Carrere, H., Delgenes, J.P., 2010. Combined ozone pretreatment and biological processes for removal of colored and biorefractory compounds in wastewater from molasses fermentation industries. J. Chem. Technol. Biotechnol. 85, 968–975. Babu, B.R., Parande, A.K., Raghu, S., Kumar, T.P., 2007. Cotton textile processing: waste generation and effluent treatment. Cotton Sci. 11, 141–153. Bijan, L., Mohseni, M., 2005. Integrated ozone and biotreatment of pulp mill effluent and changes in biodegrability and molecular weight distribution of organic compounds. Water Res. 39, 3763–3772. Baig, S., Liechti, P.A., 2001. Ozone treatment for biorefractory COD removal. Water Sci. Technol. 43, 197–204. Chu, W.H., Gao, N.Y., Yin, D.Q., Deng, Y., Templeton, M.R., 2012. Ozone–biological activated carbon integrated treatment for removal of precursors of halogenated nitrogenous disinfection by-products. Chemosphere 86, 1087–1091. Capar, G., Yilmaz, L., Yetis, U., 2008. A membrane-baesd co-treatment strategy for the recovery of print- and beck-dyeing textile effluents. J. Hazard. Mater. 152, 316–323. Chatzisymeon, E., Xekoukoulotakis, N.P., Coz, A., Kalogerakis, N., Mantzavinos, D., 2006. Electrochemical treatment of textile dyes and dyehouse effluents. J. Hazard. Mater. 137, 998–1007. Cloete, T.E., Oosthuizen, D.J., 2001. The role of extracellular exopolymers in the removal of phosphorus from activates sludge. Water Res. 35, 3595–3598. Di Iaconi, C., 2012. Biological treatment and ozone oxidation: integration or coupling. Bioresour. Technol. 106, 63–68. Einaga, H., Ogata, A., 2009. Benzene oxidation with ozone over supported manganese oxide catalysts: effect of catalyst support and reaction conditions. J. Hazar. Mater. 164, 1236–1241. Finley, J.H., 1961. Spectrophotometric determination of polyvinyl alcohol in paper coatings. Anal. Chem. 33, 1925–1927. Greenlee, L.F., Lawler, D.F., Freeman, B.D., Marrot, B., 2009. Reverse osmosis desalination: water sources, technology, and today’s challenges. Water Res. 43, 2317–2348. Garcia, J.C., Oliveira, J.L., Silva, A.E.C., Oliveira, C.C., Nozaki, J., Souza, N.E., 2007. Comparative study of the degradation of real textile effluents by photocatalytic reactions involving UV/TiO2/H2O2 and UV/Fe2+/H2O2 systems. J. Hazar. Mater. 147, 105–110.

Y. He et al. / Bioresource Technology 133 (2013) 150–157 Glaze, W.H., Kang, J.W., Chapin, D.H., 1987. The chemistry of water-treatment processes involving ozone, hydrogen-peroxide and ultraviolet-radiation. Ozone-Sci. Eng. 9, 335–352. Lotito, A.M., Fratino, U., Bergna, G., Iaconi, C.D., 2012. Integrated biological and ozone treatment of printing textile wastewater. Chem. Eng. J. 195–196, 261– 269. Liu, M.H., Lü, Z.H., Chen, Z.H., Yu, S.C., Gao, C.J., 2011. Comparison of reverse osmosis and nanofiltration membranes in the treatment of biologically treated textile effluent for water reuse. Desalination 281, 372–378. Lopez, A., Benbelkacem, H., Pic, J.S., Debellefontaine, H., 2004. Oxidation pathways for ozonation of azo dyes in a semi-batch reactor: a kinetic parameters approach. Environ. Technol. 25, 311–321. Legube, B., Karpel Vel Leitner, N., 1999. Catalytic ozonation: a promising advanced oxidation technology for water treatment. Catal. Today 53, 61–72. Lin, S.H., Lo, C.C., 1997. Fenton process for treatment of desizing wastewater. Water Res. 31, 2050–2056. Neelavannan, M.G., Revathi, M., Basha, C.A., 2007. Photocatalytic and electrochemical combined treatment of textile wash water. J. Hazard. Mater. 149, 371–378. Pérez-González, A., Urtiaga, A.M., Ibáñez, R., Ortiz, I., 2010. State of the art and review on the treatment technologies of water reverse osmosis concentrates. Water Res. 46, 267–283. Pérez, M., Torrades, F., Domènech, X., Peral, J., 2002. Fenton and photo-Fenton oxidation of textile effluents. Water Res. 36, 2703–2710.

157

Qi, L.Q., Wang, X.J., Xu, Q.K., 2011. Coupling of biological methods with membrane filtration using ozone as pre-treatment for water reuse. Desalination 270, 264– 268. Qiu, L.P., Ma, J., Zhang, L.X., 2007. Characteristics and utilization of biologically aerated filter backwashed sludge. Desalination 208, 73–80. Sangave, P.C., Gogate, P.R., Pandit, A.B., 2007. Combination of ozonation with conventional aerobic oxidation for distillery wastewater treatment. Chemosphere 68, 32–41. Selcuk, H., 2005. Decolorization and detoxification of textile wastewater by ozonation and coagulation processes. Dyes Pigment 64, 217–222. Shin, H.S., Yoo, K.S., Kwon, J.C., Lee, C.Y., 1999. Degradation mechanism of PVA and HEC by ozonation. Environ. Technol. 20, 325–330. Tizaoui, C., Grima, N., 2011. Kinetics of the ozone oxidation of Reactive Orange 16 azo-hye in aqueous solution. Chem. Eng. J. 173, 463–473. Zhou, X.J., Guo, W.Q., Yang, S.S., Ren, N.Q., 2012. A rapid and low energy consumption method to decolorize the high concentration triphenylmethane dye wastewater: operational parameters optimization for the ultrasonicassisted ozone oxidation process. Bioresour. Technol. 105, 40–47. Zhou, T., Lim, T.T., Chin, S.S., Fane, A.G., 2011. Treatment of organics in reverse osmosis concentrate from a municipal wastewater reclaimation plant: feasibility test of advanced oxidation processes with/without pretreatment. Chem. Eng. J. 166, 932–939.