Bioresource Technology 192 (2015) 177–184
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
A new flat sheet membrane bioreactor hybrid system for advanced treatment of effluent, reverse osmosis pretreatment and fouling mitigation Majid Hosseinzadeh a,⇑, Gholamreza Nabi Bidhendi a, Ali Torabian a, Naser Mehrdadi a, Mehdi Pourabdullah b a b
Water and Wastewater Engineering Group, Graduate Faculty of Environment, University of Tehran, Tehran, Iran Environmental and Water Research Center, School of Civil Engineering, Sharif University of Technology, Tehran, Iran
h i g h l i g h t s Pilot-scale hybrid and conventional membrane bioreactors was compared. Optimized electrocoagulation was used in hybrid system. Membrane fouling and sludge characteristics improved significantly in hybrid MBR. Notably better permeate quality was achieved in hybrid system. Both systems had excellent performance as reverse osmosis pretreatment.
a r t i c l e
i n f o
Article history: Received 26 March 2015 Received in revised form 16 May 2015 Accepted 19 May 2015 Available online 22 May 2015 Keywords: Electrical coagulation Membrane bioreactor Fouling Reverse osmosis
a b s t r a c t This paper introduces a new hybrid electro membrane bioreactor (HEMBR) for reverse osmosis (RO) pretreatment and advanced treatment of effluent by simultaneously integrating electrical coagulation (EC) with a membrane bioreactor (MBR) and its performance was compared with conventional MBR. Experimental results and their statistical analysis showed removal efficiency for suspended solids (SS) of almost 100% for both reactors. HEMBR removal of chemical oxygen demand (COD) improved by 4% and membrane fouling was alleviated according to transmembrane pressure (TMP). The average silt density index (SDI) of HEMBR permeate samples was slightly better indicating less RO membrane fouling. Moreover, based on the SVI comparison of two reactor biomass samples, HEMBR showed better settling characteristics which improved the dewaterability and filterability of the sludge. Analysis the change of membrane surfaces and the cake layer formed over them through field emission scanning electron microscopy (FESEM) and X-ray fluorescence spectrometer (XRF) were also discussed. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Shortage of water resources is one of the critical issues facing the world today, especially in the Middle East countries. In these countries, a significant part of the water is converted to wastewater. Using reverse osmosis (RO), water reclamation and purification of treated wastewater becomes widespread enabling production of high quality water for various purposes such as non-potable applications and industrial usage (Hosseinzadeh et al., 2013; Macedonio et al., 2012). RO systems usually use membrane technologies as pretreatment to provide appropriate quality for feed water to ⇑ Corresponding author. Tel.: +98 2161113155; fax: +98 2166407719. E-mail address: [email protected] (M. Hosseinzadeh). http://dx.doi.org/10.1016/j.biortech.2015.05.066 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
enhance RO performance and prevent premature fouling. A membrane bioreactor (MBR) is a wastewater treatment process that combines an activated sludge process (ASP) with membrane separation thereby producing reclaimed water (Wei, 2012; Xiao et al., 2014). Membrane fouling is the major obstacle to increasing MBR development because of reduction in the membrane performance and the efficiency of treatment by forming thick gel layer and cake layer that formed on and into the membrane. Several factors are involved in causing membrane fouling such as membrane characteristics, feed and biomass characteristics, operating conditions, etc.(Delrue et al., 2011; Meng et al., 2009; Tijing et al., 2015; Wu and Huang, 2009; Zhang et al., 2014b). Many studies have been conducted on membrane fouling and several methods and strategies to reduce and control the problem have been investigated in
178
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
recent decade (Huang et al., 2011; Sun et al., 2011; Tijing et al., 2015; Wang et al., 2014). Some studies have shown that coagulation can significantly reduce colloidal and soluble organic materials in water and wastewater treatment. Literature on coagulation systems coupled with membrane bioreactors have reported satisfactory results on MBR performance. Addition of coagulants can change particle characteristics such as charge, size and shape and lead to reduction in organic components, which causes a reduction in organic and colloidal fouling in MBRs (Chen and Liu, 2012; Song et al., 2008; Thang Nguyen, 2012; Tian et al., 2008). Chen and Liu confirmed the possibility and applicability of coagulation-MBR to reclaim effluent in the dairy industry (Chen and Liu, 2012). They realized that poly aluminum chloride as a coagulant played an effective role in improving MBR permeate quality and membrane fouling reduction. While the addition of chemicals for coagulation may result in production of undesirable by-products or increase the sludge amount in the reactor, applying a direct current (DC) field on the activated sludge can be considered an effective method in MBR application. In this system, usually Fe and Al are applied as sacrificial anodes to supply coagulants in the reactor (Akyol, 2012; Maha Lakshmi and Sivashanmugam, 2013; Millar et al., 2014; S ß engil et al., 2009). However, applying a DC field directly within the activated sludge reactor may be harmful to microorganism activity (Bani-Melhem and Elektorowicz, 2011; Bani-Melhem and Smith, 2012). Melhem and Smith proposed application of an electrocoagulation unit as a pretreatment step before MBR in order to prevent direct contact of the microbial community with the applied DC field (Bani-Melhem and Smith, 2012). The results demonstrated that the EC-MBR may be effective not only for increasing the quality of the treated gray water over MBR, but also for improving the overall performance of the membrane filtration process. Although limited studies have been done on the use of current field for membrane fouling reduction (Bani-Melhem and Elektorowicz, 2011; Hasan et al., 2014; Zhang et al., 2014a), to the best of our knowledge, no reports have been published for the simultaneous application of electric coagulation along with MBR in a continuous pilot project for advanced treatment of industrial wastewater treatment plant effluent and RO pretreatment, which is the main objective of this work. This study first identified best conditions for electrical coagulation and then simultaneous integration of optimized electrocoagulation, biological, and membrane processes in one reactor i.e., hybrid electro membrane bioreactor (HEMBR) was developed and its performance compared with conventional MBR based on permeate quality, RO pretreatment capability and membrane fouling was assessed. To do this, two parallel continuous pilot projects were designed, built and operated at the same conditions.
2. Methods 2.1. Effluent characteristics The pilot project was set up in the Shokouhieh industrial wastewater treatment plant (WWTP), located in the province of Qom, Iran. There are many industries in this town such as, dairy, beverage, welding, petrochemical, and metal finishing and their wastewater after pretreatment through a sewage network is routed to the wastewater treatment plant. The average COD in the influent of WWTP is about 2000 mg L1 which reduce to about 250 mg L1 in effluent. The main process units in this plant are a screen, equalization tank, anaerobic basin, aerobic unit, sedimentation tank and chlorination units. Due to problems in design, implementation and operation of the plant, a certain amount of
biodegradable organic matters will remain in the effluent after treatment. In this study, bioreactors were set up for advanced treatment of effluent and they fed with the unchlorinated effluent of wastewater treatment. The characteristics of effluent are given in Table 1. 2.2. Experimental set-up and operating conditions The study was carried out in two parallel bioreactors (Fig. 1). Each one was made of plexiglass with a volume of 32 liters. The filtration process utilized two flat sheet ultrafiltration membrane (UF) made of polyethersulfone (PES), which was placed in the center of each bioreactor. In Table 2 ultrafiltration membrane specification was shown. Control instruments were installed in the pilot for measuring temperature, dissolved oxygen (DO), pH and effluent level. Both bioreactors were operated within a temperature range of 22– 27 °C using a heat exchanger. Fine bubble diffusers were used for aeration to supply required oxygen for microbial activity. Also coarse bubble diffusers were installed under membranes to reduce premature clogging of the membrane. Using peristaltic pumps, the flux in both reactors was set to approximately 41.7 L m2 h1 and transmembrane pressure (TMP) was continuously recorded by a digital pressure gauge. Membrane cleaning was carried out when TMP increased up to 60 kPa. The membrane cleaning procedure was started by disconnecting the suction lines from the membrane modules and then the membranes were taken out from the bioreactors in order to eliminate the cake layers on the membrane by shaking the membrane in tap water. After that they were immersed in a 250 mg L1 NaOCl solution for 6 h and afterwards with 4000 mg L1 citric acid solution for at least 4 h. At the startup phase, no biomass was exited from the reactors as excess sludge to allow mixed liquor suspended solids (MLSS) concentration build up in the system to about 4500 mg L1. Assuming that the bioreactors are treated as an ideal continuously stirred tank reactor (CSTR), concentration of activated sludge in the MBRs will be the same as the concentration in the excess sludge and the SRT will determine as effective volume of bioreactor divided by excess sludge flow rate. Therefore to control of solid concentrations in the bioreactors and maintain the predetermined sludge retention time (SRT = 32 days), daily withdrawal of mixed liquor (equal to 1 L day1) was conducted from the bottom of bioreactors. 2.3. Electrocoagulation studies As shown in Fig. 1, two pairs of perforated iron plates with variable distance between them were used as anode and cathode electrodes in the HEMBR. Direct current (DC) was induced into the reactor through a DC-regulated power supply (Toshiba model, 0– 30 V and 0–5 A). Some factors affecting the performance of the electrocoagulation process includes pH, distance between anode and cathode, current density and electrolysis time (Daneshvar et al., 2006). So to achieve the highest efficiency, before startup of MBRs, the optimum amount of these parameters was
Table 1 Typical characteristics of effluent. Parameter
Unit
Ave.
Min.
Max.
pH SS COD TDS DO TN Alkalinity as CaCO3
– mg L1 mg L1 mg L1 mg L1 mg L1 mg L1
7.3 103 251 1130 3 43 212
6.9 91 213 790 1.5 25 178
7.9 115 354 1310 4 73 245
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
179
Fig. 1. Schematic diagram of MBR and HEMBR pilot plant (F1: feed stream, F2 and F3: permeate water from MBR and HEMBR, respectively, F4: air stream, F5 and F6: excess sludge from MBR and HEMBR, respectively).
2.4. Analytical methods
Table 2 Membrane specification. Process parameters
Unit
Value
Membrane configuration Cut off Pore size Dimensions (width height) Effective surface area Material pH resistance range
– kDalton lm mm m2 – –
Flat sheet 150 0.04 2 (240 200) 0.096 PES 4–11
determined by experiments. In this study, iron was used as electrode in the electrocoagulation process and the following reactions were taken place in the reactor (Zaroual et al., 2006): In the anode:
FeðsÞ ¼ Fe2þ ðaqÞ þ 2e1
ð1Þ
1 2Fe2þ ðaqÞ þ O2 ðgÞ " þ5H2 OðlÞ ! 2FeðOHÞ3 ðsÞ # þ4Hþ ðaqÞ 2
ð2Þ
Fe2þ ðaqÞ þ 2OH ! FeðOHÞ2 ðsÞ
ð3Þ
2.4.1. Analysis of samples Chemical analytical tests were carried out in order to evaluate quality of effluent, mixed liquor and permeate. Most analytical techniques used in this research followed the Standard Methods described by the APHA (American Public Health et al., 2005). All analyses were performed with at least two repetitions. SS and COD were measured according to method 2540D and 5220C, respectively. The percent removal of each parameter was determined by the difference between the concentrations in feed stream and permeate stream divided by the concentration in the feed stream. During operation, the temperature of mixed liquor in the reactor was monitored using a digital temperature probe (JENWAY, England) and TMP was continuously recorded by a digital pressure gauge. DO was measured using an oxygen electrode (JENWAY-970, England). A portable pH meter (JENWAY-370, England) was used to determine the pH in the reactor, and it was calibrated with a standard pH solution of 7 before measurement. Sludge volume index (SVI) was used to assess the settling characteristics of the biomass in two bioreactors according to method 2710D of the Standard Methods.
and in the cathode:
2H2 OðlÞ þ 2e ! H2 ðgÞ " þ2OH
ð4Þ
Electrical energy and iron consumption are two major factors affecting costs of operation in this process. Metal consumption (m) in anode is calculated from Faraday’s law which is expressed as:
m ðkgÞ ¼ ItM=nF
ð5Þ
where I is the current (A), t is electrolysis time (s), M is the relative molar mass of metal, n is the number of electrons in the oxidation/reduction reaction, and F is Faraday’s constant, 96,500 (C/mol). Also energy consumption is expressed as (Kobya et al., 2006):
E ðkWhÞ ¼ VIt
ð6Þ
where V is the voltage (v), I is current and t is electrolysis time.
2.4.2. Microbial activity Because of applying electrical current in HEMBR, biomass activity must be monitored. High current density prevents microorganism’s activity and disrupts biological treatment. Oxygen uptake rate (OUR), or respiration of activated sludge, was performed to determine microbial activity and is defined as the amount of oxygen per unit volume utilized per unit time by the microorganisms in the activated sludge (Hasar et al., 2002). This test performed by transferring 400 mL of activated sludge sample from HEMBR to respirometer cell and recording the changes of DO concentration with time. During this experiment no further aeration and no substrate were added into the activated sludge sample. 2.4.3. Silt density index The silt density index (SDI) analysis was used as indicator of fouling potential in water that is prepared for RO treatment.
180
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
According to ASTM D-4189 for measuring SDI, water is passed at a constant pressure of 30 psi through a 0.45-lm membrane filter at constant temperature. The difference between the initial and the second time required to filter 500 mL after silt build up (normally 15 min) represents the SDI value and is calculated from the rate of plugging. 2.4.4. Microscopic observations and foulants analysis At the end of experiments field emission scanning electron microscopy (FESEM) and X-ray fluorescence spectrometer (XRF) technique were used to investigate the changes in the surface morphology of membranes after fouling and chemical analysis of sediment layers, respectively. To do this, used membrane sheets were taken off from MBR module and tap water was sprayed at their surfaces to remove any particle and attached biomass. Then membranes was cut into small pieces by sterile blades under room temperature and sent to central laboratory of University of Tehran for XRF analysis and instant FESEM scanning. 2.4.5. Statistical analysis The differences between results of two bioreactors were analyzed using t-test in XLSTAT software at significance level (a-value) of 0.05 (95% confidence interval). The null hypothesis was that there were no significant differences between the analyzed data sets and it was rejected when the test result (i.e., p-value) was less than the target a-value (0.05). 3. Results and discussion
pH = 8, whole no significant difference was observed between pH = 7 and pH = 8. Regarding the neutrality of pH = 7, this pH was considered for further experiments.
3.1.4. Effect of current density change According to DC regulated power supply voltage (0–30 V), different voltage values were applied to anode and cathode electrodes, and their effects on COD removal efficiency is shown in Fig. 2d. As can be seen, the COD removal rate is high at the start of electrolysis, while removal rate gradually decreased during the electrolysis time. Also, the obtained results show that the COD rate increased by increasing the applied electrical field. On the other hand, the electrical field threshold in the MBR process should be adjusted to a value that can not disturb the microbial activity in the treatment process. Fig. 2e shows the OUR test results, which were conducted to determine the current density threshold in the bioreactors. As can be seen, the OUR decreases in higher electrical fields. The OUR reduction is very slight as current density increases from 2.5 to 10 mA cm2 which means the effects of electrical field on microbial activity is negligible. By increasing the current density by more than 10 mA cm2, the OUR rapidly reduced, meaning that the applied electrical field had disturbed microbial activity. Therefore, the maximum current density which does not disturb the biological activity of microorganisms is 10 mA cm2, which was used in HEMBR. From Eqs. (5) and (6), energy and iron consumption in this current density are 0.7 kWh and 73 g per cubic meter of permeate, respectively.
3.1. Best conditions of electrocoagulation As previously mentioned, due to dependence of EC on some parameters such as electrolysis time (t), initial pH, the distance between electrodes (d) and current density (J), by considering the initial values for the parameters under consideration, the optimal values of each parameter was specified. 3.1.1. Effect of electrolysis time Assuming an initial pH of 7, a current density of 15 mA cm2, electrodes distance of 4 cm, and an electrolysis time of 150 min, electrical coagulation was carried out on activated sludge (Fig. 2a). The results indicate a noticeable reduction of COD so that about 30% of total COD was removed during 75 min of electrolysis. Afterward COD reduction rate decreased in such a way that no considerable removal efficiency was seen in the period of 105– 150 min. Since the need for treatment during the 105–150 min was so negligible (about 3%) and because of the energy savings, 105 min was selected as the electrolysis time for other experiments. 3.1.2. Effect of distance between electrodes To select the optimum distance between anode and cathode plates, all experiments were carried out in conditions including initial pH = 7, J = 15 mA cm2, t = 105 min (Fig. 2b). The results show that the COD removal efficiency increased by increasing the distance from 4 to 5 cm but decreased when the distance was 6 cm. Thus, the optimum distance between electrode plates to reach maximum efficiency was set at 5 cm. 3.1.3. Effect of pH change The normal range of pH in an MBR operation is 6–8 (Trussell et al., 2007). The COD removal efficiency was studied considering the t = 105 min, J = 15 mA cm2, d = 5 cm, and mentioned pH range (Fig. 2c). During experiments, the efficiency of COD removal when pH = 6 was less than those efficiencies obtained when pH = 7 and
3.2. Overall performance and effluent quality Regarding the equal quality of input feed to bioreactors and the similarity of operation conditions, the permeate quality of the two bioreactors was evaluated and compared based on SS and COD removal. Results of statistical analyses by t-test are summarized in Tables 3 and 4. Also Fig. 3 shows the variation and removal efficiency of SS and COD during operation (45 days). Data summarized in Table 3 show that significant seasonal variation (p-value < 0.05) was found in concentrations of SS and COD for inlet and outlet of both bioreactors. Because of the wide range of input and output SS concentrations, the concentration measurements on the y-axis were plotted on a logarithmic scale (Fig. 3a). Inlet SS concentration varied from 91 to 115 mg L1 and its removal efficiency for two bioreactors is higher than 99%. Regarding the membrane pore size, suspended solids in the mixed liquor of both bioreactors were trapped behind the membrane pores. That’s why SS removal efficiencies for the two bioreactors were considerably high. Also from Table 4 statistical analysis showed that there was no significant difference (p-value = 0.104 > 0.05 for t-test) observed between MBR and HEMBR. Fig. 3b shows inlet COD concentrations ranging from 213 to 354 mg L1 while MBR and HEMBR permeate had an organic matter content ranging from 41 to 62 mg L1 and 27 to 55 mg L1, respectively. Some studies have reported higher efficiencies for COD removal in the MBR process compared to the present work, which is attributed to the low amounts of biodegradable organic matter and the presence of hardly degradable substances in effluent (Hoinkis et al., 2012; Qin et al., 2007). Also statistical analysis showed that HEMBR efficiency for COD removal (85%) was significantly (p-value < 0.05) higher as compare to MBR (80%). Since the electrical coagulation is carried out in HEMBR, some organic compounds are adsorbed on coagulated surfaces and trapped behind the filtration pores, which results in a lower output concentration.
181
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
Fig. 2. Effect of the (a) electrolysis time, (b) distance between electrodes, (c) initial pH, (d) current density in electrical coagulation and (e) variation of the OUR of activated sludge with time.
Table 3 Results of statistical analysis for each reactor. Parameter
Observations
Mean
Std. deviation
95% confidence interval on the mean
p-Value
SS
In OutMBR OutHEMBR
12 12 12
222.6 2.7 2.5
31.1 0.6 0.5
] 201.8; ] 2.3; ] 2.2;
243.5 [ 3.1 [ 2.9 [
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
A new flat sheet membrane bioreactor hybrid system for advanced treatment of effluent, reverse osmosis pretreatment and fouling mitigation Majid Hosseinzadeh a,⇑, Gholamreza Nabi Bidhendi a, Ali Torabian a, Naser Mehrdadi a, Mehdi Pourabdullah b a b
Water and Wastewater Engineering Group, Graduate Faculty of Environment, University of Tehran, Tehran, Iran Environmental and Water Research Center, School of Civil Engineering, Sharif University of Technology, Tehran, Iran
h i g h l i g h t s Pilot-scale hybrid and conventional membrane bioreactors was compared. Optimized electrocoagulation was used in hybrid system. Membrane fouling and sludge characteristics improved significantly in hybrid MBR. Notably better permeate quality was achieved in hybrid system. Both systems had excellent performance as reverse osmosis pretreatment.
a r t i c l e
i n f o
Article history: Received 26 March 2015 Received in revised form 16 May 2015 Accepted 19 May 2015 Available online 22 May 2015 Keywords: Electrical coagulation Membrane bioreactor Fouling Reverse osmosis
a b s t r a c t This paper introduces a new hybrid electro membrane bioreactor (HEMBR) for reverse osmosis (RO) pretreatment and advanced treatment of effluent by simultaneously integrating electrical coagulation (EC) with a membrane bioreactor (MBR) and its performance was compared with conventional MBR. Experimental results and their statistical analysis showed removal efficiency for suspended solids (SS) of almost 100% for both reactors. HEMBR removal of chemical oxygen demand (COD) improved by 4% and membrane fouling was alleviated according to transmembrane pressure (TMP). The average silt density index (SDI) of HEMBR permeate samples was slightly better indicating less RO membrane fouling. Moreover, based on the SVI comparison of two reactor biomass samples, HEMBR showed better settling characteristics which improved the dewaterability and filterability of the sludge. Analysis the change of membrane surfaces and the cake layer formed over them through field emission scanning electron microscopy (FESEM) and X-ray fluorescence spectrometer (XRF) were also discussed. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Shortage of water resources is one of the critical issues facing the world today, especially in the Middle East countries. In these countries, a significant part of the water is converted to wastewater. Using reverse osmosis (RO), water reclamation and purification of treated wastewater becomes widespread enabling production of high quality water for various purposes such as non-potable applications and industrial usage (Hosseinzadeh et al., 2013; Macedonio et al., 2012). RO systems usually use membrane technologies as pretreatment to provide appropriate quality for feed water to ⇑ Corresponding author. Tel.: +98 2161113155; fax: +98 2166407719. E-mail address: [email protected] (M. Hosseinzadeh). http://dx.doi.org/10.1016/j.biortech.2015.05.066 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
enhance RO performance and prevent premature fouling. A membrane bioreactor (MBR) is a wastewater treatment process that combines an activated sludge process (ASP) with membrane separation thereby producing reclaimed water (Wei, 2012; Xiao et al., 2014). Membrane fouling is the major obstacle to increasing MBR development because of reduction in the membrane performance and the efficiency of treatment by forming thick gel layer and cake layer that formed on and into the membrane. Several factors are involved in causing membrane fouling such as membrane characteristics, feed and biomass characteristics, operating conditions, etc.(Delrue et al., 2011; Meng et al., 2009; Tijing et al., 2015; Wu and Huang, 2009; Zhang et al., 2014b). Many studies have been conducted on membrane fouling and several methods and strategies to reduce and control the problem have been investigated in
178
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
recent decade (Huang et al., 2011; Sun et al., 2011; Tijing et al., 2015; Wang et al., 2014). Some studies have shown that coagulation can significantly reduce colloidal and soluble organic materials in water and wastewater treatment. Literature on coagulation systems coupled with membrane bioreactors have reported satisfactory results on MBR performance. Addition of coagulants can change particle characteristics such as charge, size and shape and lead to reduction in organic components, which causes a reduction in organic and colloidal fouling in MBRs (Chen and Liu, 2012; Song et al., 2008; Thang Nguyen, 2012; Tian et al., 2008). Chen and Liu confirmed the possibility and applicability of coagulation-MBR to reclaim effluent in the dairy industry (Chen and Liu, 2012). They realized that poly aluminum chloride as a coagulant played an effective role in improving MBR permeate quality and membrane fouling reduction. While the addition of chemicals for coagulation may result in production of undesirable by-products or increase the sludge amount in the reactor, applying a direct current (DC) field on the activated sludge can be considered an effective method in MBR application. In this system, usually Fe and Al are applied as sacrificial anodes to supply coagulants in the reactor (Akyol, 2012; Maha Lakshmi and Sivashanmugam, 2013; Millar et al., 2014; S ß engil et al., 2009). However, applying a DC field directly within the activated sludge reactor may be harmful to microorganism activity (Bani-Melhem and Elektorowicz, 2011; Bani-Melhem and Smith, 2012). Melhem and Smith proposed application of an electrocoagulation unit as a pretreatment step before MBR in order to prevent direct contact of the microbial community with the applied DC field (Bani-Melhem and Smith, 2012). The results demonstrated that the EC-MBR may be effective not only for increasing the quality of the treated gray water over MBR, but also for improving the overall performance of the membrane filtration process. Although limited studies have been done on the use of current field for membrane fouling reduction (Bani-Melhem and Elektorowicz, 2011; Hasan et al., 2014; Zhang et al., 2014a), to the best of our knowledge, no reports have been published for the simultaneous application of electric coagulation along with MBR in a continuous pilot project for advanced treatment of industrial wastewater treatment plant effluent and RO pretreatment, which is the main objective of this work. This study first identified best conditions for electrical coagulation and then simultaneous integration of optimized electrocoagulation, biological, and membrane processes in one reactor i.e., hybrid electro membrane bioreactor (HEMBR) was developed and its performance compared with conventional MBR based on permeate quality, RO pretreatment capability and membrane fouling was assessed. To do this, two parallel continuous pilot projects were designed, built and operated at the same conditions.
2. Methods 2.1. Effluent characteristics The pilot project was set up in the Shokouhieh industrial wastewater treatment plant (WWTP), located in the province of Qom, Iran. There are many industries in this town such as, dairy, beverage, welding, petrochemical, and metal finishing and their wastewater after pretreatment through a sewage network is routed to the wastewater treatment plant. The average COD in the influent of WWTP is about 2000 mg L1 which reduce to about 250 mg L1 in effluent. The main process units in this plant are a screen, equalization tank, anaerobic basin, aerobic unit, sedimentation tank and chlorination units. Due to problems in design, implementation and operation of the plant, a certain amount of
biodegradable organic matters will remain in the effluent after treatment. In this study, bioreactors were set up for advanced treatment of effluent and they fed with the unchlorinated effluent of wastewater treatment. The characteristics of effluent are given in Table 1. 2.2. Experimental set-up and operating conditions The study was carried out in two parallel bioreactors (Fig. 1). Each one was made of plexiglass with a volume of 32 liters. The filtration process utilized two flat sheet ultrafiltration membrane (UF) made of polyethersulfone (PES), which was placed in the center of each bioreactor. In Table 2 ultrafiltration membrane specification was shown. Control instruments were installed in the pilot for measuring temperature, dissolved oxygen (DO), pH and effluent level. Both bioreactors were operated within a temperature range of 22– 27 °C using a heat exchanger. Fine bubble diffusers were used for aeration to supply required oxygen for microbial activity. Also coarse bubble diffusers were installed under membranes to reduce premature clogging of the membrane. Using peristaltic pumps, the flux in both reactors was set to approximately 41.7 L m2 h1 and transmembrane pressure (TMP) was continuously recorded by a digital pressure gauge. Membrane cleaning was carried out when TMP increased up to 60 kPa. The membrane cleaning procedure was started by disconnecting the suction lines from the membrane modules and then the membranes were taken out from the bioreactors in order to eliminate the cake layers on the membrane by shaking the membrane in tap water. After that they were immersed in a 250 mg L1 NaOCl solution for 6 h and afterwards with 4000 mg L1 citric acid solution for at least 4 h. At the startup phase, no biomass was exited from the reactors as excess sludge to allow mixed liquor suspended solids (MLSS) concentration build up in the system to about 4500 mg L1. Assuming that the bioreactors are treated as an ideal continuously stirred tank reactor (CSTR), concentration of activated sludge in the MBRs will be the same as the concentration in the excess sludge and the SRT will determine as effective volume of bioreactor divided by excess sludge flow rate. Therefore to control of solid concentrations in the bioreactors and maintain the predetermined sludge retention time (SRT = 32 days), daily withdrawal of mixed liquor (equal to 1 L day1) was conducted from the bottom of bioreactors. 2.3. Electrocoagulation studies As shown in Fig. 1, two pairs of perforated iron plates with variable distance between them were used as anode and cathode electrodes in the HEMBR. Direct current (DC) was induced into the reactor through a DC-regulated power supply (Toshiba model, 0– 30 V and 0–5 A). Some factors affecting the performance of the electrocoagulation process includes pH, distance between anode and cathode, current density and electrolysis time (Daneshvar et al., 2006). So to achieve the highest efficiency, before startup of MBRs, the optimum amount of these parameters was
Table 1 Typical characteristics of effluent. Parameter
Unit
Ave.
Min.
Max.
pH SS COD TDS DO TN Alkalinity as CaCO3
– mg L1 mg L1 mg L1 mg L1 mg L1 mg L1
7.3 103 251 1130 3 43 212
6.9 91 213 790 1.5 25 178
7.9 115 354 1310 4 73 245
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
179
Fig. 1. Schematic diagram of MBR and HEMBR pilot plant (F1: feed stream, F2 and F3: permeate water from MBR and HEMBR, respectively, F4: air stream, F5 and F6: excess sludge from MBR and HEMBR, respectively).
2.4. Analytical methods
Table 2 Membrane specification. Process parameters
Unit
Value
Membrane configuration Cut off Pore size Dimensions (width height) Effective surface area Material pH resistance range
– kDalton lm mm m2 – –
Flat sheet 150 0.04 2 (240 200) 0.096 PES 4–11
determined by experiments. In this study, iron was used as electrode in the electrocoagulation process and the following reactions were taken place in the reactor (Zaroual et al., 2006): In the anode:
FeðsÞ ¼ Fe2þ ðaqÞ þ 2e1
ð1Þ
1 2Fe2þ ðaqÞ þ O2 ðgÞ " þ5H2 OðlÞ ! 2FeðOHÞ3 ðsÞ # þ4Hþ ðaqÞ 2
ð2Þ
Fe2þ ðaqÞ þ 2OH ! FeðOHÞ2 ðsÞ
ð3Þ
2.4.1. Analysis of samples Chemical analytical tests were carried out in order to evaluate quality of effluent, mixed liquor and permeate. Most analytical techniques used in this research followed the Standard Methods described by the APHA (American Public Health et al., 2005). All analyses were performed with at least two repetitions. SS and COD were measured according to method 2540D and 5220C, respectively. The percent removal of each parameter was determined by the difference between the concentrations in feed stream and permeate stream divided by the concentration in the feed stream. During operation, the temperature of mixed liquor in the reactor was monitored using a digital temperature probe (JENWAY, England) and TMP was continuously recorded by a digital pressure gauge. DO was measured using an oxygen electrode (JENWAY-970, England). A portable pH meter (JENWAY-370, England) was used to determine the pH in the reactor, and it was calibrated with a standard pH solution of 7 before measurement. Sludge volume index (SVI) was used to assess the settling characteristics of the biomass in two bioreactors according to method 2710D of the Standard Methods.
and in the cathode:
2H2 OðlÞ þ 2e ! H2 ðgÞ " þ2OH
ð4Þ
Electrical energy and iron consumption are two major factors affecting costs of operation in this process. Metal consumption (m) in anode is calculated from Faraday’s law which is expressed as:
m ðkgÞ ¼ ItM=nF
ð5Þ
where I is the current (A), t is electrolysis time (s), M is the relative molar mass of metal, n is the number of electrons in the oxidation/reduction reaction, and F is Faraday’s constant, 96,500 (C/mol). Also energy consumption is expressed as (Kobya et al., 2006):
E ðkWhÞ ¼ VIt
ð6Þ
where V is the voltage (v), I is current and t is electrolysis time.
2.4.2. Microbial activity Because of applying electrical current in HEMBR, biomass activity must be monitored. High current density prevents microorganism’s activity and disrupts biological treatment. Oxygen uptake rate (OUR), or respiration of activated sludge, was performed to determine microbial activity and is defined as the amount of oxygen per unit volume utilized per unit time by the microorganisms in the activated sludge (Hasar et al., 2002). This test performed by transferring 400 mL of activated sludge sample from HEMBR to respirometer cell and recording the changes of DO concentration with time. During this experiment no further aeration and no substrate were added into the activated sludge sample. 2.4.3. Silt density index The silt density index (SDI) analysis was used as indicator of fouling potential in water that is prepared for RO treatment.
180
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
According to ASTM D-4189 for measuring SDI, water is passed at a constant pressure of 30 psi through a 0.45-lm membrane filter at constant temperature. The difference between the initial and the second time required to filter 500 mL after silt build up (normally 15 min) represents the SDI value and is calculated from the rate of plugging. 2.4.4. Microscopic observations and foulants analysis At the end of experiments field emission scanning electron microscopy (FESEM) and X-ray fluorescence spectrometer (XRF) technique were used to investigate the changes in the surface morphology of membranes after fouling and chemical analysis of sediment layers, respectively. To do this, used membrane sheets were taken off from MBR module and tap water was sprayed at their surfaces to remove any particle and attached biomass. Then membranes was cut into small pieces by sterile blades under room temperature and sent to central laboratory of University of Tehran for XRF analysis and instant FESEM scanning. 2.4.5. Statistical analysis The differences between results of two bioreactors were analyzed using t-test in XLSTAT software at significance level (a-value) of 0.05 (95% confidence interval). The null hypothesis was that there were no significant differences between the analyzed data sets and it was rejected when the test result (i.e., p-value) was less than the target a-value (0.05). 3. Results and discussion
pH = 8, whole no significant difference was observed between pH = 7 and pH = 8. Regarding the neutrality of pH = 7, this pH was considered for further experiments.
3.1.4. Effect of current density change According to DC regulated power supply voltage (0–30 V), different voltage values were applied to anode and cathode electrodes, and their effects on COD removal efficiency is shown in Fig. 2d. As can be seen, the COD removal rate is high at the start of electrolysis, while removal rate gradually decreased during the electrolysis time. Also, the obtained results show that the COD rate increased by increasing the applied electrical field. On the other hand, the electrical field threshold in the MBR process should be adjusted to a value that can not disturb the microbial activity in the treatment process. Fig. 2e shows the OUR test results, which were conducted to determine the current density threshold in the bioreactors. As can be seen, the OUR decreases in higher electrical fields. The OUR reduction is very slight as current density increases from 2.5 to 10 mA cm2 which means the effects of electrical field on microbial activity is negligible. By increasing the current density by more than 10 mA cm2, the OUR rapidly reduced, meaning that the applied electrical field had disturbed microbial activity. Therefore, the maximum current density which does not disturb the biological activity of microorganisms is 10 mA cm2, which was used in HEMBR. From Eqs. (5) and (6), energy and iron consumption in this current density are 0.7 kWh and 73 g per cubic meter of permeate, respectively.
3.1. Best conditions of electrocoagulation As previously mentioned, due to dependence of EC on some parameters such as electrolysis time (t), initial pH, the distance between electrodes (d) and current density (J), by considering the initial values for the parameters under consideration, the optimal values of each parameter was specified. 3.1.1. Effect of electrolysis time Assuming an initial pH of 7, a current density of 15 mA cm2, electrodes distance of 4 cm, and an electrolysis time of 150 min, electrical coagulation was carried out on activated sludge (Fig. 2a). The results indicate a noticeable reduction of COD so that about 30% of total COD was removed during 75 min of electrolysis. Afterward COD reduction rate decreased in such a way that no considerable removal efficiency was seen in the period of 105– 150 min. Since the need for treatment during the 105–150 min was so negligible (about 3%) and because of the energy savings, 105 min was selected as the electrolysis time for other experiments. 3.1.2. Effect of distance between electrodes To select the optimum distance between anode and cathode plates, all experiments were carried out in conditions including initial pH = 7, J = 15 mA cm2, t = 105 min (Fig. 2b). The results show that the COD removal efficiency increased by increasing the distance from 4 to 5 cm but decreased when the distance was 6 cm. Thus, the optimum distance between electrode plates to reach maximum efficiency was set at 5 cm. 3.1.3. Effect of pH change The normal range of pH in an MBR operation is 6–8 (Trussell et al., 2007). The COD removal efficiency was studied considering the t = 105 min, J = 15 mA cm2, d = 5 cm, and mentioned pH range (Fig. 2c). During experiments, the efficiency of COD removal when pH = 6 was less than those efficiencies obtained when pH = 7 and
3.2. Overall performance and effluent quality Regarding the equal quality of input feed to bioreactors and the similarity of operation conditions, the permeate quality of the two bioreactors was evaluated and compared based on SS and COD removal. Results of statistical analyses by t-test are summarized in Tables 3 and 4. Also Fig. 3 shows the variation and removal efficiency of SS and COD during operation (45 days). Data summarized in Table 3 show that significant seasonal variation (p-value < 0.05) was found in concentrations of SS and COD for inlet and outlet of both bioreactors. Because of the wide range of input and output SS concentrations, the concentration measurements on the y-axis were plotted on a logarithmic scale (Fig. 3a). Inlet SS concentration varied from 91 to 115 mg L1 and its removal efficiency for two bioreactors is higher than 99%. Regarding the membrane pore size, suspended solids in the mixed liquor of both bioreactors were trapped behind the membrane pores. That’s why SS removal efficiencies for the two bioreactors were considerably high. Also from Table 4 statistical analysis showed that there was no significant difference (p-value = 0.104 > 0.05 for t-test) observed between MBR and HEMBR. Fig. 3b shows inlet COD concentrations ranging from 213 to 354 mg L1 while MBR and HEMBR permeate had an organic matter content ranging from 41 to 62 mg L1 and 27 to 55 mg L1, respectively. Some studies have reported higher efficiencies for COD removal in the MBR process compared to the present work, which is attributed to the low amounts of biodegradable organic matter and the presence of hardly degradable substances in effluent (Hoinkis et al., 2012; Qin et al., 2007). Also statistical analysis showed that HEMBR efficiency for COD removal (85%) was significantly (p-value < 0.05) higher as compare to MBR (80%). Since the electrical coagulation is carried out in HEMBR, some organic compounds are adsorbed on coagulated surfaces and trapped behind the filtration pores, which results in a lower output concentration.
181
M. Hosseinzadeh et al. / Bioresource Technology 192 (2015) 177–184
Fig. 2. Effect of the (a) electrolysis time, (b) distance between electrodes, (c) initial pH, (d) current density in electrical coagulation and (e) variation of the OUR of activated sludge with time.
Table 3 Results of statistical analysis for each reactor. Parameter
Observations
Mean
Std. deviation
95% confidence interval on the mean
p-Value
SS
In OutMBR OutHEMBR
12 12 12
222.6 2.7 2.5
31.1 0.6 0.5
] 201.8; ] 2.3; ] 2.2;
243.5 [ 3.1 [ 2.9 [
