Effects of chemical additives on filtration and rheological characteristics of MBR sludge

Bioresource Technology 117 (2012) 48–54

Contents lists available at SciVerse ScienceDirect

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

Effects of chemical additives on filtration and rheological characteristics of MBR sludge H. Koseoglu, N.O. Yigit ⇑, G. Civelekoglu, B.I. Harman, M. Kitis Department of Environmental Engineering, Suleyman Demirel University, Isparta 32260, Turkey

h i g h l i g h t s " Poly-1 and -2 additives recommended due to their good filtration characteristics. " PACl and chitosan showed average performances compared to the other additives. " Starch exhibited the worst filtration results. " Rheology and particle size data showed correlations with other fouling parameters. " Results at low shear rates were important due to possible effects on cake transport.

a r t i c l e

i n f o

Article history: Received 13 December 2011 Received in revised form 18 April 2012 Accepted 19 April 2012 Available online 26 April 2012 Keywords: Chemical additive Fouling Membrane bioreactor (MBR) Particle size distribution Rheology

a b s t r a c t The main goal of this study was to control the fouling phenomena in MBR using chemical additives. In the first phase of the study, SMP removal and bound EPS formation capacity of chemical additives were determined. Highest SMP removal (72%) was achieved by the Poly-2 additive. In the second phase of the study, short term filtration tests were conducted. Poly-1 exhibited highest performance based on membrane resistance, permeability and average TMP. According to the results obtained from constant shear rate tests in fourth phase, no significant change in viscosity with time was observed. Studies for the adaptation of rheograms to common flow models showed that chitosan and starch was not able to fit to Ostwald de Waele and Bingham models. At a shear rate of 73.4 s1 viscosities of all samples were close to each other. Chitosan and starch achieved highest viscosity values at the shear rate of 0.6 s1. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Limited natural resources of 21st century enforce the recycle, recovery and reuse applications for maximum efficiency in all processes. It is well known that the membrane bioreactors (MBR) are able to meet stringent water reuse criteria with reasonable operational costs. With an average growth rate of 10.9% per annum, the MBR market has been growing significantly faster than other advanced wastewater treatment processes and membrane technologies (Drews, 2010). Fouling, the major obstacle for the MBR process, is not only a barrier for the feasibility but also a hindrance for the process flexibility and simplicity. There is huge extent of studies about MBR fouling which can be classified in two major section: detection/identification and control of fouling. Extracellular polymeric substances (EPS) and soluble microbial products (SMP) are accepted as the major foulants in many studies. Removal of SMP is of great importance by means of internal ⇑ Corresponding author. Tel.: +90 246 211 1284; fax: +90 246 237 0859. E-mail address: [email protected] (N.O. Yigit). 0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.04.067

fouling, cake porosity, initial cake formation, and local flux increments (Hwang et al., 2007). Control methods of MBR fouling covers a large field including; surface modification and preparation of novel membranes (Maximous et al., 2009), relaxation and backwashing conditions (Wu et al., 2008), and novel module configurations (Grelot et al., 2009; Nguyen et al., 2011). Biomass modification by chemical additives is another way of fouling control. Easy implementation of this concept on existing plants is another attractive dimension of the technique. Several studies on the topic covers a wide range of additives e.g. polymers (Yoon et al., 2005); zeolite (Lee et al., 2001); metal salts (Wu and Huang, 2008); bentonite, vermiculite (Malamis et al., 2009); diatomite (Yang et al., 2010), starch and chitosan (Ji et al., 2008; Koseoglu et al., 2008). However, side-by-side comparison of the wide range of additives at regular intervals under the same operational conditions is a significant gap in literature. Side-by-side comparisons provide sound data for decision-makers. Another neglected area of fouling control by additives is their impact on the rheological characteristics of sludge. The sludge viscosity influences the MBR performance at the longitudinal head

H. Koseoglu et al. / Bioresource Technology 117 (2012) 48–54

49

Nomenclature K n

s0

fluidity coefficient fluid behavior index yield stress (dyne/cm2)

loss and the hydraulic regime in the membrane (Hasar et al., 2004). Increased sludge viscosity is one of the limiting factors for economically reasonable mixed liquor suspended solids (MLSS) concentrations in MBR. The non-Newtonian behavior of highly concentrated activated sludge has a strong impact on the operation of MBR. In areas with low convection, viscosity increases by one or two orders of magnitude. This likely forms dead zones and thus decreases the effective volume of the activated sludge compartment (Rosenberger et al., 2002). Rheological parameters also affect the energy requirements for mixing, aeration, and permeate extraction, with consequences on the operating costs. Higher solid concentrations may have an influence on sludge viscosity, circulation of bulk sludge and the shear stress on the cake surface (Laera et al., 2007; Pollice et al., 2007). However, studies specifically focused on the rheological effects of various additives on MBRs were not found in the literature. Considerations mentioned above underlines the significance of rheological investigations for modified MBR mixed liquors by chemical additives. In this study a total of six different additives (two metal salts, a starch, a biopolymer, a modified biopolymer and a cationic polymer) were tested and compared in lab-scale trials for their SMP elimination and fouling control capacities in MBR sludge samples. Changes in critical flux values were determined and short term filtration tests were conducted to detect fouling suppression capacities of additives. Batch shaker tests were conducted for six different additives at various dosages to determine SMP removal in the supernatant of the MBR sludge samples. The dosage which provided highest SMP removal was chosen as the optimum dosage for each additive. When a cationic polymer is added to the solution and adsorbs onto microbial flocs having dominantly negative surface charge, the surface charge of flocs would change from negative to neutral. The neutralized flocs may attract each other by a charge neutralization mechanism to produce larger flocs. On the other hand, when a cationic polymer is added in excess of the optimum concentration, the surface charge would be reversed to positive, and thereby the positively charged flocs would begin to de-flocculate by an electrostatic repulsion mechanism (Koseoglu et al., 2008). In addition, effects of additives on rheological characteristics which are critical parameters for MBR processes and fouling were also evaluated. 2. Methods 2.1. Pilot and lab scale MBR unit A pilot-scale MBR (ZeeWeedÒ-10 GE) unit containing an immersed hollow fiber membrane module (0.9 m2 active membrane area, 0.04 lm nominal pore size) was continuously operated aerobically during the studies. HRT was 14 h and the permeate flux was 15 L/m2 h (LMH). The MBR (HDPE tank with 227-L solution volume) was fed with screened (1-mm) raw domestic wastewater collected from the sewage system of the Suleyman Demirel University dormitories (Isparta, Turkey). This unit was used as sludge source (MLSS: 9500 mg/L, MLVSS: 6500 mg/L, mean particle size: 38.9 lm, pH: 7.61) for the batch filtration and rheological tests. A lab scale MBR unit equipped with a compact membrane module (ZeeWeedÒ-1 GE) which has the same specifications with the

gP UGr

plastic viscosity (cP) aeration velocity (m/s)

ZeeWeedÒ-10 GE module (active membrane area: 0.047 m2) was used for the filtration tests. The lab-scale MBR was filled with sludge taken from pilot-scale MBR prior to tests. Temperature was kept constant by using an external water jacket around the 8-L (d: 19 cm, h: 30 cm) reactor. Permeation and backwashing provided by a variable speed (reversible motor) peristaltic pump (Masterflex L/S, pump head: EasyLoad 7518). pH was monitored by using a pH meter. Aeration provided by an air pump (Unistar). Operational configuration of the lab-scale MBR was designed to simulate the pilot-scale MBR (e.g. backwashing duration, aeration intensity). 2.2. Chemical additives Six different chemical additives were used during the study. Stock solutions of chemical additives were prepared in deionized water. Additives were selected in a theoretical basis of SMP elimination capacity by coagulation/flocculation mechanism. Thus strongly cationic charged additives were tried to be chosen. Additives used during the study were Poly-1 a cationic polymer (NalÒ coÒ), FeCl3 a metal salt (Merck ), PACl (polyaluminium chloride) Ò a metal salt (ECS ), starch, cationic corn starch (Tate&LyleÒ), Poly-2 a modified cationic biopolymer (France ChitinÒ) and chitosan a cationic biopolymer (France ChitinÒ). Preparation of the stock solutions was performed right before each batch test. Only solution of chitosan was prepared 12 h before because of the activation period of chitosan. Solutions were mixed about 1 h for homogenization in shaking water bath. During the preparation of starch (powder) and chitosan (flake), heating and mixing were also applied. Application of additives to raw sludge was directly performed with stock solutions without any dilutions. Abbreviations for Poly-1 and -2 were used instead of MPE50Ò and BioflocÒ additives. 2.3. Batch shaker tests Batch shaker tests were conducted for each additive at various dosages in order to find the optimal dosages of additives based on maximum SMP removal. Dosage, which provided the maximum SMP removal, was determined as the optimum dosage. Five different concentrations were applied for each additive. All additives were applied in liquid form. Shaking tests were conducted with raw and additive dosed sludge samples in erlenmeyer flasks for 1 h under the conditions of 20 ± 1 °C constant temperature, pH range of 7.12–8.66 and a frequency of 130 min1 with a shaking water bath (GFL, 1086). Then, supernatant of the sludge samples were filtered (0.45 lm) prior to SMP analyses. Total SMP concentrations reported are the sum of the concentrations of protein and polysaccharide fractions. For bound EPS analyses, homogenized mixed liquor sample was extracted by cationic resin. Cationic resin was mixed at 120 min1 for 1 h in a buffer solution and then dried. After taking the supernatant away by centrifugation (3500 min1, 15 min), sludge samples were mixed in jar test with dried cationic resin at 120 min1 for 5 h. Then, supernatant of the samples was taken away by centrifuging at 3500 min1 for 15 min and analyzed for protein and carbohydrate fractions. Sum of the protein and carbohydrate concentrations of the supernatant was reported as bound EPS.

50

H. Koseoglu et al. / Bioresource Technology 117 (2012) 48–54

2.4. Short term filtration and critical flux

2.5. Rheology and particle size distribution

Optimum dosages obtained from batch shaker tests were applied during the short term filtration tests. Batch short term filtration test was conducted with sludge obtained from pilotscale MBR. Sludge for the batch lab scale MBR was collected from the pilot scale unit freshly and tests were started immediately in order to prevent the changes of sludge characteristics. Each test was completed in 2 h. 27 LMH of constant permeate flux was applied. Temperature of the reactor was 20 ± 1 °C and aeration velocity was 0.1 m/s. Because of the chemical cleaning necessity of the module after each test, raw sludge tests could not conducted each day. Since short term filtration experiments should be completed as fast as possible in order to ensure comparability, intense experimental matrix was balanced by conducting the raw sludge tests only in the beginning and at the end of the experimental matrix. First raw sludge test was used for the comparison of the first three additives, and last raw sludge test was used for the second party of additives. Two dosages, which are the optimum dosage and the selected second closest dosage to optimum dosage determined by batch shaker tests were applied. Shifting of dosage could be a matter of fact in full scale applications. Thus, behavior of additive under these conditions should be cleared. For that reason second dosage was applied for the understanding of filtration characteristics when the optimum dosage was slightly shifted. In other words, second dosage value was tested to find the effects of slight dosage swings on short term filtration characteristics. During the 2 h of operation, 1 min of backwashing was applied for every 10 min in order to simulate the pilot scale module. After each test, the membrane module was chemically cleaned. Mean TMP and fouling rate values were determined from the last 10 min of the operation. Hence, additives which effectively suppressed the fouling even in the last phase of the filtration were selected. Critical flux values of the modified mixed liquors were determined by flux step method. Duration of each step was 15 min. Starting from 5 LMH, the flux was increased by 5 LMH increments to achieve 50 LMH as final flux value. However, filtration was stopped before 50 LMH when the maximum TMP value for the membrane module was exceeded for some additives. Raw sludge tests were conducted at the beginning and at the end of the critical flux tests. Critical flux value was determined as the first flux value where actual permeability (K) was below the 90% of original (first step) permeability (K1) (K < 0.9K1).

Particle size distribution analyses were performed by Mastersizer 2000 according to the laser scattering technique. Rheological analyses were performed by a viscosimeter (Brookfield DV-II Pro) with a sample volume of 16 ml at a constant temperature of 20 °C. Two modes were tested during the rheological analyses. Time-to-stop mode was conducted in the first phase to determine the time dependent characteristics of the mixed liquor. Shear rates from 0.61 to 183.45 s1 with a linear increment in each 10 min was applied. Twenty-five measurements were recorded during each step. In the second phase (speed ramp mode), for each additive, 10 different shear rates were applied with one measurement in each 1 min. Obtained on-line data were plotted and flow regime and model fittings were determined. Ostwald de Waele and Bingham models were applied in model-fit studies. 3. Results and discussion 3.1. Batch SMP removal tests Fig. 1 shows the impacts of tested additives on SMP reductions. Increased SMP concentrations were determined when dosages above the optimum dosage were applied. While reduction in SMP concentrations at optimum dosage can be attributed to entrapment during the flocculation, de-flocculation due to excessive dosing (release of entrapped SMP from the flocs) may be responsible for increased SMP concentrations (Koseoglu et al., 2008). Also rise in the bound EPS concentration due to the entrapment of SMP was investigated. Different extents of bound EPS increments were determined due to the floc enlargement mechanism. Initial bound EPS and SMP values of raw sludge were 143.9 and 14.3 mg/L. After addition of starch for five different dosages, lowest SMP value (10.1 mg/L) was determined for 2000 mg/L dosage. Thus 2000 mg/L was selected as the optimum dosage. This approach was applied for all additives. Range of dosages applied for starch, FeCl3, Poly-2, chitosan, Poly-1 and PACl were 500–2500, 50–150, 500– 2500, 100–1000, 100–1000, 50–250 mg/L, and optimum dosages determined were 2000, 100, 1000, 250, 500, and 100 mg/L, respectively. Corresponding SMP removals were given in Fig. 1. It should be underlined that chitosan removed up to 89% carbohydrate fraction of SMP in batch tests. Carbohydrates were identified as the main foulant component which tends to clog membranes (Dvorak et al., 2011). 3.2. Short term filtration and critical flux

100 90

SMP Removal (%)

80 70 60 50 40 30 20 10 0 Starch

FeCl3

Poly-2

Chitosan

Poly-1

PACl

Fig. 1. SMP removals of all additives at optimum dosages (Starch: 2000 mg/L, FeCl3: 100 mg/L, Poly-2: 1000 mg/L, Chitosan: 250 mg/L, Poly-1: 500 mg/L, PACl: 100 mg/ L).

General performance summary of the short term filtration tests are given in Table 1. Poly-1 achieved the best performance when mean TMP and permeability values were compared with the raw sludge. Twenty-two percent TMP decrease was achieved with 500 mg/L of Poly-1 (optimum dosage), while 19% decrease was achieved with 250 mg/L dosage. Permeability values were 154 and 148 L/m2 h bar for 500 and 250 mg/L dosages, respectively. It is understood that the effect of dosage swing for Poly-1 is minor. Decrease in the fouling rate was 35%. These results were expected because of the effective performance of Poly-1 in batch shaker tests. Other research studies on MBR fouling also reported positive results for Poly-1 (Yoon et al., 2005). Guo et al. (2008) reported that the high DOC (>95%) and COD (>95%) removal were achieved while sustainable flux was raised from 25 LMH to 60 LMH by Poly-1 application. In the study of Lee et al. (2007) porosity and the volume of the biofilm on a hollow fiber module was evaluated by CLSM (confocal laser scanning microscopy) and image analysis techniques. Results showed that the biofilm porosity was raised

51

H. Koseoglu et al. / Bioresource Technology 117 (2012) 48–54 Table 1 Short term filtration test performances of raw and additive dosed sludge samples.

Raw Sludge FeCl3 (75 mg/L) FeCl3 (100 mg/L) Poly-1 (250 mg/L) Poly-1 (500 mg/L) Chitosan (100 mg/L) Chitosan (250 mg/L) Poly-2 (500 mg/L) Poly-2 (1000 mg/L) Starch (1500 mg/L) Starch (2000 mg/L) PACl (50 mg/l) PACl (100 mg/l) Raw sludge

dTMP/dt (mbar/min)

Rt (m1)

4.805 1.504 2.757 4.512 3.134 5.097 3.179 1.379 2.089 8.231 8.148 3.593 3.092 3.468

2.86E+12 2.52E+12 2.33E+12 2.31E+12 2.23E+12 3.17E+12 2.92E+12 2.74E+12 2.73E+12 4.02E+12 4.04E+12 3.22E+12 2.87E+12 3.21E+12

(±0.008) (±0.001) (±0.012) (±0.060) (±0.050) (±0.013) (±0.082) (±0.025) (±0.069) (±0.154) (±0.199) (±0.069) (±0.040) (±0.102)

(±3.4E+9) (±2.2E+10) (±1.5E+10) (±2.6E+10) (±1.2E+9) (±9.0E+9) (±3.4E+10) (±1.4E+10) (±3.2E+10) (±3.0E+9) (±5.4E+9) (±1.7E+9) (±2.6E+10) (±1.2E+10)

by floc enlargement and permeability was enhanced. Also, cake porosity showed a uniform distribution along the fiber (Lee et al., 2007). Results overall showed that the Poly-1 achieved effective filtration characteristics compared to the raw sludge even for the different module configurations and operating conditions. In this study, no sludge accumulation on the module and a sticky gel layer on the membrane surface were observed. FeCl3 was the second effective additive with permeability values close to Poly-1. Through high SMP removal rate, FeCl3 achieved 19% and 12% mean TMP decrease for 100 and 75 mg/L dosages with corresponding permeability values of 147 and 136 L/m2 h bar, respectively. Forty-three percent decrease in fouling rate was achieved at optimum dosage which is a better value than Poly-1. It is reported that small flocs in the range of 1–10 lm and SMP fractions with molecular weight above 10 kDa were eliminated by charge neutralization mechanism with FeCl3 addition (Zhang et al., 2008). Elimination of small flocs was also confirmed by our findings for particle size distribution which will be further discussed. Besides charge neutralization mechanism, according to the divalent cation bridge theory (Sobeck and Higgins, 2002), interaction of negatively charged functional groups in EPS with FeCl3 enhances the floc size. Even an increase of 2% for the mean TMP was found for chitosan at the optimum dosage of 250 mg/L while the rise was much higher, 11%, for 100 mg/L dosage. Such result was expected since chitosan obtained the second lowest SMP removal after starch. It could be speculated that very low protein removal of chitosan (29.4%) may lead to fouling related to proteins. Owing to the deteriorated filtration characteristics of starch and chitosan coupled with low protein removal (19.1% and 29.4%, respectively), a relationship between fouling suppression capacity and removal of protein fraction of SMP comes into question. Permeability values were 118 and 108 L/m2 h bar for 250 and 100 mg/L dosages. Despite unsatisfactory results for chitosan, the only positive fact was the 34% decrease in the fouling rate. Such decrease could be an indicator of fouling postponing capacity of additive and low TMP values compared to raw sludge in long term operation. However, fouling rate increased by 6% for a dosage of 100 mg/L. This fact implies the sensibility of this additive against dosage swings. It is reported that the protein and carbohydrate levels on membrane surface was lower for chitosan added sludge compared to control sludge by FTIR (Fourier transform infrared spectroscopy) analysis. GPC (gel permeation chromatography) results showed that concentration and molecular weights of the organic matters in the supernatant have significant effects on gel layer formation and low porosity in cake (Ji et al., 2008). Also, Iversen (2010) stated that chitosan addition triggered porosity increase by formation of linear structured loose flocs which results in a decrease in fractal dimension

Mean TMP (mbar)

Mean TMP change (%)

K (L/m2 h bar)

225 199 183 182 176 250 230 216 215 316 318 254 226 253

– 12 19 19 22 11 2 15 15 25 26 0 11 –

120 (±0.143) 136 (±1.155) 147 (±0.976) 148 (±1.676) 154 (±0.080) 108 (±0.304) 118 (±1.352) 125 (±0.630) 126 (±1.529) 85 (±0.063) 85 (±0.114) 106 (±0.057) 119 (±1.115) 107 (±0.411)

(±0.271) (±1.715) (±1.200) (±2.011) (±0.092) (±0.705) (±2.702) (±1.097) (±2.558) (±0.235) (±0.424) (±0.136) (±2.074) (±0.981)

(df). An inverse relationship between specific cake resistance and mean floc size was also reported (Ji et al., 2008). Another study revealed that the sustainable filtration duration of chitosan added MBR was 6.7 times higher than the control MBR (Ji et al., 2010). Overall results showed that the chitosan modified sludge samples showed morphological changes such as increase in floc size and viscosity and decrease in fractal dimension. Poly-2 addition resulted in a decrease in mean TMP, 15%, for both dosages (500 and 1000 mg/L) which is interesting when compared to the results of the rest of the additives. Same extent of TMP decrease may be an indicator of filtration stability against dosage swings. Also, same extent of TMP decrease with two different dosages creates an opportunity for lowering the optimum dosage by selecting low dosage. Permeability values were also close for the 1000 and 500 mg/L at 126 and 125 L/m2 h bar, respectively. Fouling rate decreased by 40% at optimum dosage. Mean TMP decrease achieved by PACl was 11% for 100 mg/L dosage. Permeability values were 119 and 106 L/m2 h bar for 100 and 50 mg/L of PACl, respectively. Fouling rate was decreased by 11% at optimum dosage. Wu et al. (2006) reported that the mean floc size was shifted from 75 to 135 lm by PACl application and 73% of the supernatant TOC was removed. Also initial TMP and TMP increase rate decreased 68% and 26%, respectively. It is also stated that the application of polymeric metal salts resulted in better filterability characteristics due to the longer chain molecules and more positive charges (Wu et al., 2006). However, in this study, monomeric FeCl3 showed better results than polymeric PACl. Starch achieved the highest rise in TMP and lowest permeability in short term filtration tests. For both dosages (1500 and 2000 mg/ L) the mean TMP rise was approximately 25% higher than raw sludge. Moreover, permeability values of both dosages were 85 L/ m2 h bar which corresponds to 21% decrease compared to raw sludge. In addition, fouling rate was increased by 57% even at optimum dosage. The lowest SMP removal of starch in batch shaker tests should be related to the deteriorated filtration performance. However, in the study of Koseoglu et al. (2008), significant improvements were achieved by starch addition. But results reported by Iversen et al. (2009) in a pilot scale study with starch additive are consistent with the findings of this study. Partial attachment of starch to flocs would result in permeation of starch from membranes. Starch permeation triggers the gel layer formation on the membrane surface and deteriorates the filtration performance by pore plugging. Table 1 shows the values of dTMP/dt of raw sludge and additive dosed sludge samples. After the short term filtration tests, critical flux values were determined since this parameter is important for the MBR operation. Critical flux value of the raw sludge was 30 LMH. Critical flux

52

H. Koseoglu et al. / Bioresource Technology 117 (2012) 48–54

Table 2 Mean particle sizes and fractional size distribution of raw and additive dosed sludge samples. Sample

Mean particle size (lm)

<1 lm (% vol.)

1–10 lm (% vol.)

10–50 lm (% vol.)

50–100 lm (% vol.)

>100 lm (% vol.)

Raw Sludge FeCl3 PACl Starch Chitosan Poly-1 Poly-2

38.9 43.7 39.8 54.7 206.0 81.7 45.3

0.40 0.37 0.42 0.00 0.00 0.00 0.00

7.60 6.86 7.60 5.23 1.27 2.88 6.05

68.07 61.77 66.68 49.57 5.36 22.68 61.63

22.85 28.28 23.68 37.00 21.49 48.31 30.57

1.08 2.72 1.62 8.20 71.88 26.13 1.75

values were 25, 50, 35, 25, 40 and 20 LMH and corresponding changes were 16.7%, 66.7%, 16.7%, 16.7%, 33.3% and 33.3% for chitosan, Poly-1, FeCl3, PACl, Poly-2, and starch, respectively. 3.3. Rheology and particle size Since rheological data shows intimate relations with particle size data, mean particle size analyses of the sludge samples were conducted (Table 2). In the study of Pham et al. (2010) decrease in viscosity is observed due to the decrease in floc size as a result of ultrasonication process. Effects of mean particle size on cake resistance, cake porosity (Ji et al., 2008) and viscosity pointed out the importance of rheological studies. Highest mean particle size was determined in sludge sample with chitosan additive. Chitosan added sample contains no particle below the size of 1 lm and particles of 1–10 lm range was 1.27%. Thus, it is understood that chitosan was able to remove nearly all particles below 10 lm. Hence, SMP and small-sized particles which are mainly responsible for irreversible fouling was successfully eliminated. It should be underlined that 72% of the particles were above 100 lm. This particle size is very important considering back transport mechanism and cake porosity. However, filtration characteristics of chitosan was ineffective. It is found that viscosity could be a reason of this trend and will be further discussed. After chitosan, Poly-1 achieved the highest mean particle size. Similar to chitosan, no small particles below 1 lm were observed. Also, irreversible fouling (internal pore plugging) suppression capacity should be high since 75% of the particles were above the 50 lm. In addition to its positive aspects, effective trends at filtration and rheological tests showed that Poly-1 is the most recommended additive among tested additives. Particle size data of PACl added sample was similar to the raw sludge. Also, fractional distribution of particle size for the PACl added sludge was similar to the raw sludge. Therefore, no significant change was observed in filtration characteristics. Particle size of Poly-2 and FeCl3 samples were close to each other. This trend was also observed in filtration tests for the Poly-2 and FeCl3 added

900

Shear Stress(mPa)

800

sludge and similar characteristics were obtained for the fouling rate and critical flux data for the two above mentioned additives. It should be underlined that two additives have similar particle sizes and similar rheological behaviors. Overall, results suggested that particle size distribution, filtration characteristics, SMP removal capacity and rheological behaviors showed significant correlation. First phase of the rheological experiments showed that applied constant shear rates during the tests did not cause significant changes in viscosity. This result was expected in activated sludge samples. However, starch dosed sludge showed negligible thyxotropic tendency at low shear rates. Time-dependent changes of viscosity occurred rarely in activated sludge samples but contradictory findings were also reported. In the study of Pham et al. (2010) thyxotropic tendency was observed. Overall, results suggested that no significant time-dependent change in viscosity for all sludge samples was observed. Findings of the second phase of the rheological experiments indicated that increasing the shear rates decreased the viscosity levels. All samples showed pseudoplastic behavior. This proved that all sludge samples were non-Newtonian fluids. Shear rate values against shear stress values for raw and additive dosed sludge samples are given in Fig. 2. Ostwald de Waele and Bingham models, which are frequently used models in activated sludge rheology studies, were applied. It is stated in the literature that Ostwald de Waele and Bingham models are valid for most of the MBR sludge samples (Laera et al., 2007; Pollice et al., 2007). Also, in the existence of different biomass conditions, Ostwald de Waele model would best fit the MBR sludge samples. The parameter of yield stress used in Herschel Bulkley model varies over a wide range and its determination would require specific tests based on extrusion procedures (Laera et al., 2007). Therefore, studies on this model were not performed. R2 values which are the criteria of suitability of the sludge samples to Ostwald de Waele and Bingham models and relevant model parameters are given in Table 3. Results showed that starch and chitosan were not fit to both models. Unlike other additives, starch and chitosan achieved significant increases in floc sizes. Particle size data should be taken into account with the difference of rheological behavior of chitosan. Also, a study revealed that chitosan

700 600 500

Raw Sludge

400

PACl Chitosan

300

Table 3 Rheological parameters of raw and additive dosed sludge samples for Ostwald de Waele and Bingham models. Sample

Poly-2 FeCl3

200

Poly -1

100

Starch

0 0

20

40

60

80

100

120

140

160

180

200

Shear Rate (1/s) Fig. 2. Shear rate-shear stress values of raw and additive dosed sludge samples.

Poly-2 Chitosan FeCl3 Poly-1 PACl Raw sludge Starch

Ostwald de Waele

Bingham

K

n

R2

s0

gP

R2

190.6 648.3 66.7 241.3 289.4 52.1 622.2

0.29 0.04 0.51 0.24 0.17 0.56 0.01

0.945 0.226 0.978 0.921 0.655 0.959 0.002

299.1 690.3 174.3 353.2 329.3 147.0 588.6

3.55 0.78 4.37 3.15 3.12 4.46 1.31

0.828 0.209 0.843 0.796 0.910 0.879 0.215

H. Koseoglu et al. / Bioresource Technology 117 (2012) 48–54

had higher fouling rates in comparison with FeCl3 and PACl due to the viscous nature of chitosan (Guo et al., 2010). It could be speculated that residual starch in the supernatant, with its dilatant nature, may lead to a different rheogram for starch. In addition, starch has ability to create large flocs however it does not completely bind to flocs (Iversen, 2010). High shear stress at lowest shear rate in shear rate-shear stress graphics could be interpreted as an evidence for the existence of yield stress. This situation enforces the application of Herschel Bulkley model to rheograms. However, inconsistencies detected in the yield stress values, particularly at low shear stress, for Herschel Bulkley model and recommendation and utilization of Ostwald de Waele model in a study (Rosenberger et al., 2002) should be taken into account. Finally, PACl showed weaker fitting to Ostwald de Waele model compared to other additives. In other words, fitting of the samples to Bingham model is weaker than Ostwald de Waele model except PACl (Table 3). Effectiveness of Ostwald de Waele model to determine the flow regime of MBR sludge was reported in literature. Therefore, it could be claimed that additives except starch and chitosan may lead to changes in viscosity to some extent but flow regimes did not significantly changed. Viscosity values were close to each other at 73.4 s1. The values were 9.54, 10.74, 8.75, 9.35, 9.26, 8.75, 8.06 mPa s for Poly-2, chitosan, FeCl3, Poly-1, PACl, starch and raw sludge, respectively. Hence, viscosity and related transport mechanism changes caused by additives would be negligible at turbulent and/or high shear environments. However, due to the limitations of viscosimeter, shear rates that simulates the shear velocity by air bubbles at pilot scale MBR module could not be achieved. According to the Popovic and Robinson (1984) shear rate can be calculated using the equation below (Yang et al., 2009).

c ¼ 5000:UGr

ð1Þ

This limitation prevented the detection of possible effects of the turbulence created by air sparging on sludge viscosity. In the study of Meng et al. (2007) when the sludge viscosity was smaller than 3.0 mPa s, it had small influence on the cross flow velocities, which could keep over 0.15 m/s. But, the cross flow velocity of sludge suspension decreased sharply when the sludge viscosity was larger than 3.0 mPa s. The cross flow velocity was only 0.05 m/s as the sludge viscosity increased to 4.0 mPa s. This finding underlines the importance of interactions between viscosity and air scouring. In our study, significant changes in viscosity were detected at very low shear rates. Sludge viscosity influences the MBR performance by longitudinal head loss and hydraulic regime in the membrane (Hasar et al., 2004). Thus, results obtained at low shear rates should be taken into account. In addition, the rheological properties of sludge suspension not only have major impacts on oxygen transfer and sludge conditioning, but also have strong influence on transport phenomena near the membrane surface. Thus, importance of sludge viscosity in terms of transport mechanism is critical particularly in systems with intermittent aeration. Starch and chitosan, two additives with strong floc size enhancement, which did not fit well to the Ostwald de Waele and Bingham model, achieved highest viscosity levels at 0.6 s1. PACl which had poor fitting to Ostwald de Waele model also showed high viscosity. Viscosity values were not close to each other at 0.6 s1. The values were 298.83, 1118.06, 144.27, 194.36, 504.89, 1182.95, 120.17 mPa s for Poly-2, chitosan, FeCl3, Poly-1, PACl, starch and raw sludge, respectively. Correlation between viscosity values at low shear rates and model conformation of the additives was eye catching. Another important aspect is the possible effects of high viscosity of starch and chitosan additives on the biofilm layer. It is well known that one of the most important transport and removal mechanism in MBR process is the cake filtration. Factors affecting cake filtration dynamics are also decisive elements for fouling, transport and removal mechanisms.

53

In the study of Lee et al. (2007) effects of cake porosity and morphology on MBR fouling was investigated and it was stated that morphological structure has critical importance on fouling. Another study revealed that high viscosity levels may lead to a sticky cake layer on membrane surface (Itonaga et al., 2004). These findings are also consistent with the study of Martinez-Sosa et al. (2011) who stated that an increase in viscosity of mixed liquor reduced the back transport mechanism and hence favoring particle deposition. When the rheological data were taken into account, it could be claimed that structure of the modified flocs by starch and chitosan would be different. Furthermore, enhancement of particle size would trigger the efficiency of back transport mechanism. Thus, negative effects of the increased viscosity levels would be balanced by increased particle size. Also, many researchers claimed that limiting effects of viscosity could be diminished by increased air sparging (Rosenberger et al., 2002; Wu et al., 2007; Yang et al., 2009). Thus, optimum location of air nozzles pursuant to membrane modules has critical importance. The synergistic effect of sludge morphology and module configuration may provide insight into the complexity of MBR fouling phenomenon. Overall, results suggested that changes in cake structure due to the rheological characteristics should be revealed by direct visualization and image analysis techniques. A relationship between bound EPS and viscosity were reported by Moreau et al. (2009). Meng et al. (2007) also reported the same relation but linked the viscosity increment to the existence of filamentous bacteria. Such relationship was not observed in this study. However, the highest viscosity values at low shear rates was observed for the starch and chitosan which were achieved the lowest SMP removals. In addition, a relationship between viscosity and filterability was reported (Wu et al., 2007), whereas another study denied the same relationship (Moreau et al., 2009). Statements above highlight the ambiguity about the viscosityMBR fouling topic. 4. Conclusion Additives of Poly-1 and -2 are recommended due to their effective filtration characteristics. However, a direct evaluation between cost advantages created by fouling control and additive costs should be done. Future works regarding this topic should include optimization of mixing techniques of additives, blending of additives, 3D imaging of cake architecture, and determination of irrecoverable fouling. Also, long term pilot plant tests will be useful in order to overcome the limitations (different hydrodynamicscouring conditions of module, lack of determination of long term fouling trends, lack of determination of different membrane cleaning requirements and replacement frequency) of short term labscale tests. Acknowledgement The authors acknowledge the grant of Suleyman Demirel University, BAP Unit to the Project 1629-D-08. References Drews, A., 2010. Membrane fouling in membrane bioreactors – characterisation, contradictions, cause and cures. J. Membr. Sci. 363, 1–28. Dvorak, L., Gomez, M., Dvorakova, M., Ruzickova, I., Wanner, J., 2011. The impact of different operating conditions on membrane fouling and EPS production. Bioresour. Technol. 102, 6870–6875. Grelot, A., Tazi-Pain, A., Weinrich, L., Lesjean, B., Grasmick, A., 2009. Evaluation of a novel flat sheet MBR filtration system. Desalination 236, 111–119. Guo, W., Ngo, H.H., Vigneswaran, S., Dharmawan, F., Nguyen, T.T., Aryal, R., 2010. Effect of different flocculants on short-term performance of submerged membrane bioreactor. Sep. Purif. Technol. 70, 274–279.

54

H. Koseoglu et al. / Bioresource Technology 117 (2012) 48–54

Guo, W.S., Vigneswaran, S., Ngo, H.H., Kandasamy, J., Yoon, S., 2008. The role of a membrane performance enhancer in a membrane bioreactor: a comparison with other submerged membrane hybrid systems. Desalination 231, 305– 313. Hasar, H., Kinaci, C., Unlu, A., Togrul, H., Ipek, U., 2004. Rheological properties of activated sludge in a sMBR. Biochem. Eng. J. 20, 1–6. Hwang, B.K., Lee, W.N., Park, P.K., Lee, C.H., Chang, I.S., 2007. Effect of membrane fouling reducer on cake structure and membrane permeability in membrane bioreactor. J. Membr. Sci. 288, 149–156. Itonaga, T., Kimura, K., Watanabe, Y., 2004. Influence of suspension viscosity and colloidal particles on permeability of membrane used in membrane bioreactor (MBR). Wat. Sci. Technol. 50, 301–309. Iversen, V., 2010. Comprehensive assessment of flux enhancers in membrane bioreactors for wastewater treatment, Ph.D. Thesis, Technical University of Berlin, Germany. Iversen, V., Mehrez, R., Horng, R.Y., Chen, C.H., Meng, F., Drews, A., Lesjean, B., Ernst, M., Jekel, M., Kraume, M., 2009. Fouling mitigation through flocculants and adsorbents addition in membrane bioreactors: comparing lab and pilot studies. J. Membr. Sci. 345, 21–30. Ji, J., Qiu, J., Wai, N., Wong, F.S., Li, Y., 2010. Influence of organic and inorganic flocculants on physico-chemical properties of biomass and membrane fouling rate. Water Res. 44, 1627–1635. Ji, J., Qiu, J., Wong, F.S., Li, Y., 2008. Enhancement of filterability in MBR achieved by improvement of supernatant and floc characteristics via filter aids addition. Water Res. 42, 3611–3622. Koseoglu, H., Yigit, N.O., Iversen, V., Drews, A., Kitis, M., Lesjean, B., Kraume, M., 2008. Effects of several different flux enhancing chemicals on filterability and fouling reduction of membrane bioreactor (MBR) mixed liquors. J. Membr. Sci. 320, 57–64. Laera, G., Giordano, C., Pollice, A., Saturno, D., Mininni, G., 2007. Membrane bioreactor sludge rheology at different solid retention times. Water Res. 41, 4197–4203. Lee, J.C., Kim, J.S., Kang, I.J., Cho, M.H., Park, P.K., Lee, C.H., 2001. Potential and limitations of alum or zeolite addition to improve the performance of a submerged membrane bioreactor. Water Sci. Technol. 43, 59–66. Lee, W.N., Chang, I.S., Hwang, B.K., Park, P.K., Lee, C.H., Huang, X., 2007. Changes in biofilm architecture with addition of membrane fouling reducer in a membrane bioreactor. Process Biochem. 42, 655–661. Malamis, S., Katsou, E., Chazilias, D., Loizidou, M., 2009. Investigation of Cr(III) removal from wastewater with the use of MBR combined with low-cost additives. J. Membr. Sci. 333, 12–19. Martinez-Sosa, D., Helmreich, B., Netter, T., Paris, S., Bischof, F., Horn, H., 2011. Anaerobic submerged membrane bioreactor (AnSMBR) for municipal wastewater treatment under mesophilic and psychrophilic temperature conditions. Bioresour. Technol. 102, 10377–10385. Maximous, N., Nakhla, G., Wan, W., Wong, K., 2009. Preparation, characterization and performance of Al2O3/PES membrane for wastewater filtration. J. Membr. Sci. 341, 67–75.

Meng, F., Shi, B., Yang, F., Zhang, H., 2007. New insights into membrane fouling in submerged membrane bioreactor based on rheology and hydrodynamics concepts. J. Membr. Sci. 302, 87–94. Moreau, A.A., Ratkovich, N., Nopens, I., Van der Graaf, J.H.J.M., 2009. The (in)significance of apparent viscosity in full-scale municipal membrane bioreactors. J. Membr. Sci. 340, 249–256. Nguyen, T.T., Ngo, H.H., Guo, W., Phuntsho, S., Li, J., 2011. A new sponge tray bioreactor in primary treated sewage effluent treatment. Bioresour. Technol. 102, 5444–5447. Pham, T.T.H., Brar, S.K., Tyagi, R.D., Surampalli, R.Y., 2010. Influence of ultrasonication and Fenton oxidation pre-treatment on rheological characteristics of wastewater sludge. Ultrason. Sonochem. 17, 38–45. Pollice, A., Giordano, C., Laera, G., Saturno, D., Mininni, G., 2007. Physical characteristics of the sludge in a complete retention membrane bioreactor. Water Res. 41, 1832–1840. Popovic, M., Robinson, C.W., 1984. Estimation of some important design parameters for non-Newtonian liquids in pneumatically-agitated fermenters. In: Proceedings of the 34th Canadian Chem. Engineering Congress, Canada, pp. 258–263. Rosenberger, S., Kubin, K., Kraume, M., 2002. Rheology of activated sludge in membrane bioreactors. Eng. Life Sci. 2, 269–274. Sobeck, D.C., Higgins, M.J., 2002. Examination of three theories for mechanisms of cation-induced bioflocculation. Water Res. 36, 527–538. Wu, J., Chen, F., Huang, X., Geng, W., Wen, X., 2006. Using inorganic coagulants to control membrane fouling in a submerged membrane bioreactor. Desalination 197, 124–136. Wu, Z., Wang, Z., Zhou, Z., Yu, G., Gu, G., 2007. Sludge rheological and physiological characteristics in a pilot-scale submerged membrane bioreactor. Desalination 212, 152–164. Wu, J., Le-Clech, P., Stuetz, R.M., Fane, A.G., Chen, V., 2008. Effects of relaxation and backwashing conditions on fouling in membrane bioreactor. J. Membr. Sci. 324, 26–32. Wu, J., Huang, X., 2008. Effect of dosing polymeric ferric sulfate on fouling characteristics, mixed liquor properties, and performance in a long-term running membrane bioreactor. Sep. Purif. Technol. 63, 45–52. Yang, F., Bick, A., Shandalov, S., Brenner, A., Oron, G., 2009. Yield stress and rheological characteristics of activated sludge in an airlift membrane bioreactor. J. Membr. Sci. 334, 83–90. Yang, X.L., Song, H.L., Lu, J.L., Fu, D.F., Cheng, B., 2010. Influence of diatomite addition on membrane fouling and performance in a submerged membrane bioreactor. Bioresour. Technol. 101, 9178–9184. Yoon, S.H., Collins, J.H., Musale, D., Sundararajan, S., Tsai, S.P., Hallsby, G.A., Kong, J.F., Koppes, J., Cachia, P., 2005. Effects of flux enhancing polymer on the characteristics of sludge in membrane bioreactor process. Water Sci. Technol. 51, 151–157. Zhang, H.F., Sun, B.S., Zhao, X.H., Gao, Z.H., 2008. Effect of ferric chloride on fouling in membrane bioreactor. Sep. Purif. Technol. 63, 341–347.