Tertiary membrane filtration of an industrial wastewater using granular or flocculent biomass sequencing batch reactors

Journal of Membrane Science 382 (2011) 316–322

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Tertiary membrane filtration of an industrial wastewater using granular or flocculent biomass sequencing batch reactors A. Sánchez ∗ , J.M. Garrido, R. Méndez Chemical Engineering Department, School of Engineering, University of Santiago de Compostela, Campus Sur, E-15782 Santiago de Compostela, Spain

a r t i c l e

i n f o

Article history: Received 26 May 2011 Received in revised form 10 August 2011 Accepted 11 August 2011 Available online 22 August 2011 Keywords: Tertiary membrane filtration Fouling Granular biomass

a b s t r a c t Sequencing batch reactors (SBR) are widely used for wastewater treatment. The use of granular biomass in SBR allows higher organic loading rates (OLR). Nevertheless, the main disadvantage of SBR reactors is the presence of suspended solids in the effluent. In this study, effluents from a flocculent biomass SBR (F-SBR) and a granular biomass SBR (G-SBR) were treated in tertiary membrane filtration chambers to remove suspended solids. The performances of the operating systems were compared to determine the influence of the state of the biomass on the filtration process. No significant differences were observed between the two tertiary filtration systems in terms of capacity and permeability. Tertiary filtration of the effluent from the G-SBR was similar to that of the FSBR system. The incorporation of wastewater free of suspended solids during one of the operating stages significantly worsened operation of the tertiary membrane filtration systems; permeabilities decreased by up to 40% in both systems. Additionally, other factors such as nitrification, the presence of soluble microbial products and the concentration of dissolved organic carbon seem to play an important role in tertiary membrane filtration. © 2011 Elsevier B.V. All rights reserved.

1. Introduction In recent years, granular sludge has been proposed as an alternative for high capacity SBR wastewater treatment systems. G-SBR systems can be operated at higher OLRs than F-SBR systems. Arrojo et al. [1] reported organic and nitrogen loading rates of 7 g COD/L d and 0.7 g N/L d, respectively, in a G-SBR. The OLR values recommended for flocculent SBRs in the literature range from 0.5 to 2.0 g COD/L d [2]. The basis of granulation is the continuous selection of sludge particles that occur inside a reactor. The part of the biomass that does not settle quickly enough is washed out with the effluent [2]. However, a main drawback of G-SBR systems is the presence of suspended solids that are washed out with the treated effluent.

Abbreviations: SBR, sequencing batch reactors; OLR, organic loading rates; FSBR, flocculent biomass SBR; G-SBR, granular biomass SBR; COD, chemical oxygen demand; TSS, total suspended solids; MF, microfiltration; UF, ultrafiltration; MBR, membrane bioreactors; PVDF, polyvinylidene fluoride; PLC, programmable logic controller; TMP, transmembrane pressure; SRT, sludge retention time; TOC, total organic car; SMP, soluble microbial products; EPS, extracellular polymeric substances; SRF, specific resistance to filtration; SVI, sludge volume index; SSR, sludge settling rate; VSS, volatile suspended solids; BPC, biopolymer clusters. ∗ Corresponding author. Tel.: +34 881 816 741. E-mail addresses: [email protected] (A. Sánchez), [email protected] (J.M. Garrido), [email protected] (R. Méndez). 0376-7388/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2011.08.027

Therefore, a suitable post-treatment process may be required to fulfil local requirements for the amount of total suspended solids (TSS) in the effluent, which can be accomplished using membrane filtration units, depth filters, surface filters or external settlers [1]. Tertiary filtration, especially depth filtration, has been used to remove suspended solids from secondary treated waters. However, in recent years, the use of tertiary membrane filtration systems is becoming more common. Low-pressure tertiary membranes have been proven to meet stringent standards. Membranes are a physical barrier to suspended solids that are larger than the membrane pore size. Micro- and ultrafiltration (MF/UF) membranes have been used since the early 1990s for drinking water production. They can also be used to remove particulate and colloidal matter from settled secondary effluents, which increases the effectiveness of disinfection with either ultraviolet radiation or ozone for reuse applications [3,4]. Thus, compared to depth filtration, tertiary MF or UF membrane treatments produce water of better microbiological quality that is also free of suspended solids. This should be taken into account when water will be reused or discharged into sensible areas. Fouling is the main drawback associated with the application of membrane technology for wastewater treatment [5–7]. Fouling decreases the permeability of a membrane, limits flux and shortens the life of membrane modules, thus increasing both the capital and the operating costs of filtration systems. Membrane fouling is the result of complex phenomena that are not yet completely understood.

A. Sánchez et al. / Journal of Membrane Science 382 (2011) 316–322

317

Fig. 2. Membrane module.

Fig. 1. Schematic diagram of the laboratory-scale system consisting of a sequencing batch reactor and a tertiary filtration module coupled in series.

The fluxes obtained with submerged UF membranes in the tertiary filtration of secondary wastewaters are normally below 70 L/m2 h, with a backwashing interval of 15–60 min [8–10]. Operational permeabilities of 160–250 L/m2 h bar and fluxes of 25–50 L/m2 h are typical values for pilot units that treat secondary effluents during filtration cycles of 22-27 min. Shortening the backwashing interval and decreasing the flux can slow fouling development [10]. Some studies [6,7,11,12] have quantified the fouling caused by each sludge fraction (suspended solids, colloids and solutes) and shown that colloids are of prime importance in this process. This concept is based on the fact that granular biomass produces less membrane fouling than flocculent biomass in secondary membrane bioreactors (MBR) [13]. Thus, the main objective of this study was to compare the effluents from laboratory-scale G-SBR and F-SBR tertiary filtration systems to determine if effluents with granulated suspended solids cause a different fouling pattern in tertiary membrane filtration systems. 2. Experimental 2.1. Experimental set-up and operating strategy 2.1.1. Operating system Two identical operating systems consisting of an SBR reactor coupled in series with a tertiary filtration chamber (Fig. 1) were operated in parallel. The volumes of the SBRs and filtration chambers were 1.5 L and 1.0 L, respectively. In the first stage, wastewater was treated in the SBR. The SBR effluent then moved into the filtration chamber where suspended solids were removed with an MF membrane module. The submerged membrane modules were constructed with MF polyvinylidene fluoride (PVDF) hollow-fibre membranes, with a pore size of 0.1 ␮m and a membrane area of 0.027 m2 (Fig. 2). The membrane module was prepared in the laboratory using a hollow-fibre membrane manufactured by a Spanish company (Porous Fibres). The number and length of the fibres in each module were similar, and clean water permeability was the same in both modules; tap water at 20 ◦ C produced a value of approximately 300 L/m2 h bar. The membranes were operated

in cycles, with 7 min of permeation and 0.5 min of backwashing (10 L/m2 h). Physical cleaning of the membrane module was performed by rinsing with tap water. Chemically enhanced backwashing was not used during the experiments. Chemical recovery cleaning was conducted when the pressure in the membranes was higher than 30 kPa. The tertiary filtration chambers were aerated to minimise membrane fouling. Aeration rate applied was 24 L/h. Operation of the system was controlled by a programmable logic controller (PLC) Siemens S7-200 connected to a computer. Transmembrane pressure (TMP) data were measured with an electronic pressure sensor IFM Efector500 PN-2009 with an analogue 4–20 mA output and collected on the PC via an analogue PLC module Siemens EM 235 and a data acquisition program. 2.1.2. Influent and operating strategy Industrial wastewater from a fish freezing industry was treated during the study. The raw wastewater contained 50–300 mg/L ammonia and 1000–6300 mg/L total COD. Total phosphorous ranged between 70 and 190 mg/L of which 60–70% was orthophosphate. Soluble COD represented around 85% total COD. Lots of the industrial wastewater were taken and diluted with tap water to obtain the desired OLR and COD values in the influent. The soluble COD values varied from 100 to 1100 mg/L. The total nitrogen concentration ranged from 20 to 180 mg/L, and the total phosphorus concentration ranged from 20 to 110 mg/L. An exchange volume ratio of 50% and a cycle length of 3 h were fixed as operating parameters. The two SBR reactors were operated with constant and similar HRTs and sludge retention times (SRTs) for the entire experimental period. The lengths of the feeding, reaction, settling and withdrawal phases were varied (Table 1) to promote the growth of either granular or flocculent sludge in each SBR reactor. Both SBRs were operated for 114 days prior to the start of this research. The objective of this step was to develop either granular or flocculent sludge with good settling properties prior to coupling the filtration membrane modules to the SBRs [14]. Day 0 was taken as the day on which operation of the tertiary filtration membranes started. The study lasted for 235 days that could be divided into four different experimental stages during which the applied OLR or the concentration of particulate COD was modified. Stage I (from day 0 until day 117). During this stage, both SBRs received a similar OLR of approximately 2 kg COD/m3 d. The objective was to investigate tertiary filtration, with similar operating conditions in both reactors.

318

A. Sánchez et al. / Journal of Membrane Science 382 (2011) 316–322

Table 1 Lengths of the stages during a cycle.

G-SBR (min) F-SBR (min)

Stages

Total

Feeding

Reaction

Settling

Withdrawal

Dead time

3 (aerobic) 60 (anoxic)

147 90

1 20

3 10

26 0

Later, during stages II–IV, the applied OLR in the G-SBR was higher than that of the flocculent reactor, and the same average HRT was maintained in both systems by diluting the wastewater fed to the F-SBR with tap water. Stage II (from day 118 until day 139). The OLR applied to the G-SBR was maintained at approximately 3 kg COD/m3 d. This value was greater than the rate of 1.5 kg COD/m3 d applied to the F-SBR. Stage III (from day 140 until day 210). Wastewater free of suspended solids was fed to both systems. The wastewater was prepared by first removing particulate matter in a centrifugation step and then filtering the supernatant with 0.45-␮m flat-sheet PVDF membrane filters. The objective of this stage was to investigate the effect of suspended solids on tertiary membrane filtration in both systems. Stage IV (from day 210 until day 235). The feeding of diluted wastewater with suspended solids was restarted on day 210 and continued. 2.2. Analytical methods Samples of the influents and effluents from both SBRs and the permeates from the filtration chambers were taken. COD, ammonia, nitrite, nitrate, phosphate and suspended solids concentrations were analysed according to standard methods [15]. Total organic carbon (TOC) was determined with a Shimadzu TOC-500 analyser, and total nitrogen was determined with a DN 1900 analyser (Rosemount, Dohrmann). Samples were taken at the end of the withdrawal period, when the filtration chambers had achieved their maximum operating volumes, a result of the entrance of secondary treated effluent from the SBRs. Dissolved oxygen concentrations in the reactors and filtration chambers were determined with an Oxi 340 WTW oxygen selective electrode, and pH was determined with an U-455 Ingold electrode connected to a Crison 506 pH-meter. Conductivity and temperature were monitored with a CM 35 conductivimeter. Soluble microbial products (SMP) were extracted by centrifuging the biomass for 5 min at 5000 rpm (Heraeus, Labofuge 200) and filtering the supernatant through 0.2-mm glass fibre filters (GF 52, Schleicher and Schuell). Extracellular polymeric substances (EPS) were extracted according to the heating procedure [16]. Protein and carbohydrate concentrations were determined by visible absorbance on a spectrophotometer (Cecil CE 7200), using bovine serum albumin (Sigma) and glucose standards (Merck), respectively. Carbohydrate and protein concentrations were determined according to the methods of Dubois et al. [17] and Lowry et al. [18], respectively. The specific resistance to filtration (SRF) of a sludge sample was determined with the filterability test according to the method of Wisniewski and Grasmick [19]. The test was conducted in a 180-mL stirred cell (Model 8200, Amicon) using a 0.45-␮m flat-sheet PVDF membrane filter (HVLP09050, Millipore). Particle size distributions in the filtration chamber and in the effluents were determined by laser diffraction (Mastersizer HYDRO 2000MU, Malvern Inst.). This method is based on the fact that the diffraction angle is inversely proportional to particle size.

180 180

3. Results and discussion 3.1. SBRs and tertiary filtration system performance Fig. 3 shows the evolutions of the applied OLR and the HRT, referred to the reactor volume, during the four experimental stages. The HRTs were similar in both reactors for the entire period. Moreover, during stage I (from day 0 until day 117), both units were operated with similar COD concentrations in the feed water and applied OLR rates. The OLR for both SBRs was approximately 2 kg COD/m3 d prior to the start of this research [14]. At the operating day 0, the OLR was increased from 2 to 4 kg COD/m3 d to achieve better granular biomass properties (Fig. 3). Nevertheless, this increase worsened the settling properties of the F-SBR sludge in terms of the sludge volume index (SVI) and the sludge settling rate (SSR). This led to a partial wash out of the biomass in this reactor. For this reason, on day 19, the OLR was decreased to approximately 2.5 kg COD/m3 d in both reactors and maintained at this level until the end of stage I. During stages II–IV, the applied OLR was higher in the G-SBR than in the F-SBR; the HRTs were constant in both systems. GSBRs are typically operated at higher OLRs (2.5 and 15 kg COD/m3 d [20,21]) than F-SBRs (0.5–2.0 kg COD/m3 d [2]). The wastewater fed to the F-SBR had a lower COD concentration (between 100 and 400 mg/L) than that fed to the G-SBR (between 400 and 800 mg/L). The objective of stages II–IV was to compare the behaviour of the two SBRs at their recommended OLRs and of the respective tertiary filtration chambers. Fig. 4 depicts the evolution of the overall COD removal percentage in both systems. The COD removal efficiency was 90% for both systems in stage I, during which wastewaters with similar COD values were fed to the units. The presence of either flocculent or granular biomass in SBR systems with tertiary filtration chambers did not affect COD removal efficiency, in terms of soluble COD. After stage I, the COD removal efficiency in the F-SBR was lower than that in the G-SBR because of the lower COD concentration in the feed water of the F-SBR during stages II–IV. COD removal efficiencies in the flocculent system ranged between 66 and 97%, whereas

6

6

12

←HRT

6

5

OLR (kg COD/m3·d)

Reactor

II

Stage I

III

IV

4 3 2 1 0

0

50

100

150

200

250

Time (d) Fig. 3. Evolution of the applied organic loading rate in the () F-SBR and (䊉) G-SBR systems during the four experimental stages.

A. Sánchez et al. / Journal of Membrane Science 382 (2011) 316–322 100

Removal (%)

90 80 70 Stage I

60

II

III

IV

50 0

50

100

150

200

250

Time (d) Fig. 4. Overall COD removal efficiency for the F-SBR () and G-SBR () systems.

for the granular system, these efficiencies varied between 89 and 96%. The evolution of soluble COD in the effluent of the biological reactor is depicted in Fig. 5A. In Fig. 5B, the evolution of soluble COD in the permeates of the two filtration chambers is shown. The ratio between soluble and total COD in the feed water was 0.81 ± 0.09 g/g. During stage I, similar soluble COD concentrations were observed in the effluents of the SBRs and in the permeates of both tertiary filtration chambers. The COD concentration in the SBR effluents gradually decreased from 140 mg/L to 20 mg/L (Fig. 5). A similar trend was observed for the filtration chamber permeates. During stages II–IV, the COD concentration in the permeate of the granular system was higher than that observed in the flocculent system because the former system was fed with water with a higher COD concentration. It was also observed that COD in the permeate was higher than the SBR effluent for the GSBR system. A possible reason of such behaviour is presented in Section 3.2. One additional objective of this study was to investigate the role of particulate COD in the efficiencies of the SBRs and the tertiary filtration chambers. The presence (stages I, II and IV) or absence (stage III) of particulate matter did not affect the COD removal efficiencies of the SBRs or the filtration chambers. No significant nitrification was observed in either SBR. During two periods, nitrification was initiated, but the OLR increase and the dissolved oxygen limitation interrupted this biological process. Nevertheless, ammonia was oxidised to either nitrite or nitrate in both tertiary filtration chambers, especially after operating day 100 (Fig. 6). Biomass concentration ranged between 0.5 and 5.0 and 3.0 and 10.0 g/L in the flocculent and granular SBRs, respectively. The tertiary membrane filters were operated with high concentrations

of suspended solids. Volatile suspended solids (VSS) concentrations in the tertiary filtration chambers were between 0.3 and 6.0 and 0.3 and 6.8 g/L VSS (Fig. 7) as a result of the operating strategies of the filtration systems. The permeate used to backwash the membranes accumulated in the tertiary filtration chamber and was not discharged from the filter. The filter was punctually purged to maintain similar VSS values in the SBRs. Moreover, the filtration chambers were continuously aerated to reduce membrane fouling and to provide oxygen to the suspended, washed-out biomass. The tertiary membrane filtration chambers acted as a biological polishing stage and caused variations in the COD and nitrogen concentrations (Figs. 5 and 6). As expected, the SVI and SSR sludge settling properties were significantly different for the two systems. The differences between the settling properties of the biomasses developed in both reactors were remarkable. In the F-SBR, the SVI values were between 50 and 300 mL/g, while in the G-SBR, the SVI values ranged from 30 to 100 mL/g. Moreover, the sludge settling rates were 0.7 m/h and 9.7 m/h for the flocculent and granular sludges, respectively. 3.2. Tertiary membrane filtration One of the main objectives of this study was to determine if effluent from a granular biomass bioreactor caused less membrane fouling than that from a flocculent bioreactor. Nevertheless, permeability values between 160 and 75 L/m2 h bar were observed in the two tertiary membrane filtration systems at a flux of 10 L/m2 h, indicating that tertiary membrane filtration of both effluents produced similar results. Moreover, the permeability evolutions of the two membrane modules were similar (Fig. 8). For both membranes, the maximum permeability value was achieved after performing either a physical or a chemical cleaning procedure. Permeability gradually decreased until the subsequent cleaning. These results are different from those reported by Li et al. [13], who compared two submerged MBRs with flocculent and granular sludges. Nevertheless, in that study, the granular sludge was cultivated from anaerobic granular sludge, and the bioreactors were inoculated directly. The permeability values obtained during tertiary filtration were significantly better than those obtained previously in a pilot-scale MBR treating municipal wastewater with the same fibre. The permeabilities in the MBR varied between 50 and 70 L/m2 h bar [22]. The TMP behaviour was similar for both membranes during the two periods. During stage I, the membrane in the G-SBR system performed slightly better. The objective of stage III was to assess the effect of particulate COD on the permeability of both filtration systems and on

200

200

A

180 160

StageI

III

II

B

180 160

IV

Permeate COD (mg/L)

SBR effluent COD (mg/L)

319

140 120 100 80 60 40 20

StageI

III

II

IV

140 120 100 80 60 40 20

0

0 0

50

100

Time (d)

150

200

250

0

50

100

150

200

250

Time (d)

Fig. 5. Soluble COD concentration in the effluent of both SBRs (A) and in the permeate of the filtration chambers (B) of either F-SBR () or G-SBR () systems.

A. Sánchez et al. / Journal of Membrane Science 382 (2011) 316–322

200

160

Stage I

180

III

II

B

IV

TN and N-NH4 (mg/l)

140

+

+

A TN and N-NH4 (mg/l)

320

120 100 80 60 40

200

Stage I

180

III

II

IV

160 140 120 100 80 60 40

20

20

0

0 0

50

100

150

200

250

0

50

100

Time (d)

150

200

250

Time (d)

Fig. 6. Total Nitrogen in the feeding (䊉) and ammonia concentration in the effluent of SBR () and the permeate of filtration chamber () in the F-SBR (A) and the G-SBR (B) systems.

10 Stage I

II

III

B

IV

10 Stage I

8

8

6

6

VSS (g/l)

VSS (g/l)

A

4

II

III

IV

4 2

2

0

0 0

50

100

150

200

250

Time (d)

0

50

100

150

200

250

Time (d)

Fig. 7. VSS concentration in SBR reactor () and tertiary filtration chamber () in F-SBR (A) and G-SBR (B) systems.

SBR behaviour. During this stage, the feed wastewater was first centrifuged, and the supernatant was filtered later to remove particulate matter in the influent. During this study, it was assumed that the permeability of the systems fed with wastewaters with low TSS values, such as primary treated wastewaters, would be higher than that of systems fed with raw pre-treated water. Nevertheless, the permeability values in both systems during stage III were lower than those observed during stages I, II and IV. Permeability was below 100 L/m2 h bar during this stage (grey area in Fig. 8).

Membrane fouling in both systems was clearly more severe during this stage because suspended solids were removed from the influent. The TSS concentration in the wastewater ranged from 0.05 to 0.20 g/L, depending on the COD concentration. Volatile suspended solids represented 75% of the TSS. In addition to the presence of suspended solids in the feed wastewater, nitrification also affected tertiary membrane filtration. The performance of both membranes was better when nitrification did not occur than when full ammonia oxidation took place

Fig. 8. Permeability evolution in the F-SBR () and G-SBR (䊉) systems. The grey area represents stage III, during which the feed was filtered. The observed peaks in permeability correspond to those days on which chemical recoveries were performed.

A. Sánchez et al. / Journal of Membrane Science 382 (2011) 316–322

100

200 Stage I

160

III

II

90

IV

80

Volume (%)

SMP carbohydrates (mg/L)

321

120 80

70 60 50 40 30

40

20 10

0 0

50

100

150

200

0

250

1

10

Time (d) Fig. 9. Carbohydrates SMP concentration in F-SBR () and G-SBR (䊉) filtration chambers.

3.3. Specific cake resistance and particle size distribution The particle size distributions of the solids in the chambers indicated that the particles in the flocculent system’s filtration chamber were smaller than those in the chamber of the granular system (Fig. 11). The mean particle size in the flocculent and granular filtration chambers were 80 and 250 ␮m, respectively. Moreover, 50% of the particles in the granular filtration chamber were larger than 1000 ␮m. The sludge filterability test produced similar results for both membrane chamber liquors. As shown in Table 2, the specific resistance to filtration (SRF) values were generally lower for the granular liquor. The SRF values (1011 to 1013 m/kg) are similar to those

100 80 60 40 20 0 Stage III

Stage IV

180 160

B

140

BPC (mg/L)

2

120

Stage I

10000

COD was interrupted. The BPC concentrations in the granular filtration chamber during stages III and IV were higher, which seemed to influence permeability in the flocculent system membrane. The influence of BPC concentration on permeability was analysed by comparing stages I, III and IV in the filtration chamber of the granular system. Comparing the OLRs of stages I, III and IV in the filtration chamber of the flocculent system also exposed this effect; the OLR was considerably lower during stage IV. Average values were calculated for the days on which the BPC concentration was measured, and these values were complemented with their standard deviations. Therefore, the BPC concentration in the filtration chamber might be representative of membrane fouling when the OLR was held constant (Fig. 10B). Conversely, this parameter was not indicative of membrane fouling in the flocculent system because the OLR decreased during operation (Fig. 10A). Thus, the BPC concentration might be related to membrane permeability only when the OLR was held constant.

Permeability (L/m ·h·bar)

A

140

BPC (mg/L)

2

Permeability (L/m ·h·bar)

160

1000

Fig. 11. Cumulative volume percentage distributions of the flocculent () and granular (1 mm sifted) (䊉) sludges on a sample taken during operating day 210 (period III).

(approximately 30 mg N-NOx/L), although no significant differences were observed between these scenarios. However, when ammonia was partially oxidised (approximately 10 mg N-NOx/L) in the filtration chambers, the membrane in the granular system performed better than the one in the flocculent system, reaching 50% lower pressures after the critical flux [14]. The effects of ammonia nitrification variations on membrane behaviour have been reported for secondary MBR systems [23]. Dissolved oxygen provides an oxygen source for the mineralisation of SMP, and nitrate also seems to influence SMP elimination. Elevated polysaccharide rejection and high polysaccharide concentrations of up to 150 mg/L were reported by Drews et al. [24] during a period of low nitrification activity. However, fouling was simultaneously low, which could indicate that under these conditions, SMP were too large to cause internal fouling and formed a loose cake instead [23]. The EPS concentrations in the liquor of the two tertiary filtration chambers were similar (40–100 mg/g VSS). Nevertheless, the SMP concentrations (25–75 mg/g VSS) in the tertiary filtration chamber of the granular system were significantly higher than those (10–40 mg/g VSS) in the chamber of the flocculent system (Fig. 9). Tay et al. [25] showed that aerobic granular sludges excrete more SMP than conventional bioflocs and biofilms. These smaller compounds are formed during the hydrolysis of VSS, which increases the COD concentration of the permeate. This fact could explain the higher soluble COD values in the permeate of this system (average 60 mg/L) compared to the effluent of the biological reactor (average 40 mg/L) (Fig. 5A and B). Biopolymer clusters (BPC) are a special form of organic materials formed by the affinity clustering of free EPS or SMP in the sludge cake on the membrane surface [26]. Fig. 10 shows the average permeability and standard deviation values versus the BPC concentrations during three different stages. As previously indicated, the lowest permeabilities were measured in stage III for both systems, corresponding to the period during which the supply of particulate 180

100

Particle Size (μm)

120 100 80 60 40 20 0 Stage I

Stage III

Fig. 10. Relation between permeability () and BPC () in F-SBR (A) and G-SBR (B) filtration chambers.

Stage IV

322

A. Sánchez et al. / Journal of Membrane Science 382 (2011) 316–322

Table 2 Specific cake resistance of liquors from the F-SBR and G-SBR membrane chambers. Day 70 104 112 133 156 190

Stage

Floc SRF (m/kg)

Granular SRF (m/kg)

I I I II III III

1.5 × 10 8.6 × 1012 5.1 × 1011 1.4 × 1013 8.4 × 1013 5.8 × 1012

7.0 × 10 1.3 × 1013 4.7 × 1011 4.4 × 1012 8.4 × 1012 4.0 × 1012

13

12

obtained by Pollice et al. [27] (1011 to 1013 m/kg) and to the typical values observed for CAS [8] (1–10 × 1013 m/kg). However, they were markedly lower than previously reported SRF values for MBRs of approximately 1015 m/kg [28]. The SRF values were similar for the liquors from both membrane filtration chambers. Moreover, similar SRF values were observed in period I but not in period III, during which the feed wastewater was free of solids. Theoretical and experimental studies have shown that fine particles cause severe membrane fouling [29]. Smaller aggregates increase the resistance of the cake formed, and the critical flux is reached at lower values [30]. However, despite observed differences in the particle size distributions, this phenomenon was not observed, as the SRF values were similar for both systems. This result likely indicates that other factors, such as the presence of BPCs, can impact tertiary membrane filtration more significantly than the size distribution. 4. Conclusions Two SBR reactors with tertiary filtration chambers coupled in series were operated in this study. Different forms of biomass, flocculent and granular formed in each reactor, depending on the OLR and the SBR cycle times. The main conclusions are listed below. Flocculent and granular SBR reactors provided optimal organic matter removal, with values of 90% in both systems. The permeability, transmembrane pressure and critical flux behaviours of both membranes were similar. Thus, no significant operational differences between the G-SBR and the F-SBR were observed. The experimental results indicate that the presence of suspended solids in the influent and the nitrification process more significantly affected membrane performance than the morphology of the aggregated biomass. The positive effects of suspended solids in wastewater from the fish freezing industry on membrane performance were demonstrated, and permeability was improved. Because SRF was not affected by the absence of suspended solids, the lower observed permeabilities could have been a result of changes in the soluble matrix, but the exact mechanism is unknown. This study confirmed the importance of the carbohydrate fraction of SMP as one of the most important parameters related to membrane fouling, and the observed trend is identical to that of the BPC concentration. A certain trend between BPC concentration and permeability, especially at a constant OLR, was observed. Therefore, organic carbon is suggested as an indicator of membrane fouling. To obtain the most appropriate tertiary filtration system, future investigations should focus on more specific aspects to identify accurately the differences between the systems and to determine the impact of suspended solids in the feed solution. Acknowledgements The Spanish Ministry of Education and Science (project Novedar-Consolider CSD2007-00055) and the Xunta de Galicia (PGIDT 09MDS009265PR) provided funding for this research.

References [1] B. Arrojo, A. Mosquera-Corral, J.M. Garrido, R. Méndez, Aerobic granulation with industrial wastewater in sequencing batch reactors, Water Res. 38 (2004) 3389–3399. [2] J.J. Beun, A. Hendriks, M.C.M. van Loosdrecht, E. Morgenroth, P.A. Wilderer, J.J. Heijnen, Aerobic granulation in a sequencing batch reactor, Water Res. 33 (1999) 2283–2290. [3] G. Tchobanoglous, J. Darby, K. Bourgeous, J. McArdle, P. Genest, M. Tylla, Ultrafiltration as an advanced tertiary treatment process for municipal wastewater, Desalination 119 (1998) 315–322. [4] C. Lubello, R. Gori, A.M. de Bernardinis, G. Simonelli, Ultrafiltration as tertiary treatment for industrial reuse, Water Sci. Technol.: Water Supply 3 (2003) 161–168. [5] K. Kimura, N. Yamato, H. Yamamura, Y. Watanare, Membrane fouling in pilotscale membrane bioreactors (MBRs) treating municipal wastewater, Environ. Sci. Technol. 39 (2005) 6293. [6] F. Meng, S.R. Chae, A. Drews, M. Kraume, H.S. Shin, F. Yang, Recent advances in membrane bioreactors (MBRs): membrane fouling and membrane material, Water Res. 43 (2009) 1489–1512. [7] A. Drews, Membrane fouling in membrane bioreactors – characterisation, contradictions, cause and cures, J. Membr. Sci. 363 (2010) 1–28. [8] Metcalf & Eddy, Wastewater Engineering: Treatment and Reuse, 4th ed., McGraw-Hill, New York, 2003. [9] G. Pearce, Membrane system design, in: M. Wilf (Ed.), The Guidebook to Membrane Technology for Wastewater Reclamation, Desalination Publications, Hopkinton (MA), USA, 2010, pp. 187–245. [10] X. Zheng, M. Ernst, M. Jekel, Stabilizing the performance of ultrafiltration in filtering tertiary effluent – technical choices and economical comparisons, J. Membr. Sci. 366 (2011) 82–91. [11] B. Lesjean, S. Rosenberger, C. Laabs, M. Jekel, R. Gnirss, G. Amy, Correlation between membrane fouling and soluble/colloidal organic substances in membrane bioreactors for municipal wastewater treatment, Water Sci. Technol. 51 (2005) 1. [12] W. Lee, S. Kang, H.S. Shin, Sludge characteristics and their contribution to microfiltration in submerged membrane bioreactors, J. Membr. Sci. 216 (2003) 217–227. [13] X.F. Li, F.S. Gao, Z.Z. Hua, G.C. Du, J. Chen, Treatment of synthetic wastewater by a novel MBR with granular sludge developed for controlling membrane fouling, Sep. Purif. Technol. 46 (2005) 19–25. [14] A. Sánchez, J.M. Garrido, R. Méndez, A comparative study of tertiary membrane filtration of industrial wastewater treated in a granule and a flocculent sludge SBR, Desalination 250 (2010) 810–814. [15] APHA-AWWA-WEF, Standard Methods for the Examination of Water and Wastewater, 20th ed., American Public Health Association/American Water Works Association/Water Environment Federation, Washington, DC, 1998. [16] X. Zhang, P.L. Bishop, B.K. Kinkle, Comparison of extraction methods for quantifying extracellular polymers in biofilms, Water Sci. Technol. 39 (1999) 211–218. [17] M. Dubois, K.A. Gilles, J.K. Hamilton, P.A. Rebers, F. Smith, Colorimetric method for determination of sugars and related substances, Anal. Chem. 28 (1956) 350–356. [18] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein measurement with the Folin Phenol Reagent, J. Biol. Chem. 193 (1951) 265–275. [19] C. Wisniewski, A. Grasmick, Floc size distribution in a membrane bioreactor and consequences for membrane fouling, Colloids Surf. A 138 (1998) 403–411. [20] B.Y.-P. Moy, J.-H. Tay, S.-K. Toh, Y. Liu, S.T.-L. Tay, High organic loading influences the physical characteristics of aerobic sludge granules, Lett. Appl. Microbiol. 34 (2002) 407–412. [21] Q.S. Liu, J.H. Tay, Y. Liu, Substrate concentration-independent aerobic granulation in sequential aerobic sludge blanket reactor, Environ. Technol. 24 (2003) 1235–1243. [22] P. Artiga, G. García-Torrielo, J.M. Garrido, R. Méndez, Membrane bioreactor: an advanced technology for the treatment and reuse of waste waters, Tecnología del agua 26 (2006) 54–60. [23] A. Drews, J. Mante, V. Iversen, M. Vocks, B. Lesjean, M. Kraume, Impact of ambient conditions on SMP elimination and rejection in MBR, Water Res. 41 (2007) 3850–3858. [24] A. Drews, M. Vocks, U. Bracklow, V. Iversen, M. Kraume, Does fouling in MBRs depend on SMP? Desalination 231 (2008) 141–149. [25] J.H. Tay, Q.S. Liu, Y. Liu, The role of cellular polysaccharides in the formation and stability of aerobic granules, Lett. Appl. Microbiol. 33 (2001) 222–226. [26] X.-M. Wang, X.-Y. Li, Accumulation of biopolymers clusters in a submerged membrane bioreactor and its effect on membrane fouling, Water Res. 42 (2008) 855–862. [27] A. Pollice, C. Giordano, G. Laera, D. Saturno, G. Mininni, Physical characteristics of the sludge in a complete retention membrane bioreactor, Water Res. 41 (2008) 1832–1840. [28] N. Cicek, J.P. Franco, M.T. Suidan, V. Urbain, J. Manem, Characterization and comparison of a membrane bioreactor and a conventional activated-sludge system in the treatment of wastewater containing high-molecular weight compounds, Water Environ. Res. 71 (1999) 64–70. [29] R. Bai, H.F. Leow, Microfiltration of activated sludge wastewater – the effect of system operation parameters, Sep. Purif. Technol. 29 (2002) 189–198. [30] G. Belfort, R.H. Davis, The behavior of suspensions and macromolecular solutions in crossflow microfiltration, J. Membr. Sci. 96 (1994) 1–58.