On: 26 April 2014, At: 04:01 Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Separation Science and Technology
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/lsst20
Behavior of Membrane Scaling During
Crossflow Filtration in the Anaerobic MBR
System
H. S. You ab , C. P. Huang b , J. R. Pan b & S. C. Chang a a
Industrial Technology Research Institute (ITRI)/Energy and Environment Research Laboratories , Hsinchu, Taiwan
b
Institute of Environmental Engineering, National Chiao Tung University , Hsinchu, Taiwan
Published online: 15 Feb 2007.
To cite this article: H. S. You , C. P. Huang , J. R. Pan & S. C. Chang (2006) Behavior of Membrane Scaling During Crossflow Filtration in the Anaerobic MBR System, Separation Science and Technology, 41:7, 1265-1278, DOI: 10.1080/01496390600632487
To link to this article: http://dx.doi.org/10.1080/01496390600632487
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
Behavior of Membrane Scaling During
Crossflow Filtration in the Anaerobic
MBR System
H. S. You
Industrial Technology Research Institute (ITRI)/Energy and Environment Research Laboratories, Hsinchu, Taiwan and Institute of
Environmental Engineering, National Chiao Tung University, Hsinchu, Taiwan
C. P. Huang and J. R. Pan
Institute of Environmental Engineering, National Chiao Tung University, Hsinchu, Taiwan
S. C. Chang
Industrial Technology Research Institute (ITRI)/Energy and Environment Research Laboratories, Hsinchu, Taiwan
Abstract:Fouling of membrane bioreactor (MBR) has been studied intensively. Because of the high concentration of carbonates, scaling can be a serious problem in anaerobic bioreactor, which attracts little attention. In this study, the wastewater was treated with an anaerobic process followed by either a submerged or a side-stream aerobic membrane reactor. The wastewater was spiked with calcium to investigate the effect of scaling on membrane filtration. Very little scaling was detected in the external mem-brane system (the side-stream MBR). Results from chemical cleaning of internal membrane system indicated that the flux decline caused by membrane scaling was far more severe than that by membrane fouling. However, the flux decline from membrane scaling can be effectively recovered by the chemical cleaning of EDTA and NaOCl.
Keywords:Membrane bioreactor, scaling, fouling, chemical cleaning
Received 10 January 2006, Accepted 20 March 2006
Address correspondence to C. P. Huang, Institute of Environmental Engineering, National Chiao Tung University, Hsinchu, Taiwan. Tel: þ886-3-5726463; Fax: þ886-35-725958; E-mail: cphuang@mail.nctu.edu.tw
Copyright # Taylor & Francis Group, LLC ISSN 0149-6395 print/1520-5754 online DOI: 10.1080/01496390600632487
1265
INTRODUCTION
Membrane bioreactor (MBR) is a new technology combining a bioreactor and a membrane filtration unit for the separation of biomass, of which the merits have been discussed intensively (1 – 3). Because nearly all biomass can be retained within the reactor, MBR can maintain a stable and high quality effluent (3). One more advantage of MBR is that it can be operated with long solids retention time (SRT), which is especially beneficial to slow-growing microorganisms (4). As a result, decomposition of refractory organic compounds can be achieved even under short hydraulic retention time (HRT) (5). Although MBR has many advantages, membrane fouling hinders its application tremendously (6). Membrane fouling not only results in reduced flux which demands more frequent membrane cleaning but also shortens the membrane life. Therefore, most research on MBR concentrate on controlling membrane fouling (7, 8).
Most MBRs are operated in aerobic condition. However, anaerobic MBRs have also been studied (9, 10). Unless the permeate is recycled, most anaerobic MBRs must be followed by another aerobic treatment unit to meet the discharge standard. To date, there has been little research studying aerobic MBRs associated with anaerobic units (11, 12).
Besides CO2and water, anaerobic biodegradation also produces CH4.
Because the aerobic digester is an open tank, the CO2concentration at the
top of it is identical to atmospheric CO2, 0.03%. The anaerobic tank, on
the other hand, is enclosed and is operated at 30 to 50 cm hydraulic head. Therefore, the gaseous space of the anaerobic tank contains 20 to 40% CO2 and 60 to 80% CH4. As a result of the increased CO2 partial
pressure, the solubility of CO2 is much higher than usual. Thus, in order
to maintain a neutral pH environment for microorganisms, alkalinity in the anaerobic tanks must be supplemented, increasing the concentration of carbonate ions. When the effluent of the anaerobic reactor flows into the aerobic reactor, the supersaturated CO2 is released and the pH of the
liquid phase rises substantially. In addition, the carbonates of metals such as Ca, Mg, and Fe crystallize, increasing the potential of membrane scaling (12).
The resistance-in-series model has been widely used to describe the filtration behavior of membranes (13):
J ¼ DPT ðhRtÞ
ð1Þ
Rt¼RmþRcþRf ð2Þ
where J is the permeation flux, DPTis the transmembrane pressure (TMP),his
the viscosity of the permeate; Rt is the total resistance; Rm is the intrinsic
membrane resistance; Rc is the cake resistance formed by the cake layer
resistance; and Rf is the resistance caused by solute adsorption into the
membrane pores and walls. For a MBR susceptible to scaling, the Rfand Rt
must be adjusted:
Rf ¼RorgfþRinorgs ð3Þ
Rt¼RmþRcþRorgfþRinorgs ð4Þ
where Rorg2fis the resistance caused by organic adsorption into the membrane
pores and walls; Rinorg2s is the resistance caused by inorganic scaling in
membrane pores and on walls.
Membrane fouling is a result of the reaction between the membrane and particles in the suspension. The floc-structure of the activated sludge, particle size distribution, and extracellular polymeric substances (EPS) content of the activated sludge are all related to membrane fouling (7). EPS is either produced by the microorganisms or the decomposition products of the dead cells. It bears a negative charge at neutral pH due to the functional groups of carboxyl, amino, and phosphate (4).
Carbonate compounds, on the other hand, bear a slight positive charge. The bioflocs will adsorb the carbonate crystals through charge neutraliz-ation, and the potential to foul or scale on the membrane is greatly reduced through a mechanism called bio-flocculation (14). The aim of this research is to investigate the effect of fouling and scaling on membrane filtration during MBR operation and the efficiency of chemical cleaning in the recovery of membrane flux. This MBR follows the anaerobic reactor that the carbonate crystals have the higher potential to form than normal aerobic MBR.
EXPERIMENTAL
Design of the MBR Process
The schematic of the MBR process is shown in Fig. 1, which follows an Upflow Anaerobic Sludge Bed (UASB). The influent substrates of UASB were composed of glucose and acetic acid. The influent COD of UASB ranged from 12,000 to 16,000 mg/L. The effluent of UASB, with 1500 to 2500 mg/L COD, pH of 6.8 to 7.2, and alkalinity between 1500 and 2000 mg/L as CaCO3, overflowed into a mixing tank in which the
desired amount of CaCl2 was added. Two types of MBRs were designed
to test the effect of the relative location of the membrane module to the bioreactor on scaling. One was the internal submerged MBR (internal membrane system), of which the membrane module was placed inside the bioreactor. The other one is the external submerged MBR (external membrane system), of which the membrane module was separated from the biological reactor. The volume of the biological reactor of the external membrane system was 20 L, and the membrane tank was 8 L.
The total volume of the internal membrane system was 28 L. The reactor was partitioned into two cells: the reactor area (20 L) and the membrane area (8 L). The flow rate of the sparging gas was 1 m/sec (15). No back-feeding flushing was performed during the entire period of membrane operation. The aeration rate was controlled to satisfy the biochemical reaction of the bioreactor.
The feeding during membrane operation was controlled by the liquid level controller. The permeate was accumulated in a buffer tank with balance, liquid level controller and magnetic valve to calculate the flux. The amount of the permeate and the TMP were monitored automatically by a computer with the SCADA (Supervisory Control And Data Acquisition) and data analysis. In the external membrane system the recycled flow from the membrane tank to the bioreactor was three times that of the feed into to the bioreactor. The MLSS of both systems were maintained at 8000 + 500 mg/L. The CaCl2,
50 – 350 mg/L as Ca, was added after 70 hours of operation. The feed to the
Figure 1. Schematic diagram of the experimental setup.
MBR varied with the flux. Therefore, the HRT of the MBR was lengthened with the decline in flux.
Membranes and Module Configuration
The membrane was obtained from Zenon Environmental Inc., and was an immersed ultrafiltration (UF) type of pore size 0.036 mm. The hollow fiber has an inner dia. of 0.9 mm and outer dia. of 1.9 mm. The membrane is resistant to strong acids and bases (operational pH of 3 – 12) and corrosion by chemicals like OCl2. To prevent accumulation of microorganisms between fibers, the module was modified from the ZW-1, as shown in Fig. 2. The TMP of the intrinsic membrane was 4 kPa and the membrane surface area was 0.046 m2. The initial flux was set at 14 L/m2-hr.
PHYSICAL AND CHEMICAL CLEANING
Cleaning by back-feeding was not adopted during the entire period of membrane operation. When the flux declined to 30% of the original flux, the operation was terminated and the membrane was removed from the
Figure 2. Membrane modules (Left: ZW-1 module; Right: modified module).
bioreactor and placed in the cleaning tank for cleaning, as shown in Fig. 3. During soaking cleaning, in addition to air flushing, the pump was started to remove the permeate. The TMP and the flux were monitored. When the volume of the accumulated permeate reached a predetermined value, the electric valve would open to allow the permeate to flow into the cleaning tank. The cleaning is finished when the TMP and the flux would no longer change with the cleaning time.
Membrane fouling can be attributed to three mechanisms: cake formation, inorganic scaling and organic fouling. In this study, the fouled membrane was cleaned with physical and chemical cleaning process to differentiate the con-tribution of each mechanism to membrane fouling. First, the membrane was clean with DI water and air flushing. The air rate was 2 m/sec, twice the flow rate of the sparging gas. The reduction in TMP after the physical cleaning was attributed to the cake formation. The cleaning was followed by the chemical cleaning of EDTA solution and air flushing. The flow rate of the air was 2 m/sec, and 2000 mg/L EDTA was added. The reduction in TMP after this step was attributed to inorganic scaling. The last step was the chemical cleaning by NaOCl solution and air flushing. The flow rate of the air was 2 m/sec, and 2000 mg/l NaOCl as OCl2 was added to dissolve the organic fouling. The reduction in TMP after the chemical cleaning by NaOCl was attributed to organic fouling.
Figure 3. Physical and chemical cleaning equipment.
Assessment of Membrane Performance
During MBR operation, the calcium ion concentration in the influent and the permeate were analyzed with an atomic absorbance spectrophotometer (Varian Spectra AA-30). After the operation, the membrane was removed from the membrane module and the chemical composition of the scaling was analyzed by a scanning electron micrograph—energy dispersive spec-trometry (SEM – EDS, Oxfoxd, ISIS 300); the inorganic scaling was analyzed with X-ray Diffraction (Bruker AXS D8 ADVANCE).
RESULTS AND DISCUSSION
Scaling of the Two MBR Systems
The scaling of the two MBR systems was magnified by the seeding of Ca. The influent of MBR which is generated from the UASB contained 1500 to 2500 mg/L COD, and the COD of the permeate remained between 50 to 100 mg/L during the entire course of operation. The pH of the influent was approximately 6.8 – 7.2 which increased slightly to approximately 8.2 – 8.5 in the bioreactor due to the stripping of CO2. The initial flux of both MBRs
was set at 14 L/m2-hr. At the beginning of the operation, no Ca was added. After 70 hours of operation, various amounts of CaCl2, namely, 50 and
350 mg/L as Ca, were added. The variation of Ca concentration in the permeate is shown in Fig. 4. Although the Ca concentration of the permeate increased with the increasing Ca concentration of the influent, the correlation between the two systems were similar and the permeate Ca concentration was between 20 and 80 mg/L as Ca.
The variations of flux and TMP during MBR operations are shown in Fig. 5. Both Figs. 5(a) and (b) show that the flux and the TMP stayed fairly constant in both systems before the addition of Ca. Dramatic discrepancy was observed between the two systems after the addition of calcium. In the internal membrane system, as shown in Fig. 5(a), two phases of change were observed. In the first phase, the TMP increased gradually from 7 to 20 kPa, while the flux decreased from 14 to 11 L/m2-hr. In the second phase, the TMP increased abruptly to 50 kPa with a sudden flux decline to 4 L/m2-h, which led to chemical cleaning because of the reduction of over 70%. On the other hand, the effect of CaCl2addition on the flux and TMP
of the external membrane system was minor, as shown in Fig. 5(b). It implies that the external membrane system is probably a better choice when scaling-causing metal ions are present.
Scaling was a complicated process, including crystallization and hydro-dynamic transport (16). Two mechanisms involved in crystallization are bulk (homogeneous) and surface (heterogeneous) crystallizations (16, 17). In bulk crystallization, supersaturated brine gives rise to the agglomeration
of scale-forming ions due to random collisions of the ions in motion. The cluster of ions coalesces to form a crystal inducing precipitation upon reaching a critical size. On the other hand, surface crystallization may be caused by foreign bodies such as the membrane surface itself. As a result, membrane fouling occurs by the deposition of a crystal particle formed in a bulk phase and/or the lateral growth of crystals on a membrane surface. To prevent membrane scaling, both clusters of ions or free ions and the crystals must be free from contact with the membrane.
In the membrane system, aeration in the bioreactor stripped CO2from the
liquid phase resulting in an increase in pH and the formation of Ca crystals bearing a slight positive charge. The bioflocs in the bioreactor are negatively charged due to functional groups such as – COOH, – NH2and – PO4. Through
charge neutralization, Ca crystals as well as free calcium were caught by the flocs and removed from the liquid phase. In the external membrane system, the direct contact of crystals and free cations with the membrane is avoided which prevents future membrane scaling. The bioflocculation between bioflocs and crystals has also been clarified in our previous paper (12, 14).
Furthermore, because of the difference in the aeration rate and sparging rate of the internal membrane system, internal circulation occurs inside the bioreactor. The circulation of the metal clusters or metal ions in the system continuously moves toward the surface of the membrane before they have a chance to react with the bioflocs. The scaling of the small crystals gradually blocked the membrane pore which caused declined flux. When large areas
Figure 4. Ca2þconcentration of permeate as a function of different influent Ca2þ concentration.
of the membrane were blocked, membrane fouling occurred, as seen in the sudden increase in the TMP in Fig. 5(a).
Analysis of the Membrane Surface
After each cycle of operation, the hollow fibers were removed from the membrane module of both MBRs for photographs and mineral identification.
Figure 5. Effect of CaCl2addition on permeate flux and TMP profiles under different
MBR systems. (a) internal membrane system, (b) external membrane system.
The results are shown in Fig. 6. The surface of the membrane from external membrane system is smooth and free of deposit. The photographs and results of the analysis demonstrate that the membrane from the internal membrane system was covered with crystals. The peak, as seen in Fig. 6(b-2), in the SEM – EDS diagram indicates the scale is a calcium crystal. X-ray diffraction further confirms that the scale on the membrane surface of internal membrane system consists of calcium carbonate (Figure 7).
Physical and Chemical Cleaning
Many researchers have studied the behavior of membrane filtration using the resistance-in – series model, in which potential fouling matter is separated into suspended solids, colloids, and solutes. The cake resistance and fouling resistance of each component are calculated separately (18 – 20). The disadvantage of this method is that it neglects the interactions between components. In this study, the contributions of inorganic scaling and organic fouling on flux decline were studied by examining the outcome of physical and chemical cleaning.
Figure 6. Element analysis of membrane surfaces under different MBRs operations. (a) internal membrane system: (a-1) SEM-EDS analysis and (a-2) membrane photo; (b) external membrane system: (b-1) SEM-EDS analysis and (b-2) membrane photo.
Figure 8 shows the variation of TMP and Flux of internal membrane system during physical and chemical cleaning. No change in TMP and Flux was observed during the first six hours of air flushing, indicating no cake formation on the surface of the membrane. The TMP dropped sharply and the flux increased
Figure 7. X-ray diffraction spectrum of membrane surface in the internal membrane system. (2u: 10 – 808, step scan: 0.068/step, step time: 2 second, 40 kV, 40 mA).
Figure 8. Patterns of permeate flux and TMP during physical and chemical cleaning on internal membrane system. (air flushing: 2 m/sec, EDTA concentration: 2000 mg/L, NaOCl concentration: 2000 mg/L as OCl2).
immediately after the administration of EDTA. No further change was observed four hours after EDTA cleaning. After 16 hours of EDTA administration, cleaning by NaOCl was conducted using 2000 mg/L as OCl2. In the first six hours of NaOCl cleaning, the TMP reduced gradually accompanied by a slow increase in flux. The final flux was 13.2 L/m2-hr, or a 94% recovery, indicating that chemical cleaning using a combination of EDTA and NaOCl can efficiently remove inorganic and organic fouling on the membrane surface.
The flux decline contributed by inorganic scaling and organic fouling can be calculated accordingly. Calculation results are summarized in Table 1. In the internal membrane system, 77.8% of flux decline was due to inorganic scaling and 14.1% was due to organic fouling. Out of the total flux decline, 8.1% was irreversible fouling which could not be recovered even by chemical cleaning. In the external membrane system, no chemical cleaning was performed because the flux decline was less than 70%. As shown in Fig. 6, no scaling was formed in the external membrane system. Therefore, all the flux decline was due to organic fouling. Figure 5(b) shows that after 150 hours of MBR operation the flux dropped only slightly from 13.8 to 12.6 L/m2-hr, the equivalent of a 8.7% decline.
CONCLUSIONS
1. Scaling of membrane can be effectively controlled by MBR with an external membrane system. The MBR with internal membrane system has serious scaling problem.
2. The inorganic scaling and organic fouling can be efficiently removed by chemical cleaning with EDTA and NaOCl.
ACKNOWLEDGEMENT
The authors would like to thank Dr. S.S. Chou and Dr. K.F. Chang of ITRI, and Prof. S.S. Cheng of National Cheng Kung University for their helpful dis-cussions, as well as Miss S. H. Perng for her technical assistance with
Table 1. Flux declines of MBRs with internal and external membrane system
(%) Total Cake deposited Organic fouling Inorganic scaling Internal membrane system 100 (14 to 4.1) Neglect 14.1 (13.2 – 11.8) 77.8 (11.8 – 4.1) External membrane system 100 (13.8 to 12.6) Neglect 100 (13.8 – 12.6) Neglect a 8% irreversible fouling.
chemical analysis. This study was partially supported by the project of Aim to the Top University, MOE.
REFERENCES
1. Chiemchaisri, C., Yamamoto, K., and Vigneswaran, S. (1993) Household membrane bioreactor in domestic wastewater treatment. Water Sci. Technol., 27 (1): 171 – 178.
2. Chang, I.S., Lee, C.H., and Ahn, K.H. (1999) Membrane filtration characteristics in membrane coupled activated sludge system—the effect offloc structure of activated sludge on membrane fouling. Sep. Sci. Technol., 34 (9): 1743 – 1758. 3. Jefferson, B., Laine, A.L., Judd, S.J., and Stephenson, T. (2000) Membrane
bio-reactors and their role in wastewater reuse. Water Sci. Technol., 41 (1): 197 – 204. 4. Urbain, V., Block, J.C., and Manem, J. (1993) Bioflocculation in activated sludge:
an analytic approach. Water Res., 27 (5): 829 – 838.
5. Muller, E.B., Stouthamer, A.H., van Verseved, H.W., and Eikelboom, D.H. (1995) Aerobic domestic wastewater treatment in a pilot plant with complete sludge retention by cross-flow filtration. Water Res., 29 (4): 1179 – 1189.
6. Kim, J.S., Lee, C.H., and Chang, I.S. (2001) Effect of pump shear on the perform-ance of a crossflow membrane bioreactor. Water Res., 35 (9): 2137 – 2144. 7. Chang, I.S. and Lee, C.H. (1998) Membrane filtration characteristics in membrane
coupled activated sludge system—the effect of physiological states of activated sludge on membrane fouling. Desalination, 120 (3): 221 – 233.
8. Defrance, L. and Jaffrin, M.Y. (1999) Comparison between filtrations at fixed transmembrane pressure and fixed permeate flux: Application to a membrane bioreactor used for wastewater treatment. J. Membrane Sci., 152 (2): 203 – 210. 9. Choo, K.H. and Lee, C.H. (1996) Membrane fouling mechanisms in the
membrane-coupled anaerobic bioreactor. Water Res., 30 (8): 1771 – 1778. 10. Kang, I.J., Yoon, S.H., and Lee, C.H. (2002) Comparison of the filtration
charac-teristics of organic and inorganic membranes in a membrane-coupled anaerobic bioreactor. Water Res., 36 (7): 1803 – 1813.
11. Stahl, N., Tenenbaum, A., and Galil, N.I. (2004) Advanced treatment by anaerobic process followed by aerobic membrane bioreactor for effluent reuse in paper mill industry. Water Sci. Technol., 50 (3): 245 – 252.
12. You, H.S., Tseng, C.C., Peng, M.J., Chang, S.H., Chen, Y.C., and Peng, S.H. (2005(a)) A novel application of an anaerobic membrane process in wastewater treatment. Water Sci. Technol., 51 (6 – 7): 45 – 50.
13. Chang, I.S., Bag, S.O., and Lee, C.H. (2001) Effects of membrane fouling on solute rejection during membrane filtration of activated sludge. Process Biochem., 36 (8 – 9): 855 – 860.
14. You, H.S., Huang, C.P., Cheng, H.H., and Pan, J.R. (2005(b)) Effect of biofiltration on particle characteristics and flocculation behavior. J. Chem. Technol. Biot., 80 (6): 705 – 711.
15. Vera, L., Delgado, S., and Elmaleh, S. (2000) Gas sparged cross-flow microfiltra-tion of biologically treated wastewater. Water Sci. Technol., 41 (10 – 11): 173 – 180.
16. Lee, S.H. and Lee, C.H. (2000) Effect of operating conditions on CaSO4scale
formation mechanism in nanofiltration for water softening. Water Res., 34 (15): 3854 – 3866.
17. Lee, S.H., Kim, J.H., and Lee, C.H. (1999) Analysis of CaSO4scale formation
mechanism in various nanofiltration modules. J. Membrane Sci., 163 (1): 63 – 74. 18. Defrance, L., Jaffrin, M.Y., Gupta, B., Paullier, P., and Geaugey, V. (2000) Contribution of various constituents of activated sludge to membrane bioreactor fouling. Bioresource Technol., 73 (2): 105 – 112.
19. Bouhabila, E.H., Aim, R.B., and Buisson, H. (2001) Fouling characterisation in membrane bioreactors. Sep. Purif. Technol., 22 – 23: 123 – 132.
20. Lee, W., Kang, S., and Shin, H. (2003) Sludge characteristics and their con-tribution to microfiltration in submerged membrane bioreactors. J. Membrane Sci., 216 (1 – 2): 217 – 227.
