MEMBRANE AGEING DUE TO CHEMICALCLEANING AGENTbySyed Zaki AbdullahM.A.Sc., University of British Columbia, Vancouver, Canada 2008B.Sc., Bangladesh University of Engineering and Technology, Bangladesh 2003A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFDoctor of PhilosophyinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Civil Engineering)The University of British Columbia(Vancouver)May 2014c© Syed Zaki Abdullah, 2014AbstractSodium hypochlorite is commonly used as a cleaning agent to remove adsorbedfoulants from PVDF-based micro/ultra filtration membranes in water and wastewatertreatment applications. Although effective for fouling control, extended sodiumhypochlorite exposure can affect the physical/chemical characteristics and hinderthe treatment performance of these membranes. In the present study, experimentswere conducted to comprehensively quantify the effects of sodium hypochloriteexposure on changes in the physical/chemical characteristics and the filtrationperformances of blended PVDF-based supported hollow-fiber membranes andidentifying the mechanism(s) responsible for the changes. Both the effect of thesodium hypochlorite concentration (C) and the duration of exposure (t) on themembrane characteristics are investigated. The physical/chemical characteristicsand the filtration performances of virgin and aged (i.e., weathered due to exposure tosodium hypochlorite) membranes were compared. The membranes were characteri-zed based on chemical composition (FTIR and NMR), mechanical strength (yieldstrength), surface hydrophilicity (contact angle), pore size and porosity (scanningelectron microscopy and challenge test), membrane resistance (clean water permea-tion test), and affinity of the membrane for foulants (cleaning efficiency). The resultsiiindicated that exposure dose and concentration of the sodium hypochlorite used havea significant influence on the membrane characteristics. For the exposure conditionsconsidered, the impact of sodium hypochlorite exposure on the parameters investiga-ted could be most accurately and consistently correlated to an exposure doserelationship of the form Cnt (where, C = concentration and t = exposure time)rather than the Ct relationship commonly used to define the extent of exposureto cleaning agents. For all the parameters investigated, the power coefficient nwas less than 1 indicating that time had a greater impact on the changes than didthe concentration of the sodium hypochlorite. The results suggest that the useof sodium hypochlorite for chemical cleaning, at concentrations that are higherthan those typically used for chemical cleaning would have less of an effect onthe characteristics of the membrane materials. Changes in the characteristics wereattributed to the oxidation of the hydrophilic additives (HA) present in blendedPVDF membranes. A new non-destructive membrane characterization technique toevaluate the amount of membrane ageing is proposed.iiiPrefaceThis statement confirms that the author of this thesis is the primary person responsiblefor the research contained. All experimental designs and procedures were conceivedof by the author with input from the supervisory committee, namely Dr. PierreBe´rube´, Dr. Eric Hall and Dr. David Wilkinson. The specific names of those whoassisted in conducting several of the experiments have been gratefully recognized inthe Acknowledgments section of this thesis.Several manuscripts in journals and in conference proceedings have beenpublished or are under consideration for publication pertaining to the results of thisthesis. The following is a list of such manuscripts which have been published or arecurrently under review:Journals:• Abdullah, S. Z., and Be´rube´, P. R. (2013). Assessing the effects of sodiumhypochlorite exposure on the characteristics of PVDF-based membranes.Water Research, 47 (2013) 5392 - 5399.• Abdullah, S. Z., and Be´rube´, P. R. (2013). SEM imaging of membranes:Importance of sample preparation and imaging parameters. Journal ofivMembrane Science, 463 (2014) 113125.• Abdullah, S. Z., and Be´rube´, P. R. (2013). Assessing the effects of sodiumhypochlorite ageing on the filtration characteristics of PVDF-based membranes.Research completed and manuscript being prepared, will be submitted to aJournal.Conference proceedings:• Abdullah, S. Z., and Be´rube´, P. R. (2013). Effects of Sodium HypochloriteCleaning on the Filtration Performance of PVDF-based Membrane. Proceed-ings 7th IWA Specialized Membrane Technology Conference and Exhibitionfor Water and Wastewater Treatment and Reuse, Toronto.• Abdullah, S. Z., and Be´rube´, P. R. (2012). Effects of chemical cleaning onmembrane operating lifetime. Proceedings Membrane Technology Conferenceand Exposition. AWWA/AMTA Conference, Glendale, Arizona. For thispaper I won the Best Student Paper Award.• Abdullah, S. Z., and Be´rube´, P. R. (2011). New technique to assess the effectsof chemical cleaning on membrane characteristics. Proceedings MembraneTechnology Conference and Exposition. AWWA Conference, Longbeach,California.• Abdullah, S. Z., and Be´rube´, P. R. (2011). Investigation of the factors andmechanisms that govern membrane lifetime. Presented at CAWQ Symposiumheld at the BCWWA Annual Conference, Penticton.vPoster presentation:• Abdullah, S. Z., and Be´rube´, P. R. (2014). Effects of Sodium HypochloriteAgeing on the Cleaning Requirements of PVDF-based Membrane. Posterwill be presented at the Membrane Technology Conference and Exposition.AWWA/AMTA Conference, Las Vagus, Nevada.• Abdullah, S. Z., and Be´rube´, P. R. (2012). Model development to assess theageing of polymeric membranes due to chemical cleaning. Poster presentedat Euromembrane 2012 Conference, Westminster, London.viTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xivGlossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxivDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxvi1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Objectives and Scope . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Organization of the Dissertation . . . . . . . . . . . . . . . . . . 42 Background and Literature Review . . . . . . . . . . . . . . . . . . 6vii2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2 Membrane Filtration Process . . . . . . . . . . . . . . . . . . . . 72.3 Membrane Classification . . . . . . . . . . . . . . . . . . . . . . 82.3.1 Based on Pore Size . . . . . . . . . . . . . . . . . . . . . 82.3.2 Based on Geometry . . . . . . . . . . . . . . . . . . . . . 82.3.3 Based on Configuration and Operating Condition . . . . . 92.3.4 Based on Materials . . . . . . . . . . . . . . . . . . . . . 112.4 Membrane Fouling . . . . . . . . . . . . . . . . . . . . . . . . . 112.4.1 Types of Membrane Fouling . . . . . . . . . . . . . . . . 142.4.2 Membrane Fouling Models . . . . . . . . . . . . . . . . . 172.5 Effect of Feed Matrix Characteristics and Membrane Properties onFouling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182.6 General Methods for Membrane Fouling Control . . . . . . . . . 202.6.1 Feed Pre-treatment . . . . . . . . . . . . . . . . . . . . . 202.6.2 Optimized Operating Conditions . . . . . . . . . . . . . . 212.6.3 Modification of Membrane Surface Characteristics . . . . 232.6.4 Chemical Cleaning . . . . . . . . . . . . . . . . . . . . . 242.7 Effectiveness of Chemical Cleaning . . . . . . . . . . . . . . . . 282.8 Impact of Chemical Cleaning on the Operation and Maintenance Cost 292.9 Current Knowledge on the Effects of Chemical Cleaning on MembraneSurface Properties and Filtration Performances . . . . . . . . . . 302.10 Use of SEM for Membrane Characterization . . . . . . . . . . . . 383 Objectives Of The Study . . . . . . . . . . . . . . . . . . . . . . . . 453.1 Knowledge Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . 45viii3.1.1 Objective #1 . . . . . . . . . . . . . . . . . . . . . . . . 463.1.2 Objective #2 . . . . . . . . . . . . . . . . . . . . . . . . 483.2 Selection of a Consistent SEM Imaging Approach . . . . . . . . . 494 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 524.1 Commercial Membranes . . . . . . . . . . . . . . . . . . . . . . 524.2 Lab-cast Membranes . . . . . . . . . . . . . . . . . . . . . . . . 524.3 Exposure Procedure . . . . . . . . . . . . . . . . . . . . . . . . . 534.4 Membrane Characterization Techniques . . . . . . . . . . . . . . 554.5 Assessment of the Affinity of the Membrane for Foulants . . . . . 594.5.1 Experimental Setup . . . . . . . . . . . . . . . . . . . . . 594.5.2 Filtration Operation . . . . . . . . . . . . . . . . . . . . . 614.5.3 Assessment of the Efficiency of Physical and ChemicalCleaning . . . . . . . . . . . . . . . . . . . . . . . . . . 624.6 Assessment of the Fate of HA . . . . . . . . . . . . . . . . . . . 644.7 SEM Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654.7.1 Impact of Sample Preparation and Imaging Parameters onthe Membrane Surface Properties . . . . . . . . . . . . . 654.8 QA/QC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715 SEM Imaging of Membranes: Importance of Sample Preparationand Imaging Parameters . . . . . . . . . . . . . . . . . . . . . . . . 735.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 735.2 Typical Results from SEM Imaging and Image Analysis . . . . . . 755.3 Qualitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 785.4 Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . 81ix5.5 Sensitivity of the Membrane Properties to the Parameters . . . . . 835.5.1 Coating Material . . . . . . . . . . . . . . . . . . . . . . 835.5.2 Coating Thickness . . . . . . . . . . . . . . . . . . . . . 855.5.3 Detector Type . . . . . . . . . . . . . . . . . . . . . . . . 885.5.4 Magnification . . . . . . . . . . . . . . . . . . . . . . . . 915.5.5 Tilt Angle . . . . . . . . . . . . . . . . . . . . . . . . . . 935.5.6 Impact of Individual Parameters Regardless of the Combinationof the Other Parameters . . . . . . . . . . . . . . . . . . 965.6 Conclusion and Selection of the Consistent SEM Imaging Approach 976 Impact of sodium hypochlorite exposure on the physical/chemicalcharacteristics of PVDF-based membranes . . . . . . . . . . . . . . 996.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 996.2 Impact of Exposure to Sodium Hypochlorite on Membrane Characteristics1026.2.1 Fourier Transform Infrared Spectroscopy . . . . . . . . . 1026.2.2 Nuclear Magnetic Resonance (NMR) . . . . . . . . . . . 1096.2.3 Contact Angle . . . . . . . . . . . . . . . . . . . . . . . 1166.2.4 SEM Image, Pore Size and Apparent Porosity . . . . . . . 1206.2.5 Tensile Strength . . . . . . . . . . . . . . . . . . . . . . 1266.2.6 Intrinsic Resistance of a Membrane Based on Clean WaterPermeation Test . . . . . . . . . . . . . . . . . . . . . . . 1296.2.7 Retention of Dextran Particles . . . . . . . . . . . . . . . 1326.3 Modelling the Magnitude of the Impact of Sodium HypochloriteExposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1356.4 Fate of the HA of Interest During Exposure to Sodium Hypochlorite 143x6.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1517 Impact of sodium hypochlorite exposure on the filtration and cleaningperformance of PVDF-based membranes . . . . . . . . . . . . . . . 1537.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1547.2 Impact of Exposure to Sodium Hypochlorite on the MembraneFouling Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . 1567.3 Impact of Exposure to Sodium Hypochlorite on the Cleaning Efficiency1627.4 Rate of Permeability Recovery During Chemical Cleaning . . . . 1747.5 Proposed Non-destructive Method to Assess Membrane Age . . . 1777.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1798 Conclusions and Recommendations . . . . . . . . . . . . . . . . . . 1828.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1828.2 Recommendations for Future Work . . . . . . . . . . . . . . . . . 1858.3 Engineering Significance . . . . . . . . . . . . . . . . . . . . . . 187Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190A Supporting Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 207A.1 Matlab Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207A.2 Typical SEM Images . . . . . . . . . . . . . . . . . . . . . . . . 211A.3 Typical results of the SEM image analysis . . . . . . . . . . . . . 219A.4 Typical results of the physical/chemical and fouling characteristicsof the membranes . . . . . . . . . . . . . . . . . . . . . . . . . . 224xiList of TablesTable 2.1 Typical recovery chemical cleaning protocols used by four leadingmembrane suppliers (from [87]). . . . . . . . . . . . . . . . . 27Table 2.2 Summary of the published work on the membrane ageing due toexposure to cleaning agents. . . . . . . . . . . . . . . . . . . 31Table 2.3 Summary of the information available in reviewed journal articles.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40Table 2.4 Effect of the different sample preparation and imaging parameterson SEM imaging. . . . . . . . . . . . . . . . . . . . . . . . . 41Table 4.1 Experimental conditions investigated . . . . . . . . . . . . . . 54Table 4.2 Sample preparation and imaging parameters considered . . . . 66Table 5.1 Repeatability of the typical values of membrane properties ofinterest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Table 5.2 P-values for membrane properties of interest obtained from themulti-way ANOVA analysis. . . . . . . . . . . . . . . . . . . 96xiiTable 6.1 Typical values of the empirical constants k, n and m for themembrane characteristic parameters investigated. . . . . . . . 140Table 6.2 Amount of TOC present in blended lab-cast membranes . . . . 144Table A.1 Typical results of the SEM image analysis. . . . . . . . . . . . 219Table A.2 Typical results of the physical/chemical characteristics of themembranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225Table A.3 Typical results of physical/chemical characteristics of the membranes226Table A.4 Typical results of the fouling characteristics of the membranes . 227xiiiList of FiguresFigure 2.1 Membrane filtration configurations (a) Inside-out, and (b) outside-in. 9Figure 2.2 Typical full-scale outside-in hollow-fiber membrane module . 10Figure 2.3 Accumulation of materials during membrane filtration process 13Figure 2.4 Different types of membrane fouling resistance . . . . . . . . 16Figure 4.1 Membrane module#1 (a), and bench scale filtration system (b)used for clean water permeation tests. . . . . . . . . . . . . . 58Figure 4.2 Schematic of the UBC MEBPR pilot plant. . . . . . . . . . . 60Figure 4.3 Membrane module#2 used for membrane filtration performanceexperiments. . . . . . . . . . . . . . . . . . . . . . . . . . . 61Figure 4.4 Flow chart for the SEM image analysis. . . . . . . . . . . . . 69Figure 5.1 Typical SEM images of membrane . . . . . . . . . . . . . . . 76Figure 5.2 Typical pore size distribution . . . . . . . . . . . . . . . . . . 77Figure 5.3 Typical HRSEM images of membranes . . . . . . . . . . . . 79Figure 5.4 Average values of the observed membrane properties of interestobtained from SEM images acquired with 5 nm coating, UD,50K magnification, 0 degree tilt . . . . . . . . . . . . . . . . 82xivFigure 5.5 Effects of coating metal on the average observed membraneproperties of interest . . . . . . . . . . . . . . . . . . . . . . 84Figure 5.6 Effects of coating thickness on the average observed membraneproperties of interest . . . . . . . . . . . . . . . . . . . . . . 86Figure 5.7 Effects of detector on the average observed membrane propertiesof interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89Figure 5.8 Effects of magnification on the average observed membraneproperties of interest . . . . . . . . . . . . . . . . . . . . . . 92Figure 5.9 Effects of tilt angle on the average observed membrane propertiesof interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94Figure 6.1 Typical raw and baseline-corrected FTIR spectra of virginmembrane. . . . . . . . . . . . . . . . . . . . . . . . . . . . 103Figure 6.2 Typical FTIR spectra of PVDF and HA. . . . . . . . . . . . . 104Figure 6.3 Typical baseline-corrected and scaled FTIR spectra of virginmembranes. . . . . . . . . . . . . . . . . . . . . . . . . . . . 105Figure 6.4 Typical FTIR spectra of lab-cast membranes containing 20%HA (i.e., 80% PVDF) and 30% HA (i.e., 70% PVDF). . . . . 106Figure 6.5 Typical FTIR spectra of virgin membrane and membranesexposed to sodium hypochlorite . . . . . . . . . . . . . . . . 107Figure 6.6 Effects of sodium hypochlorite on the dominant peak of themajor functional group of the HA of interest . . . . . . . . . . 108Figure 6.7 Typical NMR spectra of (a) PVDF, and (b) HA of interest. . . 110Figure 6.8 Typical NMR spectra of lab-cast membrane containing (a) 4%HA (i.e., 96% PVDF), and (b) 8% HA (i.e., 92% PVDF). . . . 112xvFigure 6.9 Typical NMR spectrum of a virgin commercial membrane. . . 113Figure 6.10 Effects of sodium hypochlorite on the HA of interest content ofthe membranes (NMR data). . . . . . . . . . . . . . . . . . . 114Figure 6.11 Comparison of the HA of interest content of the membranesanalyzed by FTIR and NMR . . . . . . . . . . . . . . . . . . 116Figure 6.12 Typical image of the contact angles measurement. . . . . . . . 117Figure 6.13 Effects of sodium hypochlorite on the contact angle of themembranes . . . . . . . . . . . . . . . . . . . . . . . . . . . 118Figure 6.14 Comparison of the HA of interest content of the membranesanalyzed by NMR to the contact angle of the membranes . . . 120Figure 6.15 Typical SEM images of a membrane sample . . . . . . . . . . 121Figure 6.16 Typical SEM images of virgin membrane (a) and membranesexposed to different exposure doses and concentrations ofsodium hypochlorite (b, c and d). . . . . . . . . . . . . . . . 121Figure 6.17 Effects of sodium hypochlorite on the average pore size of themembranes . . . . . . . . . . . . . . . . . . . . . . . . . . . 122Figure 6.18 Effects of sodium hypochlorite on 90th percentile of cumulativepore size distribution of the membranes . . . . . . . . . . . . 123Figure 6.19 Effects of sodium hypochlorite on the porosity of the membranes124Figure 6.20 Comparison of the HA of interest content of the membranesanalyzed by NMR to the apparent porosity of the membranes . 126Figure 6.21 Typical stress-strain curve of the membranes. . . . . . . . . . 127Figure 6.22 Effects of sodium hypochlorite on the tensile strength of themembranes . . . . . . . . . . . . . . . . . . . . . . . . . . . 128xviFigure 6.23 Comparison of the tensile strength of the membranes to theapparent porosity of the membranes . . . . . . . . . . . . . . 129Figure 6.24 Typical clean water filtration test result for a virgin commercialmembrane. . . . . . . . . . . . . . . . . . . . . . . . . . . . 130Figure 6.25 Effects of sodium hypochlorite on the intrinsic resistance of themembranes . . . . . . . . . . . . . . . . . . . . . . . . . . . 131Figure 6.26 Comparison of the intrinsic resistance of the membranes to theapparent porosity of the membranes . . . . . . . . . . . . . . 132Figure 6.27 Total organic carbon (TOC) content of the Dextran T70, DextranT500 and feed solution . . . . . . . . . . . . . . . . . . . . . 133Figure 6.28 Total organic carbon (TOC) content of the feed solution, andthe permeates from virgin membranes and membranes exposedto sodium hypochlorite (exposure dose and concentration are2,000,000 ppm·hr and 3600 ppm, respectively . . . . . . . . . 134Figure 6.29 Dextran retention by virgin membranes and membranes exposedto sodium hypochlorite (exposure dose and concentration are2,000,000 ppm·hr and 3600 ppm, respectively) . . . . . . . . 135Figure 6.30 Typical least squares curve fitting of the observed FTIR peaksfor the model #1 to 4. . . . . . . . . . . . . . . . . . . . . . . 139Figure 6.31 Typical response of the membrane parameters investigated withrespect to the Cnt relationship . . . . . . . . . . . . . . . . . 142Figure 6.32 Typical sodium hypochlorite consumption by lab-cast membranesduring exposure to sodium hypochlorite . . . . . . . . . . . . 145xviiFigure 6.33 Typical values of the ratio of mass of DOC in the solutionto mass of TOC originally present in the membranes, whenlab-cast membranes were exposed to sodium hypochlorite . . 147Figure 6.34 Typical values of the rate of oxidation of TOC, when lab-castmembranes were exposed to sodium hypochlorite . . . . . . . 150Figure 7.1 A typical normalized permeability trend for a virgin membrane. 156Figure 7.2 Typical normalized permeability trends for the membranesexposed to sodium hypochlorite with different concentrationsfor 500,000 ppm·hr prior to filtration. . . . . . . . . . . . . . 157Figure 7.3 Typical normalized permeability trends for the membranesexposed to sodium hypochlorite with different concentrationsfor 1,000,000 ppm·hr prior to filtration. . . . . . . . . . . . . 158Figure 7.4 Typical normalized permeability trends for the membranesexposed to sodium hypochlorite with different concentrationsfor 2,000,000 ppm·hr prior to filtration. . . . . . . . . . . . . 159Figure 7.5 Effect of sodium hypochlorite exposure dose and concentration(prior to filtration) on the filtration capacity of the membranes. 161Figure 7.6 Typical image of virgin membranes at different stages of cleaning.(a: unfouled; b: after activated sludge filtration; b: after physicalcleaning; c: after chemical cleaning.) . . . . . . . . . . . . . 163Figure 7.7 Typical image of the membrane at different stages of cleaning.(a: 500,000 ppm·hr (3600 ppm); b: 500,000 ppm·hr (22200ppm); c: 2,00,000 ppm·hr (3600 ppm); Right membranes: afterphysical cleaning, and left membranes: after chemical cleaning.)164xviiiFigure 7.8 Typical permeability trend of virgin membrane during filtrationfollowed by physical cleaning, and chemical cleaning. . . . . 165Figure 7.9 Typical physically irrecoverable permeability values for virginmembranes and membranes previously exposed to sodiumhypochlorite for different exposure doses and concentrations. . 166Figure 7.10 Typical chemically recoverable permeability values with 6hours of chemical cleaning, for virgin membranes and membranespreviously exposed to sodium hypochlorite for different exposuredoses and concentrations. . . . . . . . . . . . . . . . . . . . . 168Figure 7.11 Typical chemically irrecoverable permeability values after physical,and 6 hours of chemically cleaning, for virgin membranesand membranes previously exposed to sodium hypochloritefor different exposure doses and concentrations. . . . . . . . . 170Figure 7.12 Typical chemically irrecoverable permeability values after physicalcleaning, and 24 hours of chemical cleaning, for virgin membranesand membranes previously exposed to sodium hypochlorite fordifferent exposure doses and concentrations. . . . . . . . . . . 172Figure 7.13 Typical chemically recoverable permeability trends during chemicalcleaning for virgin membranes and membranes previously exposedto sodium hypochlorite for different exposure doses and concentrations.175Figure 7.14 Typical permeability recovery trends during chemical cleaning. 179Figure A.1 Typical SEM images of membranes acquired with 2nm coating,25K magnification and 30 degree tilt . . . . . . . . . . . . . . 211xixFigure A.2 Typical SEM images of membranes acquired with 5nm coating,25K magnification and 30 degree tilt . . . . . . . . . . . . . . 212Figure A.3 Typical SEM images of membranes acquired with 2nm coating,50K magnification and 30 degree tilt . . . . . . . . . . . . . . 213Figure A.4 Typical SEM images of membranes acquired with 5nm coating,50K magnification and 30 degree tilt . . . . . . . . . . . . . . 214Figure A.5 Typical SEM images of membranes acquired with 2nm coating,25K magnification and 0 degree tilt . . . . . . . . . . . . . . 215Figure A.6 Typical SEM images of membranes acquired with 5nm coating,25K magnification and 0 degree tilt . . . . . . . . . . . . . . 216Figure A.7 Typical SEM images of membranes acquired with 2nm coating,50K magnification and 0 degree tilt . . . . . . . . . . . . . . 217Figure A.8 Typical SEM images of membranes acquired with 5nm coating,50K magnification and 0 degree tilt . . . . . . . . . . . . . . 218xxGlossaryAu GoldAuPd Gold-palladiumAuPt Gold-PlatinumC ConcentrationCA Cellulose acetateCr Chromiumdt Major axis lengthdc Minor axis lengthEPS Extracellular polymeric substancesFESEM Field Emission Scanning Electron MicroscopyHA Hydrophilic AdditiveHRSEM High Resolution Scanning Electron MicroscopyLMH liter/m2·hrxxiMBR Membrane bioreactorMF Micro-filtrationLP PermeabilityNaOCl Sodium hypochloriteNF Nano-filtrationNOM Natural Organic MatterPA PolyamidePAN PolyacrylonitrilePE PolyethylenePEG Polyethylene glycolPES PolyethylsulfonePP PolypropylenePS f PolysulfonePSU PolysulfonePV DF Polyvinylidene difluoridePV P PolyvinylpyrrolidoneRO Reverse osmosisSEM Scanning Electron Microscopyxxiit Duration (i.e. time) of exposureT MP Transmembrane pressureUF Ultra-filtrationIr IridiumxxiiiAcknowledgmentsCompletion of this research project required a knowledge base, skill set andresources beyond what I possess. However, the support I received from manypeople made it possible to complete this project successfully. I would like to thankthe following people for their assistance, support and encouragement throughoutthe duration of my research.Firstly, I would like to thank the people of Canada for creating and supporting ahealthy research environment, without which, this study would not be possible.My research supervisor Dr. Pierre Be´rube´ was the key person, who provided mecontinuous intellectual support, project resources, and additional motivation duringdifficult times with his positive attitude and encouraging words. I am fortunate tohave a person like Dr. Be´rube´ as my mentor, who relentlessly works to make hisstudents better researchers. It is not possible for me thank him enough.I would also like to thank my committee members Dr. Eric Hall and Dr.David Wilkinson for providing me critical suggestions and guidance throughout mydoctoral tenure, especially for reviewing my 200+ page dissertation in one week,when I was under time pressure.Many thanks go out to the support stuffs of UBC for extending their helping handxxivin different stages of my research. In particular, I would like to thank Bill Leung,Paula Parkinson, Tim Ma, Frederic Koch and Derrick Horne for their continuoussupport and encouragements. I will always remember Bill Leung not only for hiscreative technical solutions, but also for the big welcoming smile in his face.I feel very lucky to have great supporting friends and colleagues throughout theentire length of my stay at UBC. I would like to thank Parvez Fattah, Colleen Chan,Alireza Abedini, Chris Lawson, Joerg Winter, Rony Das, Yafee Muntasir, IsabelLondono, Sepideh Jankah, Shona Robinson, Christina Starke, and all the peoplewhose names I am forgetting at this moment.I appreciate the suggestions and help that I received from Dr. Imran-ul Hoqueto bridge my knowledge gap on the polymer analysis. I am also grateful to myundergraduate co-op students Ralph Mercado and Cindy Wang for helping me withlaboratory analysis.I am utterly grateful to my parents, Neva and Iqbal, for all they have done forme through my entire life. During my doctoral research, our family was challengedby my mothers battle with cancer, and my fathers severe illness. Instead of askingme for support during these difficult times, my parents encouraged and supportedme to continue with my studies. Unfortunately, my father passed away last year.Dad you will always be in my mind and heart. I am also grateful to my brother andsister for their unyielding support, which gave me the strength and motivation tocomplete this research.I am blessed with a very patient life partner, Tasnuva Mariam. Tasnuva, I thankyou for your patient, encouragement, time, support and all the sacrifices you havemade during these years.Finally, this research was possible only by the grace of almighty Allah.xxvDedicationTo -My parents Syed Iqbal Ali and Syeda Gulnahar BegumMy loving and supportive wife Tasnuva MariamMy sister Syeda Shahreen and brother Syed BillahxxviChapter 1Introduction1.1 Objectives and ScopePolymer-based Micro-filtration (MF)/Ultra-filtration (UF) membrane systems areincreasingly being used for large-scale water and wastewater treatment applicationsdue to their ability to effectively remove many contaminants of concern as wellas for their relative low cost. Polyvinylidene difluoride (PV DF) is widely usedas a core material for polymeric membrane fabrication mainly due to its goodmechanical and thermal properties, and excellent chemical resistance. However,pure PVDF membranes are highly hydrophobic, making them more susceptibleto fouling in water and wastewater treatment applications than membranes madefrom hydrophilic materials [65, 111]. For this reason, hydrophobic polymer-basedmembranes are often blended or coated with Hydrophilic Additive (HA) to introducehydrophilic functional groups into the structure, which enhances the filtrationperformance of these membranes [65, 105, 106, 112]. Despite these modifications,over time, some of the retained contaminants can be adsorbed at and/or in the1membrane. Chemical cleaning is typically used to remove adsorbed foulants.Sodium hypochlorite (NaOCl) is one of the most commonly used chemicals formembrane cleaning [87].Although effective for fouling control, over extended periods of time, chemicalcleaning agents such as sodium hypochlorite can impact the physical/chemicalcharacteristics of the membrane, affecting not only the extent of fouling, but also thetreatment performance and operating lifetime of membrane systems [124]. For thesereasons, membrane manufacturers typically specify a maximum lifetime exposuredose, defined as the cumulative product of the concentration of the chemical agent(C) and the duration of the cleaning (t), above which the performance of theirmembranes cannot be guaranteed [31, 135].Despite the widespread use of sodium hypochlorite for membrane cleaning,limited research has focused on the effects of sodium hypochlorite cleaning onblended PVDF membranes [58, 119]. None of these studies those studied the effectsof sodium hypochlorite cleaning on blended PVDF membranes have comprehensivelyaddressed the effects of sodium hypochlorite exposure on the important characteristicsparameters including, the chemical composition, the important physical/chemicalcharacteristics, filtration capacity and cleaning efficiency of the blended PVDFmembranes. In addition, none of these studies has focused on identifying themechanisms responsible for the reported effects. Also, most studies have notinvestigated the effect of the concentration of sodium hypochlorite on the membranecharacteristics [8–10, 49, 88, 126].In the present study, experiments were conducted to comprehensively quantifythe effects of sodium hypochlorite exposure on changes in the physical/chemicalcharacteristics and the filtration performances of blended PVDF-based hollow-fiber2membranes and to identify the mechanism(s) responsible for the changes. Boththe effects of the sodium hypochlorite Concentration (C) and the Duration (i.e.time) of exposure (t) on the membrane characteristics are investigated. Based onthe results, recommendations were developed to optimize sodium hypochloriteexposure conditions to reduce the effect of sodium hypochlorite exposure on themembrane characteristics and performances, and as a result, increase membraneoperating life.A proper understanding of the membrane surface properties is essential tocharacterize the properties of virgin membrane and assess how these can changeover time. Scanning Electron Microscopy (SEM) and High Resolution ScanningElectron Microscopy (HRSEM), also known as Field Emission Scanning ElectronMicroscopy (FESEM) imaging, have been extensively used to characterize themembrane surface properties, both qualitatively (e.g., visual observation) andquantitatively (e.g., pore size, pore shape and porosity). A literature review ofarticles published on the topic of the membrane ageing demonstrated that a relativelylarge number of studies have used SEM and/or HRSEM to characterize the poresize, pore size distribution, pore shape, porosity, surface roughness and surfacefouling etc. of polymeric membranes. Although SEM imaging is commonly usedto characterize membrane surface properties, very limited research has focusedon the effects of sample preparation (e.g., coating material and thickness) andimaging parameters (e.g., magnification, detector type, sample orientation) on theobtained information from SEM imaging of polymeric membranes, and none havefocused on identifying the sensitivity of the parameters on the obtained membranecharacteristics from SEM imaging of polymeric membranes.As a part of the present study, experiments were conducted to quantify the3effect of sample preparation (i.e., coating metal and coating thickness) and imagingparameters (i.e., magnification, detector used, sample orientation) on the observedmembrane surface properties of polymeric membrane obtained from SEM imaging.The effects were observed both qualitatively (e.g., visual observation) and quantita-tively (e.g., pore size, pore shape and porosity). The results clarify the effect ofvarious sets of parameters on the observed membrane properties.1.2 Organization of the DissertationThe structure of this thesis is presented here. It is comprised of nine chapters:Chapter 1: IntroductionAn introduction to the thesis is given in Chapter 1, summarizing the scope ofthe research.Chapter 2: Background and Literature ReviewA comprehensive summary and discussion of past literature is given in Chapter2, explaining the importance of chemical cleaning for the membrane operation, andhow the process might affect the overall membrane performance by changing themembrane characteristics.Chapter 3: Objectives of The StudyA brief summary of the current knowledge gap and corresponding objective andresearch questions were formulated to address the knowledge gap.Chapter 4: Materials and MethodsA general description of all the materials and methods used in the present studyis provided in Chapter 4.Chapter 5: SEM Imaging of Membranes: Importance of Sample Prepara-tion and Imaging Parameters4Although it is well known that sample preparation and imaging parameters cansignificantly affect the information obtained from SEM imaging, no comprehensivestudy on the importance of these parameters for the SEM imaging of PVDF-basedmembranes has yet been conducted. Therefore, a thorough study is carried out,and in Chapter 5 the importance of sample preparation and imaging parameters arepresented.Chapter 6: Impact of sodium hypochlorite exposure on the physical/chemi-cal characteristics of PVDF-based membranesComprehensive analyses of the membranes were conducted to assess the impactof sodium hypochlorite exposure on the physical/chemical characteristics of PVDF-based membranes. The results of the analyses and corresponding outcomes arepresented in Chapter 6.Chapter 7: Impact of sodium hypochlorite exposure on the filtration andcleaning performance of PVDF-based membranesComprehensive filtration and cleaning experiments were designed to assess theimpact of sodium hypochlorite exposure on the filtration and cleaning performanceof PVDF-based membranes. The results of the experiments and correspondingoutcomes are presented in Chapter 7.Chapter 8: Conclusions and RecommendationsOverall conclusions from each chapter of this thesis are presented and discussed.Future research recommendations about what may be done to supplement this studyare suggested. Also, the engineering significance of the present study are discussed.5Chapter 2Background and LiteratureReviewA comprehensive literature review was performed to assist in defining the researchobjective for the present study. Relevant material from this review is brieflysummarized in the sections below.2.1 IntroductionMicro-filtration (MF)/ultra-filtration (UF) membrane (hereafter referred asmembranes) technologies have been used for water and wastewater applicationsfor more than three decades. With technical advances and corresponding costreductions, membrane systems are now widely used in large-scale water andwastewater treatment plants. The process has many advantages over conventionaltechnologies, including the capability of producing high quality effluent with arelatively small footprint, a reduced treatment chemical requirement, and reducedlabor requirements since the process can be easily automated [104].62.2 Membrane Filtration ProcessMembrane filtration is a physical process for the separation of a mixture of substanceswith a selective barrier. In pressure-driven membrane processes, which have beenwidely applied in water and wastewater treatment systems, material that is smallerthan the membrane pores is transported through the membrane by the pressuredriving force, while material that is larger than the membrane pores is retained atthe surface of the membrane.Typically there are three process streams in a membrane filtration process - thefeed, the permeate and the retentate. The feed, also commonly known as the influent,is the solution to be filtered by the membrane module. The permeate is the filteredeffluent from the membrane module. The retentate, also known as the concentrate,is the stream that contains the rejected or retained constituents. The permeate fluxis the rate at which permeate passes through the membrane, expressed as a unitvolume of permeate per unit area of membrane surface per time. A commonlyused unit for permeate flux is liter/m2·hr (LMH). The driving force of filtration isTransmembrane pressure (T MP), defined as the pressure difference between thefeed side and the permeate side:T MP = ∆P = Pf eed−Ppermeate (2.1)A vacuum on the permeate side, or a pressure on the feed side is typicallyapplied to create the desired TMP for filtration.72.3 Membrane Classification2.3.1 Based on Pore SizeMembranes are generally classified based on pore size (or molecular weight cut-off),such as micro-filtration (MF) membranes, ultra-filtration (UF) membranes, etc. MFand UF membranes are mainly used for water and wastewater treatments. For MFand UF membranes, pore sizes generally lie in the range of 0.1 to 0.2 µm and 0.01to 0.05 µm, respectively [5].2.3.2 Based on GeometryTypically membrane filtration media is manufactured as hollow-fibers or as flatsheet stock and then configured into different types of membrane filtration units,including hollow-fiber, tubular, flat sheet and spiral wound. In water and wastewaterapplications, hollow-fibers and flat sheet membranes are the most predominantlycommercially available for microfiltration and ultrafiltration [5]. hollow-fiber MFand UF membranes are becoming more and more popular, because of their highpacking density and relatively low cost [74].A typical hollow-fiber module may consist of few hundreds to tens of thousandsof long and narrow fibers [5]. General range for fiber length is 1 to 2 meters; fiberouter diameters generally range from 0.5 to 2.0 mm with a fiber wall thickness of0.1 to 0.6 mm [5]. The reminder of the discussion will focus predominantly onhollow-fiber membranes.82.3.3 Based on Configuration and Operating Conditionhollow-fiber membranes may be operated in either inside-out (Figure 2.1(a)) oroutside-in (Figure 2.1(b)) filtration configuration. In the inside-out configuration,feed water enters through the lumen of the membrane and permeate leaves throughthe skin of the membrane. In the outside-in configuration, feed water is outside of themembrane and permeate is collected inside the lumen. Compared to the outside-inconfiguration, the inside-out configuration is less advantageous, as the inside-outconfiguration is more susceptible to clogging and offers relatively less surface area.Figure 2.2 illustrates a typical full-scale outside-in hollow-fiber membrane module. ! "#$! "%$!&'()!*+!&'()!*+!&'()!,-.!&'()!,-.!Figure 2.1: Membrane filtration configurations (a) Inside-out, and (b)outside-in.9Figure 2.2: Typical full-scale outside-in hollow-fiber membrane modulehollow-fiber membranes may be operated in either a constant pressure orconstant flux filtration mode. The common trend in membrane design is to operatethe system at constant flux [87]. Inside-out operation is generally done underconstant pressure in a pressure-driven process, where pressurized feed water ispumped to the membrane lumen by an external pump. A typical pressure range forthis kind of operation is 20 to 280 kPa [5]. Outside-in operation is generally doneunder a vacuum-driven process, where the membrane modules are submerged in thefeed water tank and the lumen sides of the membranes are connected to a pump. Inthe vacuum-driven process the suction pressure ranges from -20 to -80 kPa [5].102.3.4 Based on MaterialsA wide variety of materials, including organic and inorganic materials are commonlyused to manufacture membranes. Organic membranes are typically made ofsynthetic polymers such as polyvinylidene difluoride (PVDF), Polyethylsulfone(PES), Polyethylene (PE), nylon, cellulose acetate and Polypropylene (PP) [103,104]. Inorganic materials, such ceramic and metals also have been used as membranematerials. Inorganic membranes do offer some advantages over organic membranes,such as higher pressure and temperature tolerance, and harsher chemical cleaningconditions can be used during membrane cleaning. However, the use of the inorganicmembranes is not as common as that of the polymeric membranes. The reasonis the higher cost of the inorganic membrane. Polyvinylidene difluoride (PVDF)is widely used as a core material for polymeric membrane fabrication mainly dueto its good mechanical and thermal properties, and excellent chemical resistance.The remainder of the discussion will focus on organic membranes, specificallyPVDF-based membranes.2.4 Membrane FoulingThe capital cost for membrane systems has become comparable to that of conventionalsystems, due to the emergence of new and lower cost membranes. However, highplant maintenance and operating costs for fouling control, limit the wide spreadapplication of membrane technologies [87]. Fouling is defined as the decrease ofmembrane performance due to the deposition or adsorption of soluble and particulatematerials onto and into the membrane [87]. Membrane fouling generally depends onthe interaction between the membrane and the feed solution, as well as the chosenoperating conditions.11The permeate flux is generally proportional to the pressure drop across themembrane (i.e., trans-membrane pressure) and inversely proportional to the productof the absolute viscosity of the solution being filtered and the total hydraulicresistance of the filtration system, as presented in the Equation 2.2.J =∆PµRt(2.2)where, J is permeate flux (L/m2·hr), ∆P is trans-membrane pressure (Pa), µ isabsolute viscosity (Pa·s) and Rt is the total hydraulic resistance (N·s/m3) of thefiltration system.Permeability (LP) is another widely used parameter to express the filterabilityof a filtration system (Equation 2.3).LP =1µRt=J∆P(2.3)where, LP is permeability (l/m2·hr·Pa).When filtering a solution, material retained by the membrane can accumulateon and/or in the membrane (Figure 2.3), leading to an increase in the hydraulicresistance of the system over time. Therefore the total resistance of the filtrationsystem is the sum of the resistance of the membrane itself, as well as the resistanceinduced by the accumulated material (i.e., foulants) as presented in the Equation 2.4[46].12Rt = Rm +R f (t) (2.4)where, Rt = resistance of the filtration system (at any time, t), Rm = resistance of themembrane, and R f (t) = resistance induced by the accumulated foulants (at any time,t).Insert your company logo or name here AWWA/AMTA© 1 Membrane pore Foulants Syed Zaki Abdullah, UBC Permeate Feed Side Figure 2.3: Accumulation of materials during membrane filtration processFrom Equations 2.2 to 2.3 it can be deduced that, an increase in the resistancedue to the accumulation of retained material results in a decline in the permeate fluxor an increase in the trans-membrane pressure for constant pressure operation andconstant flux operation, respectively.132.4.1 Types of Membrane FoulingIn general, membrane fouling is classified into two general categories: reversiblefouling and irreversible fouling. The concept of reversible fouling and irreversiblefouling can be confusing, as different definitions have been used in publishedliterature. In the present study, irreversible fouling is defined as fouling whichcannot be removed by any methods (permanent). However, it should be noted thatin much of the published literature, irreversible fouling is defined as the foulingthat cannot be controlled hydraulically. Typically no distinction is made betweenchemically reversible fouling and completely irreversible (permanent) fouling. Forthe present study fouling is classified into three different types: hydraulicallyreversible, chemically reversible and irreversible fouling.Hydraulically (also know as physically) reversible fouling, results from thebuildup of particles on the surface of the membrane or in the membrane pores, as aresult of the convective drag forces of permeate flux towards the membrane, balancedby the back-transport of the solute and particles away from the membrane due tomolecular diffusion, shear-induced forces, or inertial lift forces [13]. Hydraulicallyreversible fouling can be minimized by inducing turbulence at the membrane surfaceand/or thorough back-flushing or back-pulsing [73, 134]. Chemically reversiblefouling results from the buildup of materials which cannot be removed throughthe induced turbulence and/or back-flushing or back-pulsing, but can be removedby chemical cleaning [117]. Chemically reversible fouling occurs primarily dueto the adsorption of the foulant on or in the membrane [81]. Irreversible foulingresults from permanent changes to the membrane surface and cannot be controlledhydraulically or chemically.14Fouling resistances are normally measured by using a series of filtration experi-ments [4, 18, 42, 46, 48]. For the first experiment, clean water is filtered trough themembrane to measure the resistance of the membrane. The second experiment isdone by filtering the feed solution of interest and the total resistance of the filtrationsystem is measured over time. The resistance induced by the accumulated foulantscan be calculated by subtracting the membrane resistance from the total resistance.The membrane is then subjected to hydraulic cleaning (i.e., turbulence and/orback-flushing or back-pulsing). The hydraulically reversible fouling resistance isthe difference between the resistance induced by the accumulated foulants, beforeand after hydraulic cleaning. The membrane is then subjected to chemical cleaning.The chemically reversible fouling resistance is the difference between the resistanceinduced by the accumulated foulants before and after chemical cleaning. Irreversiblefouling resistance is the difference between the resistance after the chemical cleaningand the resistance of the membrane (Figure 2.4).15 Hydraulically reversible fouling resistance Chemically reversible fouling resistance Irreversible fouling resistance Figure 2.4: Different types of membrane fouling resistanceThe magnitude of the hydraulically reversible resistance and the sum of chemicallyreversible and irreversible resistances have been reported to range from 3×1011 to2890×1011 m−1 and 30×1011 to 2900×1011 m−1, respectively, for wastewatertreatment applications [23]. From the wide ranges it is clear that resistance valuesare highly variable. This is in part due to the fact that fouling is dependent onthe system configuration, hydrodynamic conditions, feed solution composition andvolume filtered, and most of these parameters differ between studies.162.4.2 Membrane Fouling ModelsThere are four typically used semi-empirical models that have been developedto describe membrane fouling mechanisms [64] . The standard blocking modelassumes that particles deposit on the pore walls of the membrane and, hencereduce the membrane pore diameter. The complete blocking model assumes thateach particle reaching the membrane completely blocks a membrane pore. Theintermediate blocking model is similar to complete blocking but assumes thateach particle reaching the membrane does not necessarily block a membrane pore.Particles can land across the entire membrane surface and on other particles. Thecake filtration model assumes that particles in the solution are too large to enter thepores, and are retained and accumulate on the membrane surface to form a ”cake”layer.These four classic models can provide some insight into the mechanisms ofmembrane fouling. However, these models have limited use in identifying the designand operating conditions that minimize fouling in full-scale membrane systems.This is largely because the back-transport of particles (i.e., foulants) induced by thecross-flow or gas-sparging and backwashing, that is typically applied in full-scalemembrane systems to minimize hydraulically reversible fouling, is not consideredin these classic models. In addition, these models assume that only one mechanismgoverns the fouling for a given membrane system, whereas, typically full-scalemembrane systems operate over a range of operating conditions where fouling of themembranes potentially results from more than one fouling mechanism. A numberof variations of the classic models have also been developed by many researchersto address some of the limitations of the four classic models. Some of these more17comprehensive models consider that the different fouling mechanisms can occursuccessively [19, 70, 142, 158], some consider that the different fouling mechanismscan occur simultaneously [16, 17, 76], and some consider the back-transport ofparticles from a membrane surface [14, 46].Numerous mechanistic models have also been developed to describe the suspen-sion particle transport mechanism near the membrane surface. Brownian diffusion,shear-induced diffusion, inertial lift and the surface transport are the four prevalentmechanisms which have been used to correlate the behavior of the back-transportof the particles from the membrane surface [13]. It is important to recognize thatall these models are based on highly restrictive hypotheses (i.e., foulants are inert,spherical, mono-disperse particles, and the hydrodynamics at the membrane surfaceare characterized by a laminar flow regime). Nonetheless, these mechanistic modelsprovide insight into the role of the hydrodynamics near a membrane surface onthe accumulation of retained material, and have been used to determine the size ofparticles that are likely to be accumulated at a membrane surface under specificfiltration conditions [13].2.5 Effect of Feed Matrix Characteristics and MembraneProperties on FoulingSeveral studies have tried to quantify the degree of fouling in a water or wastewaterapplication by different foulants or groups of foulants (e.g., fractions of the mixedliquor). Natural Organic Matter (NOM) [117], and soluble and colloidal fractionsof the mixed liquor are considered to be the main foulants [52] for water andwastewater, respectively. Extracellular polymeric substances (EPS), such as carbo-hydrates (polysaccharides), proteins and humic-like substances can block membrane18pores, adsorb on the membrane surface and/or form a gel structure on the membranesurface [44, 54, 69].The exact composition of the organic material that caused adsorptive fouling wasreported to vary for different membrane materials [82]. Polysaccharide-like organicmatter is considered to be the main cause of adsorptive fouling [81]. However,protein-like organic material also contributes to the adsorptive fouling. In addition,some inorganic materials, such as iron and manganese, can precipitate on themembrane and contribute to adsorptive fouling as well. It has also been reportedthat the presence of calcium can increase the extent of adsorptive fouling [15].Membrane fouling has been reported to be affected by membrane properties,such as pore size, pore size distribution, pore shape, surface roughness and membranehydrophobicity [35, 82]. If the membrane pore size is larger than the size of theparticle in the feed solution, then fouling mechanisms such as pore blocking orpore narrowing may occur during filtration [87]. In general, it has been postulatedthat increasing the pore size may decrease the deposition onto the membranes atthe expense of internal adsorption [109]. Although there are numerous studies onthe effect of membrane pore size on the extent of fouling, conflicting observationsare reported that describe this relationship [28, 48, 63, 86]. Part of the reason forthe inconsistent observations is due to the fact that the relationships between themembrane fouling and pore size (and pore size distribution) are strongly dependenton the feed solution characteristics and in particular, the particle size distribution,as well as the cross-flow velocity, the operating permeate flux, the back-washingfrequency and test duration in fouling studies [33, 79, 87].Membrane pore shape has been reported to have influence on membrane fouling.When two membranes with an identical average pore size are compared, membranes19with elliptical pores (i.e., membrane with pore aspect ratio of more than 1) exhibitedless fouling than the membrane with circular pores [87]. A higher fouling ratehas been observed for membranes with a rougher surface [161], during studies ofadsorptive fouling due to extracellular polymeric substances with ultra-filtrationmembranes.Numerous studies have demonstrated that hydrophobic membranes are moreprone to fouling than hydrophilic membranes in water and wastewater applications,as adsorption of solute happens more readily in hydrophobic membranes [32]. Thisis mainly because primary foulants are hydrophobic in nature, which favors thehydrophobic interaction between solutes and membrane material [23].2.6 General Methods for Membrane Fouling ControlMembrane fouling control is one of the most important and challenging tasks inmembrane treatment system design and operation. Strategies have been developedfor decades to avoid fouling or decrease the extent of fouling. Feed solutionpretreatment, optimization of operating parameters/conditions (e.g., permeate fluxand gas sparging), optimization of membrane characteristics and cleaning are thefour main approaches to control membrane fouling. These major membrane foulingcontrol techniques for low-pressure membranes were reviewed comprehensively byHilal et. al. [65].2.6.1 Feed Pre-treatmentPretreatment of the feed solution typically involves chemical and/or physicalprocesses which are able to remove solutes, colloidal and/or particulate materialresponsible for membrane fouling. Typically, chemical processes include coagulation,20followed by flocculation, to convert colloidal foulants into larger particles, andhence reduce the fouling due to internal adsorption inside the membrane poresor to increase the rate of back-transport [65]. Organic materials (e.g., NOM) canalso adsorb to flocs, removing the material from solution, and therefore reducingadsorption fouling [65].2.6.2 Optimized Operating ConditionsOptimization of operating parameters/conditions is an effective strategy to reducemembrane fouling. Commonly used techniques that can be optimized to reducemembrane fouling include inducing turbulence at a membrane surface (i.e., usinggas-sparging and/or high liquid cross-flow velocities), and back-flushing, back-pulsingand/or back-washing. For a submerged membrane filtration process, gas-spargingand high liquid cross-flow velocities have been extensively used to induce turbulanceand shear stress on the membrane surface. These increase the extent of back-transportof the foulant away from the membrane surface to prevent adhesion and depositionof foulant on the membrane surface [73, 87]. During back-flushing and back-pulsing,the flow through the membrane is periodically reversed to remove the foulants thataccumulate on the membrane surface and/or that plug the membrane pores.Other operating parameters, including operating modes (constant pressurevs. constant flux), permeate flux, duration of the filtration cycle, back-washingfrequency and gas-sparging have also been reported to influence fouling [33, 69].Field et al. [46] first introduced the concept of critical flux for micro filtration,hypothesizing that a critical flux exists, below which a decline of permeability withtime does not occur, and above which fouling with time is observed. The value ofcritical flux depends on a number of variables including the back-transport provided21by the turbulence generated by the imposed liquid flow and/or gas-sparging, aswell as the specific solute membrane interactions [96]. Defrance and Jaffrin [36]investigated the reversibility of membrane fouling in a Membrane bioreactor (MBR)process, and found that when the permeate flux was set below the critical flux, theTMP remained stable and the fouling was hydraulically reversible (i.e., cake fouling).On the other hand, when the filtration was done with a flux that was greater thanthe critical flux, the TMP increased and did not stabilize. Also, fouling was foundto only be partly hydraulically reversible. Selecting an appropriate permeate fluxwas concluded to be of crucial importance to obtain the best compromise betweenflux and TMP to operate a membrane filtration process economically. However, theconcept of critical flux generally cannot be applied directly, as solutions typicallycontain particles of various sizes, each with different back-transport rates, as wellas soluble compounds that can get adsorbed on the membrane surface, and whichare not influenced by hydrodynamic conditions [26, 29, 109, 158]. More recently,the concept of sustainable flux was introduced [87]. The sustainable flux can bedefined as the flux for which the TMP rises relatively slowly and frequent chemicalcleaning is not necessary [108]. By definition sustainable flux is a sub-critical fluxoperating mode.At relatively low permeate flux (i.e., below critical flux), fouling is predominantlydue to the adsorption of foulants on the membrane surface and in pores [29].Membranes operated with a low permeate flux may have a higher ratio of adsorptivefouling resistance to total fouling resistance than membranes operated with a higherpermeate flux.Cake fouling and adsorptive fouling are related to some extent. Material thatinitially loosely attaches to the membrane surface causing cake fouling, could22consolidate and generate adsorptive type fouling resistance [29]. Duration of thefiltration cycle influences the extent of the consolidation [66].Many studies reported that under sub-critical conditions, for long term filtrationoperation, when the transmembrane pressure (TMP) was plotted against filtrationtime, the TMP showed a trend of having two stages [26, 109]. In the first stage(i.e., stage 1), a relatively slow increase in TMP with respect to filtration time wasobserved (i.e., a low rate of TMP increase). After a certain permeability loss, asudden jump in the rate of TMP increase was observed (second stage). The fasterrate of TMP increase in the second stage (i.e., stage 2) leads to the low permeabilityvalue within a short time of the start of stage 2. For constant flux operation, theTMP and time can be replaced with normalized permeability and cumulative volumefiltered, respectively. Stage 1 is widely accepted to result from the adsorption offoulants onto the membrane surface and onto the pore walls. The adsorption offoulants leads to pore blocking and pore narrowing. The resulting pore blocking andpore constriction eventually generate local flux conditions that exceed the criticalflux, and therefore, lead to rapid fouling (stage 2) [71].Kim and DiGiano [79] demonstrated that an increase in the back-washingfrequency is more effective to decrease the rate of adsorptive fouling than theincrease of the back-washing time. Gas-sparging was also observed to be able toreduce adsorptive fouling to some extent, but the reduction of adsorptive foulingresistance was not as significant as that for cake fouling resistance [24].2.6.3 Modification of Membrane Surface CharacteristicsOptimization of membrane surface characteristics has also been considered to limitfouling. This can be done by careful selection of the membrane by considering23the pore size and distribution, surface charge, hydrophobicity, chemical stability,mechanical strength, and module packing density [74]. Membrane material/surfacemodification aims to change the membrane surface properties to reduce the adsorptionof foulants on the membrane surface by either reducing the attractive forces orincreasing the repulsive forces between foulants (i.e., solute) and the membrane[13]. Polyvinylidene difluoride (PVDF) is widely used as a core material forpolymeric membrane fabrication mainly due to its good mechanical and thermalproperties, and excellent chemical resistance. However, pure PVDF membranesare highly hydrophobic, making them more susceptible to fouling in water andwastewater treatment applications than membranes made from hydrophilic materials[45, 50, 65, 80, 101, 111]. For this reason, hydrophobic polymer-based membranesare often blended or coated with additives to introduce hydrophilic functional groupsinto the structure, which enhances the filtration performance of these membranes [65,105, 106, 112]. A variety of monomers and polymers has been used as hydrophilicadditives including, PVP [20, 21, 118], acrylic acid [47, 148], methacrylic acid [12,47], cellulose acetate [8], vinyl acetate [80], N-Vinylpyrrolidone [25, 114], glycidalmethacrylate [80, 155], hydroxyethyl methacrylate [146, 155], poly(ethylene glycol)methacrylate [12, 47, 143], 3-sulfopropyl methacrylate [47, 146], 2-dimethyl amino-ethyl methacrylate [143], 2-trimethylammonium-ethyl methacrylate chloride [143],and 2-Acrylamido-2-methylpropane sulfonic acid [53].2.6.4 Chemical CleaningDespite all the measures that can be taken to limit fouling, some of the retainedcontaminants can adsorb at and/or in the membrane during filtration. Chemicalcleaning is typically used for removing adsorbed foulants from the membranes. The24frequency and intensity (i.e., concentration of the cleaning agent and exposure time)of the chemical cleaning varies significantly, depending on the filtration system[117].Membrane chemical cleaning can be described as a 5-step process as listedbelow [117].1. Bulk reaction (i.e., hydrolysis) of the cleaning agents.2. Transport of the cleaning agent to the foulant layers attached to the membranesurfaces and on the membrane pores.3. Transport of the cleaning agent through the foulant layers to the membranesurface.4. Chemical reactions (i.e., oxidation) that solubilise and detach foulants.5. Transport of the foulants to the bulk solution.Types of chemical cleaningMaintenance cleaning is commonly used to maintain a desired membrane permeability,as well as to reduce the frequency of more periodic intensive chemical cleaning bypreventing excessive foulant build up. Maintenance cleaning is typically performedon a daily to weekly basis via chemically enhanced backflash. More intensiverecovery cleaning (or clean-in-place) is generally performed when permeabilitydrops below a certain level below which further filtration is no longer sustainable[87]. As recovery cleaning is done to recover permeability, longer soak times(typically 2-6 hours), higher cleaning agent concentrations, and higher soak tempe-ratures are applied compared to those used for maintenance cleaning. Recovery25cleaning is typically performed every 3 to 12 months, as needed, and can involve asequence of cleaning using different reagents (normally basic, oxidative and acidic).Very diverse cleaning practices were revealed in a survey of 87 water treatmentplants reported by AWWA [135]. More than half the plants surveyed carried outmaintenance cleaning on average less than once a week, while the recovery cleaningfrequency ranged from 0.2 to 50 per year, with a median of 4 per year.Common chemical cleaning agentsA large number of chemical cleaning agents are commercially available and manyof them are proprietary and recommended by the membrane manufacturers fortheir products. However, the most commonly used chemical cleaning agents fallinto six categories: oxidizing agents, acids, alkalis, surfactants, chelating agentsand enzymes [3]. The chemical cleaning agents have to be compatible with themembrane material. For most membranes used in water and wastewater treatment,the prevalent cleaning agents are sodium hypochlorite (for organic foulants) andcitric acid (for inorganic foulants) [87]. Oxidants (e.g., chlorine) are able to oxidizeorganic foulants (e.g., NOM). Oxidants also increase hydrophilicity of the foulnatsby increasing the amount of oxygen containing functional groups such as phenolicand carboxyl groups [93]. Table 2.1 summarizes the typical recovery chemicalcleaning protocols recommended by leading membrane suppliers.26Table 2.1: Typical recovery chemical cleaning protocols used by four leadingmembrane suppliers (from [87]).Chemicals Concentration(ppm)Protocols SupplierNaOCl 3000 Backflow through membrane MitsubishiCitric acid 2000 (2 hours) + Soaking (2 hours)NaOCl 2000 Backpulse and recirculate GE WaterCitric acid 2000-3000NaOCl 100 Recirculate through lumens MemcorCitric acid 2000NaOCl 5000 Backflow and soaking KubotaOxalic acid 10000However, sodium hypochlorite (NaOCl) is one of the most commonly usedchemicals for maintenance cleaning [87]. Although the typical duration of amaintenance cleaning cycle is considerably less than that of a recovery cleaningcycle, the cumulative exposure due to maintenance cleaning is significantly greaterthan for recovery cleaning, because maintenance cleaning is performed at a higherfrequency.272.7 Effectiveness of Chemical CleaningThe types of components that can be effectively removed by chemical cleaningare largely depended on the characteristics and strength of the chemical reagentused for the membrane cleaning and the foulants characteristics [15, 81]. Thecleaning chemicals either react with the foulant and change the morphology of thefoulant, or alter the surface chemistry of the foulant, so that it can be releasedfrom the membrane [149]. Many other important factors affect the chemicalcleaning efficiency, including cleaning time (i.e., exposure time to chemical cleaningagents), temperature, hydrodynamic conditions during the cleaning, membranecharacteristics, influent characteristics and filtration operation [93].The efficiency of chemical cleaning is normally expressed as the permeabilityrecovery in percent (i.e., removal of foulants) resulting from chemical cleaning(Equation2.5).% Permeability recovery= 100 ∗permeability o f the membrane a f ter chemical cleaningpermeability o f the virgin membrane(2.5)High permeability recovery can be achieved by the properly designed chemicalcleaning schemes. As an example, after each chemical cleaning, over 95% ofmembrane permeability recovery was reported for an application of UF membranefiltering activated sludge [51].282.8 Impact of Chemical Cleaning on the Operation andMaintenance CostThe cost associated with chemical cleaning includes the cost of chemicals andfiltration time lost during chemical cleaning. In addition, a number of studies havesuggested that chemical cleaning can reduce the operating life span of membranes,increasing the capital cost of the system for membrane replacement [117]. Resourcesrequired to replace an aged membrane include cost of the new membranes, properscheduling of the replacement, labor cost, management of the waste preservative(i.e., most of the polymeric membranes supplied with a glycerin coating to keep themembrane wet during shipment), and proper disposal of aged membranes.There is a large variation in cost of chemical cleaning. It depends on theoperational strategy (i.e., basis for scheduling the chemical cleans), the membranetype (e.g., MF vs. UF), and membrane configuration (e.g., hollow fibre vs. tubular).The overall cost associated with chemical cleaning can be ranged from 4% to 50%of the total membrane operation cost [116].It is now generally recognized that the capital cost advantage offered by higher-flux operation is more than offset by the greater operation and maintenance cost.This is mainly due to relatively higher chemical cleaning frequency and longerduration needed with higher-flux operation, which eventually affect the performanceand operating life of a membrane [135].292.9 Current Knowledge on the Effects of ChemicalCleaning on Membrane Surface Properties andFiltration PerformancesAlthough effective for fouling control, over extended periods of time, chemicalcleaning agents, such as sodium hypochlorite, can impact the physical/chemicalcharacteristics of the membrane, affecting not only the extent of fouling, but alsothe treatment performance and operating lifetime of membrane systems. For thisreason, membrane manufacturers typically specify a maximum lifetime exposuredose, defined as the cumulative product of the concentration of the chemical agent(C) and the duration of the cleaning (t), above which the performance of theirmembranes cannot be guaranteed.A number of studies have investigated the effects of chemical cleaning usingsodium hypochlorite on several non-PVDF-based polymeric membranes, such asPolysulfone (PS f ), polyether-sulfone (PES), or cellulose acetate [9, 10, 49, 88, 124,126]. Table 2.2 summarizes the major research articles on the effects of sodiumhypochlorite on polymeric membranes.30Table 2.2: Summary of the published work on the membrane ageing due to exposure to cleaning agents.Material Selectivity Agent Concentration,C (ppm)Exposuretime (hrs)Dose, Ct(ppm·hr)Use ofSEMRefPA RO NaOCl 25-625 16 0-10000 no [2]PES 10-20 kDa NaOCl 150 66-666 0-120000 no [10]CA 20 kDa NaOCl - - 2000-50000 no [9]PES, CA,PVDF10-55 kDa NaOCl - - - yes [8]PES UF Alkalinechlorideoxidant200-7600 severalmonths- yes [11]PSf/PVP - NaOCl+antioxidant100 24-384 2400-38400 no [21]PSf/PVP - NaOCl 100 24-384 2400-38400 no [20]PA NF, RO NaOCl 10-2000 1-24 10-48000 no [38]Continued on next page31Table 2.2 – Continued from previous pageMaterial Selectivity Agent Concentration,C (ppm)Exposuretime, t (hrs)Dose, Ct(ppm·hr)Use ofSEMRefPA NF NaOCl 10-2000 1-24 10-48000 no [37]PSf/PA RO NaOCl 40-4000 1 0-4000 no [43]PSU/PVP/PEG- NaOCl 400 24-2760 9600-1104000 no [49]PVDF MF NaOCl 20000 - 102000-510000 no [58]PVDF - NaOCl 10-100000 1-24 10000-2400000 yes [60]PAN/PVP - NaOCl 5000 2 10000 yes [75]PA RO NaOCl 100-2000 1 100-20000 no [83]PA RO NaOCl 100-2000 1 100-20000 no [84]PES, PVDF 30 kDa NaOCl 5000 10-240 50000-1200000 no [88]PES, PVDF 30 kDa NaOCl,NaOH1500 (NaOCl),40000 (NaOH)1-240 0-360000 no [89]Continued on next page32Table 2.2 – Continued from previous pageMaterial Selectivity Agent Concentration,C (ppm)Exposuretime, t (hrs)Dose, Ct(ppm·hr)Use ofSEMRefPVC UF NaOCl 1000-50000 24-96 1000-4800000 yes [95]PSf/PVC - NaOCl 4000 48-176 192000-704000 no [118]PVDF MF NaOCl 10000 168-3528 1680000-35280000 yes [119]CA/PVP UF NaOCl 50-200 24 120-4800 yes [122]PAN/PVP UF NaOCl - 24 - yes [121]CA/PVP UF NaOCl 4000 48 192000 yes [123]PSU/PVP UF NaOCl 2000-6000 48 96000-138000 yes [120]PSf/PVP UF NaOCl 100 24- >3072 240- >307200 yes [126]Continued on next page33Table 2.2 – Continued from previous pageMaterial Selectivity Agent Concentration,C (ppm)Exposuretime, t (hrs)Dose, Ct(ppm·hr)Use ofSEMRefPA/PSf NF Causticagents+NaOH+EDTA,NaOH+SDS- 18 - no [132]PA/PSf NF NaOH,Citric acid,SDS,EDTA- 18 - no [131]Continued on next page34Table 2.2 – Continued from previous pageMaterial Selectivity Agent Concentration,C (ppm)Exposuretime, t (hrs)Dose, Ct(ppm·hr)Use ofSEMRefPA NF NaOH,Citric acid,SDS,EDTA- 18 - no [130]PES UF NaOCl 400 0-2880 0-115200 no [140]PVC - NaOH 10000 - - yes [141]PVDF MF NaOCl 330 1.86-5.54equivalentyear of reallifeapplication31750-95100 yes [147]PVC - NaOCl 40000 0.5-14 20000-560000 no [151]Continued on next page35Table 2.2 – Continued from previous pageMaterial Selectivity Agent Concentration,C (ppm)Exposuretime, t (hrs)Dose, Ct(ppm·hr)Use ofSEMRefPES UF NaOCl 700 340-860 240000-600000 yes [154]PES UF NaOCl 1000 3-9 equivalentyear of reallifeapplication60000-250000 no [166]PSf UF NaOCl,proprietaryagent100-400 (both) 168-30249(both)16800-12099600 yes [124]Note: Polyamide (PA), Reverse osmosis (RO), Cellulose acetate (CA), Polyvinylpyrrolidone (PVP), Nano-filtration (NF)Polysulfone (PSU), Polyethylene glycol (PEG), Polyacrylonitrile (PAN)36These studies (Table 2.2) demonstrated that sodium hypochlorite exposurenegatively affects the mechanical strength [9], surface hydrophobicity [8, 10,88, 124], pore size [49, 126], and filtration performances [7] of the membranes.Limited research that focused on the effects of sodium hypochlorite cleaning onblended PVDF membranes [58, 60, 119, 147] also reported that some of themembrane characteristic parameters got negatively affected by the exposure tosodium hypochlo- rite. Studies also indicated that exposure to sodium hypochloritechanges the chemical compositions of the blended polymeric membranes [58,119, 140]. However, none of these studies those studied the effects of sodiumhypochlorite exposure on blended PVDF membranes have comprehensively addressedthe effects of sodium hypochlorite exposure on the important characteristic parametersincluding, the chemical composition, the important physical/chemical characteristics,filtration capacity and cleaning efficiency of the blended PVDF membranes. Thismakes it difficult to understand how the observed changes in the membrane characte-ristic parameters due the exposure to the sodium hypochlorite are correlated. Inaddition, none of these studies has focused on identifying the mechanism responsiblefor the reported effects.Most of the membrane ageing studies have not investigated the effect of theconcentration of sodium hypochlorite on the membrane characteristics [8–10, 49,88, 126]. Though, some studies have investigated the effect of the concentrationof sodium hypochlorite on the membrane characteristics [58, 119, 140]. In themajority of those studies, the effect of concentration for fixed exposure doseswere not reported, and the effects of the concentration of sodium hypochlorite wasreported for only few membrane characteristic parameters. These make it difficultto clearly identify the true effects of the concentration of sodium hypochlorite on37the observed changes of the membrane characteristic parameters.Although the exposure dose of chemical cleaning agents is commonly estimatedas the product of the concentration of the chemical cleaning agent during cleaning(C), and the duration of the cleaning (t), or Ct value, one study has demonstratedthat the Ct value does not adequately reflect the impact of sodium hypochloriteexposure on the physical/chemical characteristic parameters of the hollow-fiberspolysulfone (PSf) membranes [124]. Nonetheless, no alternative relationship wassuggested.2.10 Use of SEM for Membrane CharacterizationAs discussed in the previous sections, over time, the properties of the membranesurface can change due to the accumulation of foulants [40] or due to the ageing ofthe membrane material [2], both of which affect the performance of the membranesystem. Therefore, a proper understanding of the membrane surface properties isessential to assess how these can change over time. Scanning Electron Microscopy(SEM) and High Resolution SEM (HRSEM), also known as Field Emission SEM(FESEM) imaging, have been extensively used to characterize the membrane surfaceproperties, both qualitatively (e.g., visual observation) and quantitatively (e.g., poresize, pore shape and porosity). A literature review of all articles published in theJournal of Membrane Science in the volumes 351-400 (2010 to 2012) revealed thata relatively large number of studies (i.e., 37 in total) have used SEM and/or HRSEM(hereafter referred simply as SEM unless otherwise indicated) to characterize thepore size, pore size distribution, pore shape, porosity, surface roughness, foulingetc. of PVDF-based polymeric membranes [30, 34, 39, 56, 57, 59, 61, 68, 72, 77,78, 90–92, 94, 97, 98, 100, 102, 107, 110, 112, 128, 133, 136–138, 145, 152, 153,38156, 157, 157, 159, 164, 165]. Also, from Table 2.2, it can be observed that themajority of the membrane ageing studies have used SEM as a primary membranecharacterization technique.Limited information has been reported in the literature with respect to thesample preparation (i.e., whether metal coating had been used, identity of thecoating material, coating thickness) and imaging parameters (i.e., magnification,detector type, sample orientation) used, as summarized in Table 2.3.39Table 2.3: Summary of the information available in reviewed journal articles.Parameters Reported information Number of published journalsproviding information on theparameters of interest1SamplePreparationParametersCoating material 28Coating thickness 4ImagingParametersMagnification 30Detector type 2Sample orientation (i.e.,Tilt angle)01 total count; based on review of all the journal articles published in the 2010 to2012, in which SEM was used as a membrane characterization technique forPVDF-based polymeric membranes (i.e., 37 in total).Table 2.4, presents the general effects of the sample preparation and imagingparameters on the information obtained from SEM imaging.40Table 2.4: Effect of the different sample preparation and imaging parameterson SEM imaging.Parameter Effects on the SEM imagingCoatingmaterialCommonly used coating materials are Gold (Au),Gold-Platinum (AuPt), Gold-palladium (AuPd), Chromium(Cr) and Iridium (Ir).- Metal used affects structure of coating.- Gold and gold composite: often result in coating with uneventhickness, and a relatively thick layer is needed to get a uniformcoating.- Iridium/Chromium: provides smoother coating. Even coatingthickness can be achieve with a relatively thin layer.- Rate of oxidation of coating metal depends on the metal used.- Samples become unusable if the coating layer becomesoxidized.- Chromium: coating layer becomes oxidized quickly (i.e.,within few hours).- Iridium: has a relatively slow oxidation rate.Coatingthickness- Commonly used range for coating thickness is 2-5 nm.- Samples with an extremely thin layer of coating (i.e., <2 nm)can be imaged at moderately high magnification.Continued on next page41Table 2.4 – Continued from previous pageParameter Effects on the SEM imaging- However, a thin layer may result in an uneven distribution ofmetal, which can lead to charging, image drift, and/or beamdamage to membrane.- Polymeric membranes can be extremely beam sensitive andcan shrink, swell, or rupture when exposed to the electronbeam.- A thick layer can obscure detail and/or add fine-scale texture.Also a thick layer can fill in nanoscale gaps and also can bridgenanopores, which will lead to the inaccurate measurements ofthe membrane surface properties.Magnification - Magnification used to image the membranes has impact onthe resolution of the obtained images, which may influence theinformation on membrane surface properties obtained from theimages.Detector type - Typically SEM comes with two electron detectors.- Two detectors can be used independently (i.e., Upper DetectorUD, and Lower Detector - LD), or- In conjunction with one another (i.e., Mixed Detector - MD).- The choice of detector also implies a choice of signal.- Upper detector has high sensitivity, which results in a greatersignal and produce images with high resolution.Continued on next page42Table 2.4 – Continued from previous pageParameter Effects on the SEM imaging- Upper detector captures the image of only very top layer ofthe surface.- When some degree of subsurface imaging is desired, eitherlower or mixed detector mode is used.- The lower detector’s sensitivity is about half that of the upperdetector, so a higher beam energy or longer working distance istypically needed to collect the required signal. This results in anoisier or grainier image with low resolution than is normallyproduced using the upper detector.- The mixed detector mode is useful for giving a compositesurface and subsurface image, which can take on a 3D-likeeffect. The combination of signal may make imageinterpretation difficult when trying to determine true surfacefeatures.Sampleorientation(i.e., Tilt angle)- Tilting the sample relative to the electron beam increasesthe signal magnitude, and presents some vertical relief ofthe membrane surface to the beam, making the roughnessinformation more evident.- Tilting the sample highlights smaller differences in verticalrelief compared to samples that are orthogonal (i.e., 0 degreetilt angle) to the beam.Continued on next page43Table 2.4 – Continued from previous pageParameter Effects on the SEM imaging- However, the image becomes somewhat foreshortened on they-axis, while maintaining x-axis dimensionality.- Tilting of the membrane samples during SEM imagingis common, as typically membranes have minimal surfacetopography. Orthogonal viewing of the membranes results ina flat view of the surface, which may be more appropriate formeasuring pore diameter, as this provides a truer representationof the two-dimensional shape of the surface pores. However,this comes at the expense of being able to easily identify porestructure (e.g., false detection of pores, as a surface depressionmay appear as a pore).Although SEM imaging is commonly used to characterize membrane surfaceproperties, no standard approach has been adopted for SEM imaging of membranes.Also, limited information has been reported in the literature with respect to thesample preparation and imaging parameters used, as summarized in Table 2.3.However, in general, for any specimen, the sample preparation and imaging parameterscan significantly affect the information obtained from SEM imaging (Table 2.4). Nostudy has been performed to identify the sample preparation and imaging parameterset points that are the least sensitive to changes in the remaining parameters, on theobtained membrane characteristics from SEM imaging of polymeric membranes.44Chapter 3Objectives Of The Study3.1 Knowledge GapDespite the widespread use of sodium hypochlorite for membrane cleaning, limitedresearch has focused on the effects of sodium hypochlorite cleaning on blendedPVDF membranes [58, 119]. However, none of these studies those studied theeffects of sodium hypochlorite exposure on blended PVDF membranes have compre-hensively addressed the effects of sodium hypochlorite exposure on the importantcharacteristics parameters including, the chemical composition, the importantphysical/chemical characteristics, filtration capacity and cleaning efficiency of theblended PVDF membranes. Also, none of these studies has focused on identifyingthe mechanisms responsible for the reported effects. In addition, most studieshave not investigated the effect of the concentration of sodium hypochlorite onthe membrane characteristics [8–10, 49, 88, 126]. Also, no appropriate model isavailable to correlate the concentration of sodium hypochlorite and the exposuretime to the magnitude of the impact of exposure to sodium hypochlorite on the45physical/chemical characteristic parameters of the blended PVDF membranes.Although there are numerous published studies available on the effect ofmembrane characteristics (e.g., pore size and distribution, hydrophobisity etc.)on the membrane fouling for different types of membrane materials and filtrationapplications, no study has been performed to assess the effects of changes inthe membrane characteristics due to chemical cleaning, as well as membranefiltration performance in terms of cleaning requirement and cleaning efficiency(e.g., frequency of cleaning required, and efficiency of cleaning) .3.1.1 Objective #1The first objective of the present study was to comprehensively quantify the effectsof sodium hypochlorite exposure on changes in the physical/chemical characteristicsof blended PVDF-based hollow-fiber membranes, and to identify the mechanism(s)responsible for the changes.To address this objective the following hypotheses and research questions wereformulated.Hypothesis #1The exposure to sodium hypochlorite has an effect on the physical/chemicalcharacteristics (i.e., the content of the hydrophilic additive of interest, hydrophilicity,strength, pore size, pore shape and apparent porosity) of blended PVDF-basedpolymeric membranes.Research Question #1.1Does exposure to sodium hypochlorite remove some of the material components(i.e., hydrophilic additives) from blended PVDF membranes?46Research Question #1.2Does the removal of the material components (i.e., hydrophilic additives) dueto exposure to sodium hypochlorite affect the physical/chemical characteristics ofblended PVDF membranes?Research Question #1.3Does the extent of exposure to sodium hypochlorite (i.e., exposure dose)affect the amount of change of physical/chemical characteristics of blended PVDFmembranes?Research Question #1.4Does the concentration of the sodium hypochlorite affect the amount of changeof physical/chemical characteristics of blended PVDF membranes?Hypothesis #2The commonly used relationship to calculate the extent of exposure (i.e., theproduct of concentration (C) and time (t), or Ct value), does not adequately reflectthe impact of sodium hypochlorite on the physical/chemical characteristics (i.e., thecontent of the hydrophilic additive of interest, hydrophilicity, strength, pore size,pore shape and apparent porosity) of blended PVDF membranes.Research Question #2.1Can the magnitude of the impact of exposure to sodium hypochlorite on thephysical/chemical characteristics of blended PVDF membranes investigated besimply related to the commonly used relationship Ct, for any concentration and/orexposure dose?Research Question #2.2If Ct is not appropriate, what relationship can be used to accurately and47consistently model the impact of exposure to sodium hypochlorite on the physical/che-mical characteristics of blended PVDF membranes.Hypothesis#3Oxidation is the primary mechanism by which hydrophilic additives are removedfrom the membrane material during chemical cleaning.Research Question #3.1Do the hydrophilic additives of blended PVDF membranes become mineralizedwhen exposed to sodium hypochlorite?Research Question #3.2Does the initial concentration of the hydrophilic additives present in blendedPVDF membranes affect the removal rate of the hydrophilic additives?Research Question #3.3Does the concentration of sodium hypochlorite affect the removal rate of thehydrophilic additives present in blended PVDF membranes?3.1.2 Objective #2The second objective of the present study was to comprehensively quantify theeffects of sodium hypochlorite exposure on changes in the filtration performance(i.e., membrane filtration capacity and efficiency of cleaning) of blended PVDF-basedhollow-fiber membranes.To address this objective the following hypothesis and research questions wereformulated.Hypothesis #448The exposure to sodium hypochlorite has an effect on the filtration performance(i.e., frequency of cleaning required, and efficiency of physical and chemicalcleaning) of blended PVDF-based polymeric membranes.Research Question #4.1Do the changes in the physical/chemical characteristics of the membrane dueto the exposure to sodium hypochlorite affect the filtration performance of themembrane?Research Question #4.2Does the extent of exposure to sodium hypochlorite affect the filtration performanceof the membrane?Research Question #4.3Does the concentration of the sodium hypochlorite affect the filtration performanceof the membrane?3.2 Selection of a Consistent SEM Imaging ApproachAlthough SEM imaging is commonly used to characterize membrane surfaceproperties, no standard approach has been adopted for SEM imaging of membranes.Also, limited information has been reported in the literature with respect to thesample preparation (e.g., whether metal coating had been used, identity of thecoating material, coating thickness) and imaging parameters (e.g., magnification,detector type, sample orientation) used, as summarized in Table 2.3 (Chapter2). However, in general, for any specimen, the sample preparation and imagingparameters can significantly affect the information obtained from SEM imaging(see Table 2.4 in Chapter 2).Varying one or more sample preparation or imaging parameter(s) is a common49practise, and sometimes necessary, in SEM imaging to obtain a best quality imageof a sample [67]; for example, taking images at different tilt angles to obtain insightinto the 3-dimensional structure of a surface. In such cases, when varying theparameter(s) of interest, the set points for the other parameters should be selected toprovide the most consistent image quality. No study has been performed to identifythe sample preparation and imaging parameter set points that are the least sensitiveto changes in the remaining parameters, on the obtained membrane characteristicsfrom SEM imaging of polymeric membranes.The objectives of this part of the present study was to comprehensively quantifythe effects of sample preparation (i.e., coating metal and coating thickness) andimaging parameters (i.e., magnification, detector type, sample orientation) on themembrane properties of interest (i.e., the porosity (φ ), the average pore diameter(dav), the 90th percentile of cumulative pore size distribution (d90), and the poreaspect ratio (dt/dc)) of blended PVDF membranes, obtained from SEM imaging.Also, to identify the sample preparation and imaging parameter set points that arethe least sensitive to changes, and can produce consistent SEM images.HypothesisSample preparation (i.e., coating metal and coating thickness) and imagingparameters (i.e., magnification, detector type, sample orientation) has an effecton the observed membrane properties of interest of blended PVDF membranesobtained from SEM imaging.Research Questions#1Does the sample preparation and imaging parameters affect the qualitativeinformation (e.g., visual observation) of membrane properties of interest of blended50membrane obtained from SEM imaging?Research Questions#2Does the sample preparation and imaging parameters affect the quantitativeinformation of membrane properties of interest of polymeric membrane obtainedfrom SEM imaging?Research Questions#3Can optimum parameter set points be suggested for which a observed membraneproperties of interest will be least sensitive to a change of the parameters?51Chapter 4Materials and MethodsA description of all the materials and methods used in the present study is providedin the sections below.4.1 Commercial MembranesUnless otherwise indicated, all experiments were performed using virgin commercialsupported hollow-fiber membranes made of a blend of polyvinylidene difluoride(PVDF) and hydrophilic additives (HA) (ZeeWeed fibers; GE Water and Technologies,Oakville, ON, Canada). Due to the proprietary nature of the commercial membraneformulations, the type and amount of the hydrophilic additive of interest (HA) inthese membranes cannot be reported.4.2 Lab-cast MembranesFor some experiments, lab-cast membranes containing a blend of PVDF, polyvinylpyrrolidone (PVP) and the predominant hydrophilic additive of interest (HA)present in the commercial membranes, were made by the phase-inversion method.52The casting mixture was prepared by dissolving the PVDF, PVP and HA indimethylacetamide (DMAc) at a temperature of 95 ◦ C under continuous mixing.The mixture was spread over a glass plate with a casting knife of uniform thickness,and immersed immediately in a distilled water bath. Phase-inversion took place andthe membrane peeled off the glass by itself within 5 minutes. Details of the castingprocedure can be found elsewhere [85].4.3 Exposure ProcedureThe effect of sodium hypochlorite exposure on the physical/chemical characteristicsof the membrane material was investigated by exposing the commercial membranefibers to chemical cleaning solutions containing sodium hypochlorite at differentconcentrations for different durations. The concentrations, exposure times andcorresponding exposure doses (i.e., concentrations x exposure times) investigatedare listed in the Table 4.1. Each exposure experiment was performed in duplicate.53Table 4.1: Experimental conditions investigatedExposure dose (ppm·hr) Concentration (ppm) Exposure time (hr)1,5001 7502 20.002500,000 3600 139.0022200 22.5044300 11.251,000,000 3600 278.0022200 45.0044300 22.501,500,000 3600 417.0022200 67.5044300 33.752,000,000 36002 556.00222200 90.0044300 45.001 Equivalent to virgin membrane after initial chemical cleaning to removepreservative. 2 Conditions considered for challenge tests.Recommended maximum cumulative lifetime exposure doses for commercialPVDF-based membranes typically range from 500,000 ppm·hr to 1,000,000 ppm·hr,depending on the manufacturer [31, 135]. In the present study, cumulative exposure54doses ranging from 1,500 ppm·hr to 2,000,000 ppm·hr were considered to investigatethe effects of a wide range of exposure doses on the membrane characteristics. Notethat the concentrations of sodium hypochlorite used were higher than those typicallyused for chemical cleaning. This was done to reduce the time needed to achievea given exposure dose [119, 124]. By considering a range of concentrations ofsodium hypochlorite, the effects of both the exposure dose and concentration couldbe assessed independently. The pH of the cleaning solution was set to 10 by addinghydrochloric acid, as needed. For the commercial membranes used in the presentstudy, 750 ppm sodium hypochlorite was recommended as a cleaning agent, whichgenerally has a pH value close to 10. During exposure, the solution was replacedevery day to ensure that the concentration of sodium hypochlorite remained within1 percent of the targeted value. Once a target exposure dose was reached, themembranes were rinsed under flowing distilled water and stored in distilled waterprior to characterization, which was done within 2 weeks.4.4 Membrane Characterization TechniquesAttenuated Total Reflection-Fourier Transform Infrared Spectroscopy (ATR-FTIR)and Nuclear Magnetic Resonance (NMR) were used to characterize the membranein terms of its chemical components. As previously discussed in Chapter 3, thecontent of the HA present in the membrane material was of particular interest.ATR-FTIR (hereafter referred simply as FTIR) spectra were recorded using aNicolet spectrometer (model - Thermo Nicoet Nexus 870). Proton NMR spectrawere recorded using a Bruker spectrometer (model - Bruker 600 UltraShield).Prior to characterization using FTIR or NMR, the commercial membrane materialwas peeled from the inner support and dried overnight in a vacuum drier at room55temperature. lab-cast membranes with different blend ratios (i.e., HA:PVDF ratiosof 30:70, 20:80, 08:92 and 04:96 by weight) were used to confirm the wavelengthassociated with the dominant peak of the major functional group of the HA by FTIRanalysis and the concentration of the HA measurement by NMR analysis.The 0.2% offset yield strength was measured with a Multi-Purpose TensileTester (model KES-G1, Kato Tech. Co., Ltd.) to characterize the structural strengthof the membrane material [154, 166]. The yield point could not be easily definedon the stress-strain curves of the membranes. For this reason, the 0.2% offset yieldpoint was used as a measure of the structural strength of the membranes [160]. The7 cm long coupons used for testing were prepared by first slicing the membranelongitudinally, then peeling it off from its inner support, and then, drying it overnightin a vacuum drier at room temperature.Contact angle measurements were used to characterize the hydrophilic natureof the membrane surface. Contact angles of the membrane samples dried overnightat 35 ◦ C were measured using the modified sessile drop method [27]. Using a10 µL glass syringe, a 0.5 µL drop of deionized distilled water was placed onthe membrane surface and a horizontally mounted conventional digital opticalmicroscope (Motic B3 Professional Series) was used to measure the angle (MoticImages Plus 2.0) between the membrane surface and the water droplet, exactly 5seconds after the drop was placed on the membrane surface.Scanning Electron Microscopy (SEM) was used to qualitatively characterize themembrane surface in terms of apparent porosity and pore size. SEM imaging detailcan be found in Section 4.7.Clean water permeation tests were used to characterize the permeability ofthe membranes at the different stages of filtration and cleaning [43]. The tests56were performed at a constant permeate flux of 30 liters per square meter per hour(LMH) using a custom made membrane module#1 submerged in distilled water atroom temperature (20±1 ◦ C) and the trans-membrane pressure was continuouslymeasured (Omega PX243A vacuum transducer) and recorded (Onset ComputerCorporation, HOBO[R] U12 data-logger). The custom made membrane modules#1were made by potting two 33 cm long membrane fibers (i.e., effective filtration areaof 0.01447 m2) at both ends using epoxy into bulkheads as illustrated in Figure 4.1.The end of the fibers potted to the top bulkheads were connected to the permeateline through the top bulkheads while the other ends of the fibers potted to the bottombulkhead were sealed. The membranes were pre-compressed at 50 kPa pressureprior to any filtration test.57Figure 4.1: Membrane module#1 (a), and bench scale filtration system (b)used for clean water permeation tests.A challenge test with a dextran solution (Sigma-Aldrich, 50:50 w mixture ofT70 and T500 at a total concentration of 1 g/L in distilled water) was used toestimate the selectivity of the membrane, which is indirectly related to the apparentpore size of the membranes [22]. Dextran T70 and T500 were selected to covera wide range of molecular weights (approximately 1 to 10000 kDa). Membranemodules, similar to those used for the clean water permeation tests were used. The58feed and permeate were analyzed by size exclusion chromatography with a highperformance liquid chromatograph (TOSOH Bioscience, LLC model-G3000SWXLTSK gel column coupled with a GE Power and Water, Model-Sievers 900 totalorganic carbon analyzer). The difference between the total organic carbon (TOC)content of the feed and permeate from a membrane for a particular molecular weightcorresponded to the amount of dextran retained by the membrane for that apparentmolecular weight.4.5 Assessment of the Affinity of the Membrane forFoulantsFiltration of activated sludge through the membranes was conducted to characterizethe affinity of the membrane for foulants.4.5.1 Experimental SetupFiltration of activated sludge testes were conducted by submerging the custommade membrane module#2 into an aerated activated sludge tank of a membraneenhanced biological phosphorus removal (MEBPR) wastewater treatment pilot plantfacility located at the University of British Columbia, Vancouver (Figure 4.2). Thesolid retention time and the hydraulic retention time of the wastewater treatmentsystem was 12 days and 10 hours, respectively. The total suspended solids (TSS)concentration in the aerated activated sludge tank was approximately 4000 mg/l.Operational details of the pilot plant facility and the characteristics of the activatedsludge can be found elsewhere [1, 129].59HoldingTank•InfluentPO Side C: Constant inflowMixerP13 PerrieateC--)^-tankP11^i vPermeationAir sparging(A4^(A3) (A2)^(A1)OMAnoxic12SewerV P8Flow 1.4 L/min (from 1:00 AM to 7:00 AM)6.0 L/min (from 7:00 AM to 1:00 PM)3.7 L/min (from 1:00 PM to 1:00 AM)MixerAnaerobic I)P9P1P12P2Flow 3.7 L/minPrimaryClarifier P4P3P6Sewer)o^OMr^Mixer^Mixer^(A4^I II IAnaerobic'^Anoxic ^0I)^I II AerobicP5PermeationAir sparging(A3) (A2>^CADP7 PerrieatetankP10Sewer1Side V: Variable inflowlFigure 3.1. Schematic of the UBC wastewater treatment pilot plant.Al, A3 and A4: Air sparging system of operational membrane (OM) and test membrane (TM) modules, respectively.A2: Fine bubble air sparging system to maintain DO concentration.PO to P13: Pumps in the pilot plant system.HoldingTank•InfluentPO Side C: Constant inflowMixerP13 PerrieateC--)^-tankP11^i vPermeationAir sparging(A4^(A3) (A2)^(A1)OMAnoxic12SewerV P8Flow 1.4 L/min (from 1:00 AM to 7:00 AM)6.0 L/min (from 7:00 AM to 1:00 PM)3.7 L/min (from 1:00 PM to 1:00 AM)MixerAnaerobic I)P9P1P12P2Flow 3.7 L/minPrimaryClarifier P4P3P6Sewer)o^OMr^Mixer^Mixer^(A4^I II IAnaerobic'^Anoxic ^0I)^I II AerobicP5PermeationAir sparging(A3) (A2>^CADP7 PerrieatetankP10Sewer1Side V: Variable inflowlFigure 3.1. Schematic of the UBC wastewater treatment pilot plant.Al, A3 and A4: Air sparging system of operational membrane (OM) and test membrane (TM) modules, respectively.A2: Fine bubble air sparging system to maintain DO concentration.PO to P13: Pumps in the pilot plant system.HoldingTank•InfluentPO Side C: Constant inflowMixerP13 PerrieateC--)^-tankP11^i vPermeationAir sparging(A4^(A3) (A2)^(A1)OMAnoxic12SewerV P8Flow 1.4 L/min (from 1:00 AM to 7:00 AM)6.0 L/min (from 7:00 AM to 1:00 PM)3.7 L/min (from 1:00 PM to 1:00 AM)MixerAnaerobic I)P9P1P12P2Flow 3.7 L/minPrimaryClarifier P4P3P6Sewer)o^OMr^Mixer^Mixer^(A4^I II IAnaerobic'^Anoxic ^0I)^I II AerobicP5PermeationAir sparging(A3) (A2>^CADP7 PerrieatetankP10Sewer1Side V: Variable inflowlFigure 3.1. Schematic of the UBC wastewater treatment pilot plant.Al, A3 and A4: Air sparging syste of operational membrane (OM) and test membrane (TM) modules, respectively.A2: Fine bubble air sparging system to maintain DO concentration.PO to P13: Pumps in the pilot plant system.HoldingTank•InfluentPOSide C: Constant inflowMixerP13 PerrieateC--)^-tankP11^i vPermeationAir sparging(A4(A3) (A2)^(A1)OMAnoxic12 SewerV P8Flow 1.4 L/min (from 1:00 AM to 7:00 AM)6.0 L/min (from 7:00 AM to 1:00 PM)3.7 L/min (from 1:00 PM to 1:00 AM)MixerAnaerobic I)P9P1P12P2Flow 3.7 L/minPrimaryClarifierP4 P3P6Sewer)o^OMrer^Mixer^(A4^IIIIAnaerobic'^Anoxic ^0I)^III AerobicP5PermeationAir sparging(A3) (A2>^CADP7 PerrieatetankP10Sewer1Side V: Variable inflowlFigure 3.1. Schematic of the UBC wastewater treatment pilot plant.Al, A3 and A4: Air sparging system of operational membrane (OM) and test membrane (TM) modules, respectively.A2: Fine bubble air sparging system to maintain DO concentration.PO to P13: Pumps in the pilot plant system.HoldingTank•InfluentPO Side C: Constant inflowMixerP13 PerrieateC--)^-tankP11^i vPermeationAir sparging(A4^(A3) (A2)^(A1)OMAnoxic12SewerV P8Flow 1.4 L/min (from 1:00 AM to 7:00 AM)6.0 L/min (from 7:00 AM to 1:00 PM)3.7 L/min (from 1:00 PM to 1:00 AM)MixerAnaerobic I)P9P1P12P2Flow 3.7 L/minPrimaryClarifier P4P3P6Sewer)o^OMr^Mix r^Mixer^(A4^I II IAnaerobic'^Anoxic ^0I)^I II AerobicP5PermeationAir sparging(A3) (A2>^CADP7 PerrieatetankP10Sewer1Side V: Variable inflowlFigure 3.1. Schematic of the UBC wastewater treatment pilot plant.Al, A3 and A4: Air sparging system of operational membrane (OM) and test membrane (TM) modules, respectively.A2: Fine bubble air sparging system to maintain DO concentration.PO to P13: Pumps in the pilot plant system.Figure 4.2: Schematic of the UBC MEBPR pilot plant.Note: TM indicates the location of the membrane module#2.60The custom made membrane module#2 consisted of four custom made membranemodule#1 (i.e., effective filtration area of 0.05788 m2) operated in parallel, asillustrated in Figure 4.3. The open end of the membrane module#1 were connectedto the permeate line through the bottom bulkheads, while the other end of the fiberswere connected to the top bulkhead and sealed.Figure 4.3: Membrane module#2 used for membrane filtration performanceexperiments.4.5.2 Filtration OperationFiltration of activated sludge testes were performed at a constant permeate flux of25 LMH and the trans-membrane pressure was continuously measured (OmegaPX243A vacuum transducer) and recorded (Onset Computer Corporation, HOBO[R]U12 data-logger). The permeate flux of 25 LMH was selected as it is representative61of the permeate flux typical of full-scale MBRs (20 to 30 LMH) [51, 115]. Permeabi-lity and normalized permeability values were calculated by using Equations 2.3 andand 4.1. Normalized permeability values were plotted against cumulative volumefiltered.Normalized Permeability =LPLP(o)(4.1)where, LP = Permeability at any given stage of filtration, and LP(o) = Permeabilityof unfouled membrane.Once the TMP reached 70 kPa, filtration through the membrane was stopped.This limiting TMP of 70 kPa was chosen, because typically in full-scale operation,the type of membrane used in the present study would generally undergo cleaningwhen the TMP reaches to 70 kPa [51, 115].4.5.3 Assessment of the Efficiency of Physical and Chemical CleaningMembranes were cleaned physically and chemically after every filtration test withactivated sludge. Physical cleaning of the fouled membrane was conducted bybackwashing the membrane module with distilled water overnight. Following thebackwash, the membrane surfaces were wiped with wet Kim-wipe paper, and thenrinsed with a gentle stream of distilled water for 5 minutes. Chemical cleaning ofthe physically cleaned membranes was conducted by soaking the membranes in a500 ppm NaOCl solution for 24 hours.Clean water permeation tests were performed following physical and chemical62cleaning to quantify the extent of fouling that could not be reversed with eachcleaning process. The permeability recovery was calculated with respect to thepermeability of the membrane prior to the filtration tests (i.e., permeability of theunfouled membrane) (Equations 4.2, 4.3 and 4.4).LP (physically irrecoverable) =LP (physical)LP(o)(4.2)LP (irreversible) =LP (physical+chemical)LP(o)(4.3)LP (chemically recoverable) =LP (physical+chemical)−LP (physical)LP(o)(4.4)where, LP (physically irrecoverable) = physically irrecoverable permeability,LP (irreversible) = irreversible permeability = chemically irrecoverable permeability,LP (chemically recoverable) = chemically recoverable permeability,LP (physical) = permeability after physical cleaning,LP (physical+chemical) = permeability after physical and chemical cleaning, andLP(o) = permeability of unfouled membrane.The affinity of the membranes for foulants was assessed by six parameters. Thefirst parameter was the volume of activated sludge filtered per unit area of membranebefore a sudden increase in the rate of permeability loss with respect to volumefiltered during filtration tests (stage 2) was observed. The volume filtered values wererepresentative of membrane filtration capacity, as typically once the filtration tests63proceed to stage 2, TMP reached to the specified maximum value (70 kPa) withina very short period of time, and a membrane cleaning was necessary. The secondparameter was the appearance of the membrane surface (i.e., apparent amount ofmaterial accumulated on the membrane surface) due to the remaining accumulatedfoulant after physical cleaning and chemical cleaning, both with respect to virginmembrane. The third, fourth and fifth parameters were the physically irrecoverablepermeability, chemically irrecoverable permeability, and chemically recoverablepermeability. The sixth parameter was the rate at which the permeability increasedduring chemical cleaning.4.6 Assessment of the Fate of HAA fate assessment analysis was conducted to investigate the mechanism by which theHA is removed from the blended membrane material during sodium hypochloritecleaning. Blended lab-cast membranes with different blend ratios (i.e., HA:PVDFratios of 100:00, 30:70, 00:100 by weight) were soaked in solutions containingdifferent concentrations of sodium hypochlorite (i.e., 9000 ppm and 25900 ppm)in distilled water. Additional sodium hypochlorite was added to the solutions asneeded to maintain a constant concentration during the test. The pH was kept at10 by adding hydrochloric acid as needed. The concentrations of TOC and sodiumhypochlorite of the soak solutions were recorded over time to monitor the potentialrelease and mineralization of material from the membrane over time.Standard Methods 2350 B and 5310 B were used to measure the concentrationof the sodium hypochlorite and TOC in the solutions, respectively [6].644.7 SEM ImagingMembrane imaging was used to qualitatively and quantitatively characterize thesurface properties of the membranes. Qualitatively, the membranes were characte-rized based on the appearance of the images taken using SEM. Quantitatively, theimages taken using SEM were used to estimate the pore size, pore size distribution,pore shape and porosity of the membranes. Because the quantitative informationobtained using SEM is affected by the image itself, the effect of parameters thatcould impact the quantitative information of interest was assessed; and based on thethe findings the SEM imaging approach was selected.4.7.1 Impact of Sample Preparation and Imaging Parameters on theMembrane Surface PropertiesMembranesCommercial membrane fibers exposed to a chemical cleaning solution containing3600 ppm of sodium hypochlorite in distilled water for 139 hours, resulting in anexposure dose of 500,000 ppm·hr, were used. Details of the exposure procedure arediscussed in Section 4.1.Sample preparation and SEM imagingThe membranes were dried overnight in a vacuum drier at room temperature andthen coated with a metal using a sputter coater (Cressington 208HR) to achievea specific coating thickness. A quartz crystal monitor (Ted Pella Inc. MIM 10)with a resolution of 0.1 nm was used to monitor the coating thickness during thecoating process. A Field Emission Scanning Electron Microscope was used (HitachiS-4700) to acquire the images. The sample preparation (i.e., coating metals and65coating thickness) and imaging parameters (i.e., detector, magnification and tiltangle) considered for the two types of membranes are presented in Table 4.2. Thesewere selected to cover the range of parameters reported in the reviewed literature(Section 2.10).Table 4.2: Sample preparation and imaging parameters consideredSample Preparation Parameters Imaging ParametersCoatingmetalCoatingthicknessDetector1 Magnification Tilt angle ofthe sampleholder base( ◦)Gold 2 nm L 25K 05 nm M 50K 30UIridium 2 nm L 25K 05 nm M 50K 30UGold-Palladium 2 nm L 25K 05 nm M 50K 30U1: L = Lower, M = Mixed, U = UpperIn the discussion which follows later in this dissertation, images are labeledaccording to the sample preparation (i.e., coating metals and coating thickness) and66imaging parameters (i.e., detector type, magnification and sample orientation) used.As an example, an image labeled Au2nmLD25K30tilt was obtained with a goldcoating, a 2 nm coating thickness, a lower detector, a 25K magnification and a 30 ◦tilt angle.Analysis of the SEM imagesSEM imaging was used to quantitatively characterize the membranes based onthe surface porosity (φ ), the average pore diameter (dav), the 90th percentile ofcumulative pore size distribution (d90), and the pore aspect ratio (dt/dc). Matlabcodes developed to obtain the membrane surface characteristics from the SEMimages, using the Matlab and Simulink software package (MathWorks, Inc.) arepresented Appendix A.1. During the image analysis RGB images were convertedto gray scale binary images and pixels with less than a certain value (i.e. thresholdvalue) were considered belonged to the pores. The threshold values were empiricallyfound by visual observation of the efficiency of the pore identification (i.e., maximizingthe detection of the obvious pores and minimizing the detection of false pores)[125].Threshold values did not differed more than 2% for the images obtained withany given parameter combinations.The pore pixels in contact with each otherwere considered belonged to a single pore. The area of a pore was calculatedby multiplying the total number of the pixels in the pore by the equivalent area of apixel. The surface porosity (hereafter referred simply as porosity) was calculated bydividing the total cumulative number of the pixels in the pores by the total numberof the pixels in a SEM image. The pore diameter was estimated by assuming thepore area was circular. The pore aspect ratio was the average ratio of the Major axislength (dt) to the Minor axis length (dc) of an ellipse equivalent in shape to that of a67pore. Figure 4.4 summarized the work flow diagram used for the image analysis.68Figure 4.4: Flow chart for the SEM image analysis.69Sensitivity of the parametersVarying one or more sample preparation or imaging parameter(s) is a commonpractise, and sometimes necessary, in SEM imaging to obtain a best quality imageof a sample [67]; for example, taking images at different tilt angles to obtain insightinto the 3-dimensional structure of a surface. In such cases, the set points for theother parameters should be selected to provide the most consistent image quality,when varying the parameter(s) of interest. To identify the sample preparation andimaging parameter set points that were the least sensitive to changes, a comparativeanalysis was performed, where all but one of the sample preparation and imagingparameters were fixed, while the remaining parameter was varied. The effect ofchanges in the variable parameters was quantified as a percent difference in themembrane properties of interest using Equation 4.5.Percent di f f erence = 100∗x1− x2x1(4.5)where, x1 and x2 are the values of a membrane property of interest (e.g., φ ) obtainedwith an first (e.g., iridium) and a second (e.g., gold) set point, respectively, for asample preparation or imaging parameter, while the remaining sample preparationor imaging parameters were fixed. As an example, the difference in the porosity(φ ) values due to the use of different coating metal (i.e., gold and iridium) for theimages obtained with 2 nm coating, and using lower detector, 50K magnificationand 0 ◦ tilt angle was calculated using Equation 4.5 is presented in Equation 4.6.70Di f f erence(%) (Au− Ir)(2nmLD50K0tilt)= 100∗(φAu2nmLD50K0tilt −φIr2nmLD50K0tilt)φAu2nmLD50K0tilt(4.6)4.8 QA/QCAll the experiments and the analysis were repeated at least two times. The membranephysical/chemical characteristic parameters investigated were normalized withrespect to those of the virgin membranes, so that the relative changes in theseparameters could be quantified. The effect of NaOCl exposure on the membraneproperties of interest (e.g., contact angle) was expressed as a percent of that of thevirgin membrane using Equation 4.7.Percent change = 100∗y1− y2y1(4.7)where, y1 and y2 are the values of a membrane property of interest (e.g., contactangle) of the virgin membrane and the membrane exposed to NaOCl, respectively.All comparisons were assessed based on the average values and the 95% percentcoefficient interval of the results of the repeated experiments. For some of themembrane characteristic parameters investigated, it was impractical to produce alarge number of data. For those parameters, instead of the 95% percent coefficientinterval, the minimum and the maximum values of the results of the repeated71experiments were considered.72Chapter 5SEM Imaging of Membranes:Importance of SamplePreparation and ImagingParametersA version of this chapter has been submitted to a Journal.• Abdullah, S. Z., and Be´rube´, P. R. (2013). SEM imaging of membranes:Importance of sample preparation and imaging parameters. Journal ofMembrane Science, 463 (2014) 113125.5.1 IntroductionAs discussed in the Chapter 2, SEM imaging has been extensively used to characterizethe membrane surface properties, both qualitatively (e.g., visual observation) and73quantitatively (e.g., pore size, pore shape and porosity) [30, 34, 39, 56, 57, 59, 61,68, 72, 77, 78, 90–92, 94, 97, 98, 100, 102, 107, 110, 112, 128, 133, 136–138, 145,152, 153, 156, 157, 157, 159, 164, 165]. Also, from Table 2.2 (Chapter 2), it canbe observed that the majority of the membrane ageing studies have used SEM asa primary membrane characterization technique to assess the effects of chemicalcleaning on the membrane characteristics.Although SEM imaging is commonly used to characterize membrane surfaceproperties, no standard approach has been adopted for SEM imaging of membranes.Also, limited information has been reported in the literature with respect to thesample preparation (e.g., whether metal coating had been used, identity of thecoating material, coating thickness) and imaging parameters (e.g., magnification,detector type, sample orientation) used, as summarized in Table 2.3 (Chapter2). However, in general, for any specimen, the sample preparation and imagingparameters can significantly affect the information obtained from SEM imaging(see Table 2.4 in Chapter 2).Varying one or more sample preparation or imaging parameter(s) is a commonpractise, and sometimes necessary, in SEM imaging to obtain a best quality imageof a sample [67]; for example, taking images at different tilt angles to obtain insightinto the 3-dimensional structure of a surface. In such cases, when varying theparameter(s) of interest, the set points for the other parameters should be selected toprovide the most consistent image quality.In this chapter, the effect of sample preparation (i.e., coating metals and coatingthickness) and imaging parameters (i.e., detector type, magnification and sampleorientation) that could impact the qualitative and quantitative information of interestobtained from SEM images was assessed. Matlab codes were developed to obtain74the membrane properties of interest (i.e., the porosity (φ ), the average pore diameter(dav), the 90th percentile of cumulative pore size distribution (d90), and the poreaspect ratio (dt/dc)) from the SEM images.To identify the sample preparation and imaging parameter set points that werethe least sensitive to changes, a comparative analysis of the obtained membranesurface characteristics of the SEM images was performed. In the analysis, imagesacquired with all but one of the sample preparation and imaging parameters werefixed, while the remaining parameter was varied, were compared. Based on thefindings, the approach that generated the most consistent results from SEM imagingwas selected and used in the subsequent work.5.2 Typical Results from SEM Imaging and ImageAnalysisA typical image obtained by SEM imaging and the pores identified by the customMatlab application are presented in Figure 6.15. Details of the custom Matlabapplication can be found in Appendix A.1.75! "! "! #! $! %! &! '! (!(a) (b) Figure 5.1: Typical SEM images of membrane.(a: original image; b: pores identified with the custom Matlab application;Au2nmUD50k0tilt).The repeatability of the measurement performed (i.e., φ , dav, d90 and dt/dc)was assessed by quantifying the spread of the results obtained, when a randomlyselected subset of 5 images (i.e., images were collected from 3 different membranesamples) were obtained with a given combination of sample preparation andimaging parameters (e.g., Ir2nmUD50k0tilt) were considered. The coefficients ofvariation for the φ , dav, d90 and dt/dc were 0.1, 0.016, 0.025 and 0.002, respectively,indicating that the measurements were reproducible (Table 5.1). As an example, thespread of the pore size distributions for 5 of the images presented in Figure 5.2 arecomparable.76Table 5.1: Repeatability of the typical values of membrane properties ofinterest.Membrane propertiesof interestAverage Coefficient ofvariationφ 8.3% 0.1dav 0.026 µm 0.016d90 0.035 µm 0.025dt/dc 2.00 0.002Note: Values were obtained from randomly selected subset of 5 SEM images(i.e., images were collected from 3 different samples) of a typical membrane! "! "! #! $! %! &! '! (! )! *! !"+!0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 dav (micron) 20 18 16 14 12 10 8 6 4 2 0 Frequency (%) Figure 5.2: Typical pore size distribution.(Randomly selected subset of 5 images; Ir2nmUD50k0tilt).77The values of φ , dav, d90 and dt/dc obtained form the SEM images acquired withdifferent parameter combinations, ranged from 0.04 to 0.22, 0.017 to 0.042 micron,0.023 to 0.066 micron, and 1.7 to 2.3, respectively. These values are comparablewith the published values of φ (i.e. 0.06 to 0.15) and dav (i.e. 0.015 to 0.045 micron)for the type of membrane used in the present study [41, 127]. This indicated thatthe SEM imaging and the image analysis techniques used in the present study wascapable of capturing the actual membrane surface properties.5.3 Qualitative AnalysisQualitative information on the membranes was obtained based on a visual observationof the SEM images. Typical images of membranes obtained with different samplepreparation (i.e., coating metals and coating thickness) and imaging parameters (i.e.,detector type, magnification and sample orientation) are summarized in AppendixA.2. As presented in Figure 5.3, the sample preparation and imaging parametershad a substantial effect on the appearance of the SEM images.78! "! "! (a) (b) (c) (d) (e) (f) (g) (h) (i) (j) (k) (l) #!Figure 5.3: Typical HRSEM images of membranes(a:Au2nmLD25k30tilt; b:Au2nmMD25k30tilt; c:Au2nmUD25k30tilt;d:Ir2nmLD25k30tilt; e:Ir2nmMD25k30tilt; f:Ir2nmUD25k30tilt;g:AuPd2nmLD25k30tilt; h:AuPd2nmMD25k30tilt;i:AuPd2nmUD25k30tilt; j:Au5nmUD25k30tilt; k:Au5nmUD50k30tilt; l:Au5nmUD50k0tilt; Inserts are magnified (3.6x) versions of the originalimages).79Images of membranes coated with gold (Figure 5.3a-c) suggested that thesemembranes had greater surface roughness (i.e., higher density and/or relativelylarger height of the mountainous peaks) than membranes coated with iridium (Figure5.3d-f) or gold-palladium (Figure 5.3g-i), for a given combination of imagingparameters. Coating with gold-palladium resulted in images of membranes thatappeared to have the least surface roughness. The same trend was observed forall of the combinations of imaging parameters used (results are not shown for allcombinations). Images of membranes obtained with the LD appeared to have thegreatest surface roughness (Figure 5.3a,d,g). However, these images had poorlydefined features regardless of the coating metal used. Images of membranes obtainedwith the MD (Figure 5.3b,e,h) and the UD (Figure 5.3c,f,i) were similar, andsuggested that the membranes had lower surface roughness and more definedfeatures than those imaged with the LD. Images of membranes obtained witha thinner coating (i.e., 2 nm gold) (Figure 5.3c) appeared to have less surfaceroughness than those obtained with a thicker coating (i.e., 5 nm gold) (Figure 5.3j),while those obtained with a lower magnification (i.e., 25K) (Figure 5.3j) appeared tohave a lower surface roughness than those obtained with a higher magnification (i.e.,50K) (Figure 5.3k). Also, images of membranes obtained with a 30 degree tilt angle(Figure 5.3k) suggested that these membranes had a greater surface roughness anda less definite pore boundary than those imaged with a 0 degree tilt angle (Figure5.3l). These results indicate that the qualitative information (i.e., surface roughness,well-defined features, pore boundary sharpness) obtained from SEM images variesbased on the sample preparation and imaging parameters applied.805.4 Quantitative AnalysisTo expand upon the qualitative analysis, a quantitative analysis of the SEM imageswas performed. Typical values of the membrane properties of interest (i.e., theporosity (φ ), the average pore diameter (dav), the 90th percentile of cumulativepore size distribution (d90), and the pore aspect ratio (dt/dc)) for all the samplepreparation (i.e., coating metals and coating thickness) and imaging parameters (i.e.,detector type, magnification and sample orientation) are summarized in AppendixA.3. Figure 5.4 presents an example of the extent to which sample preparationparameters can have an effect on the membrane properties of interest (i.e., φ , dav,d90 and dt/dc).81! "! "! #! $! %! &! '! (! )! *! "+! ""! "#! "$!Figure 5.4: Average values of the observed membrane properties of interestobtained from SEM images acquired with 5 nm coating, UD, 50Kmagnification, 0 degree tilt.(a: Porosity; b: Pore diameter; c: 90th percentile of cumulative poresize distribution; d: Pore aspect ratio; Error bars correspond to the 95%confidence interval of the repeated measurements).825.5 Sensitivity of the Membrane Properties to theParametersTypical images of membranes obtained with all the different sample preparation(i.e., coating metals and coating thickness) and imaging parameters (i.e., detectortype, magnification and sample orientation) are summarized in Appendix A.2. Acomparative quantitative analysis of these SEM images was conducted to assess thesensitivity of membrane properties of interest to the parameters investigated. Theeffect of the variation in each parameter on the apparent membrane properties wasconsidered while all other parameters were kept constant. The effect of changesin the variable parameters was quantified as a percent difference in the membraneproperties of interest using Equation 4.5.5.5.1 Coating MaterialThe effect of changing the coating material was considered first. Presented in Figure5.5 are the differences in the membrane properties of interest (i.e., φ , dav, d90 anddt/dc) for each of the sample preparation and imaging parameter combinationsconsidered when varying the coating material.83! "! "! #! $! %! &! '! (! )! *! "+! ""! "#! "$! "%! "&! "'! "(! (a) (b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dt/dc dav d90 Figure 5.5: Effects of coating metal on the average observed membraneproperties of interest. (Difference between the observed membraneproperties for the membranes coated with (a) iridium and gold, (b)iridium and gold-palladium; Error bars correspond to the 95% confidenceinterval of the 5 repeated measurements; ? were placed for thecombinations for which the values were missing for at least one ofthe comparing conditions).84Figure 5.5a presents the results when comparing iridium (x1) and gold (x2)coating, while Figure 5.5b presents the results when comparing iridium (x1) andgold-palladium (x2) coating. Images of membranes coated with iridium suggestedthat these membranes had significantly different φ , dav, d90 and dt/dc values thanthose coated with gold (Figure 5.5a) or gold-palladium (Figure 5.5b), for mostof the sample preparation and imaging parameter combinations. For example,the dav value for the membranes with 2 nm iridium coating was approximately20 and 15% lower than those obtained with 2 nm gold (Figure 5.5a) or 2 nmgold-palladium (Figure 5.5b) coating, respectively, when the imaging parametercombination consisted of lower detector, 50K magnification and 30 degree tilt wasused. These results demonstrated that the coating metal has a significant effect onthe observed values for the membrane properties of interest. Hence, the type ofmetal should be standardized if results from different studies are to be compared. Itshould be noted that the images obtained using several parameter combinations forgold-palladium coated samples resulted in images with poorly defined features and,as a result, no quantitative information could be obtained from those images. Forthis reason, in the following section (Figures 5.6 to 5.9), results from the samplesprepared with gold-palladium coating were not considered.5.5.2 Coating ThicknessPresented in Figure 5.6 are the differences in the membrane properties of interest(i.e., φ , dav, d90 and dt/dc) for each of the sample preparation and imaging parametercombinations considered when varying the coating thickness.85! "!(a) (b) !"#$ #$ "#$%&'()(*+($,-.$ !"#$ #$ "#$%&'()(*+($,-.$!"#$ #$ "#$%&"#'#()*$%&"#'+#()*$,&"#'#()*$,&"#'+#()*$-&"#'#()*$-&"#'+#()*$%&."'#()*$%&."'+#()*$,&."'#()*$,&."'+#()*$-&."'#()*$-&."'+#()*$&/0121341$567$8498*$ :$ 8;<$ 8=#$!"#$##% #$##% "#$##%&'()*"#+,#-./%&'(0*"#+,#-./%&'(1*"#+,#-./%&'()*&"+,#-./%"'()*"#+#-./%"'()*"#+,#-./%"'(0*"#+#-./%"'(0*"#+,#-./%"'(1*"#+#-./%"'(1*"#+,#-./%"'()*&"+#-./%"'()*&"+,#-./%"'(0*&"+#-./%"'(0*&"+,#-./%"'(1*&"+#-./%"'(1*&"+,#-./%*23454'64%789%:6;:/% <% :=>% :?#%dt/dc dav d90 "! #! $! %! &! '! (! )! *! "+! ""! "#! "$! "%!Figure 5.6: Effects of coating thickness on the average observed membraneproperties of interest.(a: Difference between the observed membrane properties values forthe membranes coated with 2 nm and 5 nm (a) gold, (b) iridium;Error bars correspond to the 95% confidence interval of the 5 repeatedmeasurements; ? was placed for the combination for which the valueswere missing for at least one of the comparing conditions).Figure 5.6a presents the results when comparing 2 nm (x1) and 5 nm (x2) goldcoating, while Figure 5.6b presents the results when considering either 2 nm (x1)or 5 nm (x2) iridium coating. Images of the membranes coated with 2 nm ofgold suggested that these membranes had lower apparent dt/dc values, but higherdav, d90 values than those coated with 5 nm of gold (Figure 5.6a) for most of thesample preparation and imaging parameter combinations. Also for the images86of membranes coated with 2 nm of gold suggested that these membranes had alower φ values than those coated with 5 nm of gold (Figure 5.6a) for four of theeleven sample preparation and imaging parameter combinations considered. Forthe remaining seven sample preparation and imaging parameter combinations, thedifferences in the φ values were highly variable, and no statistically significanteffect could be reported at a 95% confidence level.Gold is known to have a tendency to agglomerate during sputter coating creatinguneven surface coverage [144]. This may have contributed to the observed higherdt/dc values, due to an increased aspect ratio with narrowing pore profiles; the sameeffect could account for lower dav, d90 for membranes coated with thicker goldcoating as some of the smaller pores could be entirely covered with thicker goldcoating. Images of membranes coated with 2 nm and 5 nm of iridium suggested thatthese membranes had similar dt/dc, φ , dav and d90 values (Figure 5.6b) for most ofthe sample preparation and imaging parameter combinations.The results demonstrate that the observed membrane properties of interestobtained from the SEM images of the membranes coated with gold were sensitiveto the coating thickness. However, the variability was statistically insignificant forthe membranes coated with iridium, for most of the sample preparation and imagingparameter combinations investigated. Hence, if gold is to be used as a coating metal,the coating thickness should be standardized if results from different studies areto be compared. On the other hand, if iridium is used as a coating metal, differentcoating thickness for different samples can be used during sample preparation, ifnecessary. However, when the φ value is the major membrane property of interest,care should be taken before varying iridium coating thickness, as many of theparameter combinations showed sensitivity of the φ values to the change in coating87thickness (i.e., IrUD25K30tilt and IrLD50K0tilt).5.5.3 Detector TypePresented in Figure 5.7 are the differences in the membrane properties of interest(i.e., φ , dav, d90 and dt/dc) for each of the sample preparation and imaging parametercombinations considered when varying the detector type.88! "!!"#$ #$ "#$$$%&'()(*+($,-.$(c) (d) !"#$ #$ "#$$$%&'()(*+($,-.$!"#$ #$ "#$%&'"#(#)*+$%&'"#(,#)*+$%&'%"(#)*+$%&'%"(,#)*+$"&'"#(#)*+$"&'"#(,#)*+$"&'%"(#)*+$"&'%"(,#)*+$$-./010&20$345$6276+$ 8$ 69:$ 6;#$(a) (b) !"#$ #$ "#$$%&'()(*+($,-.$ !"#$ #$ "#$$%&'()(*+($,-.$!"#$ #$ "#$%&'"#(#)*+$%&'"#(,#)*+$%&'%"(#)*+$%&'%"(,#)*+$"&'"#(#)*+$"&'"#(,#)*+$"&'%"(#)*+$"&'%"(,#)*+$$-./010&20$345$6276+$ 8$ 69:$ 6;#$!"#$##% #$##% "#$##%&'()*"#+,#-./%&'(0*"#+,#-./%&'(1*"#+,#-./%&'()*&"+,#-./%"'()*"#+#-./%"'()*"#+,#-./%"'(0*"#+#-./%"'(0*"#+,#-./%"'(1*"#+#-./%"'(1*"#+,#-./%"'()*&"+#-./%"'()*&"+,#-./%"'(0*&"+#-./%"'(0*&"+,#-./%"'(1*&"+#-./%"'(1*&"+,#-./%*23454'64%789%:6;:/% <% :=>% :?#%dt/dc dav d90 "! #! $! %! &! '! (! )! *! "+! ""! "#! "$! "%! "&! "'! "(! ")! "*! Figure 5.7: Effects of detector on the average observed membrane propertiesof interest.(Difference between the observed membrane properties values whenimaged with the lower detector than those obtained using the upperdetector for the samples coated with (a) gold, (b) iridium; Differencebetween the observed membrane properties values when imaged withthe mixed detector than those obtained using the upper detector for thesamples coated with (a) gold, (b) iridium; Error bars correspond to the95% confidence interval of the 5 repeated measurements; ? was placedfor the combination for which the values were missing for at least one ofthe comparing conditions).89Figure 5.7a and b present the results when comparing UD (x1) and LD (x2)detector with gold and iridium coating, respectively; while Figure 5.7c and d presentthe results when considering either UD (x1) or MD (x2) detector with gold andiridium coating, respectively. Membranes coated with either gold or iridium hadlower dt/dc values and higher dav and d90 values when imaged with the LD thanthose obtained using the UD (Figure 5.7a,b), for most of the sample preparation andimaging parameter combinations. Although no consistent trend was observed for thechanges in the φ values, the images of membranes coated with either gold or iridium,and imaged with the LD suggested that these membranes had significantly differentφ values than those obtained using the UD, for several parameter combinations(Figure 5.7a,b).When comparing the images obtained with the UD with the images obtainedwith the LD, the membrane properties of interest (i.e., φ , dav, d90 and dt/dc)were more sensitive for membranes coated with gold than membranes coated withiridium, as the differences were significantly larger for the gold coated membranes(Figure 5.7a) than those for the iridium coated membranes (Figure 5.7b). Similarly,images of membranes coated with gold suggested that these membranes hadsignificantly different dt/dc, φ , dav and d90 values when imaged with the UDthan those obtained using the MD, for most of the sample preparation and imagingparameter combinations (Figure 5.7c). In contrast to that, images of membranescoated with iridium suggested that these membranes had similar dt/dc, φ , dav andd90 values for images obtained with the UD and MD (Figure 5.7d) for most of thesample preparation and imaging parameter combinations.The results demonstrated that the observed membrane properties of interestwere sensitive to all the detectors used during imaging of the membranes coated90with gold (i.e., use of LD, MD and UD all resulted in different values for membraneproperties of interest for any given parameter combinations). Similarly, the observedmembrane properties of interest were sensitive to the detectors used during imagingof the membranes coated with iridium. However, the variability was lower for themembranes coated with iridium than that of membranes coated with gold. Also,the use of MD and UD during imaging of iridium-coated membranes resulted invery little variability of membrane properties of interest for any given parametercombination.Hence, if gold is to be used as a coating metal, the use of detectors should bestandardized if results from different studies are to be compared. On the other hand,if iridium is used as a coating metal, UD and MD can be used interchangeablyduring imaging, if necessary.5.5.4 MagnificationPresented in Figure 5.8 are the differences in the membrane properties of interest(i.e., φ , dav, d90 and dt/dc) for each of the sample preparation and imaging parametercombinations considered when varying the magnification.91! "!(a) (b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dt/dc dav d90 "! #! $! %! &! '! (! )! *! "+! ""! "#!Figure 5.8: Effects of magnification on the average observed membraneproperties of interest.(a,b: Difference between the observed membrane properties valueswhen imaged with 25K magnification than those obtained using 50Kmagnification; Confidence interval corresponds to the 95% confidenceinterval of the 5 repeated measurements; ? was placed for thecombination for which the values were missing for at least one of thecomparing conditions).Figure 5.8a presents the results when comparing 25K (x1) and 50K (x2) magnifi-cation with gold coating, while Figure 5.8b presents the results when consideringeither 25K (x1) or 50K (x2) magnification with iridium coating. The magnificationdid not affect the φ or the dt/dc values obtained from the images of membranescoated with either gold (Figure 5.8a) or iridium (Figure 5.8b), for most of thesample preparation and imaging parameter combinations. However, the imaging92suggested that membranes had significantly higher dav and d90 values with 25Kmagnification than with 50K magnification (Figure 5.8a,b). The lower pore diametervalues observed with 50K magnification is likely related to the greater number ofsmaller pores that can be detected at a higher magnification and/or to the erroneousgrouping of closely spaced pores into a single large pore which was observed toperiodically occur at a lower magnification.The results demonstrated that the dav and d90 values were very sensitive tothe magnification used during imaging of the membranes coated with either goldor iridium. However, the variability was lower for the membranes coated withiridium than that of membranes coated with gold for most of the sample preparationand imaging parameter combinations investigated. Hence, when the dav and d90values are the major membrane properties of interest, magnifications should bestandardized if results from different studies are to be compared. This is true forboth gold- and iridium-coated membranes.5.5.5 Tilt AngleThe final parameter considered was tilt angle. Presented in Figure 5.9 are thedifferences in the membrane properties of interest (i.e., φ , dav, d90 and dt/dc) foreach of the sample preparation and imaging parameter combinations consideredwhen varying the tilt angle.93! "!!"#$ #$ "#$%&'()(*+($,-.$ !"#$ #$ "#$%&'()(*+($,-.$!"#$ #$ "#$%&'()"#*$%&'+)"#*$%&',)"#*$%&'()%"*$%&'+)%"*$%&',)%"*$"&'()"#*$"&'+)"#*$"&',)"#*$"&'()%"*$"&'+)%"*$"&',)%"*$)-./0/&1/$234$51657$ 8$ 59:$ 5;#$(b) (a) !"#$##% #$##% "#$##%&'()*"#+,#-./%&'(0*"#+,#-./%&'(1*"#+,#-./%&'()*&"+,#-./%"'()*"#+#-./%"'()*"#+,#-./%"'(0*"#+#-./%"'(0*"#+,#-./%"'(1*"#+#-./%"'(1*"#+,#-./%"'()*&"+#-./%"'()*&"+,#-./%"'(0*&"+#-./%"'(0*&"+,#-./%"'(1*&"+#-./%"'(1*&"+,#-./%*23454'64%789%:6;:/% <% :=>% :?#%dt/dc dav d90 "! #! $! %! &! '! (! )! *! "+! ""! "#! "$! "%! "&!Figure 5.9: Effects of tilt angle on the average observed membrane propertiesof interest.(a,b: Difference between the observed membrane properties values whenimaged using 0 degree angle than those obtained using and 30 degree tiltangle; Confidence interval corresponds to the 95% confidence interval ofthe 5 repeated measurements; ? was placed for the combination for whichthe values were missing for at least one of the comparing conditions).Figure 5.9a presents the results when considering either 0 (x1) or 30 (x2) degreetilt angle with gold coating, while Figure 5.9b presents the results when consideringeither 0 (x1) or 30 (x2) degree tilt angle with iridium coating.Images of membranes coated with gold suggested that these membranes hadhigher dt/dc, φ , dav and d90 values when imaged with a 0 degree tilt angle than thoseimaged with a 30 degree tilt angle (Figure 5.9a), for more than half of the sample94preparation and imaging parameter combinations. The effect was more pronouncedfor the samples coated with 5 nm gold than 2 nm gold. Shielding of some proportionof the pores, which can occur in titled positions due to the rough nature of thegold-coated surfaces, likely contributed to the lower observed values, and thickercoating appears to worsen this phenomenon. For the images of membranes coatedwith iridium, there were changes in the values of the dt/dc, φ , dav and d90 for someparameter combinations with changing tilt angle; however, no consistent trend wasobserved. Also the effects of tilt angle on the membrane properties of interest weresignificantly less pronounced for iridium-coated membranes (Figure 5.9b) thanfor gold-coated membranes (Figure 5.9a). The relatively smooth surface coatinggenerated with iridium likely decreases potential shielding effect.The results demonstrated that the observed membrane properties of interestwere sensitive to the tilt angle used during imaging of the membranes coated witheither gold or iridium. However, the variability was lower for the membranes coatedwith iridium than that of membranes coated with gold for most sample preparationand imaging parameter combinations.Hence, if gold is to be used as a coating metal, the tilt angle should bestandardized if results from different studies are to be compared. Special careshould be taken for thicker gold coating. On the other hand, if iridium is used as acoating metal, different tilt angles for different samples can be used during imaging,for most of the parameter combinations. However, when the φ value is the majormembrane properties of interest, care should be taken before using different tiltangles, as several of the parameter combinations showed relatively large sensitivityof the φ values to the change in tilt angle (i.e., Ir5nmUD25K and Ir2nmLD50K).955.5.6 Impact of Individual Parameters Regardless of theCombination of the Other ParametersA multi-factor multi-way analysis of variance (ANOVA) was performed to identifythe sampling preparation and imaging parameters that significantly affected thedifferent membrane properties of interest, regardless of the combination of theparameters. As presented in the Table 5.2, the φ and dt/dc values were affected byall of the sample preparation and imaging parameters considered (based on a 95%confidence interval). Similarly, dav and d90 values were affected by the majority ofthe sample preparation and imaging parameters considered. The only exceptionswere φ , which was not affected by the detector used, and dav and d90, which werenot affected by either the coating metal or coating thickness.Table 5.2: P-values for membrane properties of interest obtained from themulti-way ANOVA analysis.Source of variation Membrane properties of interestφ dav d90 dt/dcCoating metal 0.0487 0.8164 0.3565 0.0155Coating thickness 0.0006 0.4752 0.3956 0.0001Detector 0.3647 0.0001 0.0001 0.0001Magnification 0.0005 0.0001 0.0001 0.0296Tilt Angle 0.0186 0.0172 0.0067 0.0001Note: Membrane properties with P-value 0.05 identified in bold are significantlyaffected by individual or interactive sample preparation or imaging parameters).965.6 Conclusion and Selection of the Consistent SEMImaging ApproachThe results presented in the Section 5.3 to 5.5.6 indicate that all the samplepreparation and imaging parameters considered affect the SEM images. Theconclusions that can be drawn from these results are as follows.• In general, the apparent membrane properties were less sensitive to the samplepreparation or imaging parameters when membranes were coated with iridiumthan with gold. Hence, iridium should be used as the coating metal in thestudies where coating thickness and/or, one or more imaging parameters needto be varied.• However, when the φ values are the major membrane properties of concern,care should be taken before using different iridium coating thickness ordifferent tilt angle, as several of the parameter combinations showed significantsensitivity of the φ values to the change in coating thickness or tilt angle.• Also, for iridium-coated membranes, UD and MD can be used interchangeablyduring imaging, if necessary.• When the dav and d90 values are the major membrane properties of concern,different magnifications should not be used during imaging, as this willintroduce artificial variability among the similar samples and comparisonbetween different samples will become difficult. This is true for both themembranes coated with either gold or iridium.In general, the membrane properties of interest obtained from the SEM images,were less sensitive to the parameters when membranes were coated with iridium97and imaged with UD or MD. However, any change in the coating thickness,magnification or tilt angle significantly effected the membrane properties of interestobtained from the SEM images. Hence, if possible, sample preparation and imagingparameters should not be varied during a study.For the present study, 5 nm iridium coating, upper detector, 0 ◦ tilt angle and30K magnification was selected as the SEM imaging approach for imaging of themembranes exposed to sodium hypochlorite (Chapter 6 and 7).98Chapter 6Impact of sodium hypochloriteexposure on the physical/chemicalcharacteristics of PVDF-basedmembranesA version of this chapter is published in the journal Water Research.• Abdullah, S. Z., and Be´rube´, P. R. (2013). Assessing the effects of sodiumhypochlorite exposure on the characteristics of PVDF-based membranes.Water Research, 47 (2013) 5392 - 5399.6.1 IntroductionAs discussed in the Chapter 2, although chemical cleaning is necessary, as wellas effective for fouling control, over extended periods of time, chemical cleaning99agents, such as sodium hypochlorite, can impact the physical/chemical characteristicsof the membrane [124]. For this reason, membrane manufacturers typically specifya maximum lifetime exposure dose, defined as the cumulative product of theconcentration of the chemical agent (C) and the duration of the cleaning (t), abovewhich the performance of their membranes cannot be guaranteed [31, 135].Hydrophobic PVDF-based membranes are increasingly being used for waterand wastewater treatment applications because of their ability to effectively removemany contaminants of interest. Pure PVDF has excellent mechanical and chemicalresistance to chemical cleaning agents [60]. However, in general, hydrophobicmaterials, such as PVDF, are more susceptible to fouling in water and wastewatertreatment applications than hydrophilic materials. For this reason, hydrophobicpolymer-based membrane products are typically blended (i.e., modified) or coatedwith hydrophilic additives (HA) to make the membrane surface hydrophilic [65].A number of studies have investigated the effects of chemical cleaning usingsodium hypochlorite on several non-PVDF-based polymeric membranes, such aspolysulfone (PSf), polyether-sulfone (PES) and cellulose acetate [9, 10, 49, 88, 124,126]. These studies demonstrated that sodium hypochlorite exposure negativelyaffects the mechanical strength [9], surface hydrophobicity [8, 10, 88, 124], poresize [49, 126], filtration performances [7] of these membranes.Despite the widespread use of sodium hypochlorite as cleaning agent, limitedresearch has focused on the effects of sodium hypochlorite cleaning on blendedPVDF membranes [119], and none has focused on identifying the mechanismsresponsible for the reported effects. In addition, most studies have not investigatedthe effect of the concentration of sodium hypochlorite on the membrane characteristics[8–10, 49, 88, 126]. Also, no appropriate model is available to correlate the100concentration of sodium hypochlorite and the exposure time to the magnitude of theimpact of exposure to sodium hypochlorite on the physical/chemical characteristicparameters of the blended PVDF membranes.To assess the effects of sodium hypochlorite exposure on changes in thephysical/chemical characteristics of blended PVDF-based hollow-fiber membranesand to identify the mechanism(s) responsible for the changes, PVDF-based membraneswere exposed to sodium hypochlorite at different concentrations for varying periodsof time. Details of the membrane ageing procedure can be found in Section 4.3. Thephysical/chemical characteristics of the virgin membranes and membranes exposedto sodium hypochlorite were compared. The membranes were characterized basedon chemical composition (FTIR and NMR), mechanical strength (yield strength),surface hydrophilicity (contact angle), pore size and porosity (scanning electronmicroscopy and challenge test). Details of the membrane characterization techniquesare presented in Section 4.4 and analysis of the results are presented in Section 6.2.The exposure dose to chemical cleaning agents is commonly estimated asthe product of the concentration of the chemical cleaning agent during cleaning(C), and the duration of the cleaning (t), or Ct value. However, some studieshave demonstrated that the Ct value does not adequately reflect the impact ofsodium hypochlorite exposure on the physical/chemical characteristic parametersof hollow-fiber polysulfone (PSf) membranes [124]. However, no alternativerelationship has been suggested. In the present study, different standard oxidationmodels were considered to relate the concentration of sodium hypochlorite and theexposure time to the magnitude of the changes in the physical/chemical characteristicparameters of blended PVDF membranes. The results of this analysis are presentedin Section 6.3.101A fate assessment analysis was conducted to investigate the mechanism by whichthe HA is removed from the blended membrane material during sodium hypochloritecleaning. Details of the fate assessment analysis technique are presented in Section4.6 and the results of this analysis are presented in Section 6.4.6.2 Impact of Exposure to Sodium Hypochlorite onMembrane Characteristics6.2.1 Fourier Transform Infrared SpectroscopyFourier transform infrared spectroscopy (FTIR) is a widely use method in analyticchemistry. A typical FTIR spectrum constructed from the raw data is presented inFigure 6.1.A baseline of a FTIR spectrum consists of the portions of a spectrum with nosignificant absorptions. Ideally the intensity in these portions of the spectrum haszero absorbance value [113]. However, the baseline can be affected by the ageof the device, the quality of the sample, as well as different factors that influencethe spectra collection, like the presence of stray light, temperature, pressure andhumidity. As a result, baseline drift can often be observed in different signals, andbaselines may be tilted, shifted or curved [55]. These drifted baselines significantlyinfluence the accuracy and robustness of the results of the analysis interpreted fromthe spectra. Therefore, it is necessary to remove the baseline drift in FTIR spectraanalysis. In the present study, baseline correction was done manually by adjustingthe drifted baseline to a zero absorption location. Details of the baseline correctionmethod can be found elsewhere [139]. A typical baseline-corrected FTIR spectrumis presented in Figure 6.1.102!"#$%"#&%'""%()""%$*""%*(""%*+""%,-./0-1234%5,0-670108%9267:%;1<4%=>?-40%53?!(:%@1A%B1.4C624%D/004374E%Figure 6.1: Typical raw and baseline-corrected FTIR spectra of virginmembrane.Typical baseline-corrected FTIR spectra of PVDF and the HA of interest arepresented in Figure 6.2. Figure 6.2 illustrates that the dominant peaks of the PVDFand the major functional group of the HA of interest in the membranes were locatedat positions A and B, respectively.103!"#$%"#&%'""%$(""%)*""%*$""%*+""%,-./0-1234%5,0-670108%9267:%;1<4%=>?-40%53?!$:%@ABC%5"D%E,%F1-%31.7%?4?-0124:%E,%5$""D%E,%F1-%31.7%?4?-0124:%BAFigure 6.2: Typical FTIR spectra of PVDF and HA.The peak located at position A in the FTIR spectra (i.e., the characteristic peakfor PVDF) should not be influenced by the presence of the HA of interest in theblend. Hence, absorbance values of the peaks located at position A should be samefor all the replicate samples for any membrane containing PVDF. As mentionedearlier, absorbance values of the peaks are affected by the sample quality (e.g.,presence of particulate matter), as well as different influencing factors for spectracollection, like the presence of stray light, temperature, pressure and humidity [55].For these reasons, different absorbance values for the peaks located at position Acan be observed in the FTIR spectra of replicate samples of identical membranescontaining PVDF.It is necessary to scale the FTIR spectra to make sure the absorbance values for104the characteristic peaks are identical for all the replicate samples of the identicalmembranes, as the characteristic peak generally served as an internal standard.Scaling of the FTIR spectra was done by using the ratio of the peaks to ensure thatthe absorbance values of the peaks located at the position A in the FTIR spectraof the membranes containing PVDF were the same for all the samples. Typicalbaseline-corrected and scaled FTIR spectra of the replicate samples of the identicalmembranes containing PVDF are presented in Figure 6.3.!"#$%"#&%'""%$(""%)*""%*$""%*+""%,-./0-1234%5,0-670108%9267:%;1<4%=>?-40%53?!$:%@1?AB4C$!D1.4B624%E/004374F%%@1?AB4C)!D1.4B624%E/004374F%%@1?AB4C)!D1.4B624%E/004374F%12F%@31B4F%Figure 6.3: Typical baseline-corrected and scaled FTIR spectra of virginmembranes.Typical baseline-corrected and scaled FTIR spectra of the membranes containingPVDF and different ratios of the HA of interest are presented in Figure 6.4.Increasing values for the dominant peaks of the major functional group of theHA of interest in the membranes were observed for samples with an increasing105percentage of the HA of interest in the membranes. The results confirmed that theselected dominant peaks of the PVDF and the major functional group of the HA ofinterest in the membranes were accurate. The results also confirmed that the methodapplied for the scaling of the FTIR spectra was accurate.!"#$%"#&%'""%$(""%)*""%*$""%*+""%,-./0-1234%5,0-670108%9267:%;1<4%=>?-40%53?!$:%@ABC%5"D%E,%F1-%31.7%?4?-0124:%)"D%E,%F1-%31.7%?4?-0124%*"D%E,%F1-%31.7%?4?-0124%Figure 6.4: Typical FTIR spectra of lab-cast membranes containing 20% HA(i.e., 80% PVDF) and 30% HA (i.e., 70% PVDF).FTIR analyses of the commercial membrane samples were completed to estimatethe content of the HA of interest present in the blended membranes. Typical resultsare presented in Figure 6.5.106Figure 6.5: Typical FTIR spectra of virgin membrane and membranes exposedto sodium hypochlorite. (Insert, magnification of peak of the majorfunctional group of the HA of interest).Exposure to the sodium hypochlorite solution reduced the size of the dominantpeak of the major functional group of the HA of interest in the membranes for allexposure doses investigated (Figure 6.6).107!" !#!$" !#%" !#%$"&'()'*"+,--.(+'/0"-.-1(/*."$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"%2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"%2$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"===!!"33-":/;70<"%2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"===!!"33-":/;70<"%2$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"===!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"===!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"%2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"%2$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"?@AB"C./D"6E1F,(1/*+."'*"/(1'G(/(H"I*'G<""Figure 6.6: Effects of sodium hypochlorite on the dominant peak of the majorfunctional group of the HA of interest.(Error bars correspond to minimum and maximum values from 4 repeatedexperiments).The results from the FTIR analysis suggest that sodium hypochlorite exposureremoved a portion of the HA of interest from the membranes. These results areconsistent with those reported by [119], who also observed that exposure to chemicalcleaning agents reduces the content of the HA of interest from blended PVDFmembranes. However, effects of the exposure doses and the quantification of thechanges were not reported in that work [119]. As presented in Figure 6.6, when themembranes were exposed to a fixed concentration of sodium hypochlorite, the sizeof the characteristic peak of the major functional group of the HA of interest in the108membranes decreased as the exposure dose of sodium hypochlorite increased (e.g.,the size of the characteristic peaks of the membrane exposed to 500,000 ppm·hr and2,000,000 ppm·hr of sodium hypochlorite at a concentration of 44300 ppm wereapproximately 36% and 79% lower than that of the virgin membrane, respectively).Also, when the membranes were exposed to a fixed exposure dose, the reduction inthe size of the dominant peak was greater at the lower concentration investigated(e.g., the size of the dominant peak of the membrane exposed to 2,000,000 ppm·hrof sodium hypochlorite at a concentration of 44300 ppm and 3600 ppm wereapproximately 79% and 90% lower than that of the virgin membrane, respectively).These results indicate that the removal of the HA of interest was not only affectedby the exposure dose, but also by the concentration of sodium hypochlorite in thesolution.6.2.2 Nuclear Magnetic Resonance (NMR)Although the FTIR analysis can be used to easily and rapidly confirm the presenceof a specific functional group, it cannot be used to accurately provide a quantitativeestimate of the amount of that group present in a sample. On the other hand, nuclearmagnetic resonance (NMR) analysis can be used to accurately estimate the relativeamount of functional groups [99]. For this reason, NMR analysis was used toconfirm the results obtained for some of the samples that were analyzed using FTIR.Typical NMR spectra of PVDF and the HA of interest are presented in Figure6.7. The dominant peaks of the PVDF and the major functional group of the HA ofinterest in the membranes were located at positions A and B (characteristic peaks),respectively (Figure 6.7).109(a) (b) ppm ppm BABAFigure 6.7: Typical NMR spectra of (a) PVDF, and (b) HA of interest.It is necessary to normalize the integration area of the NMR spectra to makesure that the areas under the curves for the characteristic peaks are identical for allthe replicate samples of the identical membranes, as the area under the characteristicpeak generally served as an internal standard. Scaling of the NMR spectra was doneby using the ratio of the area under the curve for the peaks to make sure that thearea under the curves of the peaks located at position A in the NMR spectra of the110membranes containing PVDF were same (i.e., 1).Typical NMR spectra of the membranes containing PVDF and different ratios ofthe HA of interest are presented in Figure 6.8. Increasing values for the area underthe curve for the dominant peaks of the major functional group of the HA of interestin the membranes were observed for samples with an increasing percentage of theHA of interest in the membranes. These results confirmed the that the selecteddominant peaks of the PVDF and the major functional group of the HA of interestin the membranes were accurate. The results also confirmed that the method appliedfor the scaling of the NMR spectra was accurate.111(a) (b) ppm ppm BABAFigure 6.8: Typical NMR spectra of lab-cast membrane containing (a) 4% HA(i.e., 96% PVDF), and (b) 8% HA (i.e., 92% PVDF).Typical NMR spectrum of a virgin commercial membrane is presented in Figure6.9.112BAFigure 6.9: Typical NMR spectrum of a virgin commercial membrane.Exposure to the sodium hypochlorite solution decreased the HA of interestcontent of the membranes for all exposure doses investigated (Figure 6.10), confirmingthat the sodium hypochlorite exposure removed a portion of the HA of interest fromthe membranes.113!" !#$" %#&" %#'" &#(" )"*+,-+."/0112,/+34"1215,3.2"6!!7!!!"8819:,";<0112,/+34"1215,3.2"+."(()!!"881"=3><4?"%7!!!7!!!"8819:,";<0112,/+34"1215,3.2"+."(()!!"881"=3><4?"%76!!7!!!"8819:,";<0112,/+34"1215,3.2"+."(()!!"881"=3><4?"&7!!!7!!!"8819:,";<0112,/+34"1215,3.2"+."(()!!"881"=3><4?"6!!7!!!"8819:,";<0112,/+34"1215,3.2"+.")$!!"881"=3><4?"%7!!!7!!!"8819:,";<0112,/+34"1215,3.2"+.")$!!"881"=3><4?"&7!!!7!!!"8819:,";<0112,/+34"1215,3.2"+.")$!!"881"=3><4?"@243AB2"CD"<0.E2.E"53F2G"0."=H@""Figure 6.10: Effects of sodium hypochlorite on the HA of interest content ofthe membranes (NMR data).(Error bars correspond to minimum and maximum values of 3 repeatedexperiments).As indicated in Figure 6.10, when the membranes were exposed to a fixedconcentration of sodium hypochlorite, the HA of interest content of the membranesdecreased as the exposure dose of sodium hypochlorite increased (e.g., the HAof interest contents of the membranes exposed to 500,000 ppm·hr and 2,000,000ppm·hr of sodium hypochlorite at a concentration of 44300 ppm were approximately26% and 76% lower than that of the virgin membrane, respectively). Also, whenthe membranes were exposed to a fixed exposure dose, the reduction in the HA ofinterest content of the membrane was greater at the lower concentration investigated(e.g., the HA of interest content of the membranes exposed to 2,000,000 ppm·hr114of sodium hypochlorite at a concentration of 44300 ppm and 3600 ppm wereapproximately 76% and 94% lower than that of the virgin membrane, respectively).These results indicate that the extent of removal of the HA of interest was affectedby the exposure dose and the concentration of sodium hypochlorite used.The results presented in Figure 6.11 also confirm that the decreasing trend ofthe HA of interest content of the membranes observed with NMR analysis wasconsistent with the trend observed with FTIR analysis. This indicated that therelatively less capital extensive FTIR analysis can be used to analyze the membranefor the HA of interest content.115!"#$#%&''()*#%&%%#%&*%#%&+%#%&,%#-&(%#%&%%# %&(%# %&)%# %&+%# %&'%# -&%%# -&(%#!./012.#34#5678.78#90:.;#67#<=>!#?@6AB0/CD.;E#!./012.#34#5678.78#90:.;#67#@F!#?@6AB0/CD.;E#Figure 6.11: Comparison of the HA of interest content of the membranesanalyzed by FTIR and NMR. (Absolute values of the HA ofinterest content of the membranes were normalized to that of virginmembranes).(x and y error bars correspond to minimum and maximum values from3 and 4 repeated experiments, respectively).6.2.3 Contact AngleContact angle is widely used to assess the hydrophilicity of the membrane surface.A typical image of the contact angles measurement is presented in Figure 6.12.116Figure 6.12: Typical image of the contact angles measurement.Exposure to sodium hypochlorite also impacted the hydrophilic characteristicsof the membranes as observed from the contact angle measurement (Figure 6.13).As presented in Figure 6.13, exposure to sodium hypochlorite increased the contactangles of the commercial membranes.117!" #$" $!" %$" &!!"'()*(+",-../),(01"./.2)0+/"$!!3!!!"44.56)"78-../),(01"./.2)0+/"(+"99:!!"44.";0<81="&3!!!3!!!"44.56)"78-../),(01"./.2)0+/"(+"99:!!"44.";0<81="#3!!!3!!!"44.56)"78-../),(01"./.2)0+/"(+"99:!!"44.";0<81="$!!3!!!"44.56)"78-../),(01"./.2)0+/"(+"###!!"44.";0<81="&3!!!3!!!"44.56)"78-../),(01"./.2)0+/"(+"###!!"44.";0<81="#3!!!3!!!"44.56)"78-../),(01"./.2)0+/"(+"###!!"44.";0<81="$!!3!!!"44.56)"78-../),(01"./.2)0+/"(+":>!!"44.";0<81="&3!!!3!!!"44.56)"78-../),(01"./.2)0+/"(+":>!!"44.";0<81="#3!!!3!!!"44.56)"78-../),(01"./.2)0+/"(+":>!!"44.";0<81="8-+?0,?"@+*1/"7A/*)//=""Figure 6.13: Effects of sodium hypochlorite on the contact angle of themembranes.(Error bars correspond to 95% confidence interval from 10 repeatedexperiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite, contact angle of the membranes increased as the exposure dose of sodiumhypochlorite increased (e.g., contact angle of the membranes exposed to 500,000ppm·hr and 2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600ppm were approximately 14% and 81% greater than that of the virgin membrane,respectively). Also, when the membranes were exposed to a fixed exposure dose, theincrease in the contact angle of the membrane was greater at the lower concentrationinvestigated (e.g., contact angles of the membranes exposed to 2,000,000 ppm·hrof sodium hypochlorite at a concentration of 44300 ppm and 3600 ppm were118approximately 40% and 81% greater than that of the virgin membrane, respectively).These results indicate that the extent of the increase in the contact angle was affectedby the exposure dose and the concentration of sodium hypochlorite used.This was expected as the FTIR and NMR results indicated that the HA of interest,which is added to increase the hydrophilic nature of the membrane, was removedduring sodium hypochlorite exposure. The contact angle data were consistent withresults from other studies that indicate that the contact angle is inversely proportionalto the content of hydrophilic additives in membranes [105, 163].The results presented in Figure 6.14 also confirm that the decreasing trend ofthe HA of interest content of the membranes observed with NMR analysis wasconsistent with the trend observed with contact angle analysis.However, it should be noted that, when the membranes were exposed to sodiumhypochlorite at a concentration of 44300 ppm, no effect of exposure dose wasobserved on the increase in the contact angle.119!"#$#%&'()*+#%&%%#%&(%#%&+%#%&,%#)&-%#%&*%# )&%%# )&-%# )&.%# )&+%# )&*%# -&%%#!/0123/#45#6789/89#:1;/<#78#=>!#?=7@A10BC/1?""8@)-*.,<""Figure 6.17: Effects of sodium hypochlorite on the average pore size of themembranes.(Error bars correspond to minimum and maximum values from 5repeated experiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite, the average pore diameter of the membranes increased as the exposure dose ofsodium hypochlorite increased (e.g., the average pore diameters of the membranesexposed to 1,000,000 ppm·hr and 2,000,000 ppm·hr of sodium hypochlorite ata concentration of 3600 ppm were approximately 10% and 22% greater thanthat of the virgin membrane, respectively). Also, when the membranes wereexposed to a fixed exposure dose, the increase in the average pore diameter of themembrane was greater at the lower concentration investigated (e.g., the average porediameters of the membranes exposed to 1,000,000 ppm·hr of sodium hypochloriteat a concentration of 44300 ppm and 3600 ppm were approximately 3% and 10%greater than that of the virgin membrane, respectively). These results indicate that122the extent of the increase in the average pore diameter was affected by the exposuredose and the concentration of sodium hypochlorite used.As presented in Figure 6.18, exposure to sodium hypochlorite increased the 90thpercentile of cumulative pore size distribution (d90) of the commercial membranes.!" !#!$%" !#!&" !#!'%"()*+),"-.//0*-)12"/0/3*1,0"$4!!!4!!!"55/67*"89.//0*-)12"/0/3*1,0"),"''&!!"55/":1;92<"%!!4!!!"55/67*"89.//0*-)12"/0/3*1,0"),"&=!!"55/":1;92<"$4!!!4!!!"55/67*"89.//0*-)12"/0/3*1,0"),"&=!!"55/":1;92<"$4%!!4!!!"55/67*"89.//0*-)12"/0/3*1,0"),"&=!!"55/":1;92<">?!"""8@)-*.,<""Figure 6.18: Effects of sodium hypochlorite on 90th percentile of cumulativepore size distribution of the membranes.(Error bars correspond to minimum and maximum values from 5repeated experiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite, the d90 of the membranes increased as the exposure dose of sodium hypochloriteincreased (e.g., the d90 of the membranes exposed to 1,000,000 ppm·hr and 2,000,000ppm·hr of sodium hypochlorite at a concentration of 3600 ppm were approximately19% and 33% greater than that of the virgin membrane, respectively). Also, whenthe membranes were exposed to a fixed exposure dose, the increase in the d90 of themembrane was greater at the lower concentration investigated (e.g., the d90 of themembranes exposed to 1,000,000 ppm·hr of sodium hypochlorite at a concentration123of 44300 ppm and 3600 ppm were approximately 6% and 19% greater than thatof the virgin membrane, respectively). These results indicate that the extent ofthe increase in the d90 was affected by the exposure dose and the concentration ofsodium hypochlorite used.As presented in Figure 6.19, exposure to sodium hypochlorite increased theapparent porosity of the commercial membranes.!" #" $!" $#"%&'(&)"*+,,-'*&./",-,0'.)-"$1!!!1!!!"22,34'"56+,,-'*&./",-,0'.)-"&)"778!!"22,"9.:6/;"#!!1!!!"22,34'"56+,,-'*&./",-,0'.)-"&)"8&?@"5A;""Figure 6.19: Effects of sodium hypochlorite on the porosity of the membranes.(Error bars correspond to minimum and maximum values from 5repeated experiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite, apparent porosity of the membranes increased as the exposure dose of sodiumhypochlorite increased (e.g., apparent porosity of the membranes exposed to 1,000,000ppm·hr and 2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600ppm were approximately 89% and 107% greater than that of the virgin membrane,respectively). Also, when the membranes were exposed to a fixed exposure dose,the increase in the apparent porosity of the membrane was greater at the lower124concentration investigated (e.g., apparent porosity of the membranes exposedto 1,000,000 ppm·hr of sodium hypochlorite at a concentration of 44300 ppmand 3600 ppm were approximately 57% and 89% greater than that of the virginmembrane, respectively). These results indicate that the extent of the increase inthe apparent porosity was affected by the exposure dose and the concentration ofsodium hypochlorite used.The removal of the HA of interest from the membranes, due to exposure tosodium hypochlorite might have contributed to the increase in apparent porosity ofthe membranes. The results presented in Figure 6.20 also confirm that the increasingtrend of the apparent porosity of the membranes was consistent with the observeddecreasing trend of the HA of interest content of the membranes due to exposure tosodium hypochlorite.125!"#$#%&'((')#%&%%#%&*%#%&+%#%&'%#,&(%#%&-%# ,&%%# ,&(%# ,&)%# ,&+%# ,&-%# (&%%# (&(%# (&)%#!./012.#34#5678.78#90:.;#67#<=!#><6?@0/AB.;C#D6?6:A8E#><6?@0/AB.;C#Figure 6.20: Comparison of the HA of interest content of the membranesanalyzed by NMR to the apparent porosity of the membranes (Absolutevalues of the parameters were normalized to that of virgin membranes).(x and y error bars correspond to minimum and maximum values of 5and 3 experiments, respectively).6.2.5 Tensile StrengthTensile strength is widely used to assess the mechanical characteristics of themembrane material [119]. A typical image of the stress-strain curve from whichtensile strength measurement was estimated is presented in Figure 6.21.126!"#"$"%"&"!" !'!$(" !'!(" !'!)(" !'#"*+,-.."/0123"*+,245"/667663"0.002 0.2% Offset yield strength Figure 6.21: Typical stress-strain curve of the membranes.Exposure to sodium hypochlorite significantly reduced the yield strength ofthe commercial membranes. As presented in Figure 6.22, when the membraneswere exposed to a fixed concentration of sodium hypochlorite, yield strength of themembranes decreased as the exposure dose of sodium hypochlorite increased (e.g.,yield strength of the membranes exposed to 500,000 ppm·hr and 2,000,000 ppm·hrof sodium hypochlorite at a concentration of 44300 ppm were approximately 29%and 67% lower than that of the virgin membrane, respectively). Also, when themembranes were exposed to a fixed exposure dose, the reduction in the yield strengthof the membrane was greater at the lower concentration investigated (e.g., yieldstrength of the membranes exposed to 2,000,000 ppm·hr of sodium hypochlorite127at a concentration of 44300 ppm and 3600 ppm were approximately 67% and 71%lower than that of the virgin membrane, respectively). These results indicate thatthe extent of decrease of the yield strength was affected by the exposure dose andthe concentration of sodium hypochlorite used.!" !#$" %#&" %#'" &#(")*+,*-"./001+.*23"0104+2-1"5!!6!!!"77089+":;/001+.*23"0104+2-1"*-"((;3?"%6!!!6!!!"77089+":;/001+.*23"0104+2-1"*-"((;3?"%65!!6!!!"77089+":;/001+.*23"0104+2-1"*-"((;3?"&6!!!6!!!"77089+":;/001+.*23"0104+2-1"*-"((;3?"%6!!!6!!!"77089+":;/001+.*23"0104+2-1"*-"&&&!!"770"=2>;3?"&6!!!6!!!"77089+":;/001+.*23"0104+2-1"*-"&&&!!"770"=2>;3?"5!!6!!!"77089+":;/001+.*23"0104+2-1"*-"<$!!"770"=2>;3?"%6!!!6!!!"77089+":;/001+.*23"0104+2-1"*-"<$!!"770"=2>;3?"&6!!!6!!!"77089+":;/001+.*23"0104+2-1"*-"<$!!"770"=2>;3?"!#&@">AB1C"D*13E"FC+1,C9":GH2?""Figure 6.22: Effects of sodium hypochlorite on the tensile strength of themembranes.(Error bars correspond to minimum and maximum values from 5repeated experiments).The reduction in membrane cross-sectional area resulting from an increase inporosity may have affected the yield strength. The results presented in Figure 6.23also confirm that the decreasing trend of the tensile strength of the membraneswas consistent with the observed decreasing trend of the apparent porosity of themembranes due to exposure to sodium hypochlorite.128!"#$#%&'(()*#%&%%#%&+%#%&(%#%&'%#,&*%#%&-%# ,&%%# ,&*%# ,&)%# ,&(%# ,&-%# *&%%# *&*%# *&)%#%&*.#/0123#45267#8392:;3<#=>?9@A65B27C#D?9?153E#=>?9@A65B27C#Figure 6.23: Comparison of the tensile strength of the membranes to theapparent porosity of the membranes. (Absolute values of the parameterswere normalized to that of virgin membranes).(x and y error bars correspond to minimum and maximum values from5 and 5 repeated experiments, respectively).6.2.6 Intrinsic Resistance of a Membrane Based on Clean WaterPermeation TestA typical clean water permeability test result is presented in Figure 6.24.129!"#!!!!!"$!!!!!!"$#!!!!!"%!!!!!!"!" #" $!" $#" %!" %#" &!" &#""'()*+(,+-"*.,+,)/(-."01")2."3.34*/(."536$ 7"8+3."59+(:).,7"Figure 6.24: Typical clean water filtration test result for a virgin commercialmembrane.Exposure to sodium hypochlorite reduced the intrinsic resistance of the commer-cial membranes. As presented in Figure 6.25, when the membranes were exposed toa fixed concentration of sodium hypochlorite, intrinsic resistance of the membranesdecreased as the exposure dose of sodium hypochlorite increased (e.g., intrinsicresistance of the membranes exposed to 500,000 ppm·hr and 2,000,000 ppm·hrof sodium hypochlorite at a concentration of 44300 ppm were approximately 9%and 16% lower than that of the virgin membrane, respectively). Also, when themembranes were exposed to a fixed exposure dose, the reduction in the intrinsicresistance of the membrane was greater at the lower concentration investigated(e.g., intrinsic resistance of the membranes exposed to 2,000,000 ppm·hr of sodiumhypochlorite at a concentration of 44300 ppm and 3600 ppm were approximately16% and 23% lower than that of the virgin membrane, respectively). These resultsindicate that the extent of decrease of the intrinsic resistance was affected by the130exposure dose and the concentration of sodium hypochlorite used.!" #!!!!!" $!!!!!!" $#!!!!!"%&'(&)"*+,,-'*&./",-,0'.)-"#!!1!!!"22,34'"56+,,-'*&./",-,0'.)-"&)"778!!"22,"9.:6/;"$1!!!1!!!"22,34'"56+,,-'*&./",-,0'.)-"&)"778!!"22,"9.:6/;"<1!!!1!!!"22,34'"56+,,-'*&./",-,0'.)-"&)"778!!"22,"9.:6/;"#!!1!!!"22,34'"56+,,-'*&./",-,0'.)-"&)"<<.?-'"@&/?'.A+)"B-C&C?-)*-"5,D$;""Figure 6.25: Effects of sodium hypochlorite on the intrinsic resistance of themembranes.(Error bars correspond to minimum and maximum values from 2repeated experiments).The apparent increases in pore size and porosity determined from SEM imagingare consistent with the results from the clean water permeation tests Figure 6.26.131!"#$#%&''%()#%&*%#%&'%#%&+%#)&%%#)&)%#%&'%# )&%%# )&(%# )&,%# )&-%# )&'%# (&%%# (&(%# (&,%#./012#31405#67/451892#!0:7:402;0#<=95>1/7?0@A#B959:74C#<=95>1/7?0@A#Figure 6.26: Comparison of the intrinsic resistance of the membranes to theapparent porosity of the membranes. (Absolute values of the parameterswere normalized to that of virgin membranes).(x and y error bars correspond to minimum and maximum values from2 and 5 repeated experiments, respectively).6.2.7 Retention of Dextran ParticlesChallenge tests with dextran solutions are widely used to estimate the selectivity ofmembranes, which is indirectly related to the apparent pore size of the membranes[150]. Figure 6.27 presents the total organic carbon (TOC) content of the DextranT70, Dextran T500 and feed solutions with respect to the molecular weight.132!"!#$%!"!#&!!"!#'!%!"!#'(!!"!#'#'!#'!!#'!!!#'!!!!#!"#$%&&'($)&*++,-.$/01,231*+$4,567.$$%89*($)*+,-*.#/012.,3-#45#6&!#7*8#6%!!#9-2.37*:;#9-2.37*:#/6&!;#9-2.37*:#/6%!!;#Figure 6.27: Total organic carbon (TOC) content of the Dextran T70, DextranT500 and feed solution analyzed by size exclusion chromatographywith a high performance liquid chromatograph.(Error bars correspond to 95% confidence interval for 4 repeatedexperiments). Details of the size exclusion chromatography and highperformance liquid chromatograph are presented in Section 4.4.Figure 6.28 present the total organic carbon (TOC) content of the feed andpermeate from the typical membranes for the range of the molecular weight ofdextran used in the study.133!"!#$%!"!#&!!"!#'!%!"!#'(!!"!#'#'!#'!!#'!!!#'!!!!#!"#$%&&'($)&*++,-.$/01,231*+$4,567.$$%89*($)*+,-*.#/012.,3-#45#6&!#7*8#6%!!#9-2.37*:;#<=,-*.#534>#?13@1*#A4>>-3A17B#>->C37*-#<=,-*.#534>#DE!!!E!!!#FF>GH3#/I4>>-3A17B#>->C37*-#1*#$J!!#FF>#K7LIB;#Figure 6.28: Total organic carbon (TOC) content of the feed solution, and thepermeates from virgin membranes and membranes exposed to sodiumhypochlorite (exposure dose and concentration are 2,000,000 ppm·hrand 3600 ppm, respectively).(Error bars correspond to 95% confidence interval for 4 repeatedexperiments).The difference between the total organic carbon (TOC) content of the feed andpermeate from a membrane for a molecular weight (e.g., 100 kDa) correspondsto the amount of dextran retained by the membrane for that apparent molecularweight. As presented in Figure 6.29, compared to that of a virgin membrane, theretention of dextran particles was significantly lower for membranes exposed tosodium hypochlorite (exposure dose and concentration are 2,000,000 ppm·hr and1343600 ppm, respectively).Figure 6.29: Dextran retention by virgin membrane and membranes exposedto sodium hypochlorite (exposure dose and concentration are 2,000,000ppm·hr and 3600 ppm, respectively).(Error bars correspond to 95% confidence interval for 4 repeatedexperiments).The apparent increases in porosity and pore size determined from SEM imaging,as well as the decrease in intrinsic resistance of the membranes are consistent withthe results from the challenge tests.6.3 Modelling the Magnitude of the Impact of SodiumHypochlorite ExposureIn general, the magnitude of the impact of exposure to sodium hypochlorite onthe membrane characteristic parameters (e.g., contact angle) is considered to be a135function of the exposure dose, defined as the cumulative product of the concentrationof the chemical agent (C) and the duration of the cleaning (i.e., time), (t), or Ctvalue [9]. Assuming that the Ct model can accurately predict the degradation of themembrane characteristics, membrane manufacturers typically specify a maximumlifetime exposure dose, in Ct, above which the performance of their membranescannot be guaranteed.The application of the Ct model assumes that the changes in the membranecharacteristics due to the exposure to the cleaning follows the relationship presentedin Equation 6.1.MMo= exp(−kCt) (6.1)where, M = value of the membrane parameter under consideration at time t (e.g.,HA of interest content of the membrane after being exposed to chemical agent fortime t), Mo = value of the membrane parameter under consideration at time zero(e.g., HA of interest content of the virgin membrane), k = an empirical constantwhich is independent of C, C = concentration of the chemical agent, and t = time.However, the magnitude of the impact of exposure to sodium hypochlorite onthe different parameters investigated varied depending on both the concentration anddose of sodium hypochlorite used, indicating that the commonly used relationshipto calculate the exposure dose which is the product of concentration (C) and time (t),or Ct value, does not adequately reflect the impact of sodium hypochlorite exposureon the physical/chemical characteristics of the membrane.136A modified exposure dose, which can be used more accurately and consistently,to correlate the impact of sodium hypochlorite exposure on the membrane character-istic parameters with concentration (C) and time (t), needs to be found. Measuredvalues of the membrane characteristic parameters investigated for different membranesamples (results presented in the Section 6.2) were modelled using the empiricalHom relationship (Equation 6.2) [62].MMo= exp(−kCntm) (6.2)where, n and m are empirical constant.The above model was fitted to the experimental data for the following fourconditions.i) Model #1: n 6= 1, and m 6= 1.MMo= exp(−kCntm) (6.3)The equation for model #1 is identical to Equation 6.2.ii) Model #2: n=1. m=1,MMo= exp(−kCt) (6.4)137The equation for model #2 is identical to Equation 6.1.iii) Model #3: n 6= 1, and m = 1.MMo= exp(−kCnt) (6.5)iv) Model #4: n = 1, and m 6= 1.MMo= exp(−kCtm) (6.6)These models were fitted to the data for FTIR, tensile strength and contact anglemeasurements. These were selected because more measurements were collectedfor these physical/chemical characteristics, which is required for curve fitting,and because these represent different characteristics of interest (i.e., chemicalcomposition, strength and hydrophilic characteristics).Least squares regression applying a Levenberg-Marquardt algorithm (LMA)was used to fit the models to the experimental data using Profit statistical software(QuantumSoft, Switzerland). For all the models considered, the average values ofthe empirical constants k, n and m, and their coefficients of variation, as well ascoefficients of determination (i.e., R2) were obtained for the best-fit curves. Theresults of the curve fitting for the FTIR peaks are presented in Figure 6.30.138!"#$#%&'())*#%&%%#%&)%#%&+%#%&,%#%&-%#.&%%#.&)%#%&%%# %&)%# %&+%# %&,%# %&-%# .&%%# .&)%#/01!#2345#6789:4;<=3>##3?@6A5BCD#5#$6-+EA-DF6-EA-D#G#$.#:#$.#!"#$#%&'(%)*#%&%%#%&+%#%&,%#%&*%#%&'%#-&%%#-&+%#%&%%# %&+%# %&,%# %&*%# %&'%# -&%%# -&+%#./0!#1234#567893:;<2=##2>?5@4ABC9D#4#$5*&+E@*DF5--E@*D#B#$%&((F%&-+#9#$-&%*F%&-,###!"#$#%&'()*+#%&%%#%&,%#%&+%#%&(%#%&'%#-&%%#-&,%#%&%%# %&,%# %&+%# %&(%# %&'%# -&%%# -&,%#./0!#1234#567893:;<2=##2>?5@4AB9 C#4#$5,DE@'CF5'E@'C#G#$-#9#$-&,'F%&%*##!"#$#%&'(%%'#%&%%#%&)%#%&*%#%&+%#%&'%#,&%%#,&)%#%&%%# %&)%# %&*%# %&+%# %&'%# ,&%%# ,&)%#-./!#0123#4567829:;1<##1=>4?3@ABC#3#$4,)D?+CE4(D?+C#A#$%&(FE%&%+#8#$,###Figure 6.30: Typical least squares curve fitting of the observed FTIR peaksfor the model #1 to 4.Table 6.1 presents a summary of the curve fitting results for different membranecharacteristics considered.139Table 6.1: Typical values of the empirical constants k, n and m for the membrane characteristic parameters investigated.Parameter Model #1 Model #2 Model #3 Model #4FTIR Peak k = 6.2E-6±11E-6 k = 84E-8±8E-8 k = 12E-6±7E-6 k = 25E-8±8E-8n = 0.77±0.12 n =1.0 n = 0.73±0.06 n = 1.0m = 1.06±0.14 m = 1.0 m = 1.0 m = 1.28±.07a R2 = 0.732 R2 = 0.870 R2 = 0.864Tensile Strength k = 39E-5±41E-5 k = 55E-8±2E-8 k = 5E-6±2E-6 k = 31E-8±7E-8n = 0.48±0.08 n = 1.0 n = 0.78±0.04 n = 1.0m = 0.66±0.08 m = 1.0 m = 1.0 m = 1.1±.05a R2 = 0.744 R2 = 0.843 R2 = 0.798Contact Angle k = 45E-6±60E-6 k = 19E-8±3E-8 k = 39E-7±10E-7 k = 52E-9±8E-9n = 0.82±0.1 n = 1.0 n = 0.69±0.03 n = 1.0m = 0.53±0.09 m = 1.0 m = 1.0 m = 1.29±.03a R2 = 0.396 R2 = 0.595 R2 = 0.561Note: average value ± coefficient of variation.140Models for which the value of the estimated empirical constant k for a model isnot statistically significantly different from zero suggests that no change in parameterunder consideration occurs when exposed to sodium hypochlorite (i.e., MMo is alwaysequal to 1). However, as previously discussed, exposure to sodium hypochloritehad a significant impact on the parameters investigated. For this reason, models forwhich the empirical constant k was not significantly different from zero was notfurther considered.As presented in Table 6.1, model #1 consistently generated values of empiricalconstant k, which were not different from zero, and was therefore not furtherconsidered (i.e., a R2 values were not presented in Table 6.1). The remaining models(i.e., models #2, #3 and #4) were compared based on the fit to the experimentaldata using their coefficient of determination (i.e., R2). As presented in Table 6.1,the coefficient of determination for model #3 was consistently greater than thosefor the other models (i.e., models #2 and #4), indicating that the impact of sodiumhypochlorite exposure on the membrane characteristics can be more accurately andconsistently correlated to a modified exposure dose of the form Cnt (Figure 6.31)than the commonly used Ct relationship.141Figure 6.31: Typical response of the membrane parameters investigatedwith respect to the Cnt relationship. (a) Normalized FTIR peak,(b) Normalized yield strength, (c) Normalized contact angle. (±corresponds to the standard error of the estimated parameter).For all the parameters investigated, the power coefficient n was less than 1,indicating that the exposure time had a greater impact on the changes in thephysical/chemical characteristics of the membrane than the concentration of the142sodium hypochlorite. These results suggest that to minimize changes to thecharacteristics of the membrane, chemical cleaning should be performed usinghigher concentrations of sodium hypochlorite for shorter durations. However,it should be noted that the selection of the maximum concentration and/or theminimum duration would also be influenced by the effectiveness of the chemicalcleaning for fouling control. Also, in the present research, concentrations of sodiumhypochlorite used were higher than those typically used for chemical cleaning infull-scale water and wastewater treatment membrane applications. Longer termexposure studies at concentrations typical of those used during full-scale membranecleaning should be performed to gain further insight into the effect of exposure tolower concentrations of chemical cleaning agents.It should be noted that, only the empirical Hom relationship (Equation 6.2),which is widely used as oxidation model [62] was fitted to the experimental data forthe four possible conditions.6.4 Fate of the HA of Interest During Exposure toSodium HypochloriteWhen blended lab-cast membranes with different blend ratios (i.e., HA:PVDF ratiosof 100:00, 30:70, 00:100 by weight) were soaked in solutions containing differentconcentrations (i.e., 9000 ppm and 25900 ppm) of sodium hypochlorite in distilledwater for extended periods (i.e., up to 83 days), significant amounts of sodiumhypochlorite were consumed by the membranes containing the HA of interest(results presented later in this section). However, almost no sodium hypochloritewas consumed by the membranes containing only PVDF (i.e., 0% HA) (results alsopresented later in this section). This indicated that only the organic carbon present143in the HA of interest was responsible for sodium hypochlorite consumption.In the discussion that follows in this section, the organic carbon present in theHA of interest of a blended lab-cast membranes is referred simply as total organiccarbon (TOC) of the membrane as the PVDF did not contribute to the organiccontent of the soak solutions. Table 6.2 summarized the calculated amount of TOCpresent in the blended lab-cast membranes used in the present study. The amount ofTOC was calculated based on the chemical composition of the HA of interest.Table 6.2: Amount of TOC present in blended lab-cast membraneslab-cast membrane HA:PVDF ratio Mass ofmembrane (g)Mass ofTOC (g)Contains only the HA ofinterest100:00 1.00 0.50(i.e., 0% PVDF)Contains 30% HA 30:70 1.00 0.15(i.e., 70% PVDF)Contains only PVDF 00:100 1.00 0.00(i.e., 0% HA)As presented in Figure 6.32, sodium hypochlorite was consumed in the presenceof the HA of interest. The consumption of sodium hypochlorite was observedto be greater for the lab-cast membranes made only of the HA of interest (i.e.,0% PVDF) than for those membranes made from a blend of PVDF and HA ofinterest. The lower sodium hypochlorite consumption rate observed for the blended144lab-cast membranes was likely due to the fact that some of the HA of interest wasat least partially shielded within the membrane matrix and not in direct contactwith the sodium hypochlorite, effectively reducing the mass of the HA of interestdirectly exposed to the sodium hypochlorite. An insignificant amount of sodiumhypochlorite was consumed by the lab-cast membranes containing only PVDF (i.e.,0% HA).!"#!"$!"%!"&!"!"#!"$!"%!"&!"!" $!!!!!!" &!!!!!!" '!!!!!!" (!!!!!!" #!!!!!!!")*++",-"./0,1234".,3+5)46"741")*++",-"#!!"8"9:;<")4)=1*34")*++",-"./0,1234".,3+5)46"741")*++",-">?@",12A23*00B""714+43C""23"C/4")4)=1*34"DE7,+514"23"@C""F7)3G/1H""#!!8"IJ"0*=".*+C")4)=1*34"23"$KL!!"77)"M*?@0"#!!8"IJ"0*=".*+C")4)=1*34"23"L!!!"77)"M*?@0"%!8"IJ"0*=".*+C")4)=1*34"23"$KL!!"77)"M*?@0"%!8"IJ"0*=".*+C")4)=1*34"23"L!!!"77)"M*?@0"#!!8"9:;<"0*=".*+C")4)=1*34"23"$KL!!"77)"M*?@0"Figure 6.32: Typical sodium hypochlorite consumption by lab-castmembranes during exposure to sodium hypochlorite.Note: The mass of sodium hypochlorite consumed is presented withrespect to the mass of TOC originally present in the membranes.The lab-cast membranes containing only PVDF (i.e., 0% HA) didnot contain any such TOC. For this reason, the mass of sodiumhypochlorite consumed is presented with respect to mass of 100%PVDF membranes.The concentration of dissolved organic carbon (DOC) in the sodium hypochlorite145solutions in which blended lab-cast membranes were submerged, was also monitoredover time to gain insight into the fate of the HA of interest during sodium hypochloritecleaning. The membrane made only of HA of interest, dissolved slowly into thesolution of sodium hypochlorite, resulting in an increase in the amount of organicmaterial (i.e., DOC) in the solution at the initial stage of exposure as presented inFigure 6.33. Similarly an increase in the amount of DOC in the solution containingmembranes made from a blend of PVDF and HA of interest was also observed(Figure 6.33). The observed increase in the amount of DOC in the solution indicatedthat the HA of interest was released from the membranes. However, over time, theconcentration of DOC in solution decreased at higher exposure doses, indicatingthat the organic material was mineralized after being released into solution.1461 !"!#!$"!#%"!#%$"!#&"!" $!!!!!" %!!!!!!" %$!!!!!" &!!!!!!" &$!!!!!" '!!!!!!" '$!!!!!" (!!!!!!")*++",-"./0"12"+,345,2"678")*++",-"./0",81912*33:""687+72;""12";<7")7)=8*27">?6,+487"12"0;"@66)#<8A"""!""#$%&$'()$*(+,$-.-)/(0.$10$234""$55-$6(78'$!""#$%&$'()$*(+,$-.-)/(0.$10$4"""$55-$6(78'$9"#$%&$'()$*(+,$-.-)/(0.$10$234""$55-$6(78'$9"#$%&$'()$*(+,$-.-)/(0.$10$4"""$55-$6(78'$2 !"!#!$"!#%"!#%$"!#&"'$$! !!!!" '( !!" '&$!!!!!!" '%!!!!!!!" $!!!!!!" & !!!!!" )$!!!!! " $!!!! "*+,,"-."/01"23",-456-3"789"*+,,"-."/01"-92:23+44;""798,83<""23"<=8"*8*>9+38"?@7-,598"23"1<"A77*#=9B "!""#$%&$'()$*(+,$-.-)/(0.$10$234""$55-$6(78'$!""#$%&$'()$*(+,$-.-)/(0.$10$4"""$55-$6(78'$9"#$%&$'()$*(+,$-.-)/(0.$10$234""$55-$6(78'$9"#$%&$'()$*(+,$-.-)/(0.$10$4"""$55-$6(78'$ Approximate value h DOC Figure 6.33: Typical values of the ratio of mass of DOC in the solution to massof TOC originally present in the membranes, when lab-cast membraneswere exposed to sodium hypochlorite.(Note: X-axis has two different scales, this is mainly because, after 72hours, sodium hypochlorite was not added routinely (i.e., not in every24 hours) to maintain a constant concentration in the solution, henceaccurate Ct value cannot be calculated).The concentration of DOC in the solution containing the blended lab-castmembranes comprised of PVDF and the HA of interest also increased, but at aslower rate and up to a lower maximum value, than the solution containing themembrane made only of the HA of interest. The observed lower extent of releaseof organic material, compared to that of the lab-cast membranes containing onlythe HA of interest (i.e., 0% PVDF) is consistent with the results for the sodiumhypochlorite consumption, and again suggest that some of the HA of interest was at147least partially shielded within the membrane matrix, and not in direct contact withthe sodium hypochlorite. Note that, similar to the sodium hypochlorite consumption,no organic material was released into the solution when soaking lab-cast membranescontaining only PVDF (i.e., 0% HA) in a solution of sodium hypochlorite (resultsnot presented).The above results suggest that the removal of HA of interest from the membranewas predominantly governed by oxidation, which initially resulted in structuralchanges to the HA of interest, enabling the oxidation products to be released fromthe membrane into the solution of sodium hypochlorite, and ultimately resulted inthe complete mineralization of the product released into the solution.The results presented in Figures 6.32 and 6.33 indicate that the effects ofexposure to sodium hypochlorite on the TOC varied, depending on both theconcentration and dose of sodium hypochlorite used. The changes in the mass ofthe TOC in a system (i.e., TOC in the membrane + DOC in the sodium hypochloritesolution) was modelled using Equation 6.7, which is similar to model #3 (presentedin Section 6.3).TOCtTOCo= exp(−kCnt) (6.7)where, TOCt = mass of the TOC in the system (i.e., TOC in the membrane + DOCin the sodium hypochlorite solution) at any time t, TOCo = mass of TOC initiallypresent in the membrane, k = an empirical constant, C = concentration of the sodiumhypochlorite, n = an empirical constant, and t = time148The TOC in a membrane at any time t cannot be measured, and therefore theTOCt cannot be measured directly. Assuming that the mineralization of the HA ofinterest is proportional to the amount of sodium hypochlorite consumed, the TOCin the system at any time t can be estimated using Equation 6.8.TOCt = TOCo(1−massCltmassCltotal) (6.8)where, massClt = mass of sodium hypochlorite consumed at time t, and massCltotal=mass of total sodium hypochlorite consumed to completely mineralize a unit massof TOC. The results indicate that 80±2 mg of sodium hypochlorite was consumedto completely mineralize 1 mg of TOC initially present in the membrane, whenmembranes were soaked in 25900 ppm sodium hypochlorite for extended periods(i.e., 83 days).As presented in Figure 6.34, the regression line satisfactorily fitted the experimen-tal data for both types of the lab-cast membranes (100% HA, as well as blendedlab-cast membranes containing PVDF and the HA of interest), for n = 0.8. The rateof oxidation (i.e., the slope of the fitted lines) was observed to be greater for thelab-cast membranes made of only the HA of interest (0% PVDF), than those madefrom a blend of PVDF and the HA of interest. The observed lower rate of oxidationfor the membranes made of a blend of PVDF and the HA of interest, compared tothat of the lab-cast membranes containing only the HA of interest (i.e., 0% PVDF)is consistent with the results for sodium hypochlorite consumption (i.e., the HA ofinterest was at least partially shielded within the membrane matrix). Also, the rate149at which TOC was oxidized is consistent with the discussion presented in Section6.3, which indicated that changes are proportional to Cnt, with n<1.!"#!$%&#!$%'#!$%(#!$%)#$#$# )$$$$$# ($$$$$# '$$$$$# &$$$$$# "$$$$$$# ")$$$$$# "($$$$$#*+,-./0-./12#34516789#:+#/+;##,55<+%=82##"$$>#?@#*AB#CA6;#<9#?@#*AB#CA6;#<9#?@#*AB#CA6;#<9#?@#*AB#CA6;#<9!!"33-"9/:70;"<2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"#>!!"33-"9/:70;"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"#>!!"33-"9/:70;"7?-?0/@A."A,0?-."B0C.(.D"1.E,(."(./+5'*)""C5."FC/)."=",E"C5."B0C(/@,*"60G-=;"!"!"Figure 7.5: Effect of sodium hypochlorite exposure dose and concentration(prior to filtration) on the filtration capacity of the membranes.(Error bars correspond to minimum and maximum values from 2 repeatedexperiments. Note that only virgin membranes had data from repeatedexperiments. * For these membranes typical 2 stage filtration behaviourwas not observed)A substantial decrease in the filtration capacity of the membranes exposed to2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600 ppm and 44300ppm prior to filtration relative to that of the virgin membrane was observed. Thedecrease in the filtration capacity likely resulted from the changes in the membranes161hydrophobicity, and/or the pore structure (i.e., pore size and shape), as discussedearlier in this section.7.3 Impact of Exposure to Sodium Hypochlorite on theCleaning EfficiencyTypical images of membranes at different stages (e.g., after physical cleaning,and after chemical cleaning) of cleaning are presented in (Figure 7.6). Fouledmembranes (Figure 7.6b,c,d) were distinct from the those of unfouled membranes(Figure 7.6a), as a dark coating could be observed on the surface of the fouledmembranes for all the conditions investigated. Based on a qualitative interpretationof the images, physical cleaning removed some of the foulants (i.e., lighter colouredmembrane surfaces) (Figure 7.6c). Following chemical cleaning, the appearance ofthe membrane was similar to that of virgin membrane, suggesting that additionalfoulants were removed (Figure 7.6d).162(a) (b) (c) (d) ! Figure 7.6: Typical image of virgin membranes at different stages of cleaning.(a: unfouled; b: after activated sludge filtration; b: after physicalcleaning; c: after chemical cleaning.)Exposure to sodium hypochlorite prior to filtration increased the apparentamount of material that accumulated on the membrane surface during filtrationof activated sludge, for all exposure doses investigated, compared to that of aunfouled membrane (Figure 7.7). When the membranes were exposed to a fixedconcentration of sodium hypochlorite prior to filtration, the apparent amount ofaccumulated material increased when the exposure dose increased (Figure 7.7 aand 7.7c). Also, when the membranes were exposed to a fixed exposure dose, theincrease in the apparent amount of accumulated material was greater at the lowerconcentrations investigated (Figure 7.7a and 7.7b). These results are consistent withthose presented in Chapter 6, and indicate that the apparent amount of materialthat accumulated on the membrane surface during filtration of activated sludge wasnot only affected by the exposure dose, but also by the concentration of sodium163hypochlorite in the solution.(a) (b) (c) ! Figure 7.7: Typical image of the membrane at different stages of cleaning.(a: 500,000 ppm·hr (3600 ppm); b: 500,000 ppm·hr (22200 ppm); c:2,00,000 ppm·hr (3600 ppm); Right membranes: after physical cleaning,and left membranes: after chemical cleaning.)Figure 7.8 presents the trends of the normalized permeability at different stagesof the filtration and cleaning cycles for a virgin membrane. When the normalizedpermeability decreased by approximately 60% (i.e., TMP reached at 70 kPa),filtration was stopped and the membrane was cleaned.Approximately half of the lost permeability could be recovered with physicalcleaning. The remaining lost permeability was considered to be physically irrecover-able permeability.To assess the extent and rate at which the permeability could be recovered164during chemical cleaning, chemical cleaning was sequenced over a number ofintervals between which clean water permeation tests were performed, enabling thepermeability recovery to be monitored over time.!"#!$!"%!$!"&!$!"'!$!"(!$!")!$*"!!$!$ +!!!$ #!!!$ &!!!$, -., -/01$234356789$:;5349$<=5>9?9@$/5.4+1$:=?A=B$494C?6B9$!"#!$!"%!$!"&!$!"'!$!"(!$!")!$*"!!$!$ #$ ($ *+$ *&$ +!$ +#$,-./012$13$456784/9$496/282:$;51-.<=$>8.:82$767?./26$!"#$%&'()&(*'+%+,)-"*.%&'((#)/*&01*2'3(*)!*2.*'3%(%4#)!"#$%&'((#)%22*&01*2'3(*)!*2.*'3%(%4#)-"*.%&'((#)%22*&01*2'3(*)!*2.*'3%(%4#)Figure 7.8: Typical permeability trend of virgin membrane during filtrationfollowed by physical cleaning, and chemical cleaning.(Details of the membrane filtration tests and cleaning procedures arepresented in Section 4.5.2 and 4.5.3, respectively.)Figure 7.9 presents the values of the physically irrecoverable permeability.Physically irrecoverable permeability ranged from 28% to 70%, depending on theconditions investigated.165!" #$" $!" %$"&'()'*"+,--.(+'/0"-.-1(/*."$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"?5@A'+/00@"'((.+,B.(/10."3.(-./1'0'C@"6D<""Figure 7.9: Typical physically irrecoverable permeability (i.e.,LP (physical)LP(o))values for virgin membranes and membranes previously exposed tosodium hypochlorite for different exposure doses and concentrations.(Error bars correspond to minimum and maximum values from 2 repeatedexperiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite prior to filtration, the physically irrecoverable permeability was larger whenthe exposure dose increased (e.g., the physically irrecoverable permeabilities ofthe membranes exposed to 1,000,000 ppm·hr and 2,000,000 ppm·hr of sodiumhypochlorite at a concentration of 3600 ppm were approximately 38% and 70%,respectively. In comparison, the physically irrecoverable permeability of the virginmembranes was approximately 28%). When the membranes were exposed to afixed exposure dose prior to filtration, the physically irrecoverable permeability166increased for the lower concentration investigated (e.g., the physically irrecoverablepermeabilities of the membranes exposed to 2,000,000 ppm·hr of sodium hypochloriteat a concentration of 3600 ppm and 44300 ppm were approximately 70% and 48%,respectively).Figure 7.10 presents the values of the chemically recoverable permeability fora 6 hour cleaning period. For the type of membranes used in the present study,the duration of a chemical recovery cleaning (i.e., CIP) in full-scale operation istypically 2 to 6 hours [87]. Chemically recoverable permeability (after 6 hours)ranged from 30% to 8%, depending on the conditions investigated.167!" #$" $!" %$"&'()'*"+,--.(+'/0"-.-1(/*."$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"75.-'+/00?"(.+,@.(/10."3.(-./1'0'A?"6B<"6/C.(">"5,D(E",F"+5.-'+/0"+0./*'*)<""Figure 7.10: Typical chemically recoverable permeability (i.e.,LP (physical+chemical)−LP (physical)LP(o)) values with 6 hours of chemical cleaning,for virgin membranes and membranes previously exposed to sodiumhypochlorite for different exposure doses and concentrations.(Error bars correspond to minimum and maximum values from 2repeated experiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite prior to filtration, the chemically recoverable permeability (after 6 hours)was lower when the exposure dose increased (e.g., the chemically recoverablepermeabilities (after 6 hours) of the membranes exposed to 1,000,000 ppm·hr and2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600 ppm wereapproximately 8% and 5%, respectively. In comparison, the chemically recoverablepermeability (after 6 hours) of the virgin membranes was approximately 21%).168When the membranes were exposed to a fixed exposure dose prior to filtration,the chemically recoverable permeability decreased for the lower concentrationinvestigated (e.g., the chemically recoverable permeabilities of the membranesexposed to 2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600ppm and 44300 ppm were approximately 5% and 20%, respectively).Figure 7.11 presents the values of the chemically irrecoverable (i.e., irreversible)permeability after 6 hours of chemical cleaning. Chemically irrecoverable permeability(after 6 hours of chemical cleaning) ranged from 0% to 65%, depending on theconditions investigated.169!" #$" $!" %$"&'()'*"+,--.(+'/0"-.-1(/*."$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"?((.@.(A'10."3.(-./1'0'BC"6D<"6/E.(">"5,F(A",G"+5.-'+/0"+0./*'*)<""Figure 7.11: Typical chemically irrecoverable permeability (i.e.,LP (physical+chemical)LP(o)) values after physical, and 6 hours of chemicallycleaning, for virgin membranes and membranes previously exposed tosodium hypochlorite for different exposure doses and concentrations.(Error bars correspond to minimum and maximum values from 2repeated experiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite prior to filtration, the chemically irrecoverable permeability (after 6 hoursof chemical cleaning) was larger when the exposure dose increased (e.g., thechemically irrecoverable permeabilities (after 6 hours) of the membranes exposed to1,000,000 ppm·hr and 2,000,000 ppm·hr of sodium hypochlorite at a concentrationof 3600 ppm were approximately 33% and 65%, respectively. In comparison,the chemically irrecoverable permeability (after 6 hours) of the virgin membranes170was approximately 7%). When the membranes were exposed to a fixed exposuredose prior to filtration, the chemically irrecoverable permeability (after 6 hoursof chemical cleaning) increased for the lower concentration investigated (e.g., thechemically irrecoverable permeabilities (after 6 hours) of the membranes exposedto 2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600 ppm and44300 ppm were approximately 65% and 28%, respectively).Figure 7.12 presents the values of the chemically irrecoverable (i.e., irreversible)permeability after 24 hours of chemical cleaning. Chemically irrecoverable permea-bility (after 24 hours of chemical cleaning) ranged from 0% to 48%, dependingon the conditions investigated. After 24 hours of chemical cleaning, the differencebetween the chemically irrecoverable permeability of virgin membranes and mem-branes previously exposed to sodium hypochlorite was reduced. However, mostof the membranes previously exposed to sodium hypochlorite exhibited greaterchemically irrecoverable permeability than that of the virgin membranes.171!" #$" $!" %$"&'()'*"+,--.(+'/0"-.-1(/*."$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"889!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"###!!"33-":/;70<"$!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"=2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"#2!!!2!!!"33-45("67,--.(+'/0"-.-1(/*."'*"9>!!"33-":/;70<"?((.@.(A'10."3.(-./1'0'BC"6D<"6/E.("#8"5,F(A",G"+5.-'+/0"+0./*'*)<""Figure 7.12: Typical chemically irrecoverable permeability (i.e.,LP (physical+chemical)LP(o)) values after physical cleaning, and 24 hours ofchemical cleaning, for virgin membranes and membranes previouslyexposed to sodium hypochlorite for different exposure doses andconcentrations.(Error bars correspond to minimum and maximum values from 2repeated experiments).When the membranes were exposed to a fixed concentration of sodium hypochlo-rite prior to filtration, the chemically irrecoverable permeability (after 24 hours ofchemical cleaning) was larger when the exposure dose increased (e.g., the chemicallyirrecoverable permeabilities (after 24 hours) of the membranes exposed to 1,000,000ppm·hr and 2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600ppm were approximately 25% and 52%, respectively. In comparison, the chemically172irrecoverable permeability (after 24 hours) of the virgin membranes was approximately2%). When the membranes were exposed to a fixed exposure dose prior to filtration,the chemically irrecoverable permeability (after 24 hours of chemical cleaning)increased for the lower concentration investigated (e.g., the chemically irrecoverablepermeabilities (after 24 hours) of the membranes exposed to 2,000,000 ppm·hrof sodium hypochlorite at a concentration of 3600 ppm and 44300 ppm wereapproximately 18% and 52%, respectively).The results presented in Figures 7.6, 7.7, 7.9, 7.10, 7.11, and 7.12 as well asthe corresponding discussions, suggest that the efficiency of membrane cleaningdepends on the prior exposure to sodium hypochlorite. The change in the parametersinvestigated (i.e., visual observation of membrane surface, physically irrecoverablepermeability, chemically irrecoverable permeability, and chemically recoverablepermeability and the trend of chemically recoverable permeability) were not onlyaffected by the exposure dose, but also by the concentration of sodium hypochloritein the solution.These results indicate that, similar to the impact of sodium hypochlorite exposureon the physical/chemical characteristics of the membrane, the commonly usedrelationship to calculate the exposure dose, which is the product of concentration(C) and time (t), or Ct value, does not adequately reflect the impact of sodiumhypochlorite exposure on the efficiency of the membrane cleaning, for the exposureconditions investigated. The results also indicate that, the exposure time had agreater impact on the changes in the membrane cleaning efficiency parameters thanthe concentration of the sodium hypochlorite, which is consistent with the Cnt modelintroduced in Chapter 6.173The observed increase in the affinity of membranes exposed to sodium hypochlo-rite to the foulants could be due the the changes in the membrane hydrophobicityresulted from exposure to sodium hypochlorite as discussed in Chapter 6. Theseobservations are consistent with those reported by Mu et al. [105], where membraneswith a low HA to PVDF ratio had a higher affinity for BSA protein than membranewith higher ratio. Proteins are considered to be one of the major foulants ofmembranes used in activated sludge filtration.Permeability recovery from chemical cleaning cleaning was lower for membranespreviously subjected to greater doses of sodium hypochlorite (i.e., membraneexposed to 2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600ppm and 44300 ppm), than that of the membranes exposed up to 1,000,000 ppm·hrof sodium hypochlorite. This was likely primarily due to an increase in the averagepore diameter, along with a greater increase in membrane hydrophobicity, resultingfrom exposure to sodium hypochlorite (as reported in Chapter 6, Figures 6.17 and6.12).7.4 Rate of Permeability Recovery During ChemicalCleaningFigure 7.13 presents the typical trends for the chemically recoverable permeabilityduring chemical cleaning, for virgin membranes and membranes exposed to sodiumhypochlorite for different exposure doses and concentrations prior to filtration. Aspresented, exposure to sodium hypochlorite prior to filtration had substantial effectson the rate of permeability recovery during chemical cleaning.174!"#!"$!"%!"!" %" &" '" #$" #(" #)" $#" $*"+,-.,/012134"5,678,-4"9:;"<=-/>7?"7@"6A,.16/2"62,/?1?B"9A7=-C;"D1-B1?".,.0-/?,"(!!E!!!"FF.GA-"91?"%&!!"FF."H/IJ2;"(!!E!!!"FF.GA-"91?"$$$!!"FF."H/IJ2;"(!!E!!!"FF.GA-"91?"**%!!"FF."H/IJ2;"#E!!!E!!!"FF.GA-"91?"%&!!"FF."H/IJ2;"#E!!!E!!!"FF.GA-"91?"$$$!!"FF."H/IJ2;"#E!!!E!!!"FF.GA-"91?"**%!!"FF."H/IJ2;"$E!!!E!!!"FF.GA-"91?"%&!!"FF."H/IJ2;"$E!!!E!!!"FF.GA-"91?"**%!!"FF."H/IJ2;"Figure 7.13: Typical chemically recoverable permeability (i.e.,LP (physical+chemical)−LP (physical)LP(o)) trends during chemical cleaning forvirgin membranes and membranes previously exposed to sodiumhypochlorite for different exposure doses and concentrations.For virgin membranes, the results presented in Figure 7.13 indicate that initially,the rate of chemically recoverable permeability was high. Most of the recoveryoccurred within first 3 hours of the chemical cleaning. After the first 3 hours,additional recovery occurred slowly. Similar chemically recoverable permeabilitytrends were observed for the membranes previously exposed to 500,000 ppm·hr ofsodium hypochlorite at a concentration of 3600 ppm, 22200 ppm and 44300 ppm.For the membranes exposed to 1,000,000 ppm·hr of sodium hypochlorite at aconcentration of 3600 ppm, 22200 ppm and 44300 ppm, the initial rates of recoverywere also high (Figure 7.13). The extents of recovery within first 3 hours were175lower than those of the virgin membranes and the membranes previously exposed to500,000 ppm·hr of sodium hypochlorite. However, after the first 3 hours, additionalrecovery occurred relatively more rapidly than those of the virgin membranes andthe membranes previously exposed to 500,000 ppm·hr of sodium hypochlorite. Asa result, after 24 hours of the chemical cleaning, the recoveries were comparableto those of the of the virgin membranes and the membranes previously exposed to500,000 ppm·hr of sodium hypochlorite.The membranes exposed at higher exposure doses (i.e., membranes previouslyexposed to 2,000,000 ppm·hr of sodium hypochlorite at a concentration of 3600ppm and 44300 ppm), exhibited substantially different chemically recoverablepermeability trends during chemical cleaning, than that of virgin membranes (Figure7.13). The rate of recovery was relatively constant for the entire duration of thechemical cleaning (i.e., from 0 to 24 hours). However, the rates of recovery weresignificantly lower for these membranes, than the initial permeability recovery ratesof the virgin membranes and the membranes previously exposed to 500,000 ppm·hrand 1,000,000 ppm·hr of sodium hypochlorite. As a result, only limited recoveryoccurred within first 3 hours of chemical cleaning. Also, with a longer chemicalcleaning period (i.e., 24 hours), the chemically recoverable permeabilities of thesemembranes were significantly lower than the chemically recoverable permeabilitiesof the virgin membranes and the membranes previously exposed to 500,000 ppm·hrand 1,000,000 ppm·hr of sodium hypochlorite.The results indicate that significantly longer chemical cleaning cycles arerequired to recover a given amount of permeability for the membranes exposed tosodium hypochlorite, than that of the virgin membranes.1767.5 Proposed Non-destructive Method to AssessMembrane AgeOver a membrane lifetime, as exposure to sodium hypochlorite increases, thefollowing behaviours are expected.1. New membranes -(a) Initially the rate of reclamation of chemically recoverable permeabilitywill be high.(b) The major portions of the chemically recoverable permeability (e.g., 2/3of the total) will be recovered within first few hours (e.g., 3 hours) ofthe chemical cleaning.(c) After the first few hours (e.g., 3 hours), the rate of reclamation ofchemically recoverable permeability will decrease significantly.2. Membranes aged with moderate exposure to sodium hypochlorite -(a) Initially the rate of reclamation of chemically recoverable permeabilitywill be high.(b) The major portions of the chemically recoverable permeability (e.g., 2/3of the total) will not recovered within first 3 hours of chemical cleaning.(c) Even after the first 3 hours, the chemically recoverable permeabilitywill be recovered relatively quickly with time.(d) However, with a longer chemical cleaning period (i.e., 24 hours), thechemically recoverable permeabilities of these membranes will becomparable to the chemically recoverable permeability of virgin mem-branes.1773. Membranes aged with significant amount of exposure to sodium hypo- chlorite-(a) The rate of reclamation of chemically recoverable permeability will besimilar for the entire duration of the chemical cleaning (i.e., from 0 to24 hours).(b) The rate of reclamation will be significantly lower for these membranes,than the initial rate of reclamation of permeability recovery for virginmembranes.(c) As a result, only a small portions of the chemically recoverable permea-bility will be recovered within first 3 hours of chemical cleaning.(d) Even with a longer chemical cleaning period (i.e., 24 hours), the chemi-cally recoverable permeabilities of these membranes will be significantlylower than the chemically recoverable permeability of virgin membranes.The rate of permeability recovery can therefore be used to evaluate the stateof a membrane in terms of its operational age. Figure 7.14 presents the typicaltrends of the permeability recovery of the new, moderately aged and extensivelyaged membranes during chemical cleaning. Monitoring the permeability recoverytrend during chemical cleaning in full-scale operation to compare that with thepermeability recovery trend of the virgin membrane can be used as an effectivenon-destructive membrane characterization technique. This will help to assess thestate of a membrane, and provide insight into the remaining operational life.178!"#$%&'(&)(*+,-.*$/(*/,$'.'0(1+&"#23(4 (((5 (((((6 (((((((7((((((((((89 ( (((((((8: ( ((9;(<+,-.*$//=(#,*&>,#$?/,(@,#-,$?./.A=(1B3(4(((((84((((((((94(((((((((54(!"#$%"%&'()"*$Figure 7.14: Typical permeability recovery trends during chemical cleaning.7.6 Conclusions• Exposure to sodium hypochlorite significantly affected the filtration andcleaning performance of blended PVDF-based membranes.• The changes in the membrane cleaning efficiency parameters investigated179(i.e., appearance the membrane surface, physically irrecoverable permeability,chemically irrecoverable permeability, and chemically recoverable perme-ability) were not only affected by the previous exposure dose, but also by theconcentration of sodium hypochlorite in the solution used during exposure.• Results indicated that, similar to the impact of sodium hypochlorite exposureon the physical/chemical characteristics of the membrane, the commonlyused relationship to calculate the exposure dose which is the product ofconcentration (C) and time (t), or Ct value, does not adequately reflect theimpact of previous sodium hypochlorite exposure on the efficiency of themembrane cleaning, or the exposure conditions investigated. The impact ofsodium hypochlorite exposure on the membrane cleaning efficiency parameterscould be more accurately and consistently correlated to a modified exposuredose of the form Cnt.• Results of the membrane cleaning tests demonstrated that exposure to thesodium hypochlorite solution significantly increased the affinity of the mem-branes for the foulants. The observed increase in the affinity of the agedmembranes for the foulants could be due the the increased hydrophobicity ofthe aged membranes, which might have contributed to the relatively strongadsorption of the foulants to the membranes. This is consistent with thedecrease in the content of the HA of interest.• Exposure to sodium hypochlorite prior to filtration significantly affectedthe chemically recoverable permeability trends of the membranes duringextended chemical cleaning.180• A new non-destructive membrane characterization technique to evaluatethe amount of membrane ageing is proposed. The permeability recoverytrends of virgin membrane and the membranes under consideration (i.e., agedmembranes), during extended cleaning chemical can be used to evaluate thestate of membrane ageing.181Chapter 8Conclusions andRecommendations8.1 ConclusionsThe study discussed in this dissertation, provides a comprehensive quantification ofthe effects of sodium hypochlorite exposure on changes in the physical/chemicalcharacteristics and the filtration performances of blended PVDF-based hollow-fibermembranes, and it identifies the mechanism(s) responsible for the changes. Totest the hypotheses of the study, PVDF-based (both commercial and las-cast)membranes were exposed to sodium hypochlorite at different concentrations fordifferent durations, and comprehensive analyses of physical/chemical characteristicsand the filtration performances of the membranes were conducted. The majorconclusions of the study are listed below.• Exposure to sodium hypochlorite removed some of the material components182(i.e., hydrophilic additives of interest (HA)) from blended PVDF membranes.• The removal of the material components (i.e., HA) due to exposure to sodiumhypochlorite affected the physical/chemical characteristics of blended PVDFmembranes.• The extent of exposure to sodium hypochlorite (i.e., exposure dose) affectedthe amount of change of physical/chemical characteristics of blended PVDFmembranes.• The concentration of the sodium hypochlorite affected the amount of changeof physical/chemical characteristics of blended PVDF membranes.• The magnitude of the impact of exposure to sodium hypochlorite on thephysical/chemical characteristics of blended PVDF membranes investigatedcould not be simply related to the commonly used relationship Ct, for anyconcentration and/or exposure dose.• Results indicated that the impact of sodium hypochlorite exposure on theparameters investigated could be most accurately and consistently correlatedto an exposure dose relationship of the form Cnt rather than to the Ctrelationship commonly used to define exposure dose, for the exposure conditionsinvestigated.• For all the parameters investigated, the power coefficient n in the Cnt relation-ship was less than 1, indicating that the duration of the exposure had agreater impact on the changes in the physical/chemical characteristics ofthe membranes than the concentration of sodium hypochlorite, suggesting183that chemical cleaning should be performed using higher concentrations ofsodium hypochlorite and shorter durations.• The HA of interest of blended PVDF membranes became mineralized whenexposed to sodium hypochlorite. Results indicated that the removal of HAfrom the membrane is predominantly governed by oxidation, which initiallyresults in structural changes to the HA, enabling the oxidation productsto be released into the solution, and ultimately results in the completemineralization of the oxidation products.• The initial concentration of the HA of interest present in blended PVDFmembranes affected the removal rate of the hydrophilic additives from themembrane.• The concentration of NaOCl affected the removal rate of the HA of interestpresent in blended PVDF membranes.• The changes in the physical/chemical characteristics of the membrane dueto the exposure to sodium hypochlorite affected the filtration performance ofthe membrane.• The extent of exposure to sodium hypochlorite affected the filtration performanceof the membrane.• The concentration of the sodium hypochlorite affected the filtration performanceof the membrane.• A new non-destructive membrane characterization technique to evaluatethe amount of membrane ageing is proposed. The permeability recovery184trends of virgin membrane and the membranes under consideration (i.e., agedmembranes), during extended cleaning chemical can be used to evaluate thestate of membrane ageing.• The sample preparation and imaging parameters affected the qualitative andquantitative information (e.g., visual observation) of membrane properties ofinterest of blended membrane obtained from SEM imaging.• Results of the SEM imaging with different sample preparation and imagingparameters demonstrate that, in general, the membrane properties of interestobtained from the SEM images were less sensitive to the parameters, whenmembranes were coated with iridium and imaged with upper detector (UD)or mixed detector (MD). However, any change in the coating thickness,magnification or tilt angle significantly effected the membrane propertiesof interest obtained from the SEM images. Hence, if possible, samplepreparation and imaging parameters should not be varied during a study.8.2 Recommendations for Future WorkResults of the present research opened up the opportunity to know more aboutthe membrane ageing. Some recommendations for future research work are listedbelow.• In the present study blended PVDF-based virgin membranes were exposedcontinuously to sodium hypochlorite for ageing without any fouling. However,in practice, membranes undergo repeated fouling and chemical cleaningcycles during operation (e.g., daily maintenance cleaning with relatively low185sodium hypochlorite concentrations and short exposure times, and monthlyrecovery cleaning with higher concentrations and longer exposure times).– Research on the effects of repeated sodium hypochlorite cleaning cycleswith different frequencies on the physical/chemical characteristics, aswell as filtration properties of the membranes needs to be conducted.– The impact of NaOCl on fouled membranes might be less damaging onthe intrinsic membrane characteristics, and might need a longer exposureto NaOCl before a significant degradation is observed. Research onthe effects of these repeated fouling and sodium hypochlorite cleaningcycles on the physical/chemical characteristics, as well as filtrationproperties of the membranes needs to be conducted. This will help tounderstand how the results obtained form the accelerated ageing withoutany fouling can be used for the real life filtration applications.• Also, in the present research, concentrations of sodium hypochlorite usedwere higher than those typically used for chemical cleaning in full-scale waterand wastewater treatment membrane applications.– Longer term exposure studies at concentrations typical of those usedduring membrane cleaning should be performed to gain further insightinto the effects of exposure to lower concentrations of chemical cleaningagents.– In addition to these, filtration and cleaning studies using higher concen-trations of sodium hypochlorite, and lower durations than those typicallyused for chemical cleaning should be performed to assess the efficiency186of cleaning.• Although the contact angle is widely used to assess the hydrophilicity of themembrane surface, the extent of the increase in the contact angle showedleast consistent trend with the exposure dose and the concentration of sodiumhypochlorite used. PVDF based blended polymeric membrane might haveconsiderable contact angle hysteresis [162]. Contact angle measurement byusing contact angle hysteresis technique might produce more consistent andreliable results for assessment of the hydrophilicity of the membrane surface.8.3 Engineering SignificanceThe main knowledge gap that leads to the present research study was the fact thatthe exact effects and mechanisms of blended PVDF-based membrane ageing dueto exposure to chemical cleaning agents were not known. As a result, there wasno reliable tool available to predict the effects concentration and exposure doseof chemical cleaning agents on the membrane characteristics. Hence, choosingan optimized chemical cleaning approach, as well as predicting the membraneoperating life is a challenge in membrane system design. The present study targetsthis knowledge gap. A comprehensive study was conducted to assess the impactof sodium hypochlorite exposure on the physical/chemical characteristics, andmembrane filtration performances of blended PVDF-based membranes.Based on the investigation, a new exposure dose calculation concept wasproposed. It was found that the impact of exposure to sodium hypochlorite canbe modeled more accurately and consistently with a Cnt relationship, where n<1,for the exposure conditions investigated. Hence, instead of using the commonly187used Ct relationship, the Cnt relationship should be used to estimate the exposuredose. The results indicate that, exposing membranes to a higher concentrationfor a shorter duration is less detrimental than exposing membranes to a lowerconcentration for a longer duration. This outcome will empower design engineers toselect the concentration of cleaning agents, the frequency and duration of differentchemical cleaning schemes, to optimize the membrane operating life. As an example,cleaning scheme comprises of less frequent maintenance cleaning (usually done withrelatively low concentration of cleaning agents) and more frequent clean-in-place(usually done with relatively high concentration of cleaning agents) might bemore beneficial than the cleaning scheme comprises of more frequent maintenancecleaning and less frequent clean-in-place.Also, a comprehensive membrane characterization technique (i.e., FTIR, NMR,contact angle, tensile strength, SEM imaging, pore size, apparent porosity, dextranretention, appearance the membrane surface, physically irrecoverable permeability,chemically irrecoverable permeability, and chemically recoverable permeabilityand the trend of chemically recoverable permeability) to assess the impact ofsodium hypochlorite exposure on the membrane characteristic parameters has beendeveloped. In addition, the data were modelled to correlate the C and t with theobserved changes in the membrane characteristics. This approach can be used toassess the membrane operating life in advance for the following cases.• Lab tests - as the results indicate that C and t could be consistently correlatedwith the observed changes in the membrane characteristics, for different Cand exposure doses, accelerated membrane-ageing tests with relatively highC and longer t can be used to estimate the membrane ageing in full-scale real188life application.• Field tests - membrane samples can be collected from the field for character-ization. The results of the characterization tests can then be compared withthe model data obtained form the lab scale accelerated ageing, to assess theremaining lifetime of the membrane.By using the proposed non-destructive membrane characterization techniquebased on the permeability recovery trends of the membranes during extendedcleaning chemical, the amount of membrane ageing can be assessed. Based on theseresults modification/optimization of the process can be done.The results also indicated that the initial amount of HA content present in themembrane significantly effected the rate of release of the HA from the membranesand subsequent mineralization of the HA during NaOCl exposure. This indicatedthat content of HA in the blended PVDF membrane might play an important role inthe change of membrane characteristics during membrane ageing. This informationopens the opportunity to develop new membranes with longer operating life.The ultimate outcome of the research will be useful for the development ofoptimal cleaning protocols that are both effective at controlling fouling and atminimizing impacts on the membrane characteristics.189Bibliography[1] S. Z. Abdullah. Investigation of effect of dynamic operational conditions onmembrane fouling in a membrane enhanced biological phosphorus removalprocess. M.a.sc. thesis, The University of British Columbia, Vancouver,2007.[2] P. Aimar and C. Causserand. Impact of chemical cleaning of filtrationmembranes on their lifetime and properties. Procedia Engineering, (44):19–20, 2012.[3] A. Al-Amoudi and R. W. Lovitt. Fouling strategies and the cleaning systemof NF membranes and factors affecting cleaning efficiency. Journal ofMembrane Science, 303(1-2):4–28, 2007.[4] M. H. Al-Malack and G. K. Anderson. Use of crossflow microfiltration inwastewater treatment. Water Research, 31(12):3064–3072, 1997.[5] S. Allgeier, B. Alspach, and J. Vickers. Membrane filtration guidancemanual. United States Environmental Protection Agency, 2005.[6] Standard methods for the examination of water and wastewater. AmericanPublic Health Association, Washington, 1998.[7] A. Antony, R. Fudianto, S. Cox, and G. Leslie. Assessing the oxidativedegradation of polyamide reverse osmosis membrane—Accelerated ageingwith hypochlorite exposure. Journal of Membrane Science, 347(1-2):159–164, 2010.[8] E. Arkhangelsky, D. Kuzmenko, N. V. Gitis, M. Vinogradov, S. Kuiry, andV. Gitis. Hypochlorite Cleaning Causes Degradation of Polymer Membranes.Tribology Letters, 28(2):109–116, 2007.[9] E. Arkhangelsky, D. Kuzmenko, and V. Gitis. Impact of chemical cleaningon properties and functioning of polyethersulfone membranes. Journal ofMembrane Science, 305(1-2):176–184, 2007.190[10] E. Arkhangelsky, U. Goren, and V. Gitis. Retention of organic matter bycellulose acetate membranes cleaned with hypochlorite. Desalination, 223(1-3):97–105, 2008.[11] L. Be´goin, M. Rabiller-Baudry, B. Chaufer, M.-C. Hautbois, and T. Doneva.Ageing of PES industrial spiral-wound membranes in acid wheyultrafiltration. Desalination, 192(1-3):25–39, 2006.[12] S. Belfer, Y. Purinson, and o. Kedem. Surface modification of commercialpolyamide reverse osmosis membranes by radical grafting: An ATR-FTIRstudy. Acta Polym, 49:574–582, 1998.[13] G. Belfort, R. Davis, and A. L. Zydney. The Behavior of Suspensions andMacromolecular Solutions in Cross-Flow Microfiltration. Journal ofMembrane Science, 96:1–58, 1994.[14] P. R. Be´rube´, H. Lin, and Y. Watai. Fouling in air sparged submerged hollowfiber membranes at sub- and super-critical flux conditions. Journal ofMembrane Science, 307(2):169–180, 2008.[15] M. Beyer, B. Lohrengel, and L. D. Nghiem. Membrane fouling and chemicalcleaning in water recycling applications. Desalination, 250(3):977–981,2010.[16] G. Bolton, D. LaCasse, and R. Kuriyel. Combined models of membranefouling: Development and application to microfiltration and ultrafiltration ofbiological fluids. Journal of Membrane Science, 277(1-2):75–84, 2006.[17] G. R. Bolton, A. W. Boesch, and M. J. Lazzara. The effects of flow rate onmembrane capacity: Development and application of adsorptive membranefouling models. Journal of Membrane Science, 279(1-2):625–634, 2006.[18] E. Bouhabila. Fouling characterisation in membrane bioreactors. Separationand Purification Technology, 22-23(1-2):123–132, 2001.[19] W. R. Bowen, J. I. Calvo, and A. Hernandez. Steps of membrane blocking influx decline during protein microfiltration. Journal of Membrane Science,101(1-2):153–165, 1995.[20] C. Causserand, S. Rouaix, J.-P. Lafaille, and P. Aimar. Degradation ofpolysulfone membranes due to contact with bleaching solution.Desalination, 199(1-3):70–72, 2006.191[21] C. Causserand, S. Rouaix, J. Lafaille, and P. Aimar. Ageing of polysulfonemembranes in contact with bleach solution: Role of radical oxidation and ofsome dissolved metal ions. Chemical Engineering and Processing: ProcessIntensification, 47(1):48–56, 2008.[22] C. Causserand, G. Pierre, S. Rapenne, J. Schrotter, P. Sauvade, andO. Lorain. Characterization of ultrafiltration membranes by tracer’sretention: Comparison of methods sensitivity and reproducibility.Desalination, 250(2):767–772, 2010.[23] I.-S. Chang, P. Le-Clech, B. Jefferson, and S. Judd. Membrane Fouling inMembrane Bioreactors for Wastewater Treatment. Journal of EnvironmentalEngineering, 128(11):1018–1029, 2002.[24] S. Chang. Filtration of biomass with axial inter-fibre upward slug flow:performance and mechanisms. Journal of Membrane Science, 180(1):57–68,2000.[25] h. Chen and G. Belfort. Surface modification of poly(ether sulfone)ultrafiltration membranes by low-temperature plasma-induced graftpolymerization. Journal of Applied Polymer Science, 72:11699–11711,1999.[26] B. D. Cho and A. G. Fane. Fouling transients in nominally sub-critical fluxoperation of a membrane bioreactor. Journal of Membrane Science, 209(2):391–403, 2002.[27] J. Cho, G. Amy, J. Pellegrino, and Y. Yoon. Characterization of clean andnatural organic matter (NOM) fouled NF and UF membranes, and foulantscharacterization. Desalination, 118(1-3):101–108, 1998.[28] H. Choi, K. Zhang, D. Dionysiou, D. Oerther, and G. Sorial. Effect ofpermeate flux and tangential flow on membrane fouling for wastewatertreatment. Separation and Purification Technology, 45(1):68–78, 2005.[29] H. Choi, K. Zhang, D. D. Dionysiou, D. B. Oerther, and G. A. Sorial. Effectof permeate flux and tangential flow on membrane fouling for wastewatertreatment. Separation and Purification Technology, 45(1):68 – 78, 2005.[30] S. Claes, P. Vandezande, S. Mullens, K. De Sitter, R. Peeters, and M. K.Van Bael. Preparation and benchmarking of thin film supportedPTMSP-silica pervaporation membranes. Journal of Membrane Science,389:265–271, 2012.192[31] P. Cote, Z. Alam, and J. Penny. Hollow fiber membrane life in membranebioreactors (MBR). Desalination, 288:145 – 151, 2012.[32] G. Crozes, C. Anselme, and J. Mallevialle. Effect of adsorption of organicmatter on fouling of ultrafiltration membranes. Journal of MembraneScience, 84(1-2):61–77, 1993.[33] G. F. Crozes, J. G. Jacangelo, and C. Anselme. Impact of ultrafiltrationoperating conditions on membrane irreversible fouling. Journal ofMembrane Science, 1997.[34] A. Cui, Z. Liu, C. Xiao, and Y. Zhang. Effect of micro-sized SiO2-particleon the performance of PVDF blend membranes via TIPS. Journal ofMembrane Science, 360(1-2):259–264, 2010.[35] M. T. P. de Amorim and I. R. A. Ramos. Control of irreversible fouling byapplication of dynamic membranes. Desalination, 192(1-3):63–67, 2006.[36] L. Defrance and M. Y. Jaffrin. Comparison between filtrations at fixedtransmembrane pressure and fixed permeate flux: application to a membranebioreactor used for wastewater treatment. Journal of Membrane Science, 152(2):203–210, 1999.[37] V. T. Do, C. Y. Tang, M. Reinhard, and J. O. Leckie. Effects of chlorineexposure conditions on physiochemical properties and performance of apolyamide membrane - mechanisms and implications. EnvironmentalScience & Technology, 46(24):13184–13192, 2012.[38] V. T. Do, C. Y. Tang, M. Reinhard, and J. O. Leckie. Degradation ofPolyamide Nano-filtration and Reverse Osmosis Membranes byHypochlorite. Environmental Science & Technology, 46(2):852–859, 2012.[39] C. Dong, G. He, H. Li, R. Zhao, Y. Han, and Y. Deng. Antifoulingenhancement of poly(vinylidene fluoride) microfiltration membrane byadding Mg(OH)2 nanoparticles. Journal of Membrane Science, 387-388:40–47, 2012.[40] A. Drews. Membrane fouling in membrane bioreactors characterization,contradictions, cause and cures. Journal of Membrane Science, 363(1):1–28,2010.[41] A. M. ElHadidy, S. Peldszus, and M. I. V. Dyke. Development of a poreconstruction data analysis technique for investigating pore size distribution193of ultrafiltration membranes by atomic force microscopy. Journal ofMembrane Science, 429(0):373 – 383, 2013.[42] S. Elmaleh and N. Ghaffor. Cross-flow ultrafiltration of hydrocarbon andbiological solid mixed suspensions. Journal of Membrane Science, 118(1):111–120, 1996.[43] A. Ettori, E. Gaudichet-Maurin, and J. C. Schrotter. Permeability andchemical analysis of aromatic polyamide based membranes exposed tosodium hypochlorite. Journal of Membrane Science, 2011.[44] H. Evenblij, B. Verrecht, J. H. J. M. van der Graaf, and B. Van der Bruggen.Manipulating filterability of MBR activated sludge by pulsed substrateaddition. Desalination, 178(1-3):193–201, 2005.[45] A. G. Fane and C. J. D. Fell. A Review of Fouling and Fouling Control inUltrafiltration. Desalination, 62(1-3):117–136, 1987.[46] R. W. Field, D. Wu, J. A. Howell, and B. B. Gupta. Critical flux concept formicrofiltration fouling. Journal of Membrane Science, 100(3):259–272,1995.[47] V. Freger, J. Gilron, and S. Belfer. TFC polyamide membranes modified bygrafting of hydrophilic polymers: an FT-IR/AFM/TEM study. Journal ofMembrane Science, 209:283–292, 2002.[48] M. Gander, B. Jefferson, and S. Judd. Aerobic MBRs for domesticwastewater treatment: a review with cost considerations. Separation andPurification Technology, 18(2):119–130, 2000.[49] E. Gaudichet-Maurin and F. Thominette. Ageing of polysulfoneultrafiltration membranes in contact with bleach solutions. Journal ofMembrane Science, 282(1-2):198–204, 2006.[50] V. Gekas, K. Persson, M. Wahlgren, and B. sivik. Contact angles ofultrafiltration membranes and their possible correlation to membraneperformance. Journal of Membrane Science, 72:293–302, 1992.[51] Z. Geng and E. R. Hall. A comparative study of fouling-related properties ofsludge from conventional and membrane enhanced biological phosphorusremoval processes. Water Research, 41(19):4329 – 4338, 2007.[52] Z. Geng and E. R. Hall. A comparative study of fouling-related properties ofsludge from conventional and membrane enhanced biological phosphorusremoval processes. Water Research, 41(19):4329–4338, 2007.194[53] J. Gilron, S. Belfer, P. Vaisanen, and M. Nystro¨m. Contact angles ofultrafiltration membranes and their possible correlation to membraneperformance. Desalination, 140:167–179, 2001.[54] P. Grelier, S. Rosenberger, and A. Tazi-Pain. Influence of sludge retentiontime on membrane bioreactor hydraulic performance. Desalination, 192(1-3):10–17, 2006.[55] P. Griffiths and J. A. De Haseth. Fourier Transform Infrared Spectrometry.John Wiley & Sons, Mar. 2007. ISBN 9780470106297.[56] M. Gryta and M. Barancewicz. Influence of morphology of PVDF capillarymembranes on the performance of direct contact membrane distillation.Journal of Membrane Science, 358(1-2):158–167, 2010.[57] L. Guo, Y. Liu, C. Zhang, and J. Chen. Preparation of PVDF-based polymerinclusion membrane using ionic liquid plasticizer and Cyphos IL 104 carrierfor Cr(VI) transport. Journal of Membrane Science, 372(1-2):314–321,2011.[58] S. Hajibabania, A. Antony, G. Leslie, and P. Le-Clech. Relative impact offouling and cleaning on PVDF membrane hydraulic performances.Separation and Purification Technology, 2012.[59] R. Halaui, A. Moldavsky, Y. Cohen, R. Semiat, and E. Zussman.Development of micro-scale hollow fiber ultrafiltration membranes. Journalof Membrane Science, 379(1-2):370–377, 2011.[60] N. A. Hashim, Y. Liu, and K. Li. Stability of PVDF hollow fibre membranesin sodium hydroxide aqueous solution. Chemical Engineering Science, 2011.[61] M. Hashino, K. Hirami, T. Ishigami, Y. Ohmukai, T. Maruyama, N. Kubota,and H. Matsuyama. Effect of kinds of membrane materials on membranefouling with BSA. Journal of Membrane Science, 384(1-2):157–165, 2011.[62] C. Hass. Oxford Handbook of Innovation, chapter General Kinetics ofDisinfection Processes. American Society for Civil Engineers, ASCE, 2002.[63] Y. He, P. Xu, C. Li, and B. Zhang. High-concentration food wastewatertreatment by an anaerobic membrane bioreactor. Water Research, 39(17):4110–4118, 2005.[64] P. H. Hermans and H. L. Bredee. Principles of the mathematical treatment ofconstant-pressure filtration. J Soc Chem Ind T, 1936.195[65] N. Hilal, O. O. Ogunbiyi, N. J. Miles, and R. Nigmatullin. MethodsEmployed for Control of Fouling in MF and UF Membranes: AComprehensive Review. Separation Science and Technology, 40(10):1957–2005, 2005.[66] S. P. Hong, T. H. Bae, T. M. Tak, S. Hong, and A. Randall. Fouling controlin activated sludge submerged hollow fiber membrane bioreactors.Desalination, 143(3):219–228, 2002.[67] D. Horne. personal communication, Bio Imaging Facility, University ofBritish Columbia, Vancouver, 2010.[68] C.-Y. Hung, S.-D. Li, C.-C. Wang, and C.-Y. Chen. Influences of a bipolarmembrane and an ultrasonic field on alkaline water electrolysis. Journal ofMembrane Science, 389:197–204, 2012.[69] C. Huyskens, E. Brauns, E. Vanhoof, and H. Dewever. A new method for theevaluation of the reversible and irreversible fouling propensity of MBRmixed liquor. Journal of Membrane Science, 323(1):185–192, 2008.[70] B. K. Hwang, W. N. Lee, P. K. Park, and C. H. Lee. Effect of membranefouling reducer on cake structure and membrane permeability in membranebioreactor. Journal of Membrane Science, 2007.[71] D5 - Literature Review on Aeration and main Operating Conditions inMembrane Bioreactors. INSA, INSA Toulouse, France, 2006.[72] A. F. Ismail and A. Mansourizadeh. A comparative study on the structureand performance of porous polyvinylidene fluoride and polysulfone hollowfiber membranes for CO2 absorption. Journal of Membrane Science, 365(1-2):319–328, 2010.[73] S. Jankhah and P. R. Brub. Power induced by bubbles of different sizes andfrequencies on to hollow fibers in submerged membrane systems . WaterResearch, 47(17):6516 – 6526, 2013.[74] T. Jiang. Characterization and modelling of soluble microbial products inmembrane bioreactors. Ghent: Ghent University, 2007.[75] B. Jung, J. K. Yoon, B. Kim, and H. W. Rhee. Effect of molecular weight ofpolymeric additives on formation, permeation properties and hypochloritetreatment of asymmetric polyacrylonitrile membranes. Journal of MembraneScience, 243(1-2):45–57, 2004.196[76] K. Katsoufidou, S. Yiantsios, and A. Karabelas. A study of ultrafiltrationmembrane fouling by humic acids and flux recovery by backwashing:Experiments and modeling. Journal of Membrane Science, 266(1-2):40–50,2005.[77] M. Khayet, C. Cojocaru, and M. C. Garc ı´a Payo. Experimental design andoptimization of asymmetric flat-sheet membranes prepared for direct contactmembrane distillation. Journal of Membrane Science, 351(1-2):234–245,2010.[78] D.-H. Kim, I.-H. Baek, S.-U. Hong, and H.-K. Lee. Study on immobilizedliquid membrane using ionic liquid and PVDF hollow fiber as a support forCO2/N2 separation. Journal of Membrane Science, 372(1-2):346–354, 2011.[79] J. Kim and F. A. DiGiano. Defining critical flux in submerged membranes:Influence of length-distributed flux. Journal of Membrane Science, 280(1-2):752–761, 2006.[80] M. Kim, K. Saito, and S. Furusaki. Water flux and protein adsorption of ahollow fiber modified with hydroxyl groups. Journal of Membrane Science,56:289–302, 1991.[81] K. Kimura, Y. Hane, Y. Watanabe, G. Amy, and N. Ohkuma. Irreversiblemembrane fouling during ultrafiltration of surface water. Water Research, 38(14-15):3431–3441, 2004.[82] K. Kimura, H. Yamamura, and Y. Watanabe. Irreversible fouling in MF/UFmembranes caused by natural organic matters (NOMs) isolated fromdifferent origins. Separation Science and Technology, 41(7):1331–1344,2006.[83] Y. Kwon and J. Leckie. Hypochlorite degradation of cross linked polyamidemembranesII. Changes in hydrogen bonding behavior and performance.Journal of Membrane Science, 282(1-2):456–464, 2006.[84] Y.-N. Kwon and J. O. Leckie. Hypochlorite degradation of cross linkedpolyamide membranes. Journal of Membrane Science, 283(1-2):21–26,2006.[85] N. Knig. Development of polymeric membranes with catalytic properties fordrinking water treatment: Phase 1 - selection and physical characterization ofmembrane materials. Bachelor thesis, Technische Universitat Berlin, 2011.197[86] P. Le-Clech, B. Jefferson, and S. J. Judd. Impact of aeration, solidsconcentration and membrane characteristics on the hydraulic performance ofa membrane bioreactor. Journal of Membrane Science, 218(1-2):117–129,2003.[87] P. Le-Clech, V. Chen, and T. Fane. Fouling in membrane bioreactors used inwastewater treatment. Journal of Membrane Science, 2006.[88] I. Levitsky, A. Duek, E. Arkhangelsky, D. Pinchev, T. Kadoshian, H. Shetrit,R. Naim, and V. Gitis. Understanding the oxidative cleaning of UFmembranes. Journal of Membrane Science, 377(1-2):206–213, 2011.[89] I. Levitsky, A. Duek, R. Naim, E. Arkhangelsky, and V. Gitis. Cleaning UFmembranes with simple and formulated solutions. Chemical EngineeringScience, 69(1):679–683, 2012.[90] C.-L. Li, D.-M. Wang, A. Deratani, D. Que´mener, D. Bouyer, and J.-Y. Lai.Insight into the preparation of poly(vinylidene fluoride) membranes byvapor-induced phase separation. Journal of Membrane Science, 361(1-2):154–166, 2010.[91] H. Li, Y.-M. Chen, X.-T. Ma, J.-L. Shi, B.-K. Zhu, and L.-P. Zhu. Gelpolymer electrolytes based on active PVDF separator for lithium ion battery.I: Preparation and property of PVDF/poly(dimethylsiloxane) blendingmembrane. Journal of Membrane Science, 379(1-2):397–402, 2011.[92] S. Liang, K. Xiao, Y. Mo, and X. Huang. A novel ZnO nanoparticle blendedpolyvinylidene fluoride membrane for anti-irreversible fouling. Journal ofMembrane Science, 394-395:184–192, 2012.[93] C. Liu, S. Caothien, J. Hayes, T. Caothuy, T. Otoyo, and T. Ogawa.Membrane chemical cleaning: from art to science. In American Water WorksAssociation, Membrane Technology Conference, pages 1–25, 2001.[94] F. Liu, M. R. M. Abed, and K. Li. Preparation and characterization ofpoly(vinylidene fluoride) (PVDF) based ultrafiltration membranes usingnano-Al2O3. Journal of Membrane Science, 366(1-2):97–103, 2011.[95] J. Lu, Z. Zhang, T. Ye, X. Qin, Y. Huang, X. Yang, W. Luo, B. Chen, W. Liu,and Y. Liang. Ageing of polyvinylchloride membranes in ultrafiltration ofdrinking water by chemical cleaning. In Multimedia Technology (ICMT),2011 International Conference on, pages 3834–3837, 2011.198[96] S. S. Madaeni, A. G. Fane, and D. E. Wiley. Factors influencing critical fluxin membrane filtration of activated sludge. Journal of Chemical Technology& Biotechnology, 74(6):539–543, 1999.[97] S. S. Madaeni, S. Zinadini, and V. Vatanpour. A new approach to improveantifouling property of PVDF membrane using in situ polymerization ofPAA functionalized TiO2 nanoparticles. Journal of Membrane Science, 380(1-2):155–162, 2011.[98] S.-K. Mah, C. P. Leo, T. Y. Wu, and S.-P. Chai. A feasibility investigation onultrafiltration of palm oil and oleic acid removal from glycerin solutions:Flux decline, fouling pattern, rejection and membrane characterizations.Journal of Membrane Science, 389:245–256, 2012.[99] F. Malz and H. Jancke. Validation of quantitative NMR. Journal ofpharmaceutical and biomedical analysis, 38(5):813–823, 2005.[100] A. Mansourizadeh, A. F. Ismail, M. S. Abdullah, and B. C. Ng. Preparationof polyvinylidene fluoride hollow fiber membranes for CO2 absorption usingphase-inversion promoter additives. Journal of Membrane Science, 355(1-2):200–207, 2010.[101] A. d. Marshall, P. A. Munro, and G. Tragardh. AThe effect of protein foulingin microfiltration and ultrafiltration on permeate flux, protein retention andselectivity: A literature review. Desalination, 91:65–108, 1993.[102] M. A. Masuelli, M. Grasselli, J. Marchese, and N. A. Ochoa. Preparation,structural and functional characterization of modified porous PVDFmembranes by gamma-irradiation. Journal of Membrane Science, 389:91–98, 2012.[103] F. Meng, S.-R. Chae, A. Drews, M. Kraume, H.-S. Shin, and F. Yang.Recent advances in membrane bioreactors (MBRs): Membrane fouling andmembrane material. Water Research, 43(6):1489–1512, 2009.[104] F. Meng, S.-R. Chae, H.-S. Shin, F. Yang, and Z. Zhou. Recent advances inmembrane bioreactors: configuration development, pollutant elimination,and sludge reduction. Environmental Engineering Science, 29(3):139–160,2012.[105] C. Mu, Y. Su, M. Sun, W. Chen, and Z. Jiang. Remarkable improvement ofthe performance of poly (vinylidene fluoride) microfiltration membranes bythe additive of cellulose acetate. Journal of Membrane Science, 350(1-2):293–300, 2010.199[106] D. A. Musale and S. S. Kulkarni. Fouling reduction inpoly(acrylonitrile-co-acrylamide) ultrafiltration membranes. Journal ofMembrane Science, 1996.[107] R. Naim, A. F. Ismail, and A. Mansourizadeh. Preparation of microporousPVDF hollow fiber membrane contactors for CO2 stripping fromdiethanolamine solution. Journal of Membrane Science, 392-393:29–37,2012.[108] C. A. Ng, D. Sun, J. Zhang, H. C. Chua, W. Bing, and S. Tay. Strategies toimprove the sustainable operation of membrane bioreactors. In Proceedingsof the International Desalination Association Conference, Singapore, 2005.[109] S. Ognier, C. Wisniewski, and A. Grasmick. Membrane bioreactor fouling insub-critical filtration conditions: a local critical flux concept. Journal ofMembrane Science, 229(1-2):171–177, 2004.[110] Y. K. Ong, N. Widjojo, and T.-S. Chung. Fundamentals of semi-crystallinepoly(vinylidene fluoride) membrane formation and its prospects for biofuel(ethanol and acetone) separation via pervaporation. Journal of MembraneScience, 378(1-2):149–162, 2011.[111] N. Pezeshk and R. M. Narbaitz. More fouling resistant modified PVDFultrafiltration membranes for water treatment. Desalination, 287:247–254,2012.[112] N. Pezeshk, D. Rana, R. M. Narbaitz, and T. Matsuura. Novel modifiedPVDF ultrafiltration flat-sheet membranes. Journal of Membrane Science,389:280–286, 2012.[113] A. J. Phillips and P. A. Hamilton. Improved detection limits in Fouriertransform spectroscopy from a maximum entropy approach to baselineestimation. Anal. Chem., 68(22):40204025, 1996.[114] j. Pieracci, j. Crivello, and G. Belfort. Water flux and protein adsorption of ahollow fiber modified with hydroxyl groups. Journal of Membrane Science,156:223–240, 1999.[115] M. Pilutti and J. Nemeth. Technical and cost review of commerciallyavailable MF/UF membrane products. In International DesalinationAssociation World Congress on Desalination and Water Reuse, pages 1–15,2003.200[116] N. Porcelli and S. Judd. Chemical cleaning of potable water membranes:The cost benefit of optimisation. Water Research, 44(5):1389 – 1398, 2010.ISSN 0043-1354.[117] N. Porcelli and S. Judd. Chemical cleaning of potable water membranes: Areview. Separation and Purification Technology, 71(2):137–143, 2010.[118] R. Prulho, S. Therias, A. Rivaton, and J.-L. Gardette. Ageing ofpolyethersulfone/polyvinylpyrrolidone blends in contact with bleach water.Polymer Degradation and Stability, 98(6):1164–1172, 2013.[119] V. Puspitasari, A. Granville, and P. Le-Clech. Cleaning and ageing effect ofsodium hypochlorite on polyvinylidene fluoride (PVDF) membrane.Separation and Purification Technology, 72:301–308, 2010.[120] J.-J. Qin, F.-S. Wong, Y. Li, and Y.-T. Liu. A high flux ultrafiltrationmembrane spun from PSU/PVP (K90)/DMF/1,2-propanediol. Journal ofMembrane Science, 211(1):139–147, 2003.[121] J.-J. Qin, Y.-M. Cao, Y.-Q. Li, Y. Li, M.-H. Oo, and H. Lee. Hollow fiberultrafiltration membranes made from blends of PAN and PVP. Separationand Purification Technology, 36(2):149–155, 2004.[122] J.-J. Qin, Y.-M. Cao, and Y. Li. Effect of hypochlorite concentration onproperties of posttreated outer-skin ultrafiltration membranes spun fromcellulose acetate/poly(vinyl pyrrolidone) blends. Journal of Applied PolymerScience, 97(1):227–231, 2005.[123] J.-J. Qin, M.-H. Oo, and Y. Li. Development of high flux polyethersulfonehollow fiber ultrafiltration membranes from a low critical solutiontemperature dope via hypochlorite treatment. Journal of Membrane Science,247(1-2):137–142, 2005.[124] C. Regula, E. Carretier, Y. Wyart, M. Sergent, G. Ge´san-Guiziou, D. Ferry,A. Vincent, D. Boudot, and P. Moulin. Ageing of ultrafiltration membranesin contact with sodium hypochlorite and commercial oxidant: Experimentaldesigns as a new ageing protocol. Separation and Purification Technology,103:119–138, 2013.[125] H. Reingruber, A. Zankel, C. Mayrhofer, and P. Poelt. A new in situ methodfor the characterization of membranes in a wet state in the environmentalscanning electron microscope. Journal of Membrane Science, 399400(0):86 –94, 2012.201[126] S. Rouaix, C. Causserand, and P. Aimar. Experimental study of the effects ofhypochlorite on polysulfone membrane properties. Journal of MembraneScience, 277(1-2):137–147, 2006.[127] D. P. Saroj, G. Guglielmi, D. Chiarani, and G. Andreottola. Subcriticalfouling behaviour modelling of membrane bioreactors for municipalwastewater treatment: The prediction of the time to reach critical operatingcondition. Desalination, 231(13):175 – 181, 2008.[128] F. Shi, Y. Ma, J. Ma, P. Wang, and W. Sun. Preparation and characterizationof PVDF/TiO2 hybrid membranes with different dosage of nano-TiO2.Journal of Membrane Science, 389:522–531, 2012.[129] M. V. Simhon. The fate of emerging contaminants in wastewater treatmentplants. M.a.sc. thesis, The university of British Columbia, Vancouver, 2013.[130] A. Simon, W. E. Price, and L. D. Nghiem. Effects of chemical cleaning onthe nanofiltration of pharmaceutically active compounds (PhACs).Separation and Purification Technology, 88:208–215, 2012.[131] A. Simon, W. E. Price, and L. D. Nghiem. Impact of membrane ageingcaused by chemical cleaning. Procedia Engineering, 2012.[132] A. Simon, J. A. McDonald, S. J. Khan, W. E. Price, and L. D. Nghiem.Effects of caustic cleaning on pore size of nanofiltration membranes andtheir rejection of trace organic chemicals. Journal of Membrane Science,447:153–162, 2013.[133] S. Simone, A. Figoli, A. Criscuoli, M. C. Carnevale, A. Rosselli, andE. Drioli. Preparation of hollow fibre membranes from PVDF/PVP blendsand their application in VMD. Journal of Membrane Science, 364(1-2):219–232, 2010.[134] N. P. D. Souza and O. D. Basu. Comparative analysis of physical cleaningoperations for fouling control of hollow fiber membranes in drinking watertreatment. Journal of Membrane Science, 436:28 – 35, 2013.[135] A. Staff. Microfiltration and Ultrafiltration Membranes for Drinking Water(M53). American Water Works Association, 2011.[136] Y. Sui, X. Gao, Z. Wang, and C. Gao. Antifouling and antibacterialimprovement of surface-functionalized poly(vinylidene fluoride) membraneprepared via dihydroxyphenylalanine-initiated atom transfer radical graftpolymerizations. Journal of Membrane Science, 394-395:107–119, 2012.202[137] P. Sukitpaneenit and T.-S. Chung. Molecular design of the morphology andpore size of PVDF hollow fiber membranes for ethanol–water separationemploying the modified pore-flow concept. Journal of Membrane Science,374(1-2):67–82, 2011.[138] P. Sukitpaneenit, T.-S. Chung, and L. Y. Jiang. Modified pore-flow model forpervaporation mass transport in PVDF hollow fiber membranes forethanol–water separation. Journal of Membrane Science, 362(1-2):393–406,2010.[139] Omnic User’s Guide. Thermo Electron Corporation, Madison WI, 2006.[140] F. Thominette, O. Farnault, E. Gaudichet-Maurin, C. Machinal, andJ. Schrotter. Ageing of polyethersulfone ultrafiltration membranes inhypochlorite treatment. Desalination, 200(1-3):7–8, 2006.[141] J.-y. Tian, Z.-l. Chen, Y.-l. Yang, H. Liang, J. Nan, and G.-b. Li. Consecutivechemical cleaning of fouled PVC membrane using NaOH and ethanol duringultrafiltration of river water. Water Research, 44(1):59–68, 2010.[142] E. M. Tracey and R. H. Davis. Protein fouling of track-etched polycarbonatemicrofiltration membranes. Journal of Colloid and Interface Science, 167(1):104–116, 1994.[143] M. Ulbricht, K. Richau, and H. Kamusewitz. Chemically andmorphologically defined ultrafiltration membrane surfaces prepared byheterogeneous photo-initiated graft polymerization. colloids and surfaces,138:353–366, 1998.[144] Structural characterization and comparison of iridium, platinum andgold/palladium ultra-thin film coatings for STM of biomolecules. U.S.Department of Energy, Office of Scientific and Technical Information, 1997.[145] X.-M. Wang, X.-Y. Li, and K. Shih. In situ embedment and growth ofanhydrous and hydrated aluminum oxide particles on polyvinylidene fluoride(PVDF) membranes. Journal of Membrane Science, 368(1-2):134–143,2011.[146] Y. Wang, J. Kim, K. Choo, Y. Lee, and C. H. Lee. Hydrophilic modificationof polypropylene microfiltration membranes by ozone-induced graftpolymerization. Journal of Membrane Science, 169:269–276, 2000.203[147] Z. Wang, Z. Wu, Q. Zhou, and D. Yang. Effect of hypochlorite cleaning onthe physiochemical characteristics of polyvinylidene fluoride membranes.Chemical Engineering Journal, 165(2):750–750, 2010.[148] D. Wavhal and E. Fisher. Hydrophilic modification of polyethersulfonemembranes by low temperature plasma-induced graft polymerization.Journal of Membrane Science, 209:255–269, 2002.[149] A. Weis, M. R. Bird, and M. Nystro¨m. The chemical cleaning of polymericUF membranes fouled with spent sulphite liquor over multiple operationalcycles. Journal of Membrane Science, 2003.[150] S. Wickramasinghe, S. Bower, Z. Chen, A. Mukherjee, and S. Husson.Relating the pore size distribution of ultrafiltration membranes to dextranrejection. Journal of Membrane Science, 340(1-2):1–8, 2009.[151] S. H. Wolff and A. L. Zydney. Effect of bleach on the transportcharacteristics of polysulfone hemodialyzers. Journal of Membrane Science,243(1-2):389–399, 2004.[152] S. Wongchitphimon, R. Wang, and R. Jiraratananon. Surface modification ofpolyvinylidene fluoride-co-hexafluoropropylene (PVDF–HFP) hollow fibermembrane for membrane gas absorption. Journal of Membrane Science, 381(1-2):183–191, 2011.[153] S. Wongchitphimon, R. Wang, R. Jiraratananon, L. Shi, and C. H. Loh.Effect of polyethylene glycol (PEG) as an additive on the fabrication ofpolyvinylidene fluoride-co-hexafluropropylene (PVDF-HFP) asymmetricmicroporous hollow fiber membranes. Journal of Membrane Science, 369(1-2):329–338, 2011.[154] K. Yadav, K. Morison, and M. P. Staiger. Effects of hypochlorite treatmenton the surface morphology and mechanical properties of polyethersulfoneultrafiltration membranes. Polymer Degradation and Stability, 94(11):1955–1961, 2009.[155] h. Yamagishi, j. Crivello, and G. Belfort. Evaluation of photochemicallymodified poly(arylsulfone) ultrafiltration membranes. Journal of MembraneScience, 105:249–259, 1995.[156] X. Yan, M. R. Bilad, R. Gerards, L. Vriens, A. Piasecka, and I. F. J.Vankelecom. Comparison of MBR performance and membrane cleaning in asingle-stage activated sludge system and a two-stage anaerobic/aerobic204(A/A) system for treating synthetic molasses wastewater. Journal ofMembrane Science, 394-395:49–56, 2012.[157] X. Yang, B. Deng, Z. Liu, L. Shi, X. Bian, M. Yu, L. Li, J. Li, and X. Lu.Microfiltration membranes prepared from acryl amide graftedpoly(vinylidene fluoride) powder and their pH sensitive behaviour. Journalof Membrane Science, 362(1-2):298–305, 2010.[158] Y. Ye, V. Chen, and A. G. Fane. Modeling long-term subcritical filtration ofmodel EPS solutions. Desalination, 191(1-3):318–327, 2006.[159] Y. Ye, L. N. Sim, B. Herulah, V. Chen, and A. G. Fane. Effects of operatingconditions on submerged hollow fibre membrane systems used aspre-treatment for seawater reverse osmosis. Journal of Membrane Science,365(1-2):78–88, 2010.[160] M. Zenkiewicz and J. Richert. Influence of polymer samples preparationprocedure on their mechanical properties. Journal of Achievements inMaterials and Manufacturing Engineering, 26:155 – 158, 2008.[161] J. Zhang, H. C. Chua, J. Zhou, and A. G. Fane. Factors affecting themembrane performance in submerged membrane bioreactors. Journal ofMembrane Science, 284(1-2):54–66, 2006.[162] W. Zhang, M. Wahlgren, and B. SIVIK. Membrane characterization by thecontact angle techniqueii. characterization of uf-membranes and comparisonbetween the captive bubble and sessile drop as methods to obtain watercontact angles. Desalination, 72:263 – 273, 1989.[163] W. Zhao, Y. Su, C. Li, Q. Shi, X. Ning, and Z. Jiang. Fabrication ofantifouling polyethersulfone ultrafiltration membranes using Pluronic F127as both surface modifier and pore-forming agent. Journal of MembraneScience, 318(1-2):405–412, 2008.[164] Z. Zhao, J. Zheng, M. Wang, H. Zhang, and C. C. Han. High performanceultrafiltration membrane based on modified chitosan coating and electrospunnanofibrous PVDF scaffolds. Journal of Membrane Science, 394-395:209–217, 2012.[165] Y.-M. Zheng, S.-W. Zou, K. G. N. Nanayakkara, T. Matsuura, and J. P. Chen.Adsorptive removal of arsenic from aqueous solution by a PVDF/zirconiablend flat sheet membrane. Journal of Membrane Science, 374(1-2):1–11,2011.205[166] E. Zondervan, A. Zwijnenburg, and B. Roffel. Statistical analysis of datafrom accelerated ageing tests of PES UF membranes. Journal of MembraneScience, 300(1-2):111–116, 2007.206Appendix ASupporting MaterialsA.1 Matlab Codes% Matlab Codes to Detect Pores on a Membrane Surface.% by Syed Zaki Abdullahclc;close all;thresholdValue = 8; % Pick something% Get a standard SEM image.filename= virginCP51dec2011e ;originalImage = imread([filename .TIF ]);% reading first 3 channels%originalImage = originalImage(:,:,1:3);subplot(1, 2, 1);imshow(originalImage, []);207title( Original image );redband = originalImage(:,:,1);% subplot(2, 3, 2);% imshow(redband, []);% title( Red Band );thresholdValue = 111;binaryImage = redband < thresholdValue;binaryImage = imfill(binaryImage, holes );%subplot(2, 3, 3);%imshow(binaryImage, []);%title( binary image );labeledImage = bwlabel(binaryImage, 8); % Label each blob so can do calc on itcoloredLabels = label2rgb (labeledImage, hsv , k , shuffle ); % pseudo randomcolor labels%subplot(2,3,4); imagesc(coloredLabels); title( Pseudo colored labels );blobMeasurements = regionprops(labeledImage, all ); % Get all the poreproperties.% Get a list of the areas.area values = [blobMeasurements.Area];% Filter the pores by area. Just keep pores in a certain area range.% Find out which pores have an area between 25 and 2500 pixels.208% Use whatever the know valid range for Pores area is.indexes = find((25 <= area values) & (area values <= 2500));poresOnlyImage = ismember(labeledImage, indexes);%subplot(1,3,2);%imshow(poresOnlyImage, []);%title( Pores-only image );% Get the subset of pores that describes only the pores.blobMeasurements = blobMeasurements(indexes); % Get the pore properties only.% % % bwboundaries returns a cell array, where each cell% contains the row/column coordinates for an object in the image.% Plot the borders of all the coins on the original% grayscale image using the coordinates returned by bwboundaries.subplot(1,2,2); imshow(originalImage); title( Pores outlined over original image );hold on;boundaries = bwboundaries(poresOnlyImage);numberOfPores = size(blobMeasurements, 1);for k = 1 : numberOfPoresthisBoundary = boundariesk;plot(thisBoundary(:,2), thisBoundary(:,1), g , LineWidth , 1);endhold off;209% hgexport(gcf,[filename .eps ]);% above line save the figure file% % % List the various parameters to the command window.outfile = 1;outfile=fopen([filename .csv ], w );fprintf(outfile, pore #, Mean Intensity, Area, Perimeter,Centroid x, Centroid y, Major axis length, Minor axis length n );for k = 1 : numberOfPores % Loop through all pores.% Find the mean of each pore.thisBlobsPixels = blobMeasurements(k).PixelIdxList; % Get list of pixels incurrent pore.meanGL = mean(originalImage(thisBlobsPixels)); % Find mean intensity (inoriginal image!)blobArea = blobMeasurements(k).Area; % Get area.blobPerimeter = blobMeasurements(k).Perimeter; % Get perimeter.blobCentroid = blobMeasurements(k).Centroid; % Get centroid.fprintf(outfile, %d, %18.1f, %11.1f, %8.1f, %8.1f, %8.1f, %8.1f, %8.1fn ,k, meanGL, blobArea, blobPerimeter, blobCentroid, blobMeasurements(k).MajorAxisLength, blobMeasurements(k).MinorAxisLength);endfclose(outfile);outfile=1;210A.2 Typical SEM ImagesFigure A.1: Typical SEM images of membranes acquired with 2nm coating,25K magnification and 30 degree tilt211Figure A.2: Typical SEM images of membranes acquired with 5nm coating,25K magnification and 30 degree tilt212Figure A.3: Typical SEM images of membranes acquired with 2nm coating,50K magnification and 30 degree tilt213Figure A.4: Typical SEM images of membranes acquired with 5nm coating,50K magnification and 30 degree tilt214Figure A.5: Typical SEM images of membranes acquired with 2nm coating,25K magnification and 0 degree tilt215Figure A.6: Typical SEM images of membranes acquired with 5nm coating,25K magnification and 0 degree tilt216Figure A.7: Typical SEM images of membranes acquired with 2nm coating,50K magnification and 0 degree tilt217Figure A.8: Typical SEM images of membranes acquired with 5nm coating,50K magnification and 0 degree tilt218A.3 Typical results of the SEM image analysisTable A.1: Typical results of the SEM image analysis.Sample ID CoatingmetalCoatingThicknessDetector Magni-ficationTiltangelD90(micron)Dav(micron)Porosity(%)dc/dtAu 2nm LD 50K 0tilt Au 2nm LD 50K 0tilt 0.041 0.026 12.3 1.7Au 2nm LD 50K 30tilt Au 2nm LD 50K 30tilt 0.043 0.026 12.2 1.7Au 2nm MD 50K 0tilt Au 2nm MD 50K 0tilt 0.034 0.023 13.0 2.0Au 2nm UD 50K 0tilt Au 2nm UD 50K 0tilt 0.028 0.020 11.5 1.9Au 2nm UD 50K 30tilt Au 2nm UD 50K 30tilt 0.029 0.020 10.5 1.9Au 5nm LD 50K 0tilt Au 5nm LD 50K 0tilt 0.038 0.023 15.4 1.9Au 5nm LD 50K 30tilt Au 5nm LD 50K 30tilt 0.040 0.025 19.3 1.9Au 5nm MD 50K 0tilt Au 5nm MD 50K 0tilt 0.032 0.021 17.2 2.1Au 5nm MD 50K 30tilt Au 5nm MD 50K 30tilt 0.031 0.021 14.9 1.9Au 5nm UD 50K 0tilt Au 5nm UD 50K 0tilt 0.033 0.022 19.1 2.1Au 5nm UD 50K 30tilt Au 5nm UD 50K 30tilt 0.023 0.017 8.7 2.1Continued on next page219Table A.1 – Continued from previous pageSample ID CoatingmetalCoatingThicknessDetector Magni-ficationTiltangelD90(micron)Dav(micron)Porosity(%)dc/dtAuPd 2nm LD 50K 30tilt AuPd 2nm LD 50K 30tilt 0.041 0.025 7.8 1.7AuPd 2nm MD 50K 30tilt AuPd 2nm MD 50K 30tilt 0.023 0.017 3.7 1.9AuPd 2nm UD 50K 30tilt AuPd 2nm UD 50K 30tilt 0.027 0.019 10.1 2.1AuPd 5nm LD 50K 0tilt AuPd 5nm LD 50K 0tilt 0.040 0.025 15.9 1.9AuPd 5nm LD 50K 30tilt AuPd 5nm LD 50K 30tilt 0.039 0.025 14.9 1.9AuPd 5nm MD 50K 0tilt AuPd 5nm MD 50K 0tilt 0.030 0.020 14.2 2.2AuPd 5nm MD 50K 30tilt AuPd 5nm MD 50K 30tilt 0.037 0.023 17.8 2.0AuPd 5nm UD 50K 0tilt AuPd 5nm UD 50K 0tilt 0.031 0.020 15.8 2.2AuPd 5nm UD 50K 30tilt AuPd 5nm UD 50K 30tilt 0.034 0.022 17.1 2.2Ir 2nm LD 50K 0tilt Ir 2nm LD 50K 0tilt 0.044 0.025 14.1 1.9Ir 2nm LD 50K 30tilt Ir 2nm LD 50K 30tilt 0.035 0.022 7.8 1.8Ir 2nm MD 50K 0tilt Ir 2nm MD 50K 0tilt 0.029 0.019 12.4 2.0Continued on next page220Table A.1 – Continued from previous pageSample ID CoatingmetalCoatingThicknessDetector Magni-ficationTiltangelD90(micron)Dav(micron)Porosity(%)dc/dtIr 2nm MD 50K 30tilt Ir 2nm MD 50K 30tilt 0.032 0.021 11.8 1.9Ir 2nm UD 50K 0tilt Ir 2nm UD 50K 0tilt 0.029 0.019 12.8 1.9Ir 2nm UD 50K 30tilt Ir 2nm UD 50K 30tilt 0.032 0.021 14.2 1.9Ir 5nm LD 50K 0tilt Ir 5nm LD 50K 0tilt 0.036 0.024 9.2 2.0Ir 5nm LD 50K 30tilt Ir 5nm LD 50K 30tilt 0.037 0.023 9.2 1.8Ir 5nm MD 50K 0tilt Ir 5nm MD 50K 0tilt 0.030 0.020 15.7 2.2Ir 5nm MD 50K 30tilt Ir 5nm MD 50K 30tilt 0.030 0.020 12.7 2.0Ir 5nm UD 50K 0tilt Ir 5nm UD 50K 0tilt 0.029 0.019 11.2 2.0Ir 5nm UD 50K 30tilt Ir 5nm UD 50K 30tilt 0.029 0.020 13.0 2.0Au 2nm LD 25K 0tilt Au 2nm LD 25K 0tilt 0.055 0.038 14.1 1.8Au 2nm LD 25K 30tilt Au 2nm LD 25K 30tilt 0.058 0.039 15.7 1.8Au 2nm MD 25K 0tilt Au 2nm MD 25K 0tilt 0.053 0.036 19.3 2.0Continued on next page221Table A.1 – Continued from previous pageSample ID CoatingmetalCoatingThicknessDetector Magni-ficationTiltangelD90(micron)Dav(micron)Porosity(%)dc/dtAu 2nm MD 25K 30tilt Au 2nm MD 25K 30tilt 0.051 0.036 19.5 1.9Au 2nm UD 25K 0tilt Au 2nm UD 25K 0tilt 0.056 0.037 21.6 1.9Au 2nm UD 25K 30tilt Au 2nm UD 25K 30tilt 0.041 0.031 9.6 1.9Au 5nm LD 25K 0tilt Au 5nm LD 25K 0tilt 0.058 0.038 19.5 2.1Au 5nm LD 25K 30tilt Au 5nm LD 25K 30tilt 0.051 0.036 16.7 1.9Au 5nm MD 25K 0tilt Au 5nm MD 25K 0tilt 0.048 0.034 18.5 2.2Au 5nm MD 25K 30tilt Au 5nm MD 25K 30tilt 0.044 0.032 15.7 2.0Au 5nm UD 25K 0tilt Au 5nm UD 25K 0tilt 0.052 0.036 21.8 2.1Au 5nm UD 25K 30tilt Au 5nm UD 25K 30tilt 0.041 0.031 12.1 2.0AuPd 2nm LD 25K 30tilt AuPd 2nm LD 25K 30tilt 0.058 0.038 8.3 1.8AuPd 2nm MD 25K 30tilt AuPd 2nm MD 25K 30tilt 0.053 0.035 11.5 2.0AuPd 2nm UD 25K 30tilt AuPd 2nm UD 25K 30tilt 0.052 0.035 14.3 2.0Continued on next page222Table A.1 – Continued from previous pageSample ID CoatingmetalCoatingThicknessDetector Magni-ficationTiltangelD90(micron)Dav(micron)Porosity(%)dc/dtAuPd 5nm LD 25K 0tilt AuPd 5nm LD 25K 0tilt 0.056 0.038 15.9 2.0AuPd 5nm LD 25K 30tilt AuPd 5nm LD 25K 30tilt 0.057 0.038 17.0 1.9AuPd 5nm MD 25K 0tilt AuPd 5nm MD 25K 0tilt 0.054 0.036 19.6 2.2AuPd 5nm MD 25K 30tilt AuPd 5nm MD 25K 30tilt 0.052 0.036 17.4 2.1AuPd 5nm UD 25K 0tilt AuPd 5nm UD 25K 0tilt 0.046 0.033 16.4 2.1AuPd 5nm UD 25K 30tilt AuPd 5nm UD 25K 30tilt 0.046 0.033 16.2 2.0Ir 2nm LD 25K 0tilt Ir 2nm LD 25K 0tilt 0.064 0.041 15.7 1.9Ir 2nm LD 25K 30tilt Ir 2nm LD 25K 30tilt 0.062 0.040 15.3 1.8Ir 2nm MD 25K 0tilt Ir 2nm MD 25K 0tilt 0.051 0.035 14.8 2.1Ir 2nm MD 25K 30tilt Ir 2nm MD 25K 30tilt 0.045 0.033 13.1 2.0Ir 2nm UD 25K 0tilt Ir 2nm UD 25K 0tilt 0.049 0.034 12.5 2.1Ir 2nm UD 25K 30tilt Ir 2nm UD 25K 30tilt 0.047 0.033 11.4 1.9Continued on next page223Table A.1 – Continued from previous pageSample ID CoatingmetalCoatingThicknessDetector Magni-ficationTiltangelD90(micron)Dav(micron)Porosity(%)dc/dtIr 5nm LD 25K 0tilt Ir 5nm LD 25K 0tilt 0.065 0.042 13.7 2.1Ir 5nm LD 25K 30tilt Ir 5nm LD 25K 30tilt 0.057 0.038 14.7 1.9Ir 5nm MD 25K 0tilt Ir 5nm MD 25K 0tilt 0.057 0.037 19.9 2.2Ir 5nm MD 25K 30tilt Ir 5nm MD 25K 30tilt 0.049 0.034 15.8 2.0Ir 5nm UD 25K 0tilt Ir 5nm UD 25K 0tilt 0.052 0.035 11.2 1.9Ir 5nm UD 25K 30tilt Ir 5nm UD 25K 30tilt 0.049 0.034 18.0 2.0A.4 Typical results of the physical/chemical and fouling characteristics of themembranes224Table A.2: Typical results of the physical/chemical characteristics of the membranesSample ID NMR(totalHAarea)FTIRpeakContactAngle(5sec)stdevNMR Minimum Maximum FTIR Minimum Maximum (Degree)Virgin commercial membrane 2.828 2.435 3.220 0.157 0.149 0.164 51.0 4.8500,000 ppm·hr (44300 ppm NaOCl) 2.087 0.101 0.094 0.108 71.3 2.31,000,000 ppm·hr (44300 ppm NaOCl) 1.083 0.075 0.074 0.075 70.1 2.71,500,000 ppm·hr (44300 ppm NaOCl) 0.803 0.061 0.060 0.0612,000,000 ppm·hr (44300 ppm NaOCl) 0.670 0.639 0.704 0.033 0.031 0.035 71.1 4.0500,000 ppm·hr (22200 ppm NaOCl) 0.111 0.108 0.114 54.4 5.51,000,000 ppm·hr (22200 ppm NaOCl) 0.079 0.061 0.097 62.4 1.41,500,000 ppm·hr (22200 ppm NaOCl) 0.065 0.061 0.0692,000,000 ppm·hr (22200 ppm NaOCl) 0.032 0.031 0.033 67.6 3.4500,000 ppm·hr (3600 ppm NaOCl) 2.624 2.488 2.720 0.099 0.096 0.102 58.1 3.51,000,000 ppm·hr (3600 ppm NaOCl) 0.515 0.021 0.020 0.022 73.5 5.01,500,000 ppm·hr (3600 ppm NaOCl) 0.0172,000,000 ppm·hr (3600 ppm NaOCl) 0.189 0.139 0.240 0.014 0.014 0.015 92.6 4.9225Table A.3: Typical results of physical/chemical characteristics of the membranesSample ID TensileStrengthstdev Dav 95% CI D90 95% CI Porosity 95%CI(Mpa) (micron) (micron) (%)Virgin commercial membrane 2.24 0.13 0.026 0.001 0.034 0.001 8.3 1.0500,000 ppm·hr (44300 ppm NaOCl) 1.59 0.151,000,000 ppm·hr (44300 ppm NaOCl) 1.28 0.05 0.027 0.001 0.036 0.001 13.0 0.91,500,000 ppm·hr (44300 ppm NaOCl) 1.04 0.172,000,000 ppm·hr (44300 ppm NaOCl) 0.73 0.02500,000 ppm·hr (22200 ppm NaOCl)1,000,000 ppm·hr (22200 ppm NaOCl) 1.32 0.071,500,000 ppm·hr (22200 ppm NaOCl)2,000,000 ppm·hr (22200 ppm NaOCl) 1.02 0.06500,000 ppm·hr (3600 ppm NaOCl) 1.47 0.09 0.026 0.001 0.040 0.001 11.0 0.81,000,000 ppm·hr (3600 ppm NaOCl) 0.80 0.05 0.028 0.001 0.040 0.001 15.6 1.11,500,000 ppm·hr (3600 ppm NaOCl)2,000,000 ppm·hr (3600 ppm NaOCl) 0.66 0.17 0.032 0.000 0.045 0.001 17.1 1.2226Table A.4: Typical results of the fouling characteristics of the membranesSample ID CleanwaterResistanceLp(physical)Lp(chemical)Lp (irreversible) Lp(chemical)Lp (irreversible)(1/m) (%) (%)(6hrs)(%)(6hrs)(%)(24hrs)(%)(24hrs)Virgin commercial membrane 1571383 29.5 5.9 7.0 26.9 2.600500,000 ppm·hr (44300 ppm NaOCl) 1429959 27.6 2.9 -1.4 30.0 -2.4001,000,000 ppm·hr (44300 ppm NaOCl) 1367103 33.4 3.3 10.4 27.5 5.9001,500,000 ppm·hr (44300 ppm NaOCl)2,000,000 ppm·hr (44300 ppm NaOCl) 1319962 47.2 3.7 28.2 30.0 17.200500,000 ppm·hr (22200 ppm NaOCl) 1414245 30.1 2.6 5.1 25.5 4.6001,000,000 ppm·hr (22200 ppm NaOCl) 1367103 32.1 4.9 17.1 23.0 9.1001,500,000 ppm·hr (22200 ppm NaOCl)2,000,000 ppm·hr (22200 ppm NaOCl) 1319962 60.7 3.6 49.7 18.5 42.200500,000 ppm·hr (3600 ppm NaOCl) 1351389 36.1 2.5 17.1 21.0 15.1001,000,000 ppm·hr (3600 ppm NaOCl) 1288534 39.7 2.7 32.2 18.0 21.7001,500,000 ppm·hr (3600 ppm NaOCl)2,000,000 ppm·hr (3600 ppm NaOCl) 1209965 62.5 3.1 57.5 16.5 46.000227