Home

Smith, M.A. "Functionalized nanoparticles as

1. 57 6 1 Diagram of reverse osmosis bench scale filtration system 60 6 2 Comparison of DI water flux Uncoated SWC5 had the highest flux with polyDADMAC coated and Ag coated fluxes slightly lower 64 6 3 Fouling of virgin polyDADMAC coated and Ag co ated membranes Somewhat less fouling occurred on the coated membranes resulting in 10 higher flux 65 xii List of Figures Continued Figure Page 6 4 Flux recovery after acid cleaning PolyDADMAC coating gives slight improvement in flux recovery but Ag coating notably increases post cleaning flux 66 6 5 Comparison of DI water flux for Ag coated SWC5 membranes The third data set dropped suddenly by approximately 5 flux after 1 5 hr which cannot be explained with the data available 67 6 6 Comparison of SWC5 fouling experiments Fouling exp eriments were more varied than DI water flux Diffe rent rates of fouling and flux decline are observed 67 6 7 Full RO experiments PolyDADMAC and Ag coatings had lower initial flux and slower fouling Ag had notable flux improvement after cleaning
2. viii LIST OF FIGURES ix CHAPTER 1 BACKGROUND AND INRODUCTION 1 1 1 Reverse osmosis membranes 3 1 2 Fouling 3 1 3 Membrane matrix modifications 5 1 4 Permanent surface modification 6 1 5 Removable coatings 7 1 6 Membrane characterization 8 2 RESEARCH OBJECTIVES 13 3 MATERIALS AND MATERIAL CHARACTERIZATION 14 3 1 Membranes 14 3 2 Coating materials 16 3 3 Membrane characterization methods 18 3 4 Membrane characterization results 22 3 5
3. 69 7 1 Expected desorption mechanisms for positive materials at low pH a and negative particles at high pH b polyDADMAC and TiO2 were expected to detach from the membrane via mechanism a while negat ively charged NPs were expected to detach from the polyDADMAC layer via mechanism b 74 1 CHAPTER ONE BACKGROUND AND INTRODUCTION Drinking water quality is a major health concern throughout the world Dwindling supplies of potable water mandate the reuse of water and purification of groundwater and surface water prior to use Shannon et al 2008 As human populations continue to expand and concentrate in coastal areas many communities around the globe only have access to briny salty or polluted water Membrane filtration systems can remove a significant amount of undesirable biological and chemical species including salt typically have a smaller footprint than distillation systems and are more effective at contaminant removal than standard filtration These attributes make membrane filtration systems a conceptually appealing solution Four levels of membrane filtration are available for water purification purposes microfiltration MF removes microbes and particles typically larger than 0 1 m ultrafiltration UF removes macromolecules such as proteins polysaccharides larger molecular weight pesticides a
4. 51 6 1 Experimental matrix for bench scale RO experiments using SWC5 membranes 59 6 2 Average flux normalized to SWC5 DO water experime nt after each stage of RO filtration 68 ix LIST OF FIGURES Figure Page 3 1 Structures of m phenylenediamine a and trimesoly chlo ride b the precursors to PA TFC surface layers 15 3 2 Structure of cross linked polyamide 15 3 3 Monomer structure of diallyldimethyl ammonium chloride 16 3 4 Sketch representation of NP surface charges imparted by polyacrylate a and polyDADMAC b 18 3 5 General representation of adjustable gap cell setup 19 3 6 Zeta potential curves for SW30HR SWC4 and SWC5 TFC PA membranes Breaks in the titration curve resulted from the separation of acid and base titrati ions which provided separate sets of data 23 3 7 Titrations of three SW30HR samples Titrations have very similar shapes but vary slightly in zeta potential as seen in the June and February data The January titr ation was performed after the membrane had been stored at 4 C for over
5. degree I can t wait to be at YOUR graduation day soon v ACKNOWLEDGMENTS I would like to acknowledge Dr David Ladner for reaching out to me and offering this research project without which I would not have attended Clemson University I would also like to thank David for the time he dedicated to mentoring both in group settings and individually James Amburgey of the University of North Carolina at Charlotte originally supplied Clarifloc C 308P polyDADMAC polymer and Polydyne Inc of Riceboro GA donated additional material when supplies ran low James Eickhoff Thomas Luxbacher and Vinod Radhakrishnan provided guidance as we learned how to use the SurPASS electrokinetic analyzer and were exceptionally responsive to troubleshooting difficulties throughout the use of the instrument their help saved us invaluable time and is very much appreciated vi TABLE OF CONTENTS Page TITLE PAGE i ABSTRACT ii DEDICATION iv ACKNOWLEDGMENTS v LIST OF TABLES
6. 1 Run HCl titration File New Measurement Document from Template Ctrl T Select Adjustable Gap Cell stf Fill appropriate sample description material pore size coating etc Select appropriate user YOU Measurement Type pH Titration Under pH Titration change Used Syringe to Syringe Right HCl Return to Sample Settings and START titration Save file year month day sample description HCl 1 E g 2011 01 01 SWC4 virgin membrane HCl srf 2 Rinse instrument 3x with DDI water pH reading should return to 5 5 3 Rinse instrument 1x with fresh KCl solution 4 Run NaOH titration using same procedure as HCl except Do NOT change Use Syringe settings Change pH Maximum to 9 Change Desired pH difference to 0 25 Save file year month day sample description NaOH 1 E g 2011 01 01 SWC4 virgin membrane NaOH srf 5 Rinse instrument with DDI water at least 3x until pH stabilizes between 5 5 Instrument shut down 1 After pH has stabilized If instrument has been run for 2 titrations set max Time for Empty cycle to 86 500s If instrument has been run for a single titration set max Time to 300s Fill 600mL Schott beaker with DDI water Disconnect the outlet hose from beaker cover and put it in the DDI beaker Place the beaker cover with inlet hose onto an empty beaker Start the Empty process in the Monitor win
7. FUNCTIONALIZED NANOPARTICLES AS REMOVABLE COATINGS FOR REVERSE OSMOSIS MEMBRANES A Thesis Presented to the Graduate School of Clemson University In Partial Fulfillment of the Requirements for the Degree Master of Science Environmental Engineering and Earth Sciences by Megan Anne Smith August 2012 Accepted by Dr David Ladner Committee Chair Dr Cindy Lee Dr Tanju Karanfil All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent on the quality of the copy submitted In the unlikely event that the author did not send a complete manuscript and there are missing pages these will be noted Also if material had to be removed a note will indicate the deletion All rights reserved This edition of the work is protected against unauthorized copying under Title 17 United States Code ProQuest LLC 789 East Eisenhower Parkway P O Box 1346 Ann Arbor MI 48106 1346 UMI 1518323 Copyright 2012 by ProQuest LLC UMI Number 1518323 ii ABSTRACT Reverse osmosis RO desalination is increasingly used to produce potable water throughout the world Despite their promising abilities membrane filtration processes are limited by fouling Fouling is a broad term for organics inorganics colloids and organisms that interact physically chemically or biologically with the membrane surface resulting in reduced flux and shortened membrane
8. 43 15 3777 3786 Yang Y Zhang H Wang P Zheng Q Li J The in uence of nano sized TiO2 llers on the morphologies and properties of PSF UF membrane J Membr Sci 2007 288 1 2 231 238 Zhou M Liu H Kilduff J E Langer R Anderson D G Belfort G High throughput membrane surface modification to control NOM fouling Environ Sci Technol 2009 43 10 3865 3871
9. 74 APPENDICES 79 A Additional Images 80 B SurPASS Adjustable Gap Cell Standard Operating Procedures 84 C Bench Scale RO for Membrane Coatings Standard Operating Procedures 87 REFERENCES 90 viii LIST OF TABLES Table Page 3 1 Calibration concentrations for ICP OES analysis 22 3 2 ICP data for Ag standards and triplicate working solu tions diluted to 0 5 28 3 3 ICP data for Ti standards and triplicate working sol utions diluted to 50 29 5 1 Comparison of single measurement electrokinetic data for coating materials Change in zeta pote ntial for acid and base cleaning is reported as the difference between the coated membrane and af ter cleaning
10. If applied in a membrane system the polyDADMAC layer would be replenished anyway Figure 5 1 Acid cleaning of polyDADMAC coated SW30HR resulted in significant but incomplete removal of the polymer PolyDADMAC was also subjected to a NaOH cleaning solution of pH 13 for 8 hours Though the cleaning conditions exceeded the manufacturer recommendations for exposure time to high pH some insights were gained from the single measurement data collected Table 5 1 While the polyDADMAC coated membrane had a zeta potential of 14 mV and the negative zeta potential of an acid cleaned sample indicated some removal of the polyDADMAC layer the base cleaned sample retained a positive zeta potential of 8 mV It is unclear from the experiments performed whether the decrease in zeta potential was caused by damage to the membrane by the caustic solution over an 51 excessive duration or if some of the polyDADMAC coating was removed If the membrane was not damaged by caustic cleaning polyDADMAC would be expected to adsorb more strongly to the membrane as the percentage of negative binding sites increased The data may support this hypothesis if a polyDADMAC chain adsorbed to the SW30HR surface at several points and the rest remained dangling from the surface a high zeta potential would result If the membrane became more negatively charged at high pH and the polymer chain became more attracted to the surface the zeta potenti
11. NPs more Ag was used in experiments than Ti The observed variation in zeta potential imparted by the same coating solution in different experiments could be attributed to charge density heterogeneity of the polyDADMAC layer heterogeneity of the virgin membrane aggregation or charge density heterogeneity of the NP coatings or a combination Uneven polyDADMAC or 42 NP coatings are more likely than significant membrane heterogeneity TiO2 NPs may have aggregated as the stock solution aged causing large clusters of NPs to adsorb to the membrane surface particle size analysis of the stock solution would allow this possibility to be confirmed or eliminated TiO2 was not visible in SEM images at the time the Ag aggregates were discovered so further SEM imaging was not pursued for TiO2 coatings Figure 4 9 TiO2 coatings on polyDADMAC membranes from two experiments Deviation in shape and zeta potential were observed 4 3 Membrane coating methods Coatings were self assembled onto the membranes in 10 mL polypropylene petri dishes which allowed membrane coupons to float easily in coating solution Two membrane coupons of approximately 1 5 cm x 2 5 cm were coated in each dish One 43 membrane was placed active side up and one active side down to prevent the coupons from sticking together and interfering with coatings Despite different orientation both membranes were completely submerged durin
12. SIX BENCH SCALE REVERSE OSMOSIS FOULING EXPERIMENTS Bench scale reverse osmosis experiments were undertaken to begin testing the viability of the coating materials Silver NPs were selected for bench scale RO experiments Membranes coated for RO experiments were coated on the bench top as done previously These experiments provide preliminary results for future work investigating fouling on RO membranes with removable coatings 6 1 Experimental design Three cycles of three experiments outlined in Table 6 1 were performed in the bench scale tests The first cycle of experiments was performed with uncoated SWC5 membrane the second cycle with polyDADMAC coated SWC5 and the third cycle with polyDADMAC and Ag coated SWC5 Three membranes were prepared for each cycle of experiments with the first membrane a three hour clean water flux test was performed the second membrane was fouled for three hours with sodium alginate and the third membrane was fouled for two hours with sodium alginate and cleaned for 30 minutes with a pH 3 HCl wash After all experiments were completed a single experiment of the complete cycle was performed for each set with a single membrane being used for clean water flux fouling cleaning and post cleaning flux recovery 59 Table 6 1 Experimental matrix for bench scale RO experiments using SWC5 membranes Experiment Coating Coating Foulant Cleaning 1 1 none none no
13. al 2009 Successful development of membrane materials followed by further modifications to increase membrane performance played a crucial role in the global application of this technology Many commercial RO membranes are thin film composite TFC polyamide PA membranes which consist of three layers a polyester supportive backing layer 120 150 m thickness a microporous polysulfone interlayer 40 m and an extremely thin polyamide active layer on the membrane surface 0 2 m Petersen amp Cadotte 1990 Commercial materials are developed for effectiveness despite heterogeneity caused by the manufacturing process and active layers can vary chemically across the membrane as well as throughout the depth of the surface layer Coronell et al 2011 1 2 Fouling While RO is low cost compared to other desalination techniques the process is hindered by fouling Goosen et al 2004 Seventy percent of seawater RO plants in the Middle East have biofouling issues and 83 of plants surveyed in the United States reported fouling problems as well Yang et al 2009 A wide range of materials can foul 4 membranes including organics e g natural organic matter NOM and humic acid inorganics e g scale forming salts like barium sulphate van de Lisdonk et al 2000 biopolymers e g proteins and polysaccharides making up extracellular polymeric substances EPS and biological organisms themselves
14. also formed in the laboratory indicating that inclusion of nanomaterials in the membrane active layer may be a promising avenue The membrane surface can also be modified by dipping the membrane in a concentrated NP solution These studies have been done with carefully controlled preparations of TiO2 NPs that were pre analyzed for particle size Bae et al 2006 Both 6 matrix incorporation and dip coating methods seem promising for application to many membrane types as well as different types of NPs Some concerns have been noted such as the aggregation of NPs when the concentration is too high resulting in a reduction of permeability and flux Soroko and Livingston 2009 Luo et al 2005 Yang et al 2007 Another concern noted by Taurozzi et al 2008 regarded significant silver NP loss from the matrix during initial wetting of the membrane There have also been encouraging results from these investigations It has been reported that addition of nano sized alumina did not affect the structure of the membrane or its pores Yan et al 2006 The mechanical properties of the membrane actually improved and later studies explored the optimal concentrations for enhancing membrane strength and fouling resistance 1 4 Permanent surface modification Anti fouling research for NF and RO membranes is currently progressing toward permanent modification by covalent bonding of coatings to the membrane surface Polymer s
15. change resulting from membrane modification or coating The sessile drop method is most commonly used for membrane analysis This method involves placement of a liquid droplet onto the membrane surface and photographic visualization with a contact angle goinometer Hurwitz et al 2010 Software is used to determine the contact angle of the drop on the membrane Titrations can be performed by 12 varying the pH of droplets allowing characterization of membrane surface chemistry throughout a range of pH values The type of liquid used in the droplet can also be varied and a wettability parameter can be identified based on the affinity of the membrane for particular liquids Brant and Childress 2004 A second technique that can be applied to membrane surface characterization is the captive bubble method in which the surface is immersed face down in a liquid and a gas bubble is released onto the membrane surface from below Drelich et al 1996 In both cases the shape of the drop bubble is metastable and can be affected by outside pressures evaporation and other factors There is a phenomenon called hysteresis well described by Gao and McCarthy 2006 In basic terms an increasing droplet will cover a defined area with changing contact angle until reaching a certain volume at which point it will increase in area with a constant contact angle a decreasing droplet will decrease in volume with a changing con
16. membrane surface in the absence of the polyDADMAC binding layer so either mechanism is possible 54 Figure 5 3 SEM images of uncoated SW30HR a Ag coated SW30HR b pH 1 acid cleaned Ag coated SW30HR c and pH 13 base cleaned Ag coated SW30HR d Electrokinetic data were collected for titrations of acid cleaned Ag coated membranes Figure 5 4 An increase in zeta potential at the low end of the titration indicated that polyDADMAC remained on the membrane The low zeta potential values at high pH drop below the SW30HR membrane suggesting that NPs may remain adsorbed to the membrane and contribute to an increase in negative character at high pH 55 Figure 5 4 Electrokinetic behavior of Ag coatings bound to a polyDADMAC layer after acid cleaning 5 2 4 TiO2 coating removal TiO2 coatings behaved quite similarly to Ag coatings for acid cleaning experiments No NP aggregates were visible in SEM images because a low concentration 3 7 mg Ti L was used throughout the course of coating and removal experiments but electrokinetic data were very similar to Ag coatings so similar conclusions regarding NP adsorption and detachment may be drawn Single measurements for TiO2 coating acid and base titrations Table 5 2 confirm the similar behavior to Ag The TiO2 coating imparted less charge decline because the concentration was much lower than that of Ag 3 7 mg Ti L
17. module that arrived pre tested and vacuum sealed in water so it may be expected that the surface charge of SWC5 will not change as significantly as a function of water soaking time Figure 3 7 Titrations of three SW30HR samples Titrations have very similar shapes but vary slightly in zeta potential as seen with the June and February data The January titration was performed after the membrane had been stored at 4 C for over 3 months When acid cleaning experiments were performed using SW30HR membrane the zeta potential curve of SW30HR decreased significantly after washing in pH 1 solution 25 for 30 minutes Figure 3 8 A significant decrease in zeta potential indicates that the surface chemistry was modified by the very low pH cleaning of the membrane One possible change is that the polyvinyl alcohol layer may have been removed to some extent This phenomenon was not observed with cleaning solutions of pH 3 or higher as Figure 3 8 shows The pH 1 cleaning appears to damage the coating layer of the SW30HR membrane and is compared to the uncoated PA of SWC5 in Figure 3 9 to determine whether the data provide a rough approximation of PVA removal Similar experiments were outlined for SWC4 and or SWC5 membrane but have not been performed yet Acid cleaning of SWC4 and SWC5 could confirm the hypothesis that PVA is somewhat removed by low pH cleaning or could indicate another phenomenon entirely Figure 3
18. the 10nm nanoparticles Based on SEM images the higher charge density on the 100 mg L Ag surface was attributed to aggregates rather than more NPs providing better coverage of the membrane surface A breakthrough aggregation concentration likely occurs between 50 100 mg L there is little difference between surface charge of the 20 mg L and 50 mg L coatings indicating the coatings may be similar in NP density while aggregates form on membranes coated with 100 mg L Ag NPs The concentration used for coatings was decreased to 20 mg Ag L to decrease likelihood of aggregation and obtain data that could be compared to TiO2 coatings 41 Figure 4 8 SEM image of Ag aggregation on SW30HR membrane Aggregates are very large up to 6 m in size Nanoparticles are 10nm 4 2 4 TiO2 NP coatings TiO2 NPs given negative charge by polyacrylate adsorbed to the polyDADMAC coated membranes Adsorption was indicated by a decrease in the zeta potential of the polyDADMAC coating Figure 4 9 TiO2 coating resulted in a similar degree of zeta potential change as the Ag coating however the variation between individual TiO2 coatings and individual Ag coatings prevented the assumption that the different NPs provided equal coverage of the surface Ag experiments used 20 mg Ag L while TiO2 experiments used 4 2 mg Ti L assuming that the metal component per NP is comparable between Ti and Ag
19. was used in recycle mode for the clean water flux time After one hour any water remaining in the feed tank was wasted until the water level was just above the pump at which point the pump was stopped Sodium alginate feed solution was prepared in a 4 L flask and added to the feed tank The feed hose was filled with water and the system set back to recycle mode The program was started and fouling occurred for three hours from the time the system reached the target pressure of 1000 psi The membrane was removed from the sample holder the sample cell was bypassed and the system was cleaned with 1 L pH 12 NaOH for five minutes followed by 1 L pH 3 HCl for 10 minutes At least 20 L DI water was flushed through the system after cleaning Membrane cleaning experiments represented in Table 6 1 by experiments X 3 consisted of one hour DI flux and two hours sodium alginate fouling as described for X 2 experiments After 2 hr of fouling any remaining solution in the feed tank was wasted until the water level was just above the pump Four liters of pH 3 0 001 M HCl solution was immediately added to the system and recycled through for 30 min The system was flushed with DI water for at least 5 minutes after acid cleaning to remove residual acid foulant and coating material from the system then the membrane was removed and the system cleaned as described for the fouling experiments 62 Finally the entire cycle of experim
20. 00 psi 7 Hz until pulsation dampens Be sure that bubbles have left the DI before running o Click Actuator valve auto control to ON o Set upper limit to 5 psi lower limit to 15 psi o Run at high pressure until system stabilizes check for leaks at joints Run membrane baseline flux rejection before each fouling experiment This will require attaching SEPA cell establishing flux rejection wasting DI water and then adding foulant solution Run the DI out until water level is just above pump don t let air get into the pump Add the fouling solution of 200 mL of 2 g L sodium alginate in 4 liters DI water Run system to mix foulant with water already present Set up membrane cell o Place membrane coupon in cell WITHOUT spacer o Active side toward O rings o Use permeate carrier from Hydranautics brackish water module o Pressurize cell to 1200 psi using hydraulic pump o Hook up to the system 88 o Membrane type and coatings To start computer data collection o Run system at 7 Hz pump speed 6 4 Hz on controller o Set actuator valve to completely open 10 Hz o Start computer data collection arrow button on top toolbar at left Write down file name and start time Turn ON actuator valve auto control Make sure upper limit is set to 5 psi and lower limit to 15 psi Run at constant 1000 psi pressure and manually input Temperature reading at least every 30 minutes If
21. 10 10 dI dp o L A 1 Here dI dp is the measured slope of streaming current versus pressure is the electrolyte viscosity is the dielectric constant of the electrolyte 0 is the vacuum permittivity L is the length of the streaming channel and A is the cross sectional area of the streaming channel 1 6 2 Scanning electron microscopy and energy dispersive X ray spectroscopy Scanning electron microscopy SEM is used to visualize the membrane surface and determine whether coatings or fouling have caused considerable changes to the surface structure Cahill et al 2008 Polymeric materials are not electrically conductive and do not contribute to the backscattering and secondary electrons that scanning electron microscopy relies upon to produce an image so membranes must be sputter coated with gold platinum or another heavy element Michler 2008 Energy dispersive X ray spectroscopy EDS is an analytical technique commonly coupled with scanning or tunneling electron microscopes to determine the elemental composition of solid samples In membrane systems EDS can be used to confirm the presence of inorganic contaminants or uncharacteristic C N O ratios For this research EDS is intended to detect the metallic components of the nanoparticles indicating the presence of the nanoparticle layer The requirement for sputter coating and the small quantities of coating mater
22. 2 762 770 Louie J S Pinnau I Reinhard M Effects of surface coating process conditions on the water permeation and salt rejection properties of composite polyamide reverse osmosis membranes J Membr Sci 2011 367 1 2 249 255 Luo M Zhao J Tang W Pu C Hydrophilic modification of poly ether sulfone ultrafiltration membrane surface by self assembly of TiO2 nanoparticles Appl Surf Sci 2005 249 1 4 76 84 Luxbacher T Electrokinetic characterization of flat sheet membranes by streaming current measurement Desalination 2006 199 376 377 Mi B Elimelech M Organic fouling of forward osmosis membranes Fouling reversibility and cleaning without chemical reagents J Membr Sci 2009 348 337 345 Michler G H Scanning electron microscopy In Electron microscopy of polymers Pasch H Ed Springer Berlin 2008 pp 87 120 Minghao G Kilduff J E Belfort G High throughput atmospheric pressure plasma induced graft polymerization for identifying protein resistant surfaces Biomaterials 2012 33 5 1261 1270 Park S H Wei S Mizaiko B Taylor A E Favero C Huang C H Degradation of amine based water treatment polymers during cloramination as n nitrosodimethylamine NDMA precursors Environ Sci Technol 2009 43 5 1360 1366 Petersen R J Cadotte J E Thin film composite reverse osmosis membrane In Handbook of Industrial Membrane Technology Porter M C Ed Noyes Publicat
23. 3 months 24 3 8 SW30HR samples in HCl solutions of varying strength for 30 min Strong acid significantly altered the zeta potential while weaker acid solutions did not 25 3 9 Zeta potential drop of pH 1 acid washed SW30HR com pared to virgin SW30HR and SWC5 membranes 26 4 1 PolyDADMAC coating on SWC4 membrane The poly mer coating significantly altered the surface chemist ry creating a positively charged membrane surface 33 4 2 PolyDADMAC coatings of 2 1 and 0 2 on dry SW30HR membranes All coatings impart a strong charge there is no notable difference between 0 2 and 1 and little difference at 2 34 x List of Figures Continued Figure Page 4 3 Effects of 2 1 mg L 4 2 mg L 10 6 mg L and 21 2 mg L TiO2 on membrane surface charge compared to the uncoated SW30HR membrane 35 4 4 TiO2 NP coatings for two sets of experiments Signifi cant differences could be attributed to changes in NP stock solution or heterogeneity of the membrane sur ace 36 4 5 Charge density of polyDADMAC coating compared to Ti
24. 5 fouling experiments are more varied than DI water experiments Figure 6 6 Experiments differ in rate of fouling and final flux after fouling 67 Figure 6 5 Comparison of DI water flux for Ag coated SWC5 membranes The third data set suddenly dropped by approximately 5 flux after 1 5hr which cannot be explained with the data available Figure 6 6 Comparison of SWC5 fouling experiments Fouling experiments were more varied than DI water flux Different rates of fouling and flux declines are observed 68 The complete cycle of experiments consisted of DI water flux for 1 3 hours sodium alginate fouling until 4 1 hours pH 3 HCl cleaning for 30 minutes and DI water flux recovery after cleaning Cleaning was performed at low pressure and flux data were not collected so flux recovery data were modified by approximately 30 minutes for continuity of Figure 6 7 In the full cycle experiment a clean water flux was run through a SWC5 membrane for six hours to evaluate membrane compaction effects Improved flux recovery was expected for the coated membranes when compared to the uncoated membrane Both coated membranes had slightly higher DI water flux than the uncoated SWC5 Fouling occurred more slowly on coated membranes than for SWC5 with 10 higher flux through coated membranes Flux recovery for the Ag coated membrane showed notable improvement over both polyDADMAC coated and uncoated SWC5 achie
25. 8 SW30HR samples in HCl solutions of varying strength for 30 min Strong acid significantly altered the zeta potential while weaker acid solutions did not 26 Figure 3 9 Zeta potential drop of pH 1 acid washed SW30HR compared to virgin SW30HR and SWC5 membranes 3 5 Polymer and nanoparticle characterization PolyDADMAC and both TiO2 NP stock solutions were prepared at the beginning of the project and lasted throughout the duration of the project Silver NP stock solution was prepared twice over the course of the project The first solution was not characterized which exacerbated erroneous calculations of NP concentrations and misleading data in the preliminary results Concentrations of NP solutions were not measured until the culmination of the project an oversight that left many questions about multiple sets of data Future work with this project should include characterization of stock solutions immediately after preparation and periodically thereafter to determine the extent to which nanoparticles degrade or aggregate over time 27 3 5 1 PolyDADMAC PolyDADMAC concentration of the hundredfold dilution was assumed to be 0 2 polymer in the aqueous working solution This assumption was based on the hydrophilic properties of the polymer Stock solution was measured into a graduated cylinder that was then rinsed several times into a volumetric flask to ensure complete transfer of the polymer during dilu
26. C polymer was not performed because a suitable method was not determined during the course of this study Total organic carbon or total nitrogen analysis of the coating solution and reverse osmosis permeate may be effective methods for detecting polyDADMAC in solution comparison to the 20 stock solution provided by the manufacturer should allow for quantification of the samples Several methods were available for quantification of NPs but only one was used Quantification of Ti NPs is particularly important because solutions were prepared from powdered NPs that were observed sticking to glassware after preparation The inorganic component of each NP was targeted during analysis in this study ICP OES determined the concentration of Ag or Ti present in stock solution samples acidified with 5 HNO3 Low Ti NP concentrations indicated that NPs were incompletely dissolved in simple acid 75 solution and digestion should be performed for further analysis Digestion was not undertaken for the solutions discussed in this study but there is confidence in the measurements of Ag NPs which was as expected A microwave digestion instrument is available for use and acid digestion SOPs were obtained comparison between the two digestion methods would be interesting ICP analysis of Ag NPs produced the expected result so further analysis is probably only necessary to produce a calibration curve at higher concentrations Ze
27. NPs on membrane surface TiO2 NPs were electrostatically adsorbed directly to the membrane surface SWC4 was initially used for TiO2 concentration experiments because the surface carries a stronger negative charge Carboxyl groups of the SWC4 surface are not shielded by polyvinyl alcohol coating as with the SW30HR membrane resulting in a higher charge density TiO2 coatings were prepared by diluting 10 L and 20 L of TiO2 stock solution into 10 mL resulting in concentrations of 0 212 mg Ti L and 0 423 mg Ti L respectively Dry membrane coupons of SWC4 were submerged in the coating solution for 24 hours at 4 C These concentrations did not alter surface charge so for SW30HR experiments higher concentrations of 21 2 mg L 10 6 mg L 4 2 mg L and 2 1 mg L were used Single measurements were taken to quickly determine the zeta potential changes imparted by varying the NP concentrations 32 4 1 3 Ag NPs on polyDADMAC coated membranes Ag NPs were adsorbed onto a 0 2 polyDADMAC binding layer Initial experiments used concentrations of 100 mg Ag L 50 mg L and 10 mg L PolyDADMAC coated membrane coupons were placed into Ag coating solutions after at least 12 hours in DDI water After 24 hours Ag coated membranes were rinsed for 30 seconds and soaked in DDI water at least overnight before analysis Soaking allowed detachment of any NPs not adsorbed to the membrane 4 1 4 TiO2 NPs on pol
28. O2 NP coating which was functionalized with polyDADMAC 37 4 6 Reduction in surface charge by Ag NPs adsorbed to polyDADMAC As expected the higher the NP con centration the lower the zeta potential Error bars re present a standard deviation from 12 measurements 38 4 7 20 mg L Ag coatings compared to virgin SW30HR 40 4 8 SEM image of Ag aggregation on SW30HR membrane Aggregates are very large up to 6 m in size NPs are 10 nm 41 4 9 TiO2 coatings on polyDADMAC membranes from two experiments Significant deviation in shape and zeta potential were observed 42 4 10 Titrations for polyDADMAC coated SW30HR show th at a dry membrane coated statically at low temperat ure had a more positive charge than a wet membrane coated dynamically at room temperature 44 4 11 Control coatings on SW30HR TiO2 did not adsorb to polyDADMAC TiO2 and Ag appear to ha ve adsorbed to SW30HR despite carboxyl groups promoting electrostatic repulsion 45 xi List of Figure
29. Polymer and nanoparticle characterization 26 4 COATING CONDITION OPTIMIZATION 30 4 1 Determining coating concentrations 30 4 2 Coating concentration results 32 4 3 Membrane coating methods 42 4 4 Coating method evaluation 43 vii Table of Contents Continued 5 REMOVAL OF COATINGS 48 5 1 Coating removal process 48 5 2 Coating removal results 49 6 BENCH SCALE RO FOULING EXPERIMENTS 58 6 1 Experimental design 58 6 2 RO methods and additional materials 59 6 3 Bench scale RO results 62 7 CONCLUSIONS AND RECOMMENDATIONS 70 7 1 Conclusions relating to research objectives 70 7 2 Future work recommendations
30. R SEVEN CONCLUSIONS AND RECOMMENDATIONS Considerable progress has been made toward the goal of determining whether the materials used in this study can be electrostatically applied and removed Coatings demonstrated success in delaying fouling of the membranes and cleaning experiments in the bench scale RO system indicated that coating removal may improve flux recovery after fouling This chapter aims to summarize the progress toward the main research goal and suggest future experiments and analyses 7 1 Conclusions relating to research objectives 7 1 1 Determine appropriate conditions for self assembly of polymer and NP coatings on RO membranes The polymer and NPs selected for this study were novel materials for RO membrane coatings PolyDADMAC is currently used in the water treatment industry as a flocculant and would be simple to implement in RO systems if effective PolyDADMAC had also been previously demonstrated to cause few negative changes to the membrane when explored in an operational context Gabelich et al 2005 NDMA formation is not a concern in drinking water treatment operations because oxidative disinfectants are not used they are detrimental to polyamide membranes However polyDADMAC has been demonstrated to decrease antiscaling membrane properties when used as a coagulant in pretreatment trains prior to RO filtration Kim et al 2009 Diluted solutions of polyDADMAC 0 2 w v effectively c
31. Standard Operating Procedures Night before Preparation 1 Cut membrane coupons 1 x0 5 and soak in DDI water overnight 2 Prepare 1 mM KCl solution Pipet 10 mL 0 1 M KCl solution into 1000 mL volumetric flask Fill to mark with DDI water Vacuum filter KCl solution o Rinse Whatman funnel with DI water o Position a 1 m 47mm membrane and assemble unit o Rinse membrane by filtering 50 100mL DDI water discard o Filter KCl solution o Clean up rinse used glassware 3x with DI water and place on drying rack Adjustable Gap Cell Assembly 1 Adhere double sided tape to blue sample holder Press out bubbles 2 Peel backing off tape and adhere backing of membrane to tape Pull overhanging edges firmly with tweezers press with Parafilm if needed 3 Trim membrane to exact size of sample holder Exercise extreme caution to not damage the sample holder with scissors 4 Insert sample holders into AGC 5 Adjust gap so both sides appear equal a very thin slit of light should be visible from each side 6 Allow knurled nuts to freely spin onto sample holder stems until they stop at the AGC Place holders and fold down arms to secure 7 Tighten knob on top of AGC until finger tight 8 Double check gap on both sides to ensure gaps are equal If gaps are not equal disassemble to step 5 and readjust before proceeding 9 Insert the leads into the perfusion holes 10 Adjust the AGC ho
32. The Adjustable Gap Cell template file in VisioLab software was used for all AGC measurements in pH titration mode modified for a maximum pH of 9 for base titrations During initial experiments on the AGC titrations were run from acidic pH 3 to basic pH 10 11 or vice versa After being advised against single titrations because of the ionic strength of the feed solution titrations were split into two halves an acid titration from ambient DDI pH 5 5 to 3 and a base titration from ambient pH to 9 Acid titrations were executed first for all samples When the acid titration finished the instrument was rinsed with three changes of DDI water to eliminate electrolyte in the system KCl solution was flushed through before the base titration was started Upon completion of the base titration the instrument was rinsed with DDI water until the baseline pH 5 5 was restored The Empty command was used to clear the tubing and conclude the experiment Used membranes were discarded 3 3 1 4 Single measurements Single measurements were performed to assess instrument variability and occasionally to perform rapid analysis of surface potential at a predetermined pH 21 Samples were prepared as for titration measurements In VisioLab software the single measurement mode was selected and 12 measurements were taken to decrease instrument noise 3 3 1 5 Instrument maintenance Routine instrument mai
33. al measurement would decrease as the membrane neutralized a higher percentage of the polymer chain Table 5 1 Comparison of single measurement electrokinetic data for coating materials Change in zeta potential for acid and base cleaning is reported as the difference between the coated membrane and after cleaning Sample polyDADMAC ZP mV Ag ZP mV TiO2 ZP mV Uncoated SW30HR 12 8 12 8 12 8 pD or pD NP coating 14 0 2 3 5 5 pH 1 cleaning 0 92 11 8 6 9 Change in ZP acid 14 92 9 5 12 4 pH 13 cleaning 7 8 1 6 9 3 Change in ZP base 6 2 3 9 3 8 5 2 2 TiO2 coating removal TiO2 NPs were effectively removed by each acid concentration considered Figure 5 2 Slight differences between titrations were assumed to be instrument variability rather than significant results since the pH 1 cleaning solution appears to have the highest zeta potential when compared to pH 2 and 3 cleaning solutions When the titration curves of the acid cleaned membrane were compared to the virgin membrane all of the NP coating appeared to be removed by acid solutions as weak as pH 3 NPs should 52 be roughly spherical in shape if this was true TiO2 NPs adsorbed to the membrane surface via a single contact point and were simpler to repel or release from the membrane than was the polyDADMAC polymer Base cleaning was not investigated for TiO2 NPs based on the results of polyDADMAC co
34. alley structure PolyDADMAC coating was observed filling in the loop and valley structure in several SEM images Figure 4 12 Figure C 5 which could prevent NPs from becoming trapped Additionally the only goal of removing the coating materials is to remove foulants if some NPs remain on the membrane after cleaning but foulants are removed the end goal has been accomplished Coatings that adsorb within the loop and valley structure also occupy active sites where undesirable materials may adsorb or accumulate such as bacteria and mineral deposits b a 74 7 1 3 Use zeta potential titrations to detect the presence of self assembled coatings Zeta potential measurements were used throughout the study to detect the presence of membrane coatings Zeta potential measurement was the only way to detect the coatings in most cases though polyDADMAC coating was not visible to ATR FTIR or low to moderate magnification SEM the adsorbed material caused significant changes to the zeta potential of the membrane confirming its presence The same was true for both TiO2 NPs and lower concentrations of Ag NPs which were not visible on SEM Comparison of zeta potential single measurements with SEM images of coated and cleaned Ag NP aggregates allowed more complex conclusions to be considered 7 2 Future work recommendations 7 2 1 Materials and material characterization Quantification of the polyDADMA
35. and physical hydrophilicity flux and roughness properties are observed and compared to virgin membrane to determine beneficial or disadvantageous effects Many techniques are available for RO membrane characterization and their use depends on the properties being investigated For chemical analysis zeta potential indicates the overall surface charge and isoelectric point attenuated total reflectance Fourier transform infrared ATR FTIR spectroscopy determines functional groups associated with the membrane surface and foulants and energy dispersive X ray spectroscopy EDS determines elemental composition For physical analysis contact angle measurement indicates hydrophilicity of the membrane material bench scale filtration determines baseline flux and flux effects of modification atomic force microscopy AFM visualizes surface roughness and scanning electron microscopy SEM visualizes the membrane surface and any coating or foulant aggregation 1 6 1 Zeta potential Electrokinetic characteristics of reverse osmosis membranes have a significant influence on fouling performance and contaminant retention Polyamide membranes 9 which have a slightly negative surface charge at neutral pH can electrostatically repel negatively charged functional groups common to natural organic matter In the presence of positively charged functional groups however the membrane charge is a hindrance and encourages electr
36. at pH 5 6 from 26 mV to a range of 1 5 5 mV Figure 4 6 As expected the higher concentrations of NPs resulted in a larger charge difference with the 10 mg Ag L coating maintaining a slightly positive surface charge and the 20 mg L 50 mg L and 100 mg L coatings imparting neutral to very slightly negative surface charges Similarly to the 38 TiO2 data the standard deviation of the zeta potential in Figure 4 5 exceeded the differences between most single measurement values for Ag concentrations The measurement with the highest standard deviation 20 mg L Ag solution was a data point from a titration curve which was averaged from four measurements while the other values were single measurement data that were averaged from 12 measurements This indicates that the instrument is more precise when more measurements are used to report each data point When 10 mg L 50 mg L and 100 mg L were compared 100 mg L coating solution was selected for future experiments because of its large zeta potential change After NP aggregates were discovered however the concentration was reduced to 20 mg L Figure 4 6 Reduction in surface charge by Ag NPs adsorbed to polyDADMAC As expected the higher the NP concentration the lower the zeta potential 39 Titrations of Ag NP coatings showed less deviation between individual titration than deviation between TiO2 coated membranes Membranes prepa
37. ated membrane cleaning it was not anticipated to be an effective means of removing the positively charged materials Figure 5 2 Acid cleaning of TiO2 coatings Acid cleaning resulted in complete removal of NPs for all acid concentrations investigated 5 2 3 Ag coating removal Aggregation of Ag NPs discovered by SEM imaging became an asset for coating removal data Along with coatings acid cleaning with pH 1 and base cleaning with pH 13 were observed using the TM3000 SEM Figure 5 3 Aggregates were observed on the coated membrane as discussed previously Fewer aggregates were observed on the acid cleaned membrane but some smaller aggregates remained No 53 aggregates were located on the base cleaned membrane When considered along with electrokinetic data for samples from the same membrane coupons Table 5 1 several conclusions can be drawn The absence of NPs in the SEM image and increase in zeta potential indicate removal of the NPs and endurance of the polyDADMAC coating Acid cleaning resulted in a significant drop in zeta potential indicating removal of polyDADMAC coating but Ag aggregates were still visible in SEM images Two possibilities exist either Ag NPs adsorbed to the membrane surface via non electrostatic interactions or aggregates remained adsorbed to the polyDADMAC polymer not removed from the membrane by cleaning Control experiments indicated that NPs had adsorbed to the
38. ave played a role in polyDADMAC adsorption after acid cleaning If binding was truly electrostatic polyDADMAC should have been repelled from the membrane surface as high proton concentrations crowded out the negative charge density on the membrane surface Figure 7 1a Negatively charged Ag and Ti NPs were removed by acid cleaning because they were bound to the polyDADMAC layer when the polyDADMAC was repelled from the membrane the NPs were removed as well The few base cleaning experiments performed indicated some removal of the negative NPs but endurance of the polyDADMAC coating This was an expected result for electrostatic interactions Figure 7 1b SEM 73 imaging of Ag aggregates indicated that NPs remained on the surface after strong acid cleaning but were removed by strong base cleaning Base cleaning was not pursued because it was not as effective in removing the polyDADMAC as acid cleaning experiments were focused on removal of all coatings so NPs were removed via removal of polyDADMAC Figure 7 1 Expected desorption mechanisms for positive materials at low pH a and negative particles at high pH b PolyDADMAC and TiO2 were expected to detach from the membrane via mechanism a while negatively charged NPs were expected to detach from the polyDADMAC layer via mechanism b There remains a question of whether NPs can be removed from the membrane if they are entrapped in the loop and v
39. benefit In RO experiments salt water should be used to determine feed water chemistry effects on coatings Coatings can be explored with other types of membranes as well to be applied to other systems though adsorption within pores will need to be considered in the case of MF and UF membranes Luxbacher 2006 79 APPENDICES 80 Appendix A Additional Images Figure A 1 SEM image of SW30HR membrane coated with 0 2 polyDADMAC and 3 7 mg L TiO2 81 Figure A 2 SEM image of SW30HR membrane coated with 0 2 polyDADMAC and 20 mg L Ag Figure A 3 SEM image of uncoated SW30HR membrane 82 Figure A 4 SEM image of SW30HR coated with 0 2 polyDADMAC The clump of material in the upper right quadrant of the image is an artifact of membrane formation Figure A 5 SEM image of SW30HR coated with 0 2 polyDADMAC This section of the membrane shows the filled in loop and valley structure 83 Figure A 6 Actual flux values for uncoated polyDADMAC coated and Ag coated SWC5 membranes While the fluxes of the coated membranes were lower when fouling began fouling occurred less dramatically overall for both coated membranes Figure A 7 Actual flux values for uncoated polyDADMAC coated and Ag coated SWC5 membranes after acid cleaning Cleaning of the Ag membrane resulted in a notable improvement 84 Appendix B SurPASS Adjustable Gap Cell
40. c reverse osmosis and nanofiltration membranes J Membr Sci 1996 119 2 253 268 Childress A E Elimelech M Relating nanofiltration membrane performance to membrane charge electrokinetic characteristics Environ Sci Technol 2000 34 17 3710 3716 Cho J Amy G Pellegrino J Yoon Y Characterization of clean and natural organic matter NOM fouled NF and UF membranes and foulants characterization Desalination 1998 118 1 3 101 108 91 Coronell O Gonzalez M I Marinas B J Cahill D G Ionization behavior stoichiometry of association and accessibility of functional groups in the active layer of reverse osmosis and nanofiltration membranes Environ Sci Technol 2010 44 17 6808 6814 Coronell O Marinas B J Cahill D G Depth heterogeneity of fully aromatic polyamide active layers in reverse osmosis and nanofiltration membranes Environ Sci Technol 2011 45 10 4513 4520 Drelich J Miller J D Good R J The effect of drop bubble size on advancing and receding contact angles for heterogeneous and rough solid surfaces as observed with sessile drop and captive bubble techniques J Colloid Interf Sci 1996 179 1 37 50 Freger V Nanoscale heterogeneity of polyamide membranes formed by interfacial polymerization Langmuir 2003 19 4791 4797 Gabelich C J Ishida K P Bold R M Testing of water treatment copolymers for compatibility with polyamide reverse osmosis membranes En
41. ch as van der Waals forces Negative NPs would be expected to show less affinity for SWC4 and SWC5 both of which are bare PA and have significantly stronger negative character than SW30HR Experiments with SWC4 and SWC5 were laid out but not performed so future work may confirm or deny this hypothesis It was assumed that the polyDADMAC binding layer carried strong enough charge to attract the negative NPs and overpower alternate adsorption pathways Figure 4 11 Control coatings on SW30HR TiO2 did not adsorb to polyDADMAC TiO2 and Ag appear to have adsorbed to SW30HR despite carboxyl groups promoting electrostatic repulsion 46 4 4 3 SEM imaging results SEM was used to investigate whether applied coatings were thick enough to be visible on the membrane surface Initial SEM analysis after coating experiments used the TM3000 instrument with a maximum resolution of 10 000 10k magnification Even with relatively low magnification of 2 5k Ag aggregates of up to 6 m were visible on the membrane surface when coated with 100 mg Ag L NP solution Figure 4 7 Titanium NPs applied in much lower concentrations were not visible on the membrane surface Figure A 1 A higher resolution instrument the SU6600 was used to achieve 100k magnification of the membrane surface Ag coatings of 20 mg Ag L were used for the high resolution imaging but were not visible on the membrane surface at the lower conce
42. ctric point pH 4 for SW30HR As the pH decreased the membrane zeta potential became slightly positive causing electrostatically unfavorable conditions for the polyDADMAC coating Theoretically as the membrane zeta potential increased the polyDADMAC would be electrostatically repelled from the membrane and the coating removed Figure 5 1 indicates that some polyDADMAC remained on the surface after cleaning with both pH 1 and pH 4 solutions This was expected because of the nature of the polyDADMAC polymer A highly charged 80 120 kiloDalton kDa linear polymer polyDADMAC likely bound to the PA surface at multiple contact points creating a strongly bound coating layer Removal of a single polyDADMAC chain would require every contact point to be electrostatically unfavorable it is likely that a percentage of the polymer remained attached to the membrane at one or several contact points per chain SW30HR was used for coating removal experiments because it was not highly charged at low pH values probably because of the PVA layer as opposed to other membranes that do not have such a layer Had the membrane been more positively charged at low pH electrostatic repulsion between it and polyDADMAC would likely have been stronger leading to greater removal Since polyDADMAC was only used as a 50 binding layer for negative NPs it was not considered a negative result for residual polymer to remain on the membrane
43. d be removed by cleaning of the NP coating alone however from an application perspective polyDADMAC is inexpensive enough to justify pursuing acid cleaning alone 78 7 2 4 Bench scale RO fouling experiments Reverse osmosis experiments were promising with the nanoparticle coated membrane having the highest relative flux recovery as well as less flux decline than the uncoated membrane during fouling Despite some flux improvement sodium alginate was visible on all fouled membrane samples removed from the bench scale unit The fouling layer may have been too thick for the acid cleaning solution to effectively repel the polyDADMAC coating from the membrane surface the quantity of foulant used represented a worst case scenario for fouling Additional RO fouling experiments should be performed in saltwater solutions so salt rejection data can be collected and the comprehensive effects of membrane coatings can be evaluated Different types of foulants should be tested as well to determine whether the coating materials are effective against multiple modes of fouling To evaluate whether coatings are regenerable and could be reapplied application of coatings in the bench scale system should be attempted Future studies of this research topic should focus on bench scale experiments Though membrane characterization is useful and method development necessary coatings must be applicable in a system to be of any
44. d from varying NP concentrations The 4 2 mg Ti L coating was selected for future work because of the relatively high zeta potential and low concentration 35 Figure 4 3 Effects of 2 1 mg L 4 2 mg L 10 6 mg L and 21 2 mg L TiO2 on membrane surface charge compared to the uncoated SW30HR membrane Error bars represent a standard deviation from 12 measurements The TiO2 coatings varied quite significantly between experiments A set of experiments was prepared in August 2011 and a second set was prepared in May 2012 The difference in titration curve between the two TiO2 coated membranes was notable The difference may be attributed to heterogeneity between membrane samples or NP aggregation over the course of the stock solution lifetime Membrane heterogeneity seems unlikely as virgin membrane comparisons did not vary nearly as much as the TiO2 coated membranes If aggregation occurred in the stock solution considerably different quantities of NPs could be acquired with each aliquot removed from solution and the coatings would not be uniform The changes observed in Figure 4 4 suggest that NP stock solutions need to be buffered sonicated or otherwise modified in the future to prevent aggregation The manufacturer provided particle size analysis for the powdered 36 form of the NP that was received but size analysis for the stock solution would be useful for detecting NP aggregates Figure 4 4 T
45. dow of VisioLab 2 If instrument is suspected to be contaminated by desorbed coatings see SurPASS instruction manual for appropriate cleaning measures pg 60 3 Turn instrument off 4 Disassemble AGC 5 Discard used membranes 87 Appendix C Bench Scale RO for Membrane Coatings Standard Operating Procedures Night before Preparation Add 200 mL of sodium alginate to 4 liters of DI to make fouling solution If cold bring water sample to room temperature overnight If membrane is not already in DI water cut coupon and place in DI water o Membranes for coated experiments should be pre coated and soaked in DI water at least overnight Clean water flux run DI should be in the system after the cleaning Calibrate conductivity meters 50 mS for on line meter 0 447 mS for bench top meter Set up the system by bypassing the membrane cell and using small plastic Nalgene feed reservoir Calibrate computer clock Tare pressure gauge Don t tare flow meter I think it drifts less than taring makes it change Start RO Control 38 vi o Set Actuator valve voltage to 10V o Run software o Set pump control to Manual 4 Hz o Turn ON pump inverter to start pump o Increase pump speed to 7 Hz o Make sure there is no air in the pump intake hose Run DI at high pressure If pulsating let it run at low pressure for several minutes or run at about 3
46. e g Mycobacterium sp Campbell et al 1999 Foulants can originate as soluble materials in the feed water that deposit on the membrane or they can be colloidal or particulate in nature before depositing Inorganic solutes can precipitate on the surface of the membrane or precipitate to form particles first then deposit Organic foulants tend to be the most common due to their prevalence in feed waters of all types and their complex interaction chemistries Organics can adsorb to the membrane surface as well as within the pores of MF UF and NF membranes Lee et al 2005 Braghetta et al 1998 Cho et al 1998 Hong et al 1997 The surface chemistry of the membrane and the chemistry of the feed solution both factor into fouling effects as well Negatively charged membrane materials tend to incur less biological fouling due to electrostatic repulsion effects but cation bridging surfactant boundary formation and pH changes can all increase fouling Childress and Elimelech 1996 The wide variety of materials that cause membrane fouling which can result in several types of fouling within a single system makes it difficult to design a membrane material impervious to all types of fouling Our research therefore aims not to develop a membrane material but a coating that acts as a snakeskin like barrier such that all types of fouling are removed when the coating is washed away The coating could then be reapp
47. e sulfonate derivative and arrived as 18 3 w w powder lot PB75 Stock solution was prepared by dissolving 0 100 g powdered TiO2 into 100 mL DDI water for a final concentration of 0 184 g Ti L Positively charged titanium dioxide TiO2 was functionalized by the manufacturer with polyDADMAC and arrived as at 20 w w powder lot PB75 Stock solution was prepared by dissolving 0 106 g powdered TiO2 into 100 mL DDI water for a final concentration of 0 212 g Ti L 18 Figure 3 4 Sketch representation of NP surface charges imparted by a polyacrylate and b polyDADMAC 3 3 Membrane characterization methods 3 3 1 Zeta potential 3 3 1 1 Instrumentation Surface zeta potential was measured with a SurPASS electrokinetic analyzer Anton Paar Graaz Austria VisioLab software version 2 10 was provided for instrument control and upgraded to version 2 20 before experiments started The instrument was purchased with a clamping cell sample holder This measuring cell required membrane coupons of 55 mm x 25 mm and used only 9 of the total membrane area while the contribution of the remaining membrane area to membrane body conductance was unaccounted for Buk ek et al 2010 The adjustable gap cell AGC sample holder which required smaller membrane coupons of 10 mm x 20 mm and used 100 of the membrane surface area for measurement was purchased in spring 2011 due to improvements in measureme
48. ent was performed on a single membrane sample involving DI water flux fouling cleaning and recovery In the full experiments membranes were compacted for one hour fouled for three hours cleaned with HCl for 30 minutes and DI water was run through the system for at least one hour to evaluate flux recovery after cleaning The system was flushed with 20 L DI water after full experiments without caustic cleaning A six hour DI water experiment was run to evaluate membrane compaction over the course of the experiments After six hours of DI flux through the membrane a 35 g L NaCl solution was used to test membrane rejection Rejection exceeded 99 throughout a one hour experiment This indicates that despite high flux values in excess of 120 L m2h the membranes used for RO experiments had acceptable integrity 6 3 Bench scale RO results The bench scale RO experiments had promising results Actual flux values are reported in standard units of L m2 hr lmh and normalized flux values are reported as a fraction When fluxes were normalized to the six hour DI water experiment and compared Table 6 2 the membrane coated with both polyDADMAC and Ag had the best flux recovery after fouling and cleaning Fluxes for experimental membranes are reported throughout this section in relation to the six hour DI water experiment which is considered to be 100 flux maximum flux achievable under experimental conditions at al
49. for electrokinetic analysis A fresh 60 membrane was used for each experiment Some repetition occurred between experiments such as fouling in two experiments of each cycle the data was used to confirm repeatability of experiments and observe variability between individual membrane coupons Feed tank High pressure pump Automated needle valve Pressure gauge Membrane cell Temperature control Pressure gauge Feed water recycle Permeate balance Computer controls Figure 6 1 Diagram of reverse osmosis bench scale filtration system A high concentration of sodium alginate was used as a model polysaccharide to foul the membranes and cause flux decline Mi and Elimelech 2009 Herzberg et al 2009 A 2 g L stock solution was prepared by combining 2 g sodium alginate powder with 1 L of DI water and stirring overnight to dissolve For fouling experiments 200 mL of sodium alginate stock solution was diluted into 4 L DI water resulting in a 95 mg L sodium alginate solution Deionized water flux tests represented in Table 6 1 by experiments X 1 used approximately 10 L of DI water in continuous recycle mode DI water experiments ran 61 for three hours to allow for membrane compaction and evaluate the effects of the different coatings on general flux Fouling experiments represented in Table 6 1 by experiments X 2 started with a one hour clean water flux to compact the membrane DI water
50. g the entire coating process Membrane coupons for initial experiments particularly concentration studies were coated in the refrigerator 4 C to prevent photodegradation of the membrane samples and inhibit any biological growth As method development progressed it was decided that room temperature coating processes better simulated real world conditions and that photodegradation and biological growth were not a concern on experimental time scales of less than 24 hours To emulate movement across the membrane surface similar to the flushing of feed water through modules in a RO plant a rocker table was used on the bench top at room temperature 22 2 C for coatings used in removal Chapter 5 and RO Chapter 6 experiments Coating Chapter 4 experiments from August 2011 also utilized the rocker table 4 4 Coating method evaluation 4 4 1 Wet vs dry membranes For the concentration experiments dry membrane coupons were placed into polyDADMAC or TiO2 solution for coating This caused inconsistency in coating conditions as the negative nanoparticles were applied to a wet membrane while the polyDADMAC layer and positive NP were applied to a dry membrane It was determined that all coatings should be applied to wet membranes to emulate real world conditions where possible Figure 4 10 shows that polyDADMAC binding was stronger on a dry membrane than on a wet one but this effect could also be attributed to whet
51. he membrane surface nor percent coverage 7 2 2 Coating condition optimization Coating kinetics were briefly examined to determine NP coating duration but kinetics did not fit the scope of this study Kinetics should be investigated more thoroughly similarly to the concentration experiments in this work seeking the shortest effective coating time It is expected that coatings assemble electrostatically on the order of minutes Reducing the duration of coating time may reduce the cost of implementing this technology Verification of coating layers by a technique other than zeta potential would be useful X ray spectroscopy is expected to be a viable option X ray photoelectron spectroscopy XPS is used for elemental characterization of membrane surfaces and can be employed to verify the presence of Ti and Ag on the membrane Siegbahn 1981 XPS requires concentrations of parts per thousand or higher which may present issues with the extremely thin NP coatings Parts per million concentrations are detectable but longer data collection times or specialized instrumentation are required Energy dispersive X ray spectroscopy EDS a technique coupled to SEM instrumentation can also be used for elemental analysis of the sample Sputter coating may interfere with detection of small quantities of Ti or Ag so samples that are not sputter coated should be analyzed Samples may then be sputter coated for SEM visual analysi
52. her the 44 coating process was dynamic wet membrane allowing less loose binding or binding via passive adsorption in a static solution dry membrane The stronger affinity of polyDADMAC for the dry membrane is assumed to be a result of static vs dynamic coating conditions rather than temperature or degree of membrane saturation Coating a wet membrane in cool static conditions could easily test this hypothesis Figure 4 10 Titrations for polyDADMAC coated SW30HR show that a dry membrane coated statically at low temperature had a more positive charge than a wet membrane coated dynamically at room temperature 4 4 2 Control coatings Control experiments were performed to determine whether coatings adsorbed via electrostatic interactions alone or a combination of factors H bonding van der Waals forces etc TiO2 did not alter the zeta potential or curve shape of the polyDADMAC coated membrane Figure 4 11 indicating that positive NPs did not adsorb to the polymer coating layer Ag and TiO2 NPs adsorbed almost equally to the bare 45 SW30HR membrane despite electrostatically repelling carboxyl groups present in both the NPs and membrane surface Atop the PA layer SW30HR had a commercially applied layer of uncharged PVA which served as a barrier between the membrane carboxyl groups and the negatively charged NPs and allowed adsorption to the membrane surface via non electrostatic interactions su
53. iO2 NP coatings for two sets of experiments Significant differences could be attributed to changes in NP stock solution or heterogeneity of the membrane surface Because TiO2 NPs were functionalized by polyDADMAC it was initially expected that the NPs would impart a surface charge comparable to that of the linear polymer Coating with NPs resulted in a significantly lesser positive charge indicating a lower charge density Figure 4 5 Upon further consideration it was concluded that the NPs could not impart a surface charge comparable to the polyDADMAC coating because each polymer chain included loops and tails that contributed significantly to the surface charge Each NP however was a roughly spherical particle contributing a single point charge and resulting in a lower charge density Surface packing may also have factored 37 in with adsorption of the dilute polymer being a more highly favored mechanism than the aggregation of NPs at the membrane surface Figure 4 5 Charge density of polyDADMAC coating compared to TiO2 NP coating which was functionalized with polyDADMAC 4 2 3 Ag NP coatings Ag NPs given negative charge by polyacrylate adsorbed to the polyDADMAC modified membranes Adsorption was indicated by a decrease in the zeta potential of polyDADMAC coating Ag coating successfully reduced the strong positive charge of the polyDADMAC coated membrane bringing the zeta potential
54. ial on the membrane surface complicate the results of EDS analysis 11 1 6 3 Attenuated total reflectance Fourier transform infrared spectroscopy ATR FTIR spectroscopy is used to confirm functional groups present in the membrane such as the carboxyl and amine groups expected for a polyamide membrane or to determine functional groups associated with natural organic matter NOM or other foulant materials Cho et al 1998 In RO research ATR FTIR is often used to confirm that the membrane surface coating or foulant is performing as expected Coronell et al 2010 Tang et al 2009a Tang et al 2007 This technique is not often the primary analysis of membranes or membrane coatings though some studies such as Belfer et al 1998 have demonstrated the benefits of utilizing peak emergence and absorbance changes for assessing membrane coatings Proper accessories and setup of the instrument such as crystal selection angle of incidence and atmosphere purging are critical for ATR FTIR use as the primary instrumentation for analysis Without extremely sensitive instrumentation it is very difficult to discern the small quantities of coating layers over the dominant vibrational bands of the polysulfone support membrane and ATR FTIR is of limited benefit Gabelich et al 2005 1 6 4 Contact angle measurements Contact angle measurement is used to assess the hydrophilic character of a membrane material and the
55. ial titrations to detect the presence of self assembled coatings The Anton Paar SurPASS electrokinetic analyzer was purchased for this project Optimization of instrument methods was a critical component of this project 14 CHAPTER THREE MATERIALS AND MATERIAL CHARACTERIZATION Significant developments have been made regarding permanent membrane modifications and improved removal of target compounds but removable coating design has not been widely reported in the literature to date Because removable coating design has not been widely studied it was necessary to determine how best to coat the membranes remove coatings and to perform concentration studies to optimize the later experiments Ultrapure water DDI from a Super Q Plus System Millipore Bedford MA was used for all solutions sample preparation and electrokinetic analysis The use of the words adsorption in this document refers to electrostatic adsorption unless otherwise noted 3 1 Membranes Reverse osmosis membranes from two manufacturers were used SW30HR from Dow Filmtec a subsidiary of the Dow Chemical Company Midland Michigan and SWC4 and SWC5 from Hydranautics a Nitto Denko company Oceanside California All membranes are thin film composite TFC polyamide PA formed by cross linking of m phenylenediamine and trimesoyl chloride shown in Figures 3 1 and 3 2 SW30HR is commercially coated with a polyvinyl alcohol surface la
56. imilar recovery Most notably Ag coated membranes had improved flux recovery over the virgin and polyDADMAC coated membranes which suggests that the NP coating improves the removal of the foulant and benefits flux recovery after washing The flux through each membrane was normalized to the flux of the six hour DI water experiment after four hours of membrane compaction When actual flux values are compared Figure A 7 a similar effect is observed Cleaning of the polyDADMAC coated membrane shows improvement when normalized flux is considered but actual flux values are very similar 66 to the uncoated membrane When the Ag coated membrane is cleaned however the resulting flux is higher than that of the uncoated membrane Figure 6 4 Flux recovery after acid cleaning PolyDADMAC coating gives slight improvement in flux recovery but Ag coating notably increases post cleaning flux Repeated parts of experiments fouling occurred on two of the three membranes in each experimental cycle for example were compared to determine the repeatability of RO flux test results Flux data in these figures are normalized to the first flux measurement of each experiment for comparison purposes DI water flux data were compared from three separate Ag coating experiments Figure 6 5 Data were comparable except for an anomalous initial flux on 5 23 2012 DI flux data for polyDADMAC coated membranes were similar SWC
57. ion New Jersey 1990 Petry M Sanz M A Langlais C Bonnelye V Durand J P Guevara D Nardes W M Saemi C H The El Coloso Chile reverse osmosis plant Desalination 2007 203 1 3 141 152 Shannon M A Bohn P W Elimelech M Georgiadis J G Marinas B J Mayes A M Science and technology for water purification in the coming decades A review Nature 2008 452 7185 301 10 93 Siegbahn K M Electron spectroscopy for atoms molecules and condensed matter In Nobel Lectures Physics 1981 1990 Fr ngsmyr T Ed World Scientific Publishing Company Singapore 1993 Soroko I Livingston A Impact of TiO2 nanoparticles on morphology and performance of cross linked polyimide organic solvent nanofiltration membranes J Membr Sci 2009 343 1 2 189 198 Subramani A Hoek E M Direct observation of initial microbial deposition onto reverse osmosis and nanofiltration membranes J Membr Sci 2008 319 1 2 111 125 Tang C Y Kwon Y N Leckie J O Probing the nano and micro scales of reverse osmosis membranes A comprehensive characterization of physicochemical properties of uncoated and coated membranes by XPS TEM ATR FTIR and streaming potential measurements J Membr Sci 2007 287 1 146 156 Tang C Y Kwon Y N Leckie J O Effect of membrane chemistry and coating layer on physicochemical properties of thin film composite polyamide RO and NF membranes I FTIR and XPS character
58. ion and characterization of fouling resistant TiO2 self assembled nanocomposite membranes J Membr Sci 2006 275 1 2 1 5 Belfer S Purinson Y Kedem O Surface modification of commercial polyamide reverse osmosis membranes by radical grafting An ATR FTIR study Acta Polym 1998 49 574 582 Braghetta A DiGiano F A Ball W P NOM accumulation at NF membrane surface Impact of chemistry and shear J Environ Eng 1998 124 11 1087 97 Brant J A Childress A E Assessing short range membrane colloid interactions using surface energetics J Membr Sci 2002 203 257 273 Brant J A Childress A E Colloidal adhesion to hydrophilic membrane surfaces J Membr Sci 2004 241 2 235 248 Buksek H Luxbacher T Petrinic I Zeta potential determination of polymeric materials using two differently designed measuring cells of an electrokinetic analyzer Acta Chim Slov 2010 57 700 706 Cahill D G Freger V Kwak S Y Microscopy and microanalysis of reverse osmosis and nanofiltration membranes MRS Bulletin 2008 33 27 32 Campbell P Srinivasan R Knoell T Phipps D Ishida R Safarik J Cormack T Ridgway H F Quantitative structure activity relationship QSAR analysis of a Mycobacterium species to cellulose acetate and polyamide reverse osmosis membranes Biotechnol Bioengineering 1999 64 527 544 Childress A E Elimelech M Effect of solution chemistry on the surface charge of polymeri
59. iterature and shows great promise with improved membrane mechanics and permeability as well as strong evidence of fouling resistance The primary focus of this research will be to develop a system of NPs and membranes where the NPs can be deposited and subsequently removed to yield a regenerable system Polyvinyl alcohol has been investigated as a removable coating paired with a positively charged NF membrane to create a coating that can be removed with a simple acid cleaning Ba et al 2010 This combination was a highly effective coating with flux recovery of nearly 100 after low pH cleaning The interaction of functionalized nanoparticles with polymeric membranes has also been investigated with the intention of removing the NPs using membrane processes Ladner et al 2012 NPs were effectively removed and demonstrated interesting adsorption characteristics which led to the development of the work performed for this thesis 8 1 6 Membrane characterization Some background on membrane characterization methods is included here to understand how the techniques used in this study fit into the literature of the field Physicochemical properties of membranes and effects of fouling coatings and modifications are characterized using an array of techniques from materials science chemical engineering and traditional chemical analysis Chemical zeta potential elemental composition and functional groups present
60. ization of polyamide and coating layer chemistry Desalination 2009a 242 1 3 149 167 Tang C Y Kwon Y N Leckie J O Effect of membrane chemistry and coating layer on physicochemical properties of thin film composite polyamide RO and NF membranes II Membrane physicochemical properties and their dependence on polyamide and coating layers Desalination 2009b 242 1 3 168 182 Taurozzi J S Arul H Bosak V Z Burban A F Voice T C Bruening M L Tarabara V V Effects of filler incorporation route on the properties of polysulfone silver nanocomposite membranes of different porosities J Membr Sci 2008 325 1 58 68 van der Bruggen B Manttari M Nystrom M Drawbacks of applying nanofiltration and how to avoid them A review Sep Purif Technol 2008 63 251 63 van de Lisdonk C A C van Paassen J A M Schippers J C Monitoring scaling in nanofiltration and reverse osmosis membrane systems Desalination 2000 132 101 108 Xie H Saito T Hickner M A Zeta potential of ion conductive membranes by streaming current measurements Langmuir 2011 27 8 4721 4727 Yan L Li Y S Xiang C B Xianda S Effect of nano sized Al2O3 particle addition on PVDF ultrafiltration membrane performance J Membr Sci 2006 276 1 2 162 167 Yang H Lin J C Huang C Application of nanosilver surface modification to RO membrane and space for mitigating biofouling in seawater desalination Water Res 2009
61. k E M V Yan Y Subramani A Huang X Hurwitz G Ghosh A K Jawor A Interfacial polymerization of thin film nanocomposites A new concept for reverse osmosis membranes J Membr Sci 2007 294 1 7 Kilduff J E Taniguchi M Belfort G Low fouling synthetic membranes by UV assisted graft polymerization Monomer selection to mitigate fouling by natural organic matter J Membr Sci 2003 222 1 2 59 70 92 Kim J Van der Bruggen B The use of nanoparticles in polymeric and ceramic membrane structures Review of manufacturing procedures and performance improvement for water treatment Environ Poll 2010 158 2335 2349 Kim M M Au J Rahardianto A Glater J Cohen Y Gerringer F W Gabelich C J Impact of conventional water treatment coagulents on mineral scaling in RO desalting of brackish water Ind Eng Chem Res 2009 48 6 3126 3135 Ladner D A Steele M Weir A Hristovski K Westerhoff P Functionalized nanoparticles interactions with polymeric membranes J Hazard Mater 2012 in press Lee S Cho J Elimelech M A novel method for investigating the influence of feed water recovery on colloidal and NOM fouling of RO and NF membranes Environ Eng Sci 2005 22 4 496 509 Louie J S Pinnau I Ciobanu I Ishida K P Ng A Reinhard M Effects of polyether polyamide block copolymer coating on performance and fouling of reverse osmosis membranes J Membr Sci 2006 280 1
62. l times the flux after four hours which was reduced from the initial by 10 due to membrane compaction was considered 100 flux after four hours of experiment time The six hour experiment accounted for membrane compaction so that experiment cycles 63 could be compared to a normal membrane subjected to the same filtration duration Results are reported as normalized flux averaged throughout the experiment timeframe one hour for DI flux one hour for DI flux after cleaning etc except for fouling experiments in which the last five minutes of flux data is averaged averaged normalized flux is then compared to that of the SWC5 six hour DI water experiment to determine what degree of flux decline or recovery was achieved after membrane compaction was addressed The initial fluxes for all membranes were similar Figure 6 2 indicating that coatings were not thick enough to cause flux decline PolyDADMAC coated SWC5 achieved 98 of uncoated membrane flux while the Ag coated membrane achieved 97 flux Some variability was observed between membrane samples 5 not shown experiments with similar flux values were compared where possible Variability was expected due to heterogeneity of the membrane and some user error with the bench scale unit Some variability on coated membranes was also attributed to heterogeneity of coatings which was exacerbated for Ag coatings if the polyDADMAC coating varied sig
63. lder to align with the grooves in the leads Press down gently Adjust if needed 11 Attach the AGC and tighten the leads with the knob on the holder until the clutch engages Instrument Startup 1 Start up VisioLab for SurPASS 2 Turn instrument on 3 Fill 600 mL Schott beaker with 500 mL 1 mM KCl Add stir bar 4 Open Pre Measurement window and run Fill cycle for 100 s 5 Run Rinse cycle to determine gap height 85 Start Rinse cycle When first ramp achieves target pressure 300mbar observe gap height readings If instrument fails to achieve target pressure narrow gap height and repeat Rinse Adjust gap height as necessary to reach 0 105 0 01mm 6 Run full Rinse cycle for 300s after acceptable gap height is achieved 7 Run Flow check at 300mbar Record final pressure and flow for each ramp If either ramp is nonlinear or does not achieve target pressure 50mbar re run Flow check o It may be necessary to take apart the AGC and check the membrane reverse the orientation of the AGC re run and closely observe the Rinse cycle or undertake other troubleshooting measures 8 If running titrations and machine has been OFF switch to beaker of 500 mL DI water and rinse syringes to clear bubbles from tubing Open Setup window and select Titration Settings Rinse both Left amp Right syringes for 2 rinse cycles into DDI water not the KCl solution Membrane Analysis
64. lied to refresh the barrier layer 5 1 3 Membrane matrix modifications One strategy for reducing fouling while maintaining membrane flux is modification of the membrane matrix before or after casting NPs have become a promising set of materials due to their small size affinity for polymer binding and attractive properties such as biocidal silver and photoactivated titanium dioxide Kim and van der Bruggen 2010 Soroko and Livingston 2009 Yan et al 2006 Polymer functionalized metal composite NPs are a type of nanomaterial with properties that could be of interest These NPs are typically given charge by a carboxyl or amino functionalized polymer however research to this point has focused on pure material NPs and only one publication could be found regarding functionalized NP use in membrane filtration Jadav et al 2010 Pure NP materials most often described in the literature are titanium dioxide TiO2 alumina Al2O3 silver Ag silica and zirconia There have been a number of self assembly poly ether sulfone and polysulfone UF membranes that use TiO2 NPs as modifiers an optimized amount of NPs is added to the casting solution before preparing the membrane itself Luo et al 2005 Jeong et al 2007 have created mixed matrix RO membranes incorporating various concentrations of 50 150 nm zeolite NPs in the casting solution These mixed matrix membranes had improved permeability over PA membranes
65. lifespan There is potential to create a snakeskin like barrier between the membrane and foulants by electrostatically binding a coating material to the membrane that can be released by pH manipulation removing foulants in the process The feasibility of using functionalized nanoparticles as removable adsorptive coatings on RO membranes was evaluated in this study Several inorganic polymer composite nanoparticles NPs were examined in this study including titanium dioxide coupled with polydiallyldimethylammonium chloride polyDADMAC to impart a positive charge TiO2 titanium dioxide incorporated with polyacrylate to impart a negative charge TiO2 and silver incorporated with polyacrylate again to incorporate a negative charge Ag PolyDADMAC was used as a positively charged binding layer atop the negatively charged membrane surface to adsorb negatively charged nanoparticles A series of concentration experiments was performed for each NP to find the lowest effective concentration for self assembled coatings The optimal concentration was then used in a series of kinetic experiments to determine the time required for coatings to assemble Coating removal experiments were performed over a range of high and low pH values with attention to indications of chemical alterations to the virgin iii membrane Titrations were run on pairs of membrane samples in an electrokinetic analyzer to measure the surface
66. like to dedicate this thesis to my Fox and Smith families both of whom have always been tremendously encouraging and supportive of my educational endeavors I am honored to be the first in either family to receive a graduate degree in a field I care about so passionately To Robyn my longest and best friend Despite our differences we have endured so much together and I look forward to a time when we will not have to Skype and phone to catch up Some people find soul sisters I am so lucky to have you as a real one and want to share my daily life with you To Toby the best surprise a family could ask for You have grown so quickly into such an amazing person I see so much of myself in you keep reaching for the stars They will deliver you to a better life than you can dream I can t wait to see you through hormones girlfriends and self discovery To Mom who passed on the inner chemist I am so glad you have found work you love and to which you can contribute so much Keep me posted about that new farm we re starting I cannot express how much I appreciate your enduring financial and emotional support though I m sure you re glad to have your basement back To Dad my student in arms I am so proud of you for stepping up and going back to school when the world decided that 30 years of experience wasn t good enough Doubtless that experience has helped tremendously with the process of getting that
67. ment is insufficient to quantify NP concentration on the membrane as the number of metal atoms present per NP is unknown but an alternative method of quantification has not yet been identified 4 1 Determining coating concentrations 4 1 1 PolyDADMAC coating layer PolyDADMAC was used as a positive binding layer bridging between negatively charged NPs and the negatively charged membrane surface Coating solutions were prepared by dilution of the stock solution using DDI water PolyDADMAC coatings of 5 2 and 0 2 w v were applied to dry SWC4 membranes to test concentration effects on a highly negative membrane Coatings of 2 1 and 0 2 were applied to dry SW30HR membrane after SWC4 data was collected All coatings for the concentration experiments were allowed to assemble for 24 hours in the refrigerator 4 C Membranes were removed from coating solution rinsed in 200 mL DDI water for 31 30 seconds and stored in DDI water in the refrigerator for at least 12 hours before analysis or further coating During the process of determining coating concentrations it was decided that using wet membranes would be more applicable to plant operating conditions so the 0 2 coatings for negative NP experiments were assembled onto wet SW30HR membranes from August 2011 onward Results of coating dry membranes with 0 2 polyDADMAC were compared to those for coating wet membranes with 0 2 polyDADMAC 4 1 2 TiO2
68. nd pharmaceuticals nanofiltration NF removes smaller organic pollutants such as chlorinated pesticides and multivalent ions like calcium magnesium and iron and reverse osmosis RO removes monovalent ions such as sodium and chloride NF and RO are both pressure driven processes requiring robust membrane materials to withstand high operating pressures Thin film composite membranes are frequently used with a thin active layer supported by a thick strong and inert backing layer This endurance requirement has until recently led to membrane 2 development focused on building stronger and longer lasting membranes Within the past several decades the focus on membrane lifetime has led to research in fouling resistance but membrane cleaning remains an issue van der Bruggen et al 2008 RO technology is widely used for desalination of seawater and brackish water both of which may contain large quantities of organic matter and marine organisms that may contribute to chemical and biological fouling of the membrane For example the sheer mass of bacteria present in red tides can cause significant fouling in desalination plants and frequently require the plant to cease operation until the bloom has dispersed Petry et al 2007 In a study using NF for seawater desalination 20 fouling was observed over a three day period making the operation unsustainable Harrison et al 2007 Fouling is a hindrance to the RO
69. ne none 1 2 none none Na Alginate none 1 3 none none Na Alginate pH 3 HCl 2 1 pDADMAC none none none 2 2 pDADMAC none Na Alginate none 2 3 pDADMAC none Na Alginate pH 3 HCl 3 1 pDADMAC Ag none none 3 2 pDADMAC Ag Na Alginate none 3 3 pDADMAC Ag Na Alginate pH 3 HCl 6 2 RO methods and additional materials A bench scale reverse osmosis system diagrammed in Figure 6 1 was used for the RO experiments An existing LabView program v38 was used for all experiments A pump circulated the water through the piping at approximately 800 mL min and a cooling system maintained the temperature of the recirculating feed water at 24 2 C An automated needle valve was used to ramp the pressure to a target of 1000 psi if the pressure dropped below 985 or above 1005 psi the two set points used the valve was automatically adjusted The clean water permeate forced through the membrane was collected into a beaker on a digital scale recording mass as input for the LabView program to calculate flux Permeate containers were 1 L in size and emptied after approximately 750 950 mL had accumulated An experiment was defined as the tests performed on a single membrane sample 3 1 or 2 2 for example while the set of experiments for a particular membrane coating was referred to as an experimental cycle After each experiment in Table 6 1 the membrane coupon was removed and prepared
70. nificantly 64 Figure 6 2 Comparison of DI water flux Uncoated SWC5 had the highest flux with polyDADMAC coated and Ag coated fluxes slightly lower In fouling tests Figure 6 3 the coated membranes fouled at approximately the same rate with an equal overall flux 75 of clean water flux The uncoated membrane fouled more quickly resulting in faster flux decline and lower overall flux 66 There is an offset in Figure 6 3 due to flux being normalized to the six hour DI water experiment rather than the actual flux of each membrane when fouling began The lower rate of fouling and higher flux for coated membranes indicates that the coatings may block foulants from adsorbing in the loops and valleys of the membrane structure When actual flux values were compared Figure A 6 it was noted that both coated membranes had lower fluxes when fouling began but fouled to a lesser extent than the uncoated membrane The flux for both coated membranes after fouling which was 65 indistinguishable in both normalized and actual flux data was slightly higher 75 lmh than the uncoated membrane 70 lmh Figure 6 3 Fouling of virgin polyDADMAC coated and Ag coated membranes Somewhat less fouling occurred on the coated membranes resulting in 10 higher flux Flux recovery experiments after pH 3 cleaning for 30 minutes are shown in Figure 6 4 with polyDADMAC coated and virgin membranes demonstrating s
71. nt reliability and was used from May 2011 onward Figure 3 5 19 Figure 3 5 General representation of adjustable gap cell setup 3 3 1 2 Chemicals for analysis Potassium chloride salt was purchased from BDH Chemicals Ltd a subsidiary of VWR International Radnor PA and used as received A stock KCl solution of 0 1 M was prepared with DDI water and kept in the dark at room temperature The stock solution was vacuum filtered with a 0 22 m polyvinylidene fluoride filtration membrane Millipore Billerica MA Working solutions of 0 001 M KCl were prepared before each titration by diluting 10 mL of stock solution into 1 L of ultrapure water and vacuum filtering through a 1 m Nylasorb nylon membrane Pall Life Sciences Ann Arbor MI to eliminate dust and other particulates Acid titrations used 0 1 M hydrochloric acid HCl prepared by diluting 8 mL of 37 laboratory grade HCl BDH Ltd Radnor PA into 1 L DDI water Base titrations VV p p electrode electrode sample electrolyte flow 20 used 0 1 M sodium hydroxide NaOH prepared by dissolving 4 g NaOH pellets EMD Millipore Billerica MA in 1 L DDI water 3 3 1 3 Titrations The AGC was assembled as instructed in the instrument user manual DDI water was used for filling and rinsing of the instrument gap adjustment and flow check KCl electrolyte solution was flushed through the instrument before each measurement
72. ntenance was performed for the SurPASS over the course of the project The conductivity meter was calibrated every three months or before use if the instrument had been unused for at least a week The pH meter was stored in 3 M KCl when not in use and was calibrated monthly Titration unit tubing was rinsed monthly to eliminate air bubbles in the lines The sample cell was fully disassembled and cleaned with dilute isopropyl alcohol after all SW30HR experiments concluded 3 3 2 Attenuated total reflectance Fourier transform infrared spectroscopy ATR FTIR spectroscopy was used to determine whether the coatings could be detected as would be expected if quantities sufficient to foul the membrane were present A Nicolet 6700 FTIR Thermo Scientific Waltham MA fitted with a diamond Smart iTR plate was used with a scanning resolution of 2 giving a total of 32 scans per spectrum 3 3 3 Scanning electron microscopy SEM was employed to visualize the effects of the coating and removal processes Membrane samples were sputter coated with gold using a Hummer 6 2 sputtering machine Anatech Ltd Battle Creek MI prior to analysis Images were taken at magnification of 2 5 k and 10 0 k with a tabletop TM3000 unit Hitachi High Technologies America Inc Dallas TX High resolution images were taken at 22 magnification 10 k and 100 k with a variable pressure Field Emission SEM SU6600 microscope Hitachi High Technologie
73. ntration Figure A 2 PolyDADMAC coatings not visible at 10k magnification were visible in 100k images as filled in valleys in the membrane structure Figure 4 12 Figure 4 12 SEM images of virgin SW30HR left and polyDADMAC coated SW30HR right PolyDADMAC coating appeared to fill in the loop and valley structure of the membrane 47 4 4 4 ATR FTIR analysis Infrared spectroscopy was used to investigate whether coatings were present in sufficient quantities to impart changes in infrared spectra A significant quantity of polyDADMAC coating was expected to result in an emerging C N peak at 1100 cm 1 while polyacrylate from the negative NP coatings was expected to result in a shift of the carboxyl C O as the aliphatic carboxyl group contributed to the peak However no changes were observed in the presence of any coating Figure 4 13 substantiating the conclusion that coatings were thin Figure 4 13 ATR FTIR of polyDADMAC and negatively charged NP coatings did not indicate any changes in surface chemistry 48 CHAPTER FIVE REMOVAL OF COATINGS One hypothesis of this work is that electrostatic interactions are the dominant mechanism for coating attachment so simple acid or base cleaning should be sufficient for coating removal 5 1 Coating removal methods For removal experiments wet membranes were coated in 0 2 polyDADMAC solution on a rocker table at room temperature for 24 hr
74. nts began dry membrane coupons were placed directly into coating solution and coated statically at low temperature 4 C As experiments progressed and implementation was considered coating conditions were modified to more closely emulate the operating conditions of an RO facility In an RO plant coating and cleaning membrane modules would be accomplished by passing the solution over the membrane surface at ambient 72 temperature 25 C Membrane modules are inspected by the manufacturer and vacuum sealed in water before purchase so all membranes being coated would already be saturated with water From August 2011 onward membrane coupons were soaked at least overnight before being immersed in the coating solution A rocker table was used to agitate the coating solution to move continuously across the membrane surface creating dynamic coating conditions expected in a RO plant 7 1 2 Determine appropriate conditions for coating removal All coatings were at least partially removed by acid cleaning as indicated by zeta potential measurements and SEM imaging Zeta potential titrations indicated that polyDADMAC remained on the membrane surface after cleaning with all acid concentrations likely due to the polymer adsorbing at multiple sites on the membrane or getting stuck in the loop and valley structure of the membrane surface Alternative binding mechanisms such as van der Waals interactions may also h
75. oated the polyamide membranes causing a large 71 increase in zeta potential throughout the pH range investigated 3 9 and negligible flux decline under normal operating conditions in DI water Functionalized inorganic polymer composite NPs are a model system useful for the fundamental work performed here though their applicability in full scale systems may be cost prohibitive Future work with this project may continue to use the NPs purchased for this study but consideration should be given to whether NP use is economically or ecologically feasible There are environmental concerns regarding behavior of nanomaterials in natural systems Auffan et al 2009 on the large scale a process would need to be implemented to remove and dispose of NPs after removal The effects of the inorganic component such as any antifouling properties should also be considered no significant difference was observed between TiO2 and Ag when measurement variability was considered suggesting that the functionalizing polymer polyacrylate dominated mechanisms of adsorption to the membrane Small quantities of NPs 3 8 mg Ti L TiO2 4 2 mg Ti L TiO2 and 20 mg Ag L Ag NPs were able to cause notable zeta potential changes on the membrane surface The percent coverage of the membrane and coating heterogeneity have not yet been determined Coating methods were developed over the course of this study When experime
76. on after deposition Gabelich et al 2005 One drawback to polyDADMAC in a water treatment context is that it has recently been implicated as a disinfection byproduct precursor it leads to n nitrosodimethylamine NDMA formation in particular Park et al 2009 The polyDADMAC used in this study was 20 w v Clarifloc C 308P donated by James Amburgey of UNC Charlotte and by Polydyne Inc Riceboro GA C 308P has an average molecular weight range of 80 120 kDa The 20 stock solution was used as received A graduated cylinder was used to measure polyDADMAC and was rinsed several times into the final solution to ensure complete transfer of the slightly viscous polymer N Cl n Figure 3 3 Monomer structure of diallyldimethyl ammonium chloride 17 3 2 2 Functionalized silver and titanium dioxide nanoparticles Functionalized nanoparticles were purchased from Vive Nano Inc Ontario Canada General representation of the NP surface is shown in Figure 3 4 for negative a and positive b charges Negatively charged silver Ag was functionalized by the manufacturer with a sodium polyacrylate derivative and received in aqueous solution of 1 g L Ag A 0 1 g L working solution of Ag was prepared by a tenfold dilution of the stock solution which was used as received Negatively charged titanium dioxide TiO2 was functionalized by the manufacturer with a polyacrylic acid derivative and a polystyren
77. ostatic binding that can be difficult to remove Ba et al 2010 Fundamental studies over the past several decades have investigated the chemical properties of the membrane surface layer Childress and Elimelech 1996 Childress and Elimelech 2000 Significant progress has been made toward understanding the complex chemistries involved in membrane surface interactions but fouling continues to plague desalination plants and other membrane filtration processes The mechanisms of fouling and means of preventing it are still being explored Zeta potential characterization of membrane cleaning studies has also been undertaken though these studies use complex cleaning agents as used in large scale operations and recommended by manufacturers Al Amoudi et al 2007 The work conducted for this thesis included a SurPASS electrokinetic analyzer manufactured by Anton Paar GMBH The electrokinetic properties of the membrane surface are determined by forcing an electrolyte solution through a sample cell containing the material of interest electrodes at each end of the sample cell measure the resulting streaming current and zeta potential is calculated Buksek et al 2010 The Fairbrother Mastin F M approach to calculating zeta potential given in Equation 1 improves upon the earlier Helmholtz Smoluchowski H S approach accounting for phenomena in the instrument that affect the reported measurements Buksek et al 20
78. process frequently requiring the system to be offline for cleaning Commercial membranes have been developed to reduce the rate of fouling or prevent certain types of fouling but no indefinitely fouling resistant membrane has been achieved Negatively charged NPs have been coupled with RO membranes to investigate the feasibility of using electrostatic interactions to attach and release self assembled coatings of these materials Electrostatic interactions between membranes and coating materials were manipulated to cause adsorption and desorption by varying the pH On negatively charged membranes such as polyamide an increase in proton concentration decreasing pH protonates the carboxyl groups on the membrane neutralizing the membrane once a critical proton concentration is reached Surface charge continues to increase beyond the isoelectric point as protons collect around electron density at surface 3 carboxyl groups The polymer polydiallyldimethylammonium chloride polyDADMAC was selected for its strong positive charge to serve as a binding layer between the negatively charged membrane and negatively charged NPs This project developed coating removal and characterization methods and examined the advantages and drawbacks of each 1 1 Reverse osmosis membranes Reverse osmosis is the most globally employed desalination technology with half of 15 000 desalination plants employing RO processes Greenlee et
79. red at different times had agreeable acid titration curves from pH 4 6 Figure 4 7 The shape of the base titration curves agreed as well but the zeta potential of one sample dropped substantially between the acid and base titrations This occurred because the experiment was left overnight before the base titration was run allowing time for the acidic solution to remove some of the coating from the membrane Had the second titration been run immediately after the first the base titrations should have lined up almost exactly as observed in other titrations Zeta potential was also measured for an Ag coated membrane that had been stored for four months The surface charge of the old membrane was considerably lower than both the fresh Ag coated membranes and the uncoated SW30HR This could be a result of the NPs dissolving detaching from the membrane or being otherwise altered by the storage solution pH 5 5 Biological growth would result in a decrease in surface charge though the biocidal properties of nanosilver may mitigate such fouling The mechanism for zeta potential decrease after long term storage has not yet been identified 40 Figure 4 7 20 mg L Ag coatings compared to virgin SW30HR SEM analysis of membranes revealed aggregation of the Ag NPs in low magnification 2 500 magnification images Figure 4 8 Aggregates of up to 6 m were visible The size is considerable when compared to
80. rmine a threshold concentration The lowest effective coating concentration was desired to keep polyDADMAC coatings thin Figure 4 2 shows zeta potential titrations for concentrations of 2 1 and 0 2 on dry coated SW30HR membranes All concentrations produced a significant increase in zeta potential with a negligible difference between 0 2 and 1 concentrations and only slightly higher zeta potential for 2 concentrations The 0 2 concentration was selected for further experiments because of the effectiveness at a low concentration 34 Figure 4 2 PolyDADMAC coatings of 2 1 and 0 2 on dry SW30HR membranes All coatings impart a strong charge there is no notable difference between 0 2 and 1 and little difference at 2 4 2 2 TiO2 NP coatings TiO2 NPs which were functionalized by the manufacturer with polyDADMAC adsorbed directly to PA membranes changing the surface charge from negative to positive The zeta potential did not change when 0 212 mg L and 0 424 mg L of TiO2 was applied to SWC4 membranes Higher concentrations were applied to SW30HR membrane and did change the surface charge indicated by changes in single measurement values in Figure 4 3 The 21 2 mg L TiO2 solution imparted the smallest change and had the tightest error bars Error bars in Figure 4 3 suggest that the instrument may not be sensitive enough to accurately evaluate small changes in zeta potential incurre
81. s 77 Both X ray techniques are critical tools as the top 1 10 nm of the material can be analyzed compared to ATR FTIR which produced a spectrum dominated by the polysulfone membrane backing these techniques are highly sensitive Tang et al 2009a No interference is expected for either Ag or Ti detection due to the organic nature of the membrane structure functionalizing polymers and polyDADMAC coating Trace amounts of other elements in the NPs analyzed and reported by the manufacturer in the technical specification sheet would produce negligible if any response SW30HR the membrane used for most of the experiments had a commercially applied PVA layer that prevented significantly high 30 mV or low 30 mV zeta potential from developing at the membrane surface Though this property is beneficial in operating conditions the shielding of the PA surface layer decreased the electrostatic strength of membrane interactions and may have prevented coatings from being fully adsorbed or removed Electrostatically unfavorable conditions would likely develop on SWC4 and SWC5 membranes effects might not be observable on SWC5 in the pH range of titrations since the isoelectric pH is 3 but comparable experiments using SWC4 could confirm conclusions drawn from SW30HR experiments 7 2 3 Removal of coatings Base solutions of various strengths should be investigated in the future to determine whether foulants coul
82. s Continued Figure Page 4 12 SEM images of virgin SW30HR left and polyDAD MAC coated SW30HR right PolyDADMAC coating appeared to fill in the loop and valley str ucture of the membrane 46 4 13 ATR FTIR of polyDADMAC and negatively charged NP coatings did not indicate any change in surface chemistry 47 5 1 Acid cleaning of polyDADMAC coated SW30HR res ulted in significant but incomplete removal of the polymer 50 5 2 Acid cleaning of TiO2 coatings Acid cleaning resul ted in complete removal of NPs for all acid concent rations investigated 52 5 3 SEM images of uncoated SW30HR a Ag coated SW30HR b pH 1 acid cleaned Ag coated SW 30HR c and pH 13 base cleaned Ag coated SW30HR d 54 5 4 Electrokinetic behavior of Ag coatings after acid cle aning 55 5 5 Electrokinetic behavior of TiO2 coatings after acid cl eaning
83. s America Inc Dallas TX 3 3 4 Inductively coupled plasma optical emission spectroscopy ICP OES Nanoparticle solution concentrations were quantified using ICP OES Three samples of each NP working solution were dissolved in 5 HNO3 for 24 hours before analysis A 1000 ppm 1000 mg L titanium standard was used for Ti NPs and a multi element standard containing 10 ppm 10 mg L silver was used for Ag NPs Titanium NPs were diluted 1 1 into HNO3 for an expected concentration of 0 1 g Ag L the 0 1 g L working solution of Ag NPs was diluted to 0 5 for an expected final concentration of 0 5 mg Ag L Calibration concentrations are shown in Table 3 1 Table 3 1 Calibration concentrations for ICP OES analysis Ti ppm Ag ppm 0 0 10 0 01 25 0 025 100 0 05 250 0 1 3 4 Membrane characterization results Each membrane had a unique zeta potential curve and isoelectric point Breaks in the titration curve were caused by the separation of acid and base titrations which resulted in two sets of data for the pH range 3 9 Each data point is the averaged value from four measurements the average standard deviation for titration curves was 2 mV and consistent throughout the experiments Figure 3 6 shows the characteristics of each membrane involved in this study SWC5 and SWC4 show similarities in zeta potential at 23 operational pH for desalination approximately pH 8 but the charge of SWC5 i
84. s much lower than SWC4 at acidic pH The isoelectric point of SWC5 is roughly 3 while SWC4 has an isoelectric point of about 4 7 This indicates a significant difference in surface functionality of the two membranes though further characterization of the chemical differences was beyond the scope of this study SW30HR is commercially coated with polyvinyl alcohol which neutralizes many of the dangling carboxyl groups of the polyamide active layer and creates a more neutrally charged membrane over the pH range used in titrations 3 to 9 Figure 3 6 Zeta potential curves for SW30HR SWC4 and SWC5 TFC PA membranes Breaks in the titration curve resulted from the separation of acid and base titrations which provided separate sets of data Some variability was noted between samples of the same membrane demonstrated in Figure 3 7 by SW30HR samples This was an expected result as surface layer heterogeneity is a well reported property of commercially prepared RO membranes 24 Coronell et al 2011 Freger 2003 From August 2011 onward samples were cut from a dry membrane sheet and soaked in water overnight before analysis Titration data from old membranes soaked for more than a month and new membranes soaked for less than a week suggest that contact time with water affects the membrane surface chemistry with the surface becoming more negatively charged over time SWC5 was purchased as a full membrane
85. soaked in DDI overnight and coated in NP solutions under the same conditions for 2 hr TiO2 NPs were simply coated for 2 hr on the rocker table Concentrations used for coating were 20 mg Ag L Ag 3 7 mg Ti L TiO2 and 4 2 mg Ti L TiO2 All coated membranes were soaked in DDI water for at least 12 hours before analysis or further coating to allow dissociation of any material not strongly attached to the membrane surface Coated membranes were exposed to hydrochloric acid HCl or sodium hydroxide NaOH solutions to investigate the appropriate conditions for coating removal Initial removal experiments used strong HCl pH 1 and NaOH pH 13 for several hours Membrane integrity limitations were addressed for subsequent experiments and acid solutions of pH 1 2 3 and 4 HCl concentration 0 1 0 01 0 001 and 0 0001 M were used for no longer than 30 minutes as per manufacturer recommendations Control samples of virgin membrane were washed at each acid and base concentration to determine whether the caustic solutions altered the properties of the membrane Figure 3 8 Strong cleaning with pH 1 HCl significantly altered the membrane so 49 polyDADMAC and NP coatings were cleaned with pH 2 3 and 4 solutions for the majority of experiments 5 2 Coating removal results 5 2 1 PolyDADMAC coating removal PolyDADMAC was effectively removed from the SW30HR membrane surface at a pH below the membrane isoele
86. ta potential titrations at different stages of the study revealed significant variation in NP coatings that were applied using the same process It is suspected that Ti NPs may have aggregated in the stock solutions after being stored for several months resulting in a large variation in concentration between aliquots of NP solution While size analysis was reported by the manufacturer for the Ag solution and powdered TiO2 NPs verification of NP size in stock solutions would be beneficial Differences between values reported by the manufacturer and laboratory measurements may indicate whether aggregation occurred when dissolving NPs in DDI water Future stock solutions may be prevented from aggregating by sonicating solutions occasionally during storage and before aliquots are removed for membrane coating The NPs adsorbed to the membrane surface should be quantified Acid digestion of the membrane coupons can be used to release the Ti and Ag from the functionalizing polymer and the membrane surface these concentrations can be used to infer the quantity of inorganic components on the membrane Because both sides of the membrane are exposed to NPs during coating it must be taken into account that Ti or Ag values from 76 ICP analysis will include NPs bound to the backing layer as well as the active layer This method does not address how many inorganic components are present in each NP or how NPs are distributed on t
87. tact angle until reaching a critical volume at which point the area will decrease with a constant contact angle To account for both aspects of this phenomenon contact angle measurements often include additions and subtractions of drops bubbles to find the constant contact angle Drelich et al 1996 Apolar compounds can also be used to serve as an indicator of hydrophobicity for RO membranes large contact angles indicating a decrease in hydrocarbon affinity would be desired Brant and Childress 2002 Subramani and Hoek 2008 13 CHAPTER TWO RESEARCH OBJECTIVES This project was designed to evaluate the interactions between reverse osmosis membranes and nanoparticle coatings and to determine whether the coatings could be removed from the membrane The specific objectives were 1 Determine appropriate conditions for self assembly of polymer and nanoparticle coatings on reverse osmosis membranes Experimental conditions such as working concentrations of polymer and nanoparticles time required for coating and methods for applying coatings needed to be identified 2 Determine appropriate conditions for coating removal Two possible desorption scenarios were anticipated desorption of the nanoparticles from the polyDADMAC coating and desorption of the polyDADMAC coating from the membrane Experiments investigated whether high or low pH cleaning was more effective for coating removal 3 Use zeta potent
88. ted a concentration of 5 5 mg L for TiO2 and 55 mg L for TiO2 Table 3 2 29 Disagreement between expected values and ICP results was most likely caused by incomplete dissolution of TiO2 NPs An acid or microwave digestion would be needed to completely dissolve Ti NPs Data suggest that TiO2 NPs are more soluble in HNO3 than TiO2 NPs Table 3 3 ICP data for Ti standards and triplicate working solutions diluted to 50 Sample Signal Intensity Ti mg L Average Ti mg L TiO2 1 286325 5 07 5 4 1 5 TiO2 2 368288 6 96 TiO2 3 242744 4 06 TiO2 1 2525300 56 7 54 9 1 8 TiO2 2 2441280 54 8 TiO2 3 2370068 53 1 Ti Std 0ppm 101 0 0 N A Ti Std 10ppm 456409 10 0 Ti Std 25 ppm 1150798 25 0 Ti Std 100 ppm 4583184 100 0 Ti Std 250 ppm 10836125 250 0 30 CHAPTER FOUR COATING CONDITION OPTIMIZATION The goal of this chapter is to describe the conditions required for effectively coating RO membranes with NPs Each coating material was prepared at several different concentrations and the effects were examined by electrokinetic measurements in the SurPASS Once optimal concentrations were determined coating times were varied to determine how quickly coatings would assemble NP concentrations are reported as mass of the metal component Ti or Ag per liter as the manufacturer did not provide more specific particle analysis A mass based measure
89. ted sample concentration of 0 5 mg Ag L A miscalculation regarding the final anticipated Ag concentration resulted in a calibration curve 10x lower than necessary The calibration curve had a high R2 value despite being too low in concentration and was used for preliminary Ag concentration analysis Based on the calibration curve and instrument response for Ag samples Table 3 3 the NPs appear to have dissolved completely in 5 HNO3 ICP results indicate that Ag concentrations were 0 47 0 004 mg Ag L which was in excellent agreement with the expected 0 5 mg Ag L Table 3 2 ICP data for Ag standards and triplicate working solution diluted to 0 5 Sample Signal Intensity Ag Average Ag Ag 1 9911 0 46 0 47 0 004 Ag 2 10077 0 47 Ag 3 10042 0 47 Ag Std 0ppm 76 13 0 00 N A Ag Std 0 01 ppm 120 04 0 01 Ag Std 0 025ppm 449 09 0 025 Ag Std 0 05ppm 1001 10 0 05 Ag Std 0 1ppm 2078 54 0 10 A 1 g L titanium standard was used for calibration allowing Ti NP stock solutions to be diluted 1 1 in 5 HNO3 with a final concentration of 2 5 HNO3 for ICP analysis Concentrations were expected to be similar for the positive and negative NPs stock solutions were calculated as 212 mg Ti L and 184 mg Ti L for TiO2 and TiO2 respectively For the diluted solutions ICP data were expected to yield approximately 100 mg L for TiO2 and 90 mg L for TiO2 However ICP results indica
90. temperature has not changed input different temperature and then return to actual temperature 1 minute later When you are reading data this confirms that temperature was monitored although not changing o Check and record gauge pressure just prior to Sepa cell _______________________________________________________________ ____ Turn off program o Set actuator valve to completely open 10 Hz o Set pump to zero in Labview o Turn OFF pump inverter to stop pump o Use the STOP button on Labview screen to stop data recording Cleaning must be done after fouling The system should be sitting in DI from the previous run Run the DI out until water level is just above pump After fouling Fill small feed tank with 1 liter of pH 12 sodium hydroxide solution 0 4 g into 1 L Run through for 10 minutes in recycle mode to remove foulants then drain to just above the level of the pump Run through at least 20 L of DI water without recycle Make sure conductivity of the exit stream ends at close to conductivity of DI water o Example input of 25 mS DI water and output of 29 mS General maintenance Fill small feed tank with 1 liter of 10 phosphoric acid H3PO4 solution 117 ml of 85 H3PO4 added to DI water to make 1 liter Run through for 20 minutes 10 minutes for H3PO4 in recycle mode to remove rust and particulates Note that after adding to DI the acid concentration is lo
91. tion The method was significantly more reliable than using a pipette due to the viscous nature of the concentrated stock solution PolyDADMAC is not UV active or fluorescent manufacturers could not recommend a simple laboratory experiment to quantify polyDADMAC so reported concentrations are based on dilution volumes 3 5 2 Nanoparticle concentration determination using ICP OES NP working solution concentrations were measured by ICP OES using Aristar Plus multielement standard for trace metal analysis which contained 10 mg L Ag BDH Ltd Radnor PA lot C2 MEB296061 Excellent calibration curves were obtained for both materials with an R2 value of 0 99991 for Ag calibration and 0 99946 for Ti calibration Silver NPs were received in aqueous suspension and diluted into DDI water to make working solutions Early use of a high concentration of Ag NPs resulted in aggregation which was observable on low magnification SEM images TiO2 NP concentrations were prepared from powdered NPs diluted into DDI water These NPs were observed adhered to glassware used for stock solution preparation so not all of the material was fully dissolved into the stock solutions 28 The 0 1 g L silver NP working solution was first diluted by 50 resulting in an expected concentration of 50 mg Ag L This sample concentration was far greater than the highest standard so the 50 mg L Ag samples were diluted by 100 for a final anticipa
92. urface modification typically makes the surface more hydrophilic since lower fouling potential is generally correlated to greater hydrophilicity Ba et al 2010 Louie et al 2006 Over the past decade researchers such as Kilduff and Belfort have made significant developments in surface modification developing a graft polymerization technique that employs UV light or plasma to create reactive sites on the membrane surface that are then exposed to a coating material creating permanent chemical bonds between a membrane and the polymer coating Zhou et al 2009 Kilduff et al 2003 Minghao et al 2012 These modifications tend to be highly specialized targeting certain 7 chemical or biological species types instead of serving as a one stop filtration unit and reduction of membrane fouling has been questioned van der Bruggen et al 2008 1 5 Removable coatings Permanent surface modification is a useful means of changing fouling behavior but even the most resistant membranes can eventually be fouled This is especially true for feed waters like wastewater and surface water which contain a wide variety of organic and inorganic material An ideal removable membrane coating would reverse fouling no matter the foulant characteristics This would reduce the impact of membrane cleaning leading to longer lasting membranes and lower overall system cost The use of NPs in membrane systems has been reported in the l
93. ving 83 flux after pH 3 HCl cleaning compared to 74 and 69 of DI water flux respectively With respect to Figure 6 7 the average flux of the SWC5 DI flux is considered to be 100 throughout the experiment such that the 83 flux of the Ag coated membrane is relative to the SWC5 membrane after four hours of membrane compaction Table 6 2 Average flux normalized to SWC5 DI water experiment after each stage of RO filtration Averaged Normalized Flux Membrane Initial Flux lmh DI water Fouling Post cleaning Final SWC5 122 0 907 0 664 0 639 0 690 SWC5 pD 113 0 958 0 751 0 684 0 739 SWC5 pD Ag 112 0 954 0 751 0 766 0 827 69 Figure 6 7 Full cycle RO experiments PolyDADMAC and Ag coatings had lower initial flux and slower fouling Ag had notable flux improvement after cleaning Overall the coating concept showed promise in the RO experiments PolyDADMAC and Ag coated membranes did not significantly decrease the flux of the membrane Fouling occurred to a lesser extent with less flux decline on coated membranes Nanoparticle coated membranes showed notable flux improvement after simple acid cleaning Future experiments need to be performed in saltwater conditions so salt rejection the most important aspect of desalination membranes can be considered Coatings should be re applied in the RO system to determine feasibility of the regenerable coating concept 70 CHAPTE
94. viron Prog 2005 24 4 410 416 Gao L McCarthy T J Contact angle hysteresis explained Langmuir 2006 22 14 6234 6237 Goosen M F A Sablani S S Al Hinai H Al Obeidani S Al Belushi R Jackson D Fouling of reverse osmosis and ultrafiltration membranes A critical review Sep Sci Technol 2005 39 10 2261 2297 Greenlee L F Lawler D F Freeman B D Marrot B Moulin P Reverse osmosis desalination Water sources technology and today s challenges Water Res 2009 43 2317 2348 Harrison C J Le Gouellec Y A Cheng R C Childress A E Bench scale testing of nanofiltration for seawater desalination J Environ Eng ASCE 2007 133 11 1004 1014 Herzberg M Kang S Elimelech M Role of extracellular polymeric substances EPS in biofouling of reverse osmosis membranes Environ Sci Technol 2009 43 4393 4398 Hong S K Elimelech M Chemical and physical aspects of natural organic matter NOM fouling of nanofiltration membranes J Membr Sci 1997 132 2 159 81 Hurwitz G Guillen G R Hoek E M V Probing polyamide membrane surface charge zeta potential wettability and hydrophilicity with contact angle measurements J Membr Sci 2010 349 349 357 Jadav G L Aswal V K Singh P S SANS study to probe nanoparticle dispersion in nanocomposite membranes of aromatic polyamide and functionalized silica nanoparticles J Colloid Interf Sci 2010 351 1 304 314 Jeong B H Hoe
95. vs 100 mg Ag L Acid cleaning resulted in a zeta potential drop while base cleaning resulted in an increase in zeta potential When the quantitative change in zeta potential was compared for acid and base cleaning of Ag and TiO2 56 coatings the results were very similar These similarities suggested that the functionalizing polymer component of the composite NP polyacrylate rather than the inorganic component Ti or Ag contributed most significantly to NP behavior on the membrane surface Unlike with other coatings electrokinetic behavior after acid cleaning of TiO2 coatings appeared to be directly related to acid strength Figure 5 5 A pH 4 cleaning solution did not change the acid titration zeta potential values and a drop in zeta potential for the base titration was attributed to leaving the membrane in acid overnight before running the second half of the titration A pH 3 cleaning solution decreased the zeta potential slightly for both titrations A pH 2 cleaning solution produced the largest change but still did not return the acid titration curve to the SW30HR values the zeta potential remained high at low pH and dropped far below the SW30HR zeta potential during the base titration indicating that NPs were still bound to the membrane and contributing additional negative charge at high pH 57 Figure 5 5 Electrokinetic behavior of TiO2 coatings after acid cleaning 58 CHAPTER
96. wer than 10 and that is fine Do not leave running much more than 20 minutes 15 minutes for H3PO4 Run through at least 28 l of DI water without recycle Make sure conductivity of the exit stream ends at close to conductivity of DI water 89 o Example input of 25 mS DI water and output of 29 mS o Example input of lt 5 mS DI and output of 8 9 mS Leave system full of DI water Shutdown Turn off program then shut down system Remove membrane from cell Cut membrane samples for electrokinetic analysis and refrigerate in DDI water If it s late and you won t be doing cleaning today run 20 L DI water through the system Leave system in DI water over night Rinse conductivity probe with DI water Leave to dry 90 REFERENCES Al Amoudi A Williams P Mandale S Lovitt R W Cleaning results of new and fouled nanofiltration membrane characterized by zeta potential and permeability Sep Purif Technol 2007 54 234 240 Auffan M Rose J Bottero J Y Lowry G V Jolivet J P Wiesner M P Towards a definition of inorganic nanoparticles from an environmental health and safety perspective Nature Nanotechnol 2009 4 10 634 641 Ba C Ladner D A Economy J Using polyelectrolyte coatings to improve fouling resistance of a positively charged nanofiltration membrane J Membr Sci 2010 347 1 2 250 259 Bae T Kim I Tak T Preparat
97. yDADMAC coated membranes TiO2 NPs were adsorbed onto a 0 2 polyDADMAC binding layer Initial experiments used 18 4 mg Ti L 9 2 mg L and 1 8 mg L As with Ag NPs polyDADMAC coated membrane coupons were coated with NP coating solution at least 12 hours after polyDADMAC coating Samples were coated with TiO2 for 24 hours then rinsed for 30 seconds before soaking in DDI water at 4 C at least overnight 4 1 5 Coating controls Coating controls were prepared to confirm that binding of the coating layers was electrostatic Controls included TiO2 exposed to a polyDADMAC coated membrane TiO2 exposed to a virgin membrane and Ag exposed to a virgin membrane NPs were not expected to bind in control experiments due to electrostatic repulsion 4 2 Coating concentration results 4 2 1 PolyDADMAC coatings PolyDADMAC coatings adsorbed strongly to the PA membranes causing a significant increase in zeta potential Figure 4 1 The polyDADMAC coating resulted in 33 a positive zeta potential over the entire pH range of the titration 3 to 9 The increase in zeta potential at the low end of the pH range can be attributed to high proton density at the membrane surface Figure 4 1 PolyDADMAC coating on SWC4 membrane The polymer coating significantly altered the surface chemistry creating a positively charged membrane surface Varying concentrations of polyDADMAC were applied to SW30HR membranes to dete
98. yer while SWC4 and SWC5 are uncoated Tang et al 2009b 15 NH2 H2N a O O O Cl Cl Cl b Figure 3 1 Structures of m phenylenediamine a and trimesoyl chloride b the precursors to PA TFC surface layers C O H N C O C NH O H N C O C n O H N C O O O 1 n Figure 3 2 Structure of cross linked polyamide SW30HR and SWC4 were obtained as flat sheets prior to 2009 and stored dry in cardboard tubes to protect from light Membrane material was cut from the flat sheet wetted with DDI water and stored at 4 C in DDI water that was changed at least biweekly SWC5 was purchased as a complete tested RO module in late 2011 and stored at 4 C in DDI water once opened Coupons for experiments were cut and placed in DDI water then stored at 4 C at least overnight and up to several weeks with DDI water changed weekly 16 3 2 Coating materials 3 2 1 PolyDADMAC positively charged polymer PolyDADMAC was selected for its low cost and widespread availability as well as its strong positive charge It is widely used as a flocculent in water treatment and if effective would be easy to implement The monomer structure is shown in Figure 3 3 PolyDADMAC has been demonstrated to interact with polyamide membranes in favorable ways regarding these research goals including deposition on the membrane in thin layers and minimal flux reducti
99. zeta potential over a range of pH values While virgin SW30HR had an isoelectric point 4 the polyDADMAC coated membrane was completely positive over the pH range tested 3 to 9 When low pH cleaning was used polyDADMAC was partially removed from the membrane indicated by an isoelectric point of 4 5 NP coatings also resulted in changes to the membrane isoelectric point which was more closely recovered after low pH cleaning A secondary method for confirming coating and removal was not found which is an excellent dilemma coatings are thin enough to be invisible to scanning electron microscopy SEM and infrared spectroscopy analysis but are definitely present as indicated by significant zeta potential changes induced by extremely small quantities of material It is expected that X ray photoelectron spectroscopy XPS would effectively detect and quantify coatings Bench scale RO experiments were run to test coating efficacy in an applied system Flux was monitored for deionized water sodium alginate fouling and acid cleaning A second deionized water run was performed to determine flux recovery after fouling and cleaning Membranes were coated similarly to those for electrokinetic studies but on a larger coupon RO experiments resulted in 69 recovery of the DI water flux after cleaning a fouled membrane with no coating compared to a flux of 83 recovery when a NP layer was used iv DEDICATION I would

Smith, M.A. "Functionalized nanoparticles as

Contents

Download Pdf Manuals

Related Search

Related Contents

Smith, M.A. &quot;Functionalized nanoparticles as

Contents

Download Pdf Manuals

Related Search

Related Contents

Smith, M.A. "Functionalized nanoparticles as