Repositorio Institucional de Documentos

0 Erasmus Mundus Master in Membrane Engineering 2011-2013 4th semester Membrane Filtration for wastewater treatment applications Papoutsoglou Dimitra Project supervisor: prof. Chris. Paraskeva Module supervisor: prof. Reyes Mallada Patras, July 2013 1 The EM3E Master is an Education Programme supported by the European Commission, the European Membrane Society (EMS), the European Membrane House (EMH), and a large international network of industrial companies, research centres and universities Visit EM3E internet site: www.em3e The EM3E education programme has been funded with support from the European Commission. This publication reflects the views only of the author, and the Commission cannot be held responsible for any use which may be made of the information contained therein. 1. Contents Prologue .............................................................................................................................................. 5 Abstract ............................................................................................................................................... 5 Nomenclature ..................................................................................................................................... 6 PART A: Dye industry wastewater treatment ..................................................................................... 7 1. Theoretical background .............................................................................................................. 7 1.1 Paint industry wastewater .................................................................................................. 7 1.2 Coagulation – Flocculation .................................................................................................. 7 1.3 Polyelectrolytes ................................................................................................................... 9 2. Experimental procedure ............................................................................................................. 9 2.1 Characterization of wastewater .......................................................................................... 9 2.2 Coagulation/Flocculation .................................................................................................. 10 2.3 PAC optimization ............................................................................................................... 11 2.4 PDADMAC optimization .................................................................................................... 12 2.5 Membrane Filtration ......................................................................................................... 16 2.6 Ultrafiltration .................................................................................................................... 17 2.7 Reverse Osmosis ............................................................................................................... 18 2.8 Conclusions of Part A ........................................................................................................ 19 PART B: Carbon nanotubes and polymeric membranes ................................................................... 21 3. Theoretical background ............................................................................................................ 21 3.1 Carbon nanotubes ............................................................................................................. 21 3.2 Modification ...................................................................................................................... 22 3.3 The four mechanisms model............................................................................................. 22 2 3.4 Computer simulations ....................................................................................................... 24 3.5 Sonication.......................................................................................................................... 24 3.6 Carbon nanotubes dispersion ........................................................................................... 25 4. Experimental Procedure ........................................................................................................... 26 4.1 Carbon nanotubes ............................................................................................................. 26 4.2 Dispersion of carbon nanotubes solutions ....................................................................... 26 4.3 Deprotonation ................................................................................................................... 28 4.4 Ultra sonication ................................................................................................................. 29 5. Membranes embedded with carbon nanotubes ...................................................................... 30 5.1 Polyvinylidene Fluoride (PVDF) membranes ..................................................................... 30 5.2 Sonication effect ............................................................................................................... 33 5.3 PES/PET membranes ......................................................................................................... 33 5.4 Optimization of PES/PET embeddement .......................................................................... 35 5.5 Permeability of deprotonated CNTs incorporated from PES and PET sides of UP150 membranes ................................................................................................................................... 38 5.6 PES membranes with CNTs with Spin coating method ..................................................... 40 5.7 Conclusions of Part B ........................................................................................................ 47 Literature .......................................................................................................................................... 48 Index of Figures Figure 1 Jar test samples varying PAC concentration, sedimentation (PAC varying flocculant 10 mg/L, 2h, pH 12) .............................................................................................................................................. 11 Figure 2 ζ-potential, COD and TS variation in function of polyelectrolyte concentration (PAC varying flocculant 10 mg/L, 2h, pH 12) .............................................................................................................. 12 Figure 3 Structure formula of poly (dimethyl diallyl ammonium chloride) (6) ..................................... 12 Figure 4 Proton binding isotherms of PDADMAC at five different ionic strengths: ○ 0.01M □ 0.05M  0.10M 0.5M ◊1.00M (6) .................................................................................................................... 13 Figure 5 COD in function of pH for PDADMAC optimization experiments (500mg/L PDADMAC, 2h sedimentation, varying pH) ................................................................................................................... 14 Figure 6 ζ-potential in function of pH for PDADMAC optimization experiments (500mg/L PDADMAC, 2h sedimentation, varying pH) .............................................................................................................. 14 Figure 7 Particle size in function of pH for PDADMAC optimization experiments (500mg/L PDADMAC, 2h sedimentation, varying pH) .............................................................................................................. 15 Figure 8 Total solids size in function of pH for PDADMAC optimization experiments (500mg/L PDADMAC, 2h sedimentation, varying pH)........................................................................................... 15 Figure 9 Dye industry wastewater (sample 3) treated with 500mg/L PDADMAC, stirring and after 2 hours of sedimentation ......................................................................................................................... 16 Figure 10 Image Laboratory ultrafiltration unit .................................................................................... 17 3 Figure 11 Laboratory reverse osmosis unit ........................................................................................... 19 Figure 12 Different types of pristine carbon nanotubes and inner diameter (9) ................................. 22 Figure 13 Evolution of UV–vis spectra of an aqueous 0.1 wt.% MWCNT–0.15 wt.% SDS solution as a function of sonication time at continuous power of 20 W (solutions are diluted by a factor of 150) . 27 Figure 14 CNTs dispersion in water solutions after 5months (Thin Multi Walled CNTs where 1) ThinMW-COO2)Thin-MW-COOH 3) Thin-MW-COOH and surfactant 4) Thin-MW-COOH 5) Thin-MW and surfactant & Singe Walled CNTs 1) SW-COOH and surfactant 2) SW-COO0 and surfactant 3) SW-COO4) SW-COOH .......................................................................................................................................... 28 Figure 15 Deprotonation method by (a) smashing dried CNTs -COOH (b) washing CNTs with water (d) rinsing with NaOH solution (d) filtrated CNTs in ultrafiltration unit with polycarbonate filter............ 29 Figure 16 Sonication pin over ultrafiltration unit in laboratory............................................................ 30 Figure 17 PVDF membranes embedded with carbon nanotubes solution of concentration: 0.1μg/L (b) 0.3 μg/L (c) 0.5 μg/L (d) 0.7 μg/L (f) 0.9 μg/L (e) 1.1 μg/L (f) 1.3 μg/L .................................................. 31 Figure 18 Ultrafiltration unit in laboratory (pressure supply, vessel, and membrane) ........................ 31 Figure 19 Flux increase (%) in function of CNTs concentration in solution of embeddement ............. 32 Figure 20 UP150 PES/PET membrane geometry of pores and PES/PET layers in SEM image .............. 34 Figure 21 UP150 Active layer PES side cross section image by SEM. Thickness of active layer measured to 2.488μm........................................................................................................................... 34 Figure 22 Tip sonication and ultrafiltration configuration. Rectangular vessel contains CNTs solution which by pressure difference is driven through membrane pores ...................................................... 35 Figure 23 UP150 PES/PET membranes infiltrated with (a) SWCNTs and (b) Thin MW-COOby PES side .............................................................................................................................................................. 36 Figure 24 Images of cross sections of PET side of membranes infiltrated through the support, with Thin-MW-COOH or SW CNTs (10) ......................................................................................................... 37 Figure 25 Images of the surface exposed to the feed from membranes infiltrated through the thin selective layer side, with Thin-MW-COOH or Thin-MW-COO- ............................................................. 37 Figure 26 UP150 membrane infiltrated with Thin-MW-COOCNTs 0.312 μg/mL (1.6 μg/cm2 ) from PES side ................................................................................................................................................. 38 Figure 27 Permeability vs time plot of commercial and CNTs embedded MN membranes (com:commercial UP150, CNTs_PES:UP150 with CNTs incorporated from PES side) .......................... 39 Figure 28 Spin coater and phase inversion in water for PES-CNTs membrane preparation (a) drop by drop PES-CNTs mixture over spin coater (b) fast round movement 300m/min for some seconds (d) & (c) phase inversion in water .................................................................................................................. 41 Figure 29 Lab made membrane with phase inversion and spin coating method with bare PES 10%w/v (left side) and SWCNTs 0.5%wt. (right side) ......................................................................................... 42 Figure 30 Mixed matrix membrane PES/Thin MW-COO0.312μg/mL ................................................. 43 Figure 31 Mixed matrix membrane PES/SW-COOH 0.312μg/mL ......................................................... 44 Figure 32 Water Vapour Transmission Permeability Apparatus in Laboratory .................................... 45 Figure 33 GPC results for molecular weight cut off value of mixed matrix PES-CNTs membranes using 200, 100, 35 and 10 kDa ....................................................................................................................... 46 Index of Tables Table 1 Dye removal using inorganic coagulants with organic polymers as flocculant aids (2) ............. 8 Table 2 Properties of Wastewater Samples ............................................................................................ 9 4 Table 3 General requirement for reused water (2) .............................................................................. 10 Table 4 Experimental Conditions for PAC optimization ........................................................................ 11 Table 5 Summary of experiments for PDADMAC optimization in function of pH ................................ 13 Table 6 Experimental conditions and results for ultrafiltration ........................................................... 18 Table 7 Experimental conditions and results for reverse osmosis filtration ........................................ 19 Table 8 Carbon nanotube types, physical and cost parameters used in the project ........................... 26 Table 9 Measured permeability for distilled water through bare PVDF membranes and PVDF membranes embedded with different concentration of Thin MW-COOH CNTs .................................. 32 Table 10 Results of permeability change for distilled water in ultrafiltration unit before and after tip sonication .............................................................................................................................................. 33 Table 11 Suspension concentration and composition for PES/PET membranes embeddement ......... 36 Table 12 Water permeability results for PES-CNTs membranes made by phase inversion method.... 42 5 Prologue I well over remember the moment when my application to the EM3E was accepted and EM3E master committee announced a favourably advise on it. Since two years, me and all my colleagues did experienced an offspring program becoming reality and membrane engineering is ending to give finally its first alumni. Working on my master thesis, I mainly worked on dye industry wastewater treatment with conventional methods and membrane processes. We handled several unexpected difficult tasks mainly on membranes embedded with carbon nanotubes, a quite innovative field. In this manuscript, successful experiments and conclusions are reported. The subject raises high expectations and I conclude that it should steadily get across lab scale to large scale applications. I am grateful to professors Christakis Paraskeva, George Voyiatzis and Reyes Mallada who offered me the opportunity making my project on this subject and undoubtedly to all the people involved to my project, especially phD candidates of Patras University, Giannis Anastasopoulos, Spyros Kontos, Zagklis Dimitris, Eleni Moschopoulou. My warmer acknowledgements also are addressed to my EM3E colleagues, all my professors and master organizers and secretaries. Last, but not least, both my family and my husband who stand by me during Erasmus experience. This is the line dropped to express my gratitude to them. Abstract Membranes are used for particles or molecules separation in plethora processes of extraction, adsorption and filtration for liquid-liquid, gas-liquid, gas-solid, liquid-solid or gasgas separation. In sector of wastewater works, membranes are already commercial applied for reverse osmosis, ultrafiltration, nanofiltration and bioreactors. Crucial drawbacks in membrane filtration technology are both irreversible fouling of membranes and high energy dissipation usually applied for pressure drop. Innovative membranes with embedded carbon nanotubes can be an optimal solution to overpass these drawbacks. In industry, the first water treatment involves processes of a purely physical, mechanical and chemical nature to reduce the solid content. The technological evolution has led to 6 widespread mechanisation of the systems. The quality of water supplies has gradually declined largely because of excessive consumption of natural water and the abuse of ground soil as a recipient of wastewater. Pollution has also contributed to this effect. An existing physicochemical treatment is tried to be optimized controlling different parameters of sedimentation part and raw materials use. Wastewater from the industry is collected and characterised. The first batch is treated using a coagulation-flocculation process to remove large particles remaining in the sludge (primary treatment). The product is completely separated in supernatant solution and sludge. Supernatant is sampling for different flocculants concentration. A new polyelectrolyte (PDADMAC) is tested in terms of alkalinity and the results are compared with current electrolyte (PAC). All samples before and after the process are characterised using zeta size and particle size measurements together with COD and total solids measurements. The supernatant from primary treatment process is fed to ultrafiltration unit (UF) (secondary treatment). The permeate outlet from UF is fed to nanofiltration (NF) where almost all organics and solids are removed. The final NF permeate shows very high purity containing only mono and divalent ions of salts. A tertiary treatment refines the water product of nanofiltration membranes using commercial polymeric membranes with embedded aligned carbon nanotubes (CNTs). Before applying the hybrid membranes a full detailed study has been carried out in terms of CNTs embedment, water and wastewater flux permeability and SEM microscopy. Polymeric membranes embedded with CNTs have been made in the laboratory with spin coating method and by ultrafiltration unit size exclusion is tested using PEG solutions. Nomenclature DWCNTs Double Walled Carbon Nanotubes CNTs Carbon Nanotubes CVD Carbon Vapor Deposition GPC Gel Permeation Chromatography MD Monte Carlo Simulation MTWW Model Textile Waste Water MWCO Molecular Weight Cut Off PAC Poly(aluminum chloride) 7 PDADMAC Poly(diallyldimethylammonium chloride) PEG Polyethylene glycol SDS Sodium Dodecyl Sulfate TS Total Solids PART A: Dye industry wastewater treatment 1. Theoretical background 1.1 Paint industry wastewater The main sources of wastewater in dye industry are tanks of dye solution (baths), reflux water, washing water and run-off rain water. In the current work, treated waste is mainly the washing waters of the polymerization tanks and contains the products of the industry diluted, which are polyvinyl acetate and poly(acrylic esters). The concentration of pollutants can widely vary and quantities generally differ also depending on the batch. The different sorts of pollutants can be divided in organic, inorganic molecules and metal ions. Some compounds are highly toxic and mutagenic and many studies have shown relation between this kind of wastes and carcinogenic amines (1), (2). The presence of residual chemicals is strongly undesirable especially if wastewater is disposed in natural sources due to depletion of the dissolved oxygen. Dye industry wastewater can contain impurities, dispersed solids, additives such as sodium chloride, sodium carbonate, sulphate cellulose etc changing in batch mode. Organic molecules containing in the dye wastewater can be chromophore, auxochrome and also heavy metals as chromium, lead, iron, aluminium. In common practice, it is not a matter of only one step dye wastes purification and more than two steps are usually used (2). 1.2 Coagulation – Flocculation Coagulation is the destabilization of solution via minimizing zeta potential. Coagulants can be classified into two main categories: metal coagulants and polymers. Flocculation is the process of whereby destabilized particles from larger agglomerates due to surficial tendencies (2). Both practices are well known since hundreds of years (4). In dye industry 8 coagulation/flocculation offers a low cost treatment method and several coagulants/flocculants are applied (see Table 1). Table 1 Dye removal using inorganic coagulants with organic polymers as flocculant aids (2) Type of dyes Inorganic coagulants Type of polymer Condition Performance Real textile wastewater Aluminium based (2g/l) Cationic polymer – Cyanoguanidine – formaldehyde (500mg/L) Final pH: 5 40oC Mixing time =11min Settling time = 30min Colour removal=60% Turbidity removal =80% COD removal= 28% 100mg/L reactive blue STE Polyferric chloride Cationic (polyDADMAC*) (dosage of composite=20mg/L) Initial pH:7 Ambient temperature Mixing time=15min Settling time=12min Colour removal=90% Real wastewater from fabric dyeing industry Aluminium oxide, Al2O3 (1800mg/L) Cationic (polyDADMAC) (dosage of composite=30mg/L) Initial pH:5.75.9 Ambient temperature Mixing time=11min Settling time=30min Colour removal=69% Turbidity removal =99% Optimization of coagulation/flocculation process is an intriguing target given that industrial dyes prove high water solubility and consist of complex usually not well known substances. All the methods and systems must be designed for large scale applications. Many references show large differences above ferrous or aluminium flocculants. Aboulhassan et al. (5) have achieved a COD reduction of 91% with the use of FeCl3 combined with high molecular weight polyelectrolytes (flocculants). Other coagulants like FeSO4, Al2(SO)4, and 15 Figure 7 Particle size in function of pH for PDADMAC optimization experiments (500mg/L PDADMAC, 2h sedimentation, varying pH) In pH 6, particle size even increased to 824.3 nm. For pH, COD is still high (2157 mg/L) when TS decrease to 0.6mg/L and particle size is relatively well control (122.9nm). Interest is pointed to higher pH values, 10 and 12 where PDADMAC works in the industry. For pH 10, COD is 1370 mg/L slightly higher than 1295mg/L for pH 12. Particle size cannot be taken into account given that for both pH are measured higher than initial solution particle size (376.5nm and 356.3nm respectively) probably made by electrolyte molecules. Figure 8 Total solids size in function of pH for PDADMAC optimization experiments (500mg/L PDADMAC, 2h sedimentation, varying pH) 16 Results regarding to total solids and particle size cannot be totally trustable and organic load need to be the main benchmarking for electrolyte evaluation. In case of tertiary waste treatment, pH 10 can be a balanced choice requiring smaller amount of NaOH and saving chemicals in industrial scale applications. Figure 9 Dye industry wastewater (sample 3) treated with 500mg/L PDADMAC, stirring and after 2 hours of sedimentation Comparing PDADMAC and PAC results, polymeric polyelectrolyte and metal coagulant, organic load decrease 94% for PDADMAC instead of 80% for PAC coagulant, ζ potential is significantly destabilized for both cases. Particle size is, though, higher for PDADMAC case (356nm) than 187nm for PAC because of larger organic agglomerates and high molecular size of polymeric electrolyte. Total solids appear less for PDADMAC (2.3g/l) than PAC case (4.7g/l). In overview, polymeric electrolyte gives better results for wastewater purification via coagulation/flocculation process. 2.5 Membrane Filtration In literature, review paper present in abundance successful use of nanofiltration membranes in textile industry (1). Organic load and total solids can readily be removed by membrane modules given that waste is primary treated by coagulation/flocculation systems. The final target of a tertiary treatment as filtration membranes usually is related to recycling or disposal of the effluent. Organic load might differ from 30 to 100 mg/L and thus coagulation flocculation process does not fulfil disposal and reuse requirements (7). In next paragraphs, the results of membrane filtration are described. The pilotic systems are illustrated in the image below. The inlet of membrane step is provided by industrial wastewater treatment method. 17 2.6 Ultrafiltration Initially the waste (sample 3), which is the final effluent of the existing physicochemical treatment process of the industry with the use of poly(aluminium chloride) and anionic poly(acrylamide), is treated with an ultrafiltration unit, in different transmembrane pressures (TMP) from 2.5 to 4 bar. Concentration of total solids is about 5 mg/L, and the mean particle size is close to 80 nm thanks to the efficient coagulation/flocculation process that is preceded. Figure 10 Image Laboratory ultrafiltration unit UF (Figure 10 Image Laboratory ultrafiltration unit) is implemented only to play an auxiliary role before the use of a more effective membrane (NF or RO). Because the mean size of suspended particles is 80 nm, and the mean pore of UF membrane modulus is 100 nm, TS reduction is practically zero and only particles with size larger than 100 nm are removed. However, UF led to the removal of 20% of the organic matter. An unpleasant result is that the flux that is acquired is low for an ultrafiltration process, which might have been caused by clogging of membrane pores by the suspended polymers. A high flux rate is recovered 18 after a thorough cleaning of the membranes with a NaOH solution for half an hour (85% of the initial flux rate). Table 6 Experimental conditions and results for ultrafiltration Parameter Initial solution Transmembrane pressure (ΔP) in filtrated solution 2.5 bar 3 bar 3.5 bar 4 bar COD mg/L 950 ±17.3 818 ±15 765 ±55 828 ±10.8 792 ±148 T.S. g/l 4.93 ±0.07 4.82 ±0.28 5.18 ±0.08 5.14 ±0.23 5.04 ±0.06 2.7 Reverse Osmosis The permeate stream of ultrafiltration is fed to a reverse osmosis unit. The results are very encouraging with COD and TS reduction being around 90−100%. Initial COD values at 960 mg/L are reduced to 16 mg/L after treatment with the RO process. TS is reduced dramatically and their value in permeate stream do not exceeded the value of 0.5 mg/L. On the other hand, irreversible fouling phenomena are observed, as the flux of clear water is not the same before and after the treatment of the waste, even after chemical cleaning of the membrane module. Implementation of membranes shall take place after a complete evaluation of all operational parameter values and of the problems related to their performance (flux decline, cleaning procedure, long-term behavior, etc). 19 Figure 11 Laboratory reverse osmosis unit Table 7 Experimental conditions and results for reverse osmosis filtration Parameter Initial Solution Transmembrane pressure in filtrated solution (bar) 10 bar 20 bar 30 bar 40 bar COD mg/L 792±148 16±25 76±34 83±32 43±21 T.S. g/L 5.04±0.06 -0.09±0.07 0.23±0.04 0.39±0.18 0.42±0.25 2.8 Conclusions of Part A Through a parametric study, the optimization of the existing physicochemical treatment process of a paint industry has been carried out, in terms of coagulant concentration and pH. In the first set of experiments, the coagulants and flocculants implemented are the ones that are currently in use in the wastewater treatment unit of the factory. It is found that higher removal efficiency could be achieved if lower concentration, compared to the one used by the industry, of polyelectrolytes is used (400 mg/L instead of 1000 mg/L). By working at higher concentrations of positive polyelectrolytes the negative ζ potential value of suspended particles is inverted to positive, stabilizing again the solution and the particles 20 remained under suspension. It is proved that the coagulant, if used in excess, can have negative effects on the separation. Moreover, a new polyelectrolyte that can cause the coagulation and flocculation of the suspended solids is proposed, which leads to higher COD and TS reduction than the reductions possible with the existing process. An extra step is taken for the treatment of the waste with the implementation of membrane technology. Ultrafiltration did not alter significantly the waste but removed all the larger suspended solids, preparing it for the step of reverse osmosis. With the use of a reverse osmosis membrane, the organic content of the waste is dramatically reduced to a value of around 30 mg/L, and the final effluent is suitable for recycling, irrigation, or disposal to water banks. Fouling phenomena are apparent, but further experiments must take place in order to find the experimental conditions that minimize such problems. 21 PART B: Carbon nanotubes and polymeric membranes 3. Theoretical background Since thirty years of research, carbon nanotubes abbreviated by CNTs acronym are dominating in the new materials world and they are revealing an enormous potential of hundred different possibilities and applications. Andre Geim and Konstantin Novoselov were awarded 2010 Nobel Prize for their work on carbon sheet of graphene and this is not a matter of chance but fact of a real new age beginning for material engineering. Membranes of different polymers embedded with clouds of single or multi walled carbon nanotubes raise new efficiency standards and expectations. The innovative membranes can be considered as a good alternate solution for both tap water and wastewater treatment via already applied processes as ultrafiltration, nanofiltration and reverse osmosis. 3.1 Carbon nanotubes Carbon nanotubes (CNTs) are hollow and more than 50.000 times thinner than a human hair (15). CNT is simply a nanometer-sized rolled-up atomically smooth graphene sheet that forms a perfect seamless cylinder capped at the ends by fullerene caps (8). A single-walled carbon nanotube (SWCNT) is a single graphene sheet rolled into a seamless cylinder with either open or closed ends. Multi-walled carbon nanotubes (MWCNTs) are two or more concentric cylinders of graphene sheets of successively larger diameter, forming a layered composite tube bonded together by van der Waals forces, with a distance of approximately 0.34 nm between layers. In the market, average diameter of a single-wall carbon nanotube typically ranges of 0.6 nm to 100 nm. The aspect ratio, i.e., length to diameter, typically ranges from 100 to 1000. A nanotube of 2 nm diameter has a length of 100 to about 1000 nm. In preferred embodiments, the average length is from about 200 nm to about 1000 nm. 22 Figure 12 Different types of pristine carbon nanotubes and inner diameter (9) For reverse osmosis, and notably for water desalination, a preferred inner diameter range is about 0.4 nm to about 5 nm (Figure 12), and a most preferred range is from about 0.4 nm to about 1.2 nm. For nanofiltration membranes, a preferred size range is from about 1 nm to about 10 nm. For ultrafiltration membranes, a preferred size ranges from approximately 5 nm to about 200 nm (8). 3.2 Modification Carbon nanotubes are modified by alcoholic group –COOH in their external part leading to hydrophilicity nature and improving water contact on the surface of the membrane. Therefore modified CNTs offer improved electrostatic effect with metal ions containing in water and wastewater and more chirality effect with functionalized groups are appeared (11). Certain improvements are observed as a function of CNT functionalization and most importantly CNT volume fraction. It is proved in the literature that CNTs with larger diameter are more effective to remove organic molecules (11), (18). For water the pore diameter is related to the higher flow through the pores. Deprotonation reaction modifies COOH to COOwith negative charge. CNTs with COOfunctional group causes charge effects, improves stabilization of solution, has better water dispersion and probably rejection factor for some sorts of foulants and pollutants. 3.3 The four mechanisms model Mixed matrix membranes are created on the cross point of different physical phenomena enable to provide them with great separation properties. The inner cavity of CNTs forms a natural pore with very small diameter that can in some instances be smaller than 2 nm. This pore works as a both side passage for fluids and retain greater molecules by size exclusion mechanism and diffusion solution model explains water movement inside the pores. 23 Moreover, smooth hydrophobic surfaces of the nanotubes lead to nearly frictionless flow of water through them, enabling transport rates orders of magnitude higher than transport in conventional pores (11), (15). The structure of CNTs permits targeted specific modifications of the pore entrance without destroying the unique properties of the inner nanotube surface. Charge effects occur between functional groups attached carbon nanotubes edges contributing increasing molecules interactions (12). The combination of these three factors: size exclusion, hydrophobic environment and charge effects make mixed matrix membranes or membranes embedded with CNTs enable membranes efficiency in numerous applications improving rejection factor, extending lifetime and improving flux permeability (13). Another mechanism which appears quite often is sorption of molecules inside CNTs empty sites depending on the nature of molecules of pollutants. In details, the four mechanisms which act simultaneously are described below:  Size exclusion: Ultrafiltration applications The membrane is in this case a filter plate with holes (pores) that are too small for the particles to pass but big enough for the fluid to permeate easily. Permeation of a molecule through an ideally permeable membrane occurs without energy dissipation. In ultrafiltration, the carbon nanotube wall acts as a filter for all particles larger than 0.01 micron: pollen, algae, parasites, bacteria, viruses, germ and large organic molecules (15). In literature, CNTs membranes prove selectivity close to 100% for different molecules (15).  Surface interactions: Adsorption effects Filtration by CNTs pores occurs for the larger sized macromolecules, but sorption dominating for the medium molecular weight organics. On the surface of CNTs, free spaces provide adsorption sites for small molecules in solution (13). The data for the hydrocarbons show some evidence for sorption effects occurring inside the pores. In wastewater of paint industry metallic ions can be adsorbed or retained in these pores (12).  Ion exclusion: Charge effect Polar groups (eg –COOH) are attached on the surface of membrane or in the end of carbon nanotubes by chemical modification (see modification part) in order to improve charge interactions in case of charged molecules (metal ions). Polar groups enhances adsorption desorption and ion exchange in high pH solution (17) and increase the flux. Mesoporous membranes that have a charged pore surface in salt solutions may exhibit significant ion 24 retention by a space charge effect if pore size is smaller than the Debye length of the solution.  Water transition: Hydrophilicity-Hydrophobicity effects In case of mixed matrix membranes or membranes embedded with CNTs the water molecules just fit inside the pore and apparently have significant mobility with respect to the pore wall (18). By simulation models, water appears to move like a solid (19) inside the tube making a hydrogen bond wire wherein the hydrogen bonds literally move between the hydrophobic walls and they try desperately to escape. Water molecules inside and outside the nanotubes are in thermodynamic equilibrium (15). 3.4 Computer simulations The flux of charged and neutral molecules is studied thoroughly using modelling tools by Luca and Voyiatzis group in Institute of Chemical Engineering Sciences in Patras (19). The first question one may ask is why does water wet CNTs. The study showed that water flow is limited mainly by particle entry and exit events, and that tube length had hardly any effect. For far small SWNT studied by Hummer et al., carbon nanotubes of very small diameter (0.8nm) have so narrow passage that only a single water molecule could be inserted, forming a single file water chain (19). Hummer et al. further noted that this wall friction appears to be exceedingly small, as in the gas-diffusion case. Indeed, graphite is an industrial-grade solid lubricant. Monte Carlo simulations show that a defining feature of the water structure in CNTs is the formation of the hydrogen-bonded “water wires” oriented along the nanotube axis (12). 3.5 Sonication Sonication is the process of converting an electrical signal into a physical vibration that can be directed toward a substance and it is the act of applying sound (usually ultrasound) energy to agitate particles in a sample (21). Sonication effect enhances formation, growth, and implosive collapse of bubbles in a liquid. In the laboratory, it is usually applied using an ultrasonic bath or an ultrasonic probe, colloquially known as a sonicator. The primary part of a sonication device is the ultrasonic electric generator. This device creates a signal (usually around 20 KHz) that powers a transducer. This transducer converts the electric signal by using piezoelectric crystals, or crystals that respond directly to the electricity by 31 Figure 17 PVDF membranes embedded with carbon nanotubes solution of concentration: 0.1μg/L (b) 0.3 μg/L (c) 0.5 μg/L (d) 0.7 μg/L (f) 0.9 μg/L (e) 1.1 μg/L (f) 1.3 μg/L Figure 18 Ultrafiltration unit in laboratory (pressure supply, vessel, and membrane) The membranes have been embedded with CNTs solution of different concentrations from 0.1μg/l to 1.3μg/l. Embeddement with several concentrations of thin MW-COOH carbon nanotubes has been made at 0.2bar. The results are grouped in next figure. 32 Table 9 Measured permeability for distilled water through bare PVDF membranes and PVDF membranes embedded with different concentration of Thin MW-COOH CNTs Concentration (μg/mL) Permeability before CNTs [L/m2·h·bar] Permeability after CNTs [L/m2·h·bar] 0.1 5141 5126 0.3 5340 5340 0.5 5423 5066 0.7 4957 4773 0.9 5149 5103 1.1 5149 5315 1.3 5141 4901 Figure 19 Flux increase (%) in function of CNTs concentration in solution of embeddement By Figure 19, highest flux is observed at 0.5μg/L carbon nanotubes. Τhe flux increases 27% using carbon nanotubes in PVDF pores by tip sonication embeddement and water permeability increases 6.5%. By literature (23), low or super low concentration proves better result and main cause can be fouling effects inside the pores of polymeric membranes. Although 27% higher flux of PVDF membrane, the result cannot get beyond normal polymeric membranes efficiency or either in terms of cost and environmental risks. Repetition of experiment for 0.3μg/L tried to optimize the process and verify CNTs contribution. Under same experimental conditions, 0.2bar and thin MW-COOH CNTs three more PVDF membranes have been tested for 0.3μg/L CNTs concentration. The flux reduction is measured in ultrafiltration configuration and new results show no increase of -5 0 5 10 15 20 25 30 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Increase in permeability [%] CNTs concentration [μg/L] 33 flux in any case and either decrease of flow, approximately 4%. Fouling effects and typical measurement errors might be the source of insufficient results. 5.2 Sonication effect In order to evaluate tip sonication impact in pores geometry, a series of experiment with two different duration of sonication has been carried out. PVDF membranes have been sonicated for 2min and 7min. Flux has been measured in both cases and in both cases flux after sonication is lower. Consequently, there is not positive impact in flux increase by the tip probably because of spreading effect of sonication in membranes pores or crack effects. Table 10 Results of permeability change for distilled water in ultrafiltration unit before and after tip sonication Permeability before tip sonication [L/m2·h·bar] Permeability after tip sonication [L/m2·h·bar] 2min of sonication 4245 3995 7min of sonication 4233 4150 PVDF membranes shown poor results and further study on PVDF membranes embeddement is not imposed at this point. However other type of polymers need be tested in terms of CNTs fouling effect, permeability and pollutants exclusion. 5.3 PES/PET membranes Commercial PES/PET membranes are used for embeddement with CNTs. Microdyn Nadir Membranes UP150 purchased by Microdyn Nadir company have nominal properties as following: Ultrafiltration, PES side, 150kDa molecular weight cut off (MWCO), 40nm pore size, 200μm thickness and they combine PES and PET layers. PES side works like filter and PET is the support layer. Nominal water flux of UP150 is 200L/(m2h).Geometry of PES/PET in terms of pores is conical where PET has large diameter pores and PES smaller one. 34 Figure 20 UP150 PES/PET membrane geometry of pores and PES/PET layers in SEM image Target pore size is defined to 1.5-2nm and 0.1-3kDa MWCO. Appropriate carbon nanotubes need to have diameter in scale of target pore size as Thin MWCNTs functionalized with OH, COOH 1.0-6.5nm of internal diameter, DWCNTs 1-2nm of internal diameter and SWCNTs of 0.8-1.6nm internal diameter. Target pore size need to be in same size with internal diameter of carbon nanotubes. Due to UP150 conical structure, CNTs attached at PES are able to stay on PES side or might stack there. Otherwise, if CNTs move on PET side and PET side embeddement take place, CNTs cross the path inside PET-PES pore and appear on the PES side of the membrane. In the laboratory, experiments have already proved last argument. Figure 21 UP150 Active layer PES side cross section image by SEM. Thickness of active layer measured to 2.488μm The optimal case is to infiltrate CNTs across PES selective thin layer. However, due to anisotropic, sponge like character of the porosity of this thin layer, infiltration is quite demanding and thus optimization of the method is required. 35 The configuration of rectangular membranes embeddement consists of lab made vessel in scale of UP150 membranes 1600mL, pressure supply, ultrafiltration configuration, membrane, metallic sieve, sealer made by silicone. Figure 22 Tip sonication and ultrafiltration configuration. Rectangular vessel contains CNTs solution which by pressure difference is driven through membrane pores 5.4 Optimization of PES/PET embeddement Experiments have been carried out in the past also by PET side. In that case carbon nanotubes have emerged in PES side with 2μm distance and this is what called the active side of the membrane. In older experiments of the group, PES side and SWCNTs 2.5 μg/ml gave 14.6% increase in water flux permeance when for PET side and SWCNT’s 2.5 μg/ml percentage raised to 32.40%. In this project, different concentrations of CNTs are applied aim to optimize water flux and rejection properties of UP150 membranes. Thin MW COOCNTs and SW have shown fine dispersion behaviour, suspension concentration is 100μg/mL and CNT density is 384μg/cm2. Isopropanol water solution (1:4) is prepared to enhance opening pore of membranes. 36 Figure 23 UP150 PES/PET membranes infiltrated with (a) SWCNTs and (b) Thin MW-COOby PES side Before tip sonication of CNTs solution, SWCNTs are mixed with Sodium Dodecyl Sulphate (SDS) surfactant 1:1.5 and for Thin MW-COOdispersion experiments have proved no surfactant is needed. For SWCNTs tip sonication took 5 min in order to open CNTs paths and for Thin MW-COOtip sonication took 7 min. Pressure supply for embeddement is 0.5bar. Tip sonicator stand slightly high and CNTs solution is adding drop by drop in order to achieve better dispersion, opening and homogeneity of solution. Tap water is used (10mL containing 667μg CNTs, hence final solution is 0.312 μg/mL at 1600mL). Ultrafiltrating membrane with 300mL dispersion is embedded, thus final density on membrane surface is 1.6μg/cm2. Ultrafiltration of 300mL takes 2.5min and PES side starts change to grey because of CNTs. Table 11 Suspension concentration and composition for PES/PET membranes embeddement CNT type Surfactant Suspension concentration [μg/mL] CNT [μg/cm2] REMARKS Infiltration parameter from PES side Rectangular UP150 membrane 6.5x12cm Thin MWCOOSDS 0.312 1.6 5 items SWCNTS 0.312 1.6 5 items 37 Figure 24 Images of cross sections of PET side of membranes infiltrated through the support, with Thin-MW-COOH or SW CNTs (10) Figure 25 Images of the surface exposed to the feed from membranes infiltrated through the thin selective layer side, with Thin-MW-COOH or Thin-MW-COOIn Figure 25 the embeddement of carbon nanotubes inside the pores and over the PES area is well illustrated. A homogenous distribution of carbon nanotubes is also achieved. 38 Figure 26 UP150 membrane infiltrated with Thin-MW-COOCNTs 0.312 μg/mL (1.6 μg/cm2 ) from PES side Membranes incorporated with Thin MW-COOCNTs 50μg/mL can be covered with PVDF layer 0.1% w/v in MeOH have shown tendency to increase pure water permeability in comparison with commercial UP150 PES/PET membrane. Experiments are repeated in order to evaluate Thin MW-COOH CNTs geometry over PES layer. Tip sonication to the mixture took 25min in order to separate sufficiently carbon nanotubes. Surfactant used was PF127 1:1. Isopropanol solution (1:4 in water) removed glycerine layer over PES/PET membranes before ultrafiltration. PVDF coating tried to apply over PES/PET embedded membranes with spin coating – phase inversion method (see below). However process is simple and promising, PVDF does not show enough adhesion over PES and other methods might need to be studied. 5.5 Permeability of deprotonated CNTs incorporated from PES and PET sides of UP150 membranes The UP150 membranes with PES on the surface and PET as support layer were infiltrated with deprotonated thin walled CNTs (Thin MW-COO-) and were tested on laboratory OSMOTA test unit at HSKA, Karlsruhe, Germany for the pure water flux as well as for the flux with model textile wastewater (MTWW). The flux from the deprotonated CNTs incorporated inside PES and PET surface were compared to the commercial membranes under similar operating conditions of 1.5-2 bar pressure and 0.4 L/min of cross-flow velocity. For functionalized CNTs incorporated from PES side of UP150 results verified an increase of permeability in pure water relatively high (above 1400 L/m2·h·bar) initially. Need to point 39 out that a great loss of deprotonated CNTs was noticed while treating with glycerol as well in the solution from these set of CNTs embedded membranes. The pure water flux as well as the flux from the MTWW, tested on our laboratory unit was compared to the pure water flux from commercial membranes as shown in Figure 27. Figure 27 Permeability vs time plot of commercial and CNTs embedded MN membranes (com:commercial UP150, CNTs_PES:UP150 with CNTs incorporated from PES side) The permeability of deprotonated CNTs incorporated through PES side was around 409 L/m2·h·bar while that of commercial membranes is 323 L/m2·h·bar when treated and operated under similar conditions. The permeability with MTWW was around 160 L/m2·h·bar while that for commercial membrane was 80 L/m2·h·bar. Infiltration of functionalized CNTs from PES side showed significantly higher permeability when compared to the commercial membranes. However, the rejections of red and blue colour from UV-vis spectrometer measurements display 0.8 and 1.65 % respectively while that of commercial membranes was noticed to be 17.55 % for red while 22.35 % for blue respectively. Carbon nanotubes incorporated from PES side (small pore hydrophilic surface) give promising results in permeability. Losses of CNTs during permeability experiments need high concern in case of membrane use in potable water purification applications. SEM images 0 200 400 600 800 1000 1200 1400 1600 0:00 0:10 0:20 0:30 0:40 0:50 1:00 1:10 Perm. Pure water (com) Perm. Color (com) Perm. Pure water (CNTs_PES) Perm. Color (CNTs_PES) Time (hh:mm) Permeability (l/m2hbar) 40 pose important questions on CNTs penetration inside PES pores and thus further optimization is needed. Thin MW COOCNTs have slightly better consistency than SWCNTs given imaging characterization. Low concentration of embeddement solution is indeed an optimal choice at this point of study in terms also of cost saving in raw materials. 5.6 PES membranes with CNTs with Spin coating method Single walled and thin multi-walled COOcarbon nanotube and polyethersulfone blend membranes (also called as mixed matrix membranes) are synthesized via the phase inversion method (25). The resultant membranes are then characterized by scanning electron microscopy (SEM), gel permeation chromatography (GPC), ultrafiltration flux and water vapour permeability method. The mixed matrix membranes appeared to be more hydrophilic, with a higher pure water flux than the polyethersulfone (PES) membranes and give better size exclusion results for model foulants of PEGs. Therefore, it was noticed that the amount of CNTs in the blend membranes was an important factor affecting the morphology and their permeation properties. Polyethersulfone (PES) membranes are synthesized via the phase inversion method using spin coater to create PES thin layer. A PES solution is prepared with 7%w/v polyethersulfone in dimethylformamide (DMF) solvent. The mixture was well stirring for 2h at room temperature. Two different solutions of SWCNTs and Thin MW-COOare prepared with NMethyl-2-pyrrolidone (NMP) solvent to obtain 0.5% wt. CNTs respect to PES (0.035gr of CNTs for 7gr of PES in 100mL of solution). CNTs solution are mixed on stirring apparatus at room temperature and mixed are sonicated for good dispersion of CNTs. After dispersing CNTs in solvent, PES (20 wt. %) was dissolved in the dope solution by continuous stirring and heating at 60oC until the solution became completely dissolved and homogenous. Mixtures are finally mixed and well stirring overnight. 47 Three different membranes have been picked up by laboratory made PES membranes and tested in GPC apparatus. The results show an overall promising retention of high molecular weights and even the sample cannot be totally representative a clear tendance is illustrated (Figure 33). The retention of high MW PEGs is close to 100% for PES/Thin MW CNTs membranes (sample (a) and (b)). Sample (b) has a significant cut off on 35kDa. One only value cannon express retention it can though contribute to plan new experiments for further study. 5.7 Conclusions of Part B Measurements results report the performance of mixed matrix and asymmetric polyethersulfone ultrafiltration flat sheet membranes with carbon nanotubes. Three different types of membranes are prepared. Pure PES membranes, PES blended with SWCNTs membranes and PES blended with Thin MW-COOmembranes. The membranes are prepared by phase inversion process containing polyethersulfone (PES) as polymer (10% wt., N,N-dimethylformamide (DMF) as solvent of polymer, SWCNTs in N-Methyl-2-pyrrolidone (NMP) as a solvent (0.5%wt. respect to polymer) or Thin MW-COOCNTs in NMP (0.5%wt. respect to polymer). Blend membranes displayed a higher flux and slower fouling rate than the PES membranes. Subsequent analyses of the desorbed foulants showed that the amount of foulant on bare PES membranes was higher than the blend membrane for 0.5% SWCNTs content. Thus, the carbon nanotube content of membranes is shown to alleviate the membrane fouling caused by natural water. SWCNTs show also far higher water permeability than Thin MW-COOincorporated in PES membrane and in all the cases mixed matrix membranes had better efficiency than pure PES membrane. As the method can considered as comparatively easy and low cost, further study is needed to optimized the different amount of components and evaluate the scale up possibility. 48 Literature (1) D. Zagklis, P. Koutsoukos, C. Paraskeva, A Combined Coagulation/Flocculation and Membrane Filtration Process for the Treatment of Paint Industry Wastewaters, Ind. Eng. Chem. Res., 51, 15456−15462 (2012). (2) T. Deblonde, C. Cossu-Leguille, P. Hartemann, Emerging pollutants in wastewater: A review of the literature, Journal of Hygiene and Environmental Health, 214, 442−448 (2011). (3) A.Y. Zahrim, C. Tizaoui, N. Hilal, Coagulation with polymers for nanofiltration pretreatment of highly concentrated dyes: A review, Desalination, 266, 1–16 (2011) (4) N. Tzoupanou, Preparation, properties and applications of innovative aluminum reagents for coagulation for water purification and wastewater treatment, PhD Thesis, Aristotle University of Thessaloniki, Greece (2009). (5) M.A. Aboulhassan, S. Souabi, A. Yaacoubi, M. Baudu, Improvement of paint effluents coagulation using natural and synthetic coagulant aids. Journal of Hazardous Materials, 138, 40−45 (2006). (6) M. Hammer, Water and Wastewater Technology, Wiley, New York, United States (1986). (7) B. Durham, M. Bourbigot, T. Pankratz, Membranes as pretreatment to desalination in wastewater reuse: operating experience in the municipal and industrial sectors, Desalination, 138, 83−90 (2001). (8) T. Ratto, J. Holt, A. Szmodis, Asymmetric Nanotube Containing Membranes, U.S. Patent 20110290730 (2011). (9) T. K. Karachalios, S. F. Nitodas, Influencing factors towards a scalable synthesis of aligned carbon nanotubes by chemical vapor deposition, ECS Transactions 25, 749-756 (2009). (10) J. Anastasopoulos, BioNexGen Report, University of Patras, Greece (2013). (11) O.Bkajin, A.Noy, F.Fornasiero, H.G. Park, S.Kim, Membranes with functionalized carbon nanotube pores for selective transport, U.S. Patent 20110253630, issued in November 2011, (2011). (12) A. Noy, H. G. Park, F. Fornasiero, J.K. Holt, C. P. Grigoropoulos and O. Bakajin, Nanofluids in CNTs, Nanotoday 2, 22-29 (2007). 49 (13) V. Upadhyayula, S. Deng, M. Mitchell, G. Smith, Application of carbon nanotube technology for removal of contaminants: A review, Science of Environment 408, 1 (2009). (14) C. Tang, Q. Zhang, K, Wang, Q. Fu, and C. Zhang, Water transport behavior of chitosan porous membrane containing MWNTs, J. Membr. Sci. 337, 240-247 (2009). (15) M. Majumder, N. Chopra, B.J. Hinds, Nanoscale Hydrodynamcs: Enhanced flow in carbon nanotubes, Nature 438, 44 (2005). (16) B. Corry, Designing Carbon Nanotube Membranes for Efficient Water Desalination, J. Phys. Chem. B 112, 1427-1434 (2008). (17) J. Holt, H. Park, Y. Wang, M. Stadermann, A. Artyukhin, C. Grigoropolous, A. Noy, O. Bakajin, Fast mass transport through sub-2-nanometer carbon nanotubes, Science 312, 1034-103 (2006). (18) M. Majumder, N. Chopra, B.J. Hinds, Effect of tip functionalization on transport through vertically oriented carbon nanotube membranes, J. Am. Chem. Soc. 127, 9062-9070 (2005). (19) A. Alexiadis and S. Kassinos, Molecular Simulation of Water in Carbon Nanotubes, Chem. Rev. 108, 5014-5034 (2008). (20) S.Kar, R.C. Bindal, P.K. Tewari, Carbon nanotube membranes for desalination and water purification: Challenges and opportunities Nano Today 7, 385 (2012). (21) G.Voyiatzis, Sonication workshop presentation, University of Patras, Greece (2010). (22) http://www.millipore.com/catalogue/item/vvlp04700 (2013). (23) R.Fries, S.Gebler, M.Simko, Carbon nanotubes – Part II: Risks and Regulations, Nanodossier 24, 1 (2012). (24) J.Yu, N.Grossiord, C.Konin, J.Loos, Controlling the dispersion of multi-wall carbon nanotubes in aqua surfactant solution, Carbon 45, Issue 3, 618-623 (2007). (25) E.Celik, H.Park, H.Choi, H.Choi, Carbon nanotube blended polyethersulfone membranes for fouling control in water treatment, Water Research 45, 274-282 (2011). (26) E.Moschopoulou, Water Vapor Permeability, University of Patras, Greece (2013). (27) N.Hilal, M. Al-A bri, H. Al-Hinai, Characterization and retention of UF membranes using PEGs, Desalination 206, 568-578 (2007). (28) D.Sklavounaki, Embeddement of porous polymeric membranes with carbon nanotubes, Master Thesis, University of Patras, Greece (2012). (29) http://www.malvern.com/labeng/technology/gel_permeation_chromatography_theory /gpc_sec_theory.html (2013).