纳滤膜的发展

FINAL PROJECT REPORT Development and Testing of “Smart” Nanofiltration Membranes Prepared By: Gregory R. Guillen and Eric M.V. Hoek University of California, Los Angeles NWRI Final Project Report Development and Testing of “Smart” Nanofiltration Membranes Prepared by: Gregory R. Guillen and Eric M.V. Hoek, Ph.D. Department of Civil & Environmental Engineering University of California, Los Angeles Los Angeles, CA Published by: National Water Research Institute 18700 Ward Street P.O. Box 8096 Fountain Valley, California 92728-8096 USA July 2010 About NWRI A 501c3 nonprofit organization, the National Water Research Institute (NWRI) was founded in 1991 by a group of California water agencies in partnership with the Joan Irvine Smith and Athalie R. Clarke Foundation to promote the protection, maintenance, and restoration of water supplies and to protect public health and improve the environment. NWRI’s member agencies include Inland Empire Utilities Agency, Irvine Ranch Water District, Los Angeles Department of Water and Power, Orange County Sanitation District, Orange County Water District, and West Basin Municipal Water District. For more information, please contact: National Water Research Institute 18700 Ward Street P.O. Box 8096 Fountain Valley, California 92728-8096 USA Phone: (714) 378-3278 Fax: (714) 378-3375 www.nwri-usa.org Jeffrey J. Mosher, Executive Director Gina Melin Vartanian, Editor © 2010 by the National Water Research Institute. All rights reserved. Publication Number NWRI-2010-03. This NWRI Final Project Report is a product of NWRI Project Number 07-TM-003. i Acknowledgments This Final Project Report was prepared by Gregory R. Guillen and Eric M.V. Hoek, Ph.D., of the University of California, Los Angeles (UCLA) and sponsored by the National Water Research Institute of Fountain Valley, California. Special thanks are extended to members of the UCLA Nanomaterials and Membrane Technology Research (NanoMeTeR) Laboratory and Dr. Christina Baker for their help and advice. ii Contents 1. Statement of Problem and Significance ………………………………………..…...… 1 1.1. Particle Filtration ………...……………………………………………….. 1 1.2. Membrane Filtration …………...……………………………………….… 1 2. Background and Related Research …………………………………………………… 2 2.1. Fundamentals of Porous Filtration Membranes ……………...…………… 2 2.2. Separation Mechanisms for Porous Filtration Membranes ……..………… 5 2.3. Analysis of Membrane Pore Size or Molecular Weight Cut-Off ……...…. 7 2.4. Pressure-Driven Flow through Porous Filtration Membranes …….……… 9 2.5. Synthesis of Porous (Particle Filtration) Membranes ….….…………..….. 12 2.6. Polyaniline ……………………………………………………………..…. 17 2.7. Polyaniline Synthesis ……………...…………………………………..….. 18 3. Preliminary Work ……………………………………………………………………... 22 3.1. Polyaniline Membrane Synthesis and Characterization ……..………..….. 22 3.1.1. Membrane Thickness, Pure Water Permeability, and SiO2 Rejection …………………..…………………………………..….. 22 3.1.2. Scanning Electron Microscopy-Focused Ion Beam (SEM-FIB) …. 24 3.1.3. Atomic Force Microscopy ……………...…………………...…..... 26 3.1.4. Membrane Electrical Resistance as a Function of pH and Polyaniline Content …………………………………..…………… 28 3.2. Crossflow Membrane Electrofiltration Flow Cell ………………………... 29 3.3. Electrofiltration Using Polysulfone Membrane with Porous Stainless Steel Permeate Support …………………………………….…… 30 3.3.1. Constant Potential ………………………………………………… 30 3.3.2. Pulsed Potential …………………………………………………… 31 3.4. Electrofiltration Using Polyaniline Membrane ………………….………... 33 4. Conclusions …………………………………………………………………………… 34 5. References ……………………………………………………………………….……. 36 6. Publications Based on This Project …………………………………………………… 40 Tables 2.1. Common Test Solutes Used to Characterize UF Membranes ………...…………….. 9 2.2. Kozeny Coefficient as a Function of Particle Volume Fraction …..………………… 11 2.3. Solvents Compatible with Polysulfone/Water System ...…...……………………….. 13 2.4. Pure Water Flux Through Polysulfone Membranes ……………...…………..……... 16 3.1. Polyaniline-Polysulfone Blend Membrane Thickness, Pure Water Permeability, and 40-nm Silica Rejection …………………………………………………..……... 22 3.2. Water Contact Angles on PANI-PSf Membranes ………………………………...…. 23 3.3. Membrane Conductivity as a Function of pH and Polymer Composition ….…...…... 28 Figures 2.1. Cross-flow membrane filtration schematic ………………………………………….. 2 2.2. Dead-end membrane filtration schematic ………………………………………...…. 3 iii 2.3. Application range of MF, UF, NF, and reverse osmosis membranes…………...…… 4 2.4. Particle rejection mechanism, according to Ferry’s model ………………………….. 5 2.5. Particle capture mechanism in filtration of liquid solutions by depth microfilters ….. 7 2.6. Relationship between pore size, molecular weight of ideal solutes, and ratings of ideal and real membranes …………………………….……………… 8 2.7. Membrane modeled as an array of cylindrical channels and a pore with diameter, dp, and length (membrane thickness), l …………………………..………. 10 2.8. Cross section of a filtration membrane modeled as a packed bed of spherical particles ………………………………………………………………… 11 2.9. Immersion precipitation membrane formation ……………………………….…...… 13 2.10. Demixing delay for cellulose acetate/water/solvent system ……….………….……. 14 2.11. Asymmetric structure of a UF membrane ………………………....………………... 15 2.12. Chemical structure of polyaniline …………………………….….…………………. 17 2.13. Various oxidation states of polyaniline …………………………….……..………… 18 2.14. The oxidative polymerization of aniline in an acidic solution ……..….……………. 19 2.15. TEM images of polyaniline powders made by traditional chemical polymerization using 1.0 HCl showing a small portion of nanofibers in the sample ……...………... 19 2.16. The morphological evolution of polyaniline during chemical polymerization is explored by electron microscopy …………………………………….…………... 20 2.17. Schematic illustration showing a rapidly mixed reaction in which the initiator and monomer are rapidly mixed together all at once ………………...……………... 21 3.1. Permeability and rejection for PANI-PSf blended membranes ……………..………. 23 3.2. Plan view SEM image of a pure PANi membrane …………………………..……… 24 3.3. Cross-section image of a pure PANi membrane …………………………………….. 25 3.4. Pores within the walls of a void …………………………………………..…………. 26 3.5. Plan view AFM image of a pure PANI membrane ………………………...………... 27 3.6. Three-dimensional AFM image of a pure PANI membrane …………...…………… 27 3.7. Crossflow membrane electrofiltration flow cell system and opened flow cell ……… 29 3.8. Normalized transmembrane pressure over time for different field strengths …..…… 30 3.9. Silica nanoparticle rejection over time for different constant field strengths ……….. 31 3.10. Silica nanoparticle rejection and permeate flux over time for different pulsed field strengths ……………...……………………………………………………………... 32 3.11. Transmembrane pressure and applied field strength over time …….……….……… 33 3.12. Pressure, silica nanoparticle rejection, and applied potential over time for a pure PANI membrane …………………………………...…………………….. 34 iv 1. Statement of Problem and Significance 1.1. Particle Filtration Particle filtration relies on a selective barrier to remove particles from a liquid phase. Filtration has applications in drinking water production, wastewater treatment, desalination pretreatment, food and beverage production, protein separations, pharmaceutical purification, and analytical separations, as well as many other industrial applications. Two basic filter types exist: media (depth) filters and membrane (sieving) filters. Media filters are an established technology; the first recorded media filters were used in India around 2,000 B.C. to purify water (Crittenden et al., 2005). Depth filters are most widely used in water and wastewater treatment, relying on cheap, natural media, such as sand, anthracite, crushed magnetite, garnet, and others (Droste, 1997). Membrane filtration has been used to purify drinking water as early as World War II (Crittenden et al., 2005); however, the vast majority of modern water and wastewater treatment plants still use granular media filters to this day. Membrane filtration is primarily used for industrial, biological, and analytical separations. In general, particle concentration in the product stream of granular media filters is proportional to the particle concentration in the feed; moreover, particle removal in depth filters is limited by the effectiveness of coagulation. Media is not fixed within the filter bed, which allows for preferential flow paths and the passage of particles. Relatively recent incidents such as the Cryptosporidium outbreaks of 1993 (Milwaukee) and 2000 (Ontario, Canada) have led to large epidemics of waterborne disease and exposed the Achilles heel of media filtration – that is, poor removal of stable particles in the micrometer (µm) size range (Corso et al., 2003). The Milwaukee Cryptosporidium outbreak forced utilities to consider replacing granular media filters with membrane filtration, which (relative to media filtration) offers an absolute barrier to pathogens – protozoa, bacteria, and viruses – depending on the pore size of the membrane. In other separations, membranes offer a much greater ability to tailor filtration selectivity by modifying the surface pore structure and chemistry. 1.2. Membrane Filtration Membrane filtration has been the top choice for industrial particle-liquid separations for many years, but more recently has gained traction in water and wastewater treatment. Particle separation primarily depends on the ratio of particle size to membrane pore size and, to a lesser extent, feed water chemistry, particle properties, and membrane chemistry. Today’s membranes have fixed physical-chemical properties. Once membranes are cast, their chemical functionality and pore structure are fixed; thus, their selectivity is static. To cope with rising energy costs, emerging contaminants, stricter environmental regulations, and industrial demand, the next generation of filtration membranes must be more selective and robust, while requiring lower chemical and energy inputs. New membrane materials must be explored to help meet these goals in applications such as water and wastewater treatment, in addition to other well-established industrial, biomedical, and analytical separations. 1 2. Background and Related Research 2.1. Fundamentals of Porous Filtration Membranes In membrane filtration, a pressure difference forces liquid to flow through the membrane. Solutes (dissolved or particulate) in the feed are rejected based on size; solutes larger than the filter pores are blocked, while smaller particles pass through the filter pores. There are three liquid streams in a membrane filtration process: feed, retentate, and permeate. The feed stream enters the membrane module and is composed of liquid and particles. The liquid that passes through the membrane is called permeate (or product). The rejected solutes and liquid form a concentrated stream called the retentate (or concentrate). These streams are shown in Figure 2.1. Figure 2.1: Cross-flow membrane filtration schematic. There are two flow configurations in membrane filtration: cross-flow and dead-end. In cross- flow filtration, liquid flows tangentially to the membrane surface and through the membrane, as shown in Figure 2.1. Solutes in the feed stream are concentrated as they are rejected by the membrane and eventually exit the module. The retentate stream can be recycled or wasted. In dead-end filtration, liquid flows normal to the membrane surface and through the membrane, as shown in Figure 2.2. There is no retentate stream in dead-end filtration. Filtration membranes can be made from inorganic or polymeric materials. Inorganic membranes are classified as ceramic, metallic, glass, or zeolitic membranes, and generally have greater thermal and chemical stabilities than typical polymeric membranes (Mulder, 2003). Excellent thermal and chemical stability allow inorganic membranes to be used in high temperature and extreme pH applications. Inorganic membranes can generally be used in any organic solvent (Mulder, 2003). However, inorganic membrane mechanical stability can be an issue in high- pressure systems because materials such as ceramics are brittle and easily broken (Mulder, 2003). Inorganic membranes are also expensive to produce per unit area of membrane. Polymeric membranes are relatively cheap, easy to fabricate in large quantities, and stable in a wide range of physical-chemical conditions. Current polymeric filtration membranes are resistant to temperatures over 100°Celcius (C), a wide pH range (1 to 14), and to some organic solvents (Mulder, 2003). Polymeric membranes are not brittle like inorganic membranes, which allow polymeric membranes to be packaged in high surface area modules. Thus, the polymeric membrane plant footprint is lower when compared to plants using inorganic membrane modules. 2 Figure 2.2: Dead-end membrane filtration schematic. Filtration membranes are very thin, often less than 150 µm, allowing membranes to be packed into high surface area membrane modules. High packing density translates into a smaller footprint in large-scale applications (i.e., water and wastewater treatment). Membrane systems are modular, which allows easy design of large plants. Membrane filtration is now a proven separation technology in large-scale water treatment applications. Filtration membranes provide an absolute physical barrier for the separation of pathogens, such as protozoa, viruses, and bacteria. Membrane filtration systems provide continuous high-quality product water regardless of feed water quality, unlike media filters (Crittenden et al., 2005). Membrane filtration is separated into three classes: microfiltration (MF), ultrafiltration (UF), and nanofiltration (NF). Filtration membranes reject solutes based on solute size and shape relative to membrane pore size and shape. Generally, in membrane filtration processes, osmotic pressures are negligible, internal fouling is possible, and over 90-percent recovery in a single pass is achievable. The transport of solvent through filtration membranes is directly proportional to transmembrane pressure. UF and NF membranes are typically asymmetric in structure, with a thin, dense skin layer over a porous supporting layer. Solute rejection and hydraulic resistance occur almost entirely across this thin layer. MF membranes are typically symmetric in structure, with particle rejection and hydraulic resistance occurring throughout the entire thickness of the membrane (Mulder, 2003). The three classes of membranes are separated based on pore size or solute rejection. As membrane type moves across the range of MF to NF, pore size, solute passage, and solvent permeability decrease. NF membranes have pores on the order of a nanometer (nm) (Mulder, 2003). UF membranes have pores of a few nanometers to a few tens of nanometers in diameter, 3 while MF membranes have pores ranging in diameter from 0.1 µm to several micrometers (Mulder, 2003). There is some overlap of solute rejection between MF, UF, and NF membranes, as shown in Figure 2.3. Figure 2.3: Application range of MF, UF, NF, and reverse osmosis membranes (Mulder, 2003). Membrane filtration processes are limited by several factors, including fixed selectivity, fouling, and degradation. Once they are fabricated, filtration membranes have fixed selectivity. It is difficult to adapt existing membrane processes to emerging contaminants without the addition of chemical agents or new membranes. Like all filtration processes, membrane filtration is prone to fouling. Filtration membranes can experience both internal and external fouling, which can occur on top of the membrane surface and within the pore structure, respectively. Membrane fouling increases energy consumption and operating costs in the forms of increased applied pressure, air scouring, cleaning chemicals, or other anti-fouling measures. Filtration membranes are cleaned by frequent hydraulic backwashing and periodic chemical washing to remove adsorbed foulants. These frequent process interruptions reduce productivity and create a residual waste stream. Repeated chemical cleaning can reduce membrane permeability and selectivity over time. Membrane integrity must be continuously monitored to ensure that membrane degradation has not compromised the filtration process. 4 2.2. Separation Mechanisms for Porous Filtration Membranes The main mechanisms by which particles are rejected by porous filtration membranes include: (1) size exclusion or sieving, (2) charge exclusion, and (3) specific chemical interactions. Depth filtration may play a minor role, but it is generally neglected due to the thin cross-section of a membrane. The capture of particles and solutes internally by a membrane leads to a rapid, often irreversible loss of permeability (see discussion below on fouling). Size exclusion is the dominant filtration mechanism. Particles larger than membrane pores are largely rejected, while much smaller particles mostly pass through. The transport of particles similar in size to membrane pores is more difficult to track and sensitive to additional factors. Real membranes exhibit a distribution of pore sizes and shapes. Solutes also vary in size, shape, and chemistry – sometimes dynamically. For example, solutes may be amphoteric or amphiphilic and can change shape when forced through a membrane pore. Ferry developed a model analogous to the interception mechanism in the isolated collector model for granular filtration to describe particle retention by pore exclusion (Crittenden et al., 2005; Ferry, 1936). However, this model does not account for attachment efficiency based on membrane-particle interfacial forces. Ferry’s model is based on the assumption that any particle contacting the membrane surface is retained. In laminar flow conditions, particles move parallel to streamlines towards cylindrical membrane pores. Particles impacting pore edges are rejected, while those that follow streamlines through the center of pores pass as shown in Figure 2.4. Figure 2.4: Particle rejection mechanism, according to Ferry’s model. 5 According to the Ferry mechanical-sieving model, it is possible for particles smaller than the pore to be rejected. Particle passage (p) is a function of particle diameter (dp) relative to membrane pore diameter (dpore), according to (Wiesner and Buckley, 1996) ⎭⎬ ⎫ ⎩⎨ ⎧ > ≤−−−= 1;0 1;])1(2[)1( 22 λ λλλ Gp , (2.1) where λ = dp/dpore and G is the lag coefficient empirically determined by Zeman and Wales (1981) to be ( )27146.0exp λ−=G . (2.2) Pore size may be altered during the filtration process by pore blocking. Spherical particles larger than spherical pores (λ > 1) can cause complete pore blocking. Smaller non-spherical particles can partially block pores at the membrane surface or on the pore walls, resulting in pore restriction and altering the rejection properties of the membrane. However, this model cannot be taken literally as particles are not hard spheres. For example, flexible proteins can deform under stress and change shape as they pass through the pore of a membrane (Mulde