Membranes Coupled with Nanotechnology for Daily Drinking Water: an Overview

It is well known that current global development is not sustainable over the long term. Every major ecosystem is under threat at different timescales, impacting water, food, energy, biodiversity, and mineral resources—all exacerbated by the population growth and climate change.Potable water is a threatened but critical resource, the scarcity of which isdevastating for the developing world. Water-related nanotechnology researchhas the potential to make safe drinking water inexpensive and accessible todeveloping countries. Therefore, the relevant nanotechnologies on membranes for filtrating drinking water are reviewed. At the same time, it is pointed out that for the foressable future membranes must exhibit a number ofcharacteristics such as high water flux, high salt rejection, mechanical stability,resistance to fouling, and low cost. With new breakthroughs in membrane technology, constituent specificmembranes should be realized andthe promising reverse osmosis membranes coupled with new nanotech will be applicable tonanofiltration membranes in the future. Membranes Coupled with Nanotechnology for Daily Drinking Water: an Overview


Introduction
It is well known that current global development is not sustainable over the long term. Every major ecosystem is under threat at different timescales, impacting water, food, energy, biodiversity, and mineral resources-all exacerbated by the population growth and climate change. It has been estimated that about 1.1 billion people are now at risk from a lack of clean water and about 35 percent of people in the developing world die from water-related problems [1,2]. Also, aquifers throughout around the world are suffering from declining water levels, saltwater intrusion, contamination from surface waters, and inadequately replenished fresh groundwater. Major rivers and watersheds are also being overdrawn, while return flows are contributing to downstream nutrient loading and salinity problems. Lakes and wetlands are also experiencing increased salting in many areas. Furthermore, as increases in acreage and irrigation are needed for the production of biofuels, further dramatic increases in demand will strain existing water supplies and new water sources will be needed. If lack of water limits growth in new energy supplies, every aspect of the global economy will be affected, increasing costs to the consumer. Clearly, the costs to upgrade infrastructure and create new supplies likely cannot be met without revolutionary improvements in the science and technology of water purification. Because water is a ubiquitous facilitator of civilization, de facto essential for almost everything, ranging from maintaining human health, to ensuring plant growth, and enabling the transport of merchandise, ensuring adequate supplies of potable water for human use must continue to be a research priority, the environmental impacts of increasing human water consumption will also demand attention. Moreover, better research on the availability, detection of contaminants, and strategies for remediation can increase utilization of currently available sources as well as facilitate development of new water sources such as brackish aquifers. At the same time, depending on the new generation and cooling technology deployed, fresh water consumption is expected to increase more. Concurrent efforts to reduce dependence on imported oil through new fuels such as biomass, syngas, and hydrogen are expected to expand the overall water footprint of the energy sector even further. New sources of water, and new purification technologies to enhance water reuse, will be needed to keep pace with energy demands. Also, food production is even more closely linked to the availability of water than energy production. New technologies that can dramatically enhance agricultural water conservation and increase the recovery and reuse of irrigation runoff and livestock waste water could have the largest impact on future water availability [3][4][5].
water are growing so quickly that incremental improvements in current methods of water purification could leave much of the world with inadequate supplies of clean water in mere decades. The challenges to overcome in science, technology, and society require a long term vision of what needs to be solved [9][10][11][12][13].
Therefore, ensuring the availability of clean, abundant fresh water for human use is among the most pressing issues facing the world. It is stated over the next two decades, the average supply of water per person will drop by a third, possibly condemning millions of people to an avoidable premature death, and environmental stress resulting from climate change and population growth and migration is expected to increase over next two decades.
There is urgent that satisfying humankind's demand for water in a sustainable manner requires visionary new approaches to management and conservation of water resources augmented by new technologies capable of dramatically reducing the cost of supplying clean fresh water-technologies that can best be derived from tightly coupled basic and applied research.
Science and technology are turbulent dynamic fields where coherent structures appear and break down. Nanotechnology promises to dominate the landscape for many of these fields over the next several decades. Research and development at this scale can answer major challenges for society, from improved comprehension of nature and increased productivity in manufacturing, to molecular medicine and extending the limits for sustainable development [14][15][16].
Actually, for nanotechnology, a key objective is to develop materials, devices, and systems with fundamentally different properties by exploiting the unique properties of molecular and supramolecular systems at the nanoscale. In recent years, research in this field has grown exponentially as scientists and engineers continue to develop nanomaterials with unique and enhanced properties. Nearly every field of science has been affected by the tools and ideas of nanotechnology, and breakthroughs have been made in computing, medicine, sensing, energy production, and environmental protection. Recent advances strongly suggest that many of the current problems involving water quality can be addressed and potentially resolved using nanosorbents, nanocatalysts, bioactive nanoparticles, nanostructured catalytic membranes, and nanoparticle enhanced filtration, among other products and processes resulting from the development of nanotechnology [17][18][19][20].
Moreover, nanotechnology solutions are essential because the abiotic and biotic impurities most difficult to separate in water are in the nanoscale range. By its control at the foundation of matter, nanoscale science and engineering may bring breakthrough technologies possible for improving water quality. Also, on the other end, nanotechnology may offer efficient manufacturing with less resources and waste to reduce pollution at its site of origin [21][22][23][24].
It is by now presumably well known that nanotechnology has the potential to contribute novel solutions to an enormous range of problems currently facing the world. Ensuring the availability of potable water ranks as one of the more important and urgent of those problems, and nanotechnology is clearly a candidate for helping to solve it. Furthermore, nanotechnology is specifically appropriate because the problems of water purification, from the viewpoint of what needs to be removed from contaminated water, crucially mostly involve the nanoscale.
In addition, water filtration and desalinization have been relatively less-explored topics in nanotechnology, but there has been a new trend in this direction in the last few years. This is an area of importance for life, it is of significant interest to the productive engine and the public at large, and there are many stakeholders. It should be pointed out that research and development have been lagging behind advances in areas such as electronics, materials science, and pharmaceuticals, which have relatively shorter term returns. Infrastructure for water resources requires a relatively larger investment for a longer period of time and the diverse potential sponsoring sources need to be better coordinated.
Considering the problems of current drinking water, research on membranes coupled with nanotechnologyis reviewed as follows.

Membranes with functional nanoparticles
Polymers and ceramics, the two main classes of membrane materials, have distinct advantages as platforms for the design of multifunctional membranes. Due to their high chemical and thermal stability, ceramic membranes can support additional functions and processes in highly oxidizing environments and under conditions of extreme pH and temperatures. Polymers, on the other hand, offer great design flexibility and are generally less expensive. Polymeric multifunctional membranes can be prepared by modifying materials traditionally used in membrane manufacture. Surface functionalization of existing membranes and the use of block copolymers and hybrid polymers as membrane materials are examples of such methods. Another approach is that of developing new hybrid membrane materials by incorporating (metal oxide, metal, carbon, or polymeric) functional nanoparticles into the polymeric matrix [25][26][27][28][29][30].
At the same time, nanotechnology has enabled the development of a new class of atomic scale materials capable of fighting waterborne disease-causing microbes. The explosive growth in nanotechnology research has opened the doors to new strategies using nanometallic particles for oligodynamic disinfection [31][32][33]. The excellent microbicidal properties of the oligodynamic nanoparticles qualify their use as viable alternatives for water disinfection. Oligodynamic metallic nanoparticles such as silver, copper, zinc, titanium, nickel and cobalt are among the most promising nanomaterials with bactericidal and viricidal properties owing to their charge capacity, high surfaceto-volume ratios, crystallographic structure, and adaptability to various substrates for increased contact efficiency. This new class of nanometallic particles produces antimicrobial action referred to as oligodynamic disinfection for their ability to inactivate microorganisms at low concentrations. When oligodynamic metals with microbicidal, bactericidal, and viricidal properties are reduced the size of the metals to the nanoscale, they show tremendous advantages in disinfection capacity due to the greater surface area, contact efficiency, and often better elution properties. These qualities enable these materials to be considered asviable alternative disinfectants, such as silver (Ag), copper (Cu), zinc (Zn), titanium (Ti), and cobalt (Co).New combinatorial oligodynamic materials consisting of these nanometallic particles have been deployed among a number of substrates for their use in water disinfection [34][35][36]. Such materials as Ag deposited on titanium oxide, and Ag-coated iron oxide had displayed faster kinetics and greater efficiency in eliminating bacteria.
Nowadays, silver is the most widely studied oligodynamic material due to its wide range in microbicidal effectiveness, low toxicity, and ease of incorporation on various substrates in a host of dynamic disinfection applications. Furthermore, the systems supported with nanometallic silver particles are effective in reducing the presence of target microorganisms in a wide variety of water disinfection applications, except the main known negative health effect from silver is argyria, which is an irreversible darkening of the skin and mucous membrane resulting from overexposure to ionic silver (Ag(I), Ag+) [37].
Typically, the silver nanoparticles are derived from silver salts (silver nitrate (AgNO 3 ), silver chloride (AgCl), silver bromide (AgBr), and silver iodide (AgI)), and a variety of substrates that silver deployed on such as activated carbon, activated carbon fibers (ACF), polyurethane, zeolites, and ceramics in POE and POU applications displays the effective inactivation of pathogens in water [38][39][40][41]. Wang et al. [42] prepared viscose-based activated carbon fiber supporting silver (ACF(Ag)) by pretreatment, carbonization, activation, vacuum impregnation and decomposition processes, which ACFs were successively subjected to a vacuum impregnation treatment in unsaturated silver nitrate(analytical grade) aqueous (AgNO 3 ) solutions (NH 4 H 2 PO 4 3.3 gl -1 , (NH 4 ) 2 SO 4 6.7 gl -1 ) with varying concentrations for different times, and were varying concentrations for different times, and were finally heated to different temperatures for decomposition, thus producing ACF(Ag). Thereafter, the ACF (Ag) samples were ashed at 800°C for6 hours in atmosphere, the resulting ashes were then dissolved in 50 ml of 10% (vol.) HNO 3 solutions (90°C) to determine the silver content. It reveals that the silver particle size is influenced by the concentration of AgNO 3 solution, immersion time and decomposition temperature, of which AgNO 3 content is the most remarkable factor (as shown in Figure 1). Moreover, the ACF (Ag) containing as low as 0.065 wt% of silver exhibits the strong antibacterial property against Escherichia coli and Staphylococcus aureus.
Though many forms of silver have found use in disinfection applications, which include swimming pools and hospital hot water systems, silver nanoparticles find the most extant usage in POU applications including activated-carbon-based and ceramic water purification filters. POU filters composed of granular activated carbon impregnated with silver have received ample attention in the past decade owing to their high surface area and pore size distribution that allow silver to be easily entrapped in the pores and later desorbed [42,43]. Carbon-based substrates lower the impact of the silver nanoparticles. With a loading of 0.05 wt percent Ag impregnation in an ACF with extremely high surface area (1200 m 2 /g), the fastest time achieved for complete bacterial elimination is 30 minutes. The silver impregnated carbon-based filters displays only bacteriostatic performance since they are not able to completely eliminate microbial regrowth in POU devices. Such performance related issues can be addressed by deploying silver nanoparticles on inorganic-based substrates and by using combinations of oligodynamic nanoparticles [43].

Nitrogen-doped titanium-oxide
Titanium oxide (TiO 2 ) is modified by nanoparticles of transition metal oxides and made into nanoparticles, nanoporous fibers, and nanoporous foams. The nanostructured photocatalysts show very fast photocatalytic degradation rates in organics, bacteria, spores, and virus, and thus have great potential in water disinfection and removal of organic contaminants in water. The basis for photocatalytic control is the production of highly reactive oxidants, such as OH radicals, for oxidization of organic pollutants, disinfection of microorganisms, and degradation of hazardous disinfection by-products (DBPs) and disinfection by-product precursors (DBPPs) [44][45][46].
With the newly developed TiON-based photocatalysts, photocatalytic degradation and disinfection can be implemented with visible light. The replacement of UV by visible light offers potential for low-cost environmental measures, especially for water treatment, where UV access is rather limited. The removal of organic contaminations has been demonstrated by the photodegradation of HA by TiON/PdO under visible-light illumination. The addition of PdO allows the electron transfer process on the photocatalyst to be "regulated" by storing and releasing electrons to minimize electron ehole recombination [47] or to produce a long-lasting photocatalytic 'memory' effect after light is turned off. While the hydroxyl radicalsgenerated by the visible light photocatalysis are believed to be the working species in bacterial inactivation [48,49]. Escherichia coli is a model gram negative bacterium, which is widely used as a bacterial indicator. TiON/PdO shows the high disinfection efficiency against E. coli. After 30min visible-light illumination, the survival ratio of Escherichiacoli drops to approximately 10 -8 . Wu et al. [50] characterized the cellular responses of Escherichia coli to visible light photocatalysis by chemical,optical, electron-beam, and surface-force techniques, to elucidate the mechanisms of photocatalytic inactivation of E. coli on PdO/TiON fiber. PdO/TiON-deposited fiber photocatalyst was prepared by a mixture of titanium tetraisopropoxide and tetramethyl ammonium hydroxide (mol ratio 4:1) in absolute ethanol. Then a proper amount of Pd (acac) 2 dissolved in CH 2 Cl 2 was added. Activated carbon-coated glass fiber (ACGF) was soaked in the precursor mixture for 24 hat room temperature. After wash and dry, fine crystallites of PdO/TiON nanoparticles deposited on fibers were obtained by calcination (400ºC, 3 h), followed by removal of carbon at 500ºC for 1 h in air. For viability assays, a typical procedure of photocatalytic treatment was as follows. Overnightcultivated E. coli AN387 was washed and resuspended in buffer solution (0.05 MKH 2 PO 4 and 0.05 M K 2 HPO 4 , pH 7.0) to ca. 10 9 colony-forming units per ml (cfu/ml).The cell suspension was pipetted onto a sterile petri dish which was illuminated by a visible-light source (ca.1.6 mW/cm 2 ) in the presence of the PdO/TiON fiber photocatalyst. At regular time intervals, 20 μL aliquots of the irradiated cell suspensions were withdrawn. After appropriate dilutions in buffer, aliquots of 2μL together with 2.5 ml top agar were spread onto agar medium plates and incubated at 37 ºC for 18-24 h. The number of viable cells in terms of colony forming units was counted. The E. colirec, a mutant strain AS224 was used for a comparative study following the same viability assay. The hypersensitivity ofAS224 mutants to DNA damage was confirmed with in-house UV irradiation prior to use of the bacteria. Through the chemical assay and several microscopic techniques (such as scanning electron microscopy (SEM), transmission electron microscopy (TEM), fluorescence microscopy, atomic force microscopy (AFM)), the cellular responses of E. coli tovisible light photocatalysis at different treatment intervals were characterized in detail. and (c) of treated cells. Second, in nearly every cell, the appearance of rumples and a high degree of disconfigurations are observed. Images in Figure 2(b) and (c) show that many E. coli cells are missing parts of the cell wall and the cell membrane or even material inside, so that deep 'holes' appear. These images obtained in two separate sets of experiments verify that photocatalysis causes oxidative damages on bacteria. It is interesting to note that after treatment for 30 min (with ~90% cell population "destroyed" as indicated by the viability assay), the cell morphology and its surface structure change little as shown in Figure 2(d). It indicates that at this disinfection stage, when bacteria cells have just lost their viability, the exterior damage in morphology has occurred; yet the damage is too subtle to be observed under SEM. Figure 3(a) is a representative TEM image ofuntreated E. coli cells that have a fluffy boundary. The fluffy outer layer is considered to be the outer membrane of E. coli cell. It can benoted that after photocatalytic disinfection, the outer membrane is completely decomposed. In Figure  4(b)-(d) after treatment for 2 h, a noteworthy difference from the untreated sample is that every treated cell has lost its outer membrane, i.e., the fluffy boundary. Some treated cells show a clearly-cut edge, which indicates that the plasma membrane may have been exposed after theo uter membrane has decomposed. The other cells completely or partially lose this edge, which is a severely damaged stage with plasma membrane also gone. An extensively damaged 'ghost' cell is apparent in the center of Figure 4(b). There remains no more cell wall or cell membrane in this cell; instead, many dark granules appear.
TEM images also show remarkable interior damage of the cells after photocatalytic disinfection. Normal E. coli cells exhibit a homogeneous microstructure in Figure 3(a). This image shows that a healthy E. coli cell has a well-defined cell wall and a uniform interior material distribution, which corresponds to an inner zone full of proteins and DNA molecules. In contrast, dark mass aggregates appeared in the 2h well-treated cells in Figure 3     substances are coiled or twisted, like fingerprints. There are also some electron-dense granules/mass aggregates near the cell wall and membrane boundaries. In Figure 4(b), an E. coli cell is more severely damaged, as its outer membrane seems to be decomposed (lacking the fluffy edge) and its inner structure badly disturbed. A few individual cells appear less damaged, except for a few material-dense dark granules. Results indicate that the delicate cell membrane is subject to oxidative damage first by the very reactive hydroxyl radicals. The initial membrane damage is so subtle that it is not detectable in SEM and TEM until the damage advances to a greater level. But as soon as the cell's first line of defense (which is the cell wall and membrane) is broken, some interior substances (including the sensitive DNA molecules) are damaged almost immediately and readily observable in TEM.
Furthermore, in the previous works [51], TiONis also fabricated into both powder and thin film.Compared with TiON thin films, TiON powder photocatalysts offer theadvantages of high surface area, low cost, and suitability for large-scaleproduction. Among various synthesis methods for preparing TiON powders,sol-gel based processes [52,53] seem to have the most potential. However, a systematic study of the precursor ratio effect on thestructure, composition, and optical properties of sol-gel derived TiONnanoparticles [54] should be focused on.

Porous titanium oxide (TiO 2 ): Sol-gel synthesis
Research efforts in photo catalysis have dramatically expanded since the discovery of the photocatalytic properties of TiO 2 and the demonstration of its effectiveness to generate hydroxyl radicals in the presence of UV. TiO 2 photo catalysis is of particular interest because of its environmentally friendly features. The process can completely oxidize virtually all organic contaminants (nonselective) without addition of any other chemicals for the reaction and thus produce no harmful end products in most cases. Especially, TiO 2 photocatalysis forms no disinfection by-products unlike other chemical oxidation processes when sufficient time is allowed for organic mineralization. In general, the photo catalytic process has features of a green engineering process. Although various materials (oxides: TiO 2 , ZnO, ZrO 2 , CeO 2 , SnO 2 , Fe 2 O 3 , SbrO 4 ; sulfides: CdS, ZnS) have been used for photo catalysis, generallyTiO 2 is the most promising photo catalyst, considering its energy efficiency, durability, photo stability, water insolubility and nontoxicity [55][56][57][58][59][60].
Among various synthesis methods of TiO 2 , sol-gel technology, which involves the formation of solid inorganic materials from liquid molecular precursors, is popular for the fabrication of TiO 2 inorganic materials with engineered properties because of (i) room temperature wet chemistry based synthesis, (ii) wide-range selection of precursors and support materials, (iii) precise control of the properties of TiO 2 at the molecular level, and (iv) easy doping with other metals or nonmetals in TiO 2 matrix. During the sol-gel synthesis of TiO 2 , surfactant molecules as a pore directing agent in the TiO 2 inorganic matrix are introduced, as demonstrated in Figure 5 [61]. Under certain conditions, surfactants are known to self-assemble into various structures (micellar, hexagonal, lamellar) in a water-rich environment [62]. For example, when liquidphase TiO 2 precursors are added in the surfactant micellar aqueous solution, the TiO 2 precursors are hydrolyzed and condense to form a solid TiO 2 inorganic network around the micelles, forming a surfactant organic/TiO 2 inorganic composite. During thermal treatment, the surfactant templates are pyrolyzed, leaving the TiO 2 inorganic matrix with a porous structure. In the case of water-poor conditions, the situation is reversed toform reverse micellar structures, then TiO 2 inorganic core/surfactant organic composites, and finally well-defined TiO 2 nanoparticles [63].

Adsorption of chemicals to TiO 2
It is well known that someb of the electron and hole pairs migrating to the surface get involved in redox reactions even during their short lifetime (on the order of nanoseconds). Titanium (IV) is reduced to titanium (III), which is finally transformed totitanium(IV) combined with superoxide radical anions if electron acceptors such as oxygen are available on the surface listed as follows: At the same time, the generated holes are utilized for the generation of hydroxyl radicals and direct oxidation of organics, R (shown as follows) or they can be combined with the electron from a donor species, depending on the reaction mechanism: Because of the short lifetime of photo carriers, the prerequisite for above-mentioned reactions is the adsorption of substances such as water and organic molecules on the TiO 2 surface and lattice oxygen ( 2 L O − ) shown as follows. This facilitates the redox reaction at the interface of TiO 2 solid and the water: Figure 5: Synthesis approaches of engineered TiO 2 via sol-gel method employing surfactant self assembly as (a)aporetemplate and (b)particle growth template (a)Synthesis of TiO 2 with mesoporous inorganic network: (i)surfactant molecules are self-organized in water-rich environment forming surfact and the head group outside towards water molecules and its tail group inside free from water, (ii)titanium alkoxide precursor ishydrolyzed and condensed to form TiO 2 inorganic network around the self-assembled surfactant, forming a surfactant organic template-embedded TiO 2 inorganic matrix and (iii)porousTiO 2 inorganic network is formed after removal of the organic template by thermal treatment or organic extraction. (b)Synthesis of TiO 2 nanoparticles: (i)surfactant molecules are self-organized in water-poor environment (bulk hydrophobic solvent(HS) with small portion of water) forming surfactant head group inside towards water molecules and its tail group outside towards HS, (ii)titanium alkoxide precursor is hydrolyzed and condensed to form TiO 2 in organic network in the waterphase, inside of self-assembled surfactant, forming TiO 2 inorganic core/ surfactant organic shell structure and (iii)well-defined TiO 2 nanoparticles are formed after removal of the organic template [61].

Radical attack on organics
The hydroxyl radicals, the primary oxidizing species in the photo catalytic system, initiate chain reactions leading to the generation of other radicals and subsequently oxidation of organics [63]. Even though it is not necessary for the reaction that hydroxyl radicals and organics are adsorbed at the TiO 2 sites, the adsorbed forms are much more helpful to increase the overall reactivity, compared to those free from TiO 2 sites: • ' • ' • ' • ' HO , H 2 O 2 ) are also generated and are involved in redox reactions to decompose organic contaminants in water: After the photo excitation process and the generation of reactive species, a series of reactions lead to complete mineralization of the parent compound. These reactions include hydroxyl radical attack, hydroxylation (e.g., OH • addition, reaction with O 2 , and elimination of ( • HOO ), dihydroxylation, hydration, hydrogen abstraction, deprotonation, decarboxylation and one-electron transfer reactions [64].
Nanoparticles can also be introduced as components of polyelectrolyte multilayer films (PMFs). Importantly, when supported by a porous membrane, PMFs are known to have water permeabilities and ion rejections typical for nanofiltration membranes. The possibility to control the composition of the PMF in terms of its polymeric constituents and nanoparticles fillers and to regulate PMF's separation properties presents unique opportunities for the design of nanoparticleenabled membranes [81][82][83].
Metal-polymer nanocomposite membranes can be prepared by the in situ reduction of unbound ionic precursors in the process of phase inversion, such as silver-polymer membranes were synthesized by reducing ionic silver chemically, by ultraviolet irradiation, or by heat [84][85][86]. Also, two silver incorporation approaches were adapted for NP-Ag-DMF (dimethylformamide). In the first approach, Ag nanoparticles were synthesized ex situ and were added to the casting solutionas Ag-DMF organosol. The organosol was prepared by adding AgNO 3 to DMF(reducing agent) and heating the solution under intense stirring conditions [87].The second approach involved an in situ reduction of ionic Ag + by DMF in the membrane casting mixture.
In this case, AgNO 3 was first dissolved in DMF at room temperature to minimize Ag reduction.

Nanocomposite membranes
Nanocomposite membranes above-mentioned exhibit one to three times the water permeability with the same rejection as commercial reverse osmosis (RO) membranes, and can be imparted with antimicrobial and photo-reactive functionality.
Since Maxwell conceived the ideal membrane [88], membranebased water purification processes are now among the most important and versatile technologies for conventional drinking water production, wastewater treatment, ultrapure water production, desalination, and water reuse.
Commercially available membrane processes for water purification include electrodialysis (ED), electro-deionization (EDI), reverse osmosis (RO), nanofiltration, ultrafiltration and microfiltration (MF). Nanofiltration, ED and EDI find some use in demineralization, softening, and organic separations, but RO and MF membranes are the workhorse technologies for desalination and water reuse. Other membrane based processes such as forward osmosis, membrane distillation, and pervaporation are emerging, but have found limited application in practice.
In principle, intrinsic advantages of membrane processes include continuous, chemical-free operation, low energy consumption, easy scale-up and hybridizationwith other processes, high process-intensity (i.e., small land area per unitvolume of water processed), and highly automated process control. General disadvantages of membrane processes are short membrane lifetime, limitedchemical selectivity, concentration polarization, and membrane fouling [88]. Polarization and fouling of RO membranes require extensive physical and chemical feed water pretreatment (i.e., filtration, acidification, antiscalant addition, disinfection), low flux operation, extensive chemical cleaning, and frequent operator intervention. Reverse osmosis processes further suffer from high intrinsic energy consumption, environmental issues associated with feed water intake and brine discharge, and the need for chemical conditioning of product water.
Nanotechnology promises to dramatically enhance many water purification technologies such as adsorption, ion exchange, oxidation, reduction, filtration, membranes, and disinfection processes [89]. However, one of the key issues related to nanotechnology is the question of how to apply it. Specifically, it isnot clear how to interface nanoparticles with contaminants. At present, many expensive nanoparticles cannot be added to water like commodity chemicals and some nanoparticles could present new hazards to human health and theenvironment [90]. Thus, additional separation processes are required recover nanomaterials for risk avoidance and reuse. A promising approach is to immobilize nanomaterials on or within a solid matrix, such as a membrane. The resulting membrane may exhibit improved separation performance, chemical, thermal, or mechanical stability, interfacial properties, or advancedfunctionality depending on the nanomaterial selected.

Inorganic-organic nanocomposite membranes
In general, nanocomposite materials are created by introducing nanoparticulate materials (the "filler") into a macroscopic sample material (the "matrix") [91]. The resulting nanocomposite material may exhibit drastically enhanced properties such as mechanical properties (e.g., strength, modulus, and dimensional stability); chemical and thermal stability; permeability for gases, water, and hydrocarbons; electrical and thermal conductivity; surface properties, optical properties, or dielectric properties. For example, dispersing molecular sieve nanoparticles into polymers canproduce mixed matrix membrane materials with improved gas mixture perm selectivity [92].
In the literature [93], the synthesis and characterization of zeolitepolyamidethin film nanocomposite (TFN) membranes formed by interfacial polymerization are expressed. The general approach to TFN membrane formation is similar to that of traditional polyamide thin film composite (TFC) membranes, but nanoparticles are dispersed in the initiator solution prior to interfacial polymerization as depicted schematically in Figure 6. Figure 6 shows that the embedded molecular sieve nanoparticles throughout the polyamideth in film layer of an interfacial composite RO membrane. Synthesized NaA zeolite nanoparticles, characterized by a super-hydrophilic and negatively charged three-dimensional molecular sieve pore network, are used as the dispersed nanophase. Thin film nanocomposite membranes offer new degrees of freedom in designing NF and RO membranes because the nanoparticle and polymer phases can be independently designed to impart a wide array of separation performance and novel functionality. Where, the zeolite nanoparticles (NaA-type) were synthesized from the Na 2 O-SiO 2 -Al 2 O 3 -H 2 O system with the use of an organic template (tetramethyl ammonium hydroxide) by a hydrothermal reaction. The as-synthesized zeolite A nanoparticles are porefilled because the presence of the template inside the zeolite pore structures. Pore-opened zeolite nanoparticles were obtained from the pore-filled particles by removing the template by calcinations, assisted by a polymer network as designed as atemporary barrier to prevent nanoparticle aggregation during the calcination process.
Both TFC and TFN membranes were hand-cast on preformed polysulfone ultrafiltration (UF) membranes (provided by KRICT, Korea) through interfacial polymerization. A UF membrane taped to a glass plate was placed in an aqueous solution of 2% (w/v) m-phenylenediamine(MPD, >99%,Sigma-Aldrich) for approximately 2 min, and MPD soaked support membranes were then placed on a rubber sheet and rolled with a rubber roller to remove excess solution. The MPDsaturated UF membrane was then immersed in a solution of 0.1% (w/v) trimesoyl chloride (TMC, 98%, Sigma-Aldrich) inhexane. After 1 min of reaction, the TMC solution waspoured off and the resulting membranes were rinsed with an aqueous solution of 0.2% (w/v) sodium carbonate (Na 2 CO 3 , HPLC grade, Fisher Scientific). Nanocomposite membranes were made by dispersing 0.004-0.4% (w/v) of synthesized zeolite A nanoparticles in the hexane-TMC solution. Nanoparticle dispersion was obtained by ultrasonication for 1 h at room temperature immediately prior to interfacial polymerization. Figure 7 shows the morphology of the final synthesized zeolite nanoparticles.  Figure  8(f) around features are believed to be zeolite nanoparticles because of their more "cubic" shape, which is consistent with the morphology of zeolite A nanoparticles. Also, EDX analysis confirms nanoparticle presence at discrete locations within the film layer and at the interface.
Results show that the formation of zeolite-polyamide nanocomposite thin films by interfacial polymerization are resulted in reverse osmosis membranes with dramatically improved permeability and interfacial properties when compared to similarly formed pure polyamide thin films. This new concept combines important properties of conventional membrane polymers(flexibility, ease of manufacture, high packingdensity modules)with the unique functionality of molecular sieves (tunable hydrophilicity, charge density, pore structure, and antimicrobial capability along with better chemical, thermal, and mechanical stability).Water molecules appear to flow preferentially through super-hydrophilic, molecular sieve nanoparticle pores, while solute rejection remains comparable to pure polyamide membranes. Pendergast et al. [94] assess the compaction behavior of hand-cast nanocomposite supported polyamide composite membranes relative to polysulfone supported olyamide composite membranes to help understand and control irreversible, internal fouling of RO membranes by physical compaction. Support membrane preparation was prepared with addition of N-methyl pyrrolidone (NMP) (Acros Organics, Morris Plains, New Jersey, USA) to a mass of transparent polysulfone  The resulting composite membranes were heat cured for 10 min at 82 °C, washed thoroughly with deionized water, and stored in deionized water until performance testing. Subsequently, membrane samples were placed into the crossflow membranemodules and compacted with a 10 Mm NaCl feed solution at pressuresof 1700 and 3400 kPa (250 and 500 psi). The compaction tests continued until a steady-state flux was obtained for both membranes (typically, 16-20 h), at which point the membranes were removedand stored in a desiccator for subsequent characterization. Observed permeate water flow rate was recorded every 30 min. Figures 9-15 show the cross-sectional thicknesses of uncompacted (as-cast) and compacted RO membranes analyzed by SEM. It indicates that membranes containing nanoparticles undergo less compaction, while the pure polysulfone membrane experiences a drastic change in thickness and support structure. In general, TFN supported membranes undergo less compaction, and all the membranes containing nanoparticles appear to maintain their uncompacted porous structure following compaction, while pure polysulfone supported membrane (TFC) macrovoid morphology changes after compaction. Hence, the addition of inorganic nanoparticles increases the mechanical stability and, therefore, decreases physical compaction of the nanocomposite supported membranes. The performance advantage of nanocomposite supported membranes is greater at higher applied pressure, and nanocomposite-supported RO membranes represent one potential approach to mitigate internal, irreversible fouling due to membrane compaction, particularly in high-pressure applications like brackish and ocean water desalination.
Fathizadeh et al. [95] also used the nano-NaX zeolite synthesized via the hydrothermal method to investigate the effect of nano-NaX zeolite dispersed in the zeolite-polyamide thin film nanocomposite on the membrane performance, the matrix structure, film thickness and surface hydrophilicitity.
Results show that nano-NaX zeolite increases physical and chemical stability properties. Also, the surface properties such as contact angle, RMS roughness and interfacial free energy as well as water permeability improved by increasing the content of nano-NaX in the polyamide structure. An enhancement in the concentration of MPD and TMC monomers lead to production of TFN membrane with high water permeability and low solute rejection. The TFN membrane with high concentration of TMC (0.15%, w/v), MPD (3%, w/v) and nano-NaX zeolite (0.2%, w/v) (B4 sample) has the highest waterflux and the lowest salt rejection. The A4 TFN membrane demonstrates      a good separation efficiency, productivity flux and thermal stability between the TFN and TFC membranes.

Biomimetic membranes
Biomimetic membranes are designed to mimic the highly selective transport of water and solutes across biological membranes [96][97][98][99]. The general approach to forming biomimetic membranes is to incorporate active transport proteins (isolated from cell cultures) within a vesicular or planar lipid bilayer or a more stable synthetic analog. Liu et al. [100] developed a novel biomimetic absorbent containing the lipid triolein for removing persistent organic pollutants (POPs) from water in 2006 using the concept of bioaccumulation. A cellulose acetate (CA) polymer was chosen for preparing hybrid materials because it can be easily molded into different forms such as membranes, fibers and spheres. Furthermore, its hydrophilicity improves the accessibility of aqueous solutions to the surface of the film. The absorbent was prepared by embedding triolein into cellulose acetate (CA) spheres and the temperature ranges of the CA solution and liquid olefin were kept at 30-35ºC and 15-20ºC respectively. Finally, composite absorbents were washed with distilled water to remove all soluble impurities. It shows that the biomimetic absorbent can be prepared with these pollutionfree and environmentally friendly raw materials and the spherical absorbent is easy to use in water treatment processes. Figure 16 presents SEM micrographs of the absorbent cross sections containing no triolein and those containing 2.0 % (w/w) triolein. Figures 16(a) and (c) show that both absorbents present a very thin layer from the top surface. Triolein dropletscannot be seen in the layer that should be the CA membrane.This could be explained by the aggregating rate of CA into theair being faster than that of triolein. In addition, triolein loadingis likely to affect the process of formation of the CA membrane.A careful inspection of the images of Figure 16(a) and (c) reveals that the layer of the CA-triolein absorbent is thinner than that of CA. The thickness of the CA and CA-triolein absorbentlayers is about 10 mm and 5 mm, respectively. The image of the CA-triolein absorbent shows fewer pores and triolein droplets that have been embedded into the hole formed by the CA fiber. Triolein is wrapped in the thin CA polymer, and constitute a significant fraction of the lipid pool in the absorbent. This biomimetic absorbent should be most effective in its capacity to accumulate lipophilic substances. In addition, the thin CA polymer canprevent triolein leakage. Results reveal a higher affinity of CA and CA-triolein absorbents for the most hydrophobic and less polar compounds of the group. The role of triolein is probably similar to that of an organic solvent in a liquid-liquid extraction. The distribution of a neutral organic compound between the biomimetic absorbent and water can be correlated with the octanol-water partition coefficient. It also indicates that triolein is effectively embedded into CA spheres and that a thin film of CA surface layer containing no triolein is formed, thus ensuring that virtually no triolein leaks out through the CA membrane. The absorbent are promising for the removal of non-polar compounds. The CA-trioleinabsorbent can efficiently and quickly accumulate hydrophobic POPs from water. Lower residual concentrations of selectedPOPs are obtained by the CA-triolein absorbent in comparison with the CA absorbent.
Moreover, James et al. [101] propose an optrode based on the use of ion imprinted polymers (IIPs) particles prepared via chemical immobilization afterthe formation of mixed ligand complex, uranium-4-vinylphenylazo-2-naphthol (VPAN)-4-vinyl pyridine (VP) and then polymerization with matrix forming monomers. The predetermined selectivity, operational stability ofthe sensing material (IIP) is more advantageous compared to conventional optical sensors.
Membranes incorporating bacterial Aquaporin Z proteins have been reported to show superior water transport efficiency relative to   conventional RO membranes. Aquaporins were incorporated into the walls of self-assembled polymer vesicles constituted of tri-block copolymer, poly (2-methyl-2-oxazoline)-blockpoly (dimethylsiloxane)block-poly (2-methyl-2-oxazoline). An initial permeability test was carried out on the aquaporin-triblock polymer vesicles by stopped-flow light-scattering experiments. The results reported at least an order of magnitude improvement in permeability compared to commercially available TFC RO membranes. Although a salt separation test has yet to be reported,extremely high salt rejection is expected from aquaporins sincetheir functional biological performance is to only allow the passageof water molecules. Hence, there is an ideal opportunity forthe production of ultrapure water. These studies haveso far been limited to investigating water permeability propertiesacross a barrier layer composed of aquaporins and triblock polymers [102].

Carbon nanotubes (CNTs) membranes
It is distinct that the availability of safe, clean, and inexpensive water has emerged as an issue that defines global problem in the twenty-first century. Water short ages are some of the root causes of societal disruptions such as epidemics, environmental disasters, tribal and ethnic conflicts, growth shortfalls, and even countrywide political destabilization. Membrane-based filtration is the current leading energy-efficient technology for cleanup and desalination of brackish water, recycled water, and seawater. Membrane-based filtration offers other advantages as filtration through the tight membrane pores can also remove dangerous impurities, such as As, as well as toxic large organic compounds. Factors that limit the efficiency of the membrane purification technologies include the membrane resistance to the flow, membrane fouling and membrane imperfections that lead to incomplete rejection or to a drop in the membrane rejection properties over time. The technological developments and high-efficiency energy recovery systems in particular have pushed the current efficiency of reverse osmosis (RO) membranes to a very impressive 4 kWh/m 3 [103,104]; However, this number is still well above the theoretical minimum energy cost of 0.97 kWh/m 3 for 50 percent recovery [105,106]. To move further, the transformative membrane technologies that utilize fundamentally new transport and filtration mechanisms for drastic gains in transport efficiency need to be developed.
Carbon Nanotubes (CNTs) [107,108] are unique nanosystems with extraordinary mechanical and electronic properties, which derive from their unusual molecular structure. An ideal carbon nanotube can be thought of as a single graphite layer (graphene sheet), rolled up to make a seamless hollow cylinder. These cylinders can be tens of microns long, with diameters as small as 0.7 nanometers and are closed at both ends by fullerene-like caps. CNTs having wall thickness of one carbon sheet are named single-wall carbon nanotubes (SWCNTs). In consequence of the Vander Waals interactions between nanotubes, they often aggregate in large ropes: ordered arrays of SWCNTs arranged on a triangular lattice. SWCNTs can be considered as the building blocks of multi-wall carbon nanotubes (MWCNTs), which consist of a coaxial array of SWCNTs with increasing diameter. MWCNTs are also usually long many microns, with the external diameter that ranges from two to several tens of nanometers, providing very high aspect ratio structures, as shown in Figure 17 [108].
By now the CNT has firmly established itself as the iconic molecule of nanoscience [109][110][111][112][113]. Several methods of CNT production currently exist. In the laboratory environment, catalytic chemical vapor deposition (CVD) is preferred over other methods such as arc discharge and laser ablation because it produces higher quality CNTs. CVD reactors can produce individual isolated nanotubes as well as densely packed vertically aligned arrays. Unfortunately, the ultimate goal of the CNTs synthesis-producing a uniform population of nanotubes with a given chirality-still remains elusive. Several studies indicated that the size of the catalyst particle during the growth stage determines the size of the CNT to less than 10 percent; yet efforts to control the size of the CNTs with greater precision have been largely unsuccessful. Thus, synthesizing a vertically aligned CNT array with a narrow distribution of sizes still remainsa difficult endeavor requiring considerable process development and optimization efforts [114][115][116].
In the last decade, the unique geometry and internal structure of carbon nanotubes (CNTs)give rise to the newly discovered phenomena of the ultraefficient transport of water through these ultra-narrow molecular pipes. Water transport in nanometer-size nanotube pores is orders of magnitude faster than transportin other pores of comparable size [111,112,[115][116][117][118].

Simulation/dynamic analysis on water transport in carbon nanotubes (CNTs)
It is extremely important to understand the mechanism that hydrophilic liquids, especially water, enter and fill very narrow and hydrophobic CNTs. If the water does enter the CNTs, what influence does crucial confinement have on the water structure and properties, such as how these changes in structure influence the rate, efficiency, and selectivity of the transport of liquids and gases through the CNTs. As the traditional method, molecular dynamics (MD) simulations or dynamic analysis will provide the exact answers to these questions.
It is well known that confined matter on the nanometer scale differs significantly from bulk matter. The special properties of confined water can influence molecular transport inside membrane pores. Confinement of water can be reached by limiting the size of water agglomerates in one, two, or three dimensions. In an infinite ice crystal each water molecule, complying with the "bulk ice rule" [119], is tetrahedrally coordinated, simultaneously donating and accepting two hydrogen atoms, forming a hydrogen bonded network. However this arrangement is disrupted in agglomerates of crystalline water of finite sizes, leading to a variety of shapes for small water clusters with different types of OH groups ranging from bulk-like to essentially free OH groups. This diversity can be readily detected by means of vibrational spectroscopy that provides one of the most insightful means for OH characterization. In order to express the details of water phase inside single-walled carbon nanotubes (SWCNTs), Byl et al. [120] studied the unusual hydrogen bonding in water-filled carbon nanotubes. They had performed classical molecular simulations (MD) for water confined in SWNTs and also carried out quantum mechanical density functional theory (DFT) calculations for water in ring structures in vacuum and inside a SWNT. Where: classical molecular simulations were used to identify structural, energetic, and vibrational properties of water in SWNTs and in the bulk phase at low temperatures. The average energies for water confined in armchair (8,8), (9,9), (10,10), and (11,11) SWNTs are shown in Figure 18. Atlow temperatures water forms stacked ring structures in all of the nanotubes considered. See, for example, the water structure in a (10, 10) nanotube shown in Figure  19. The number of water molecules in a ring depends primarily on the diameter of then anotube. However, the (8,8) and (10,10) nanotubes can support different polymorphs. Five-and four-membered rings are observed in the (8,8) nanotube, while the (10, 10) nanotube can have both eight-and seven-membered rings, as shown in Figure 18. Order-to-disorder structural transitions, indicated by rapid rises in the potential energy with temperature ( Figure 18), occur for water confined in (10,10) and (11,11) SWNTs; the water remains well-ordered in the smaller diameter nanotubes, even at 298 K. The nanotubes were filled at room temperature for computational efficiency.
Results show that H 2 O molecules confined inside of SWNTs form ring structures that involve hydrogen bonds of two types. Hydrogen bonds within the ring structure exhibit frequencies like those found in bulk H 2 O and O-OH angles of near 5°. Hydrogen bonds formed between neighboring rings exhibit an unusual stretching frequency at 3507 cm -1 and are associated with larger O-OH angles near 17°. The strained angles and unusual IR mode are a direct result of the confinementinduced stacked ring structures, which would not be stable in the bulk. It is possible that water in other confined environments will exhibit similar distinct stretching frequencies, showing that IR spectroscopy, coupled with atomistic modeling, is proving to be a powerful tool for probing the structure and energetics of confined water.
Because open-ended CNT membranes, which are formed by filling the interstitial region between 1 and 10 nm diameter CNTs with a non-porous matrix material, have been proposed as next-generation desalination, results from molecular dynamics (MD) simulations suggest that flow enhancement in CNTs is caused by liquid slip at the water/carbon boundary and confinement-related changes in the liquid viscosity [121][122][123]. Thomas et al. [124,125] used MD simulation to study water flow in CNTs with diameters between 4.99 nm and 1.66 nm. It is predicted the variation in flow enhancement with CNT diameter and identified how it is related to the water viscosity and the water/CNT slip length. Results show that liquid flow through CNTs with diameters as small as 1.66 nm can be described using the slip modified Poiseuille relation. In addition, predictions from MD simulations indicate that (i) a continuum description of water flow is appropriate inside CNTs larger than 1.66 nm, (ii) diameter-related changes to the viscosity and slip length must be considered when modeling liquid flow, and (iii) the flow enhancement factor decreases monotonically with increasing diameter for smooth walled tubes. The variation in flow enhancement predicted using MD is qualitatively consistent with the majority of the experimental data. Since the MD simulations considered flow though atomically smooth and defect-free CNTs, discrepancies between simulation and experiment are likely caused by differences in the structure and/or chemistry of the carbon surface.
Furthermore, since the properties of nanotubes conveying fluids are highly sensitive to the vibrational characteristics, investigation of CNTs'dynamical properties is very important. Hashemnia et al. [126] described the dynamical analysis of carbon nanotubes conveying water considering carbon-water bond potential energy and nonlocal effects. Where: it was assumed that fluid media was not continuum and the interaction between carbon atoms and water molecules was modeled using carbon-water bond potential energy. Also, the effects of different parameters such as flow velocity, CNT's chirality, CNT's diameter, CNT's aspect ratio, elastic foundation stiffness and various end conditions on the fundamental frequencies of SWCNT's vibration were studied. Results show that by increasing the flow velocity, the natural frequencies are decreased. The molecular based water mass per unit length of CNT assumption predicts much larger natural frequencies than those predicted by continuum based one. Also, the  (8,8), (9,9), (10,10),and (11,11) SWNTs at temperatures ranging from 123 to 318 K fromparallel tempering NVT Monte Carlo simulations. wall shear force calculation using continuum or molecular model results in similar fundamental frequencies.
Verification of these seemingly exotic predictions of fast transport through CNTs that emerged from the MD simulations requires fabrication of a robust test platform: a CNT membrane. Such membranes typically consist of an aligned array of CNTs encapsulated by a filler (matrix) material, with the nanotube ends opened at the top and bottom. Carbon nanotube (CNT) membranes are promising candidates for one such solution primarily because of their transport characteristics. The inner cavity of a CNT forms a natural pore with very small diameter that can in some instances be smaller than 1 nm. Moreover, smooth hydrophobic surfaces of the nanotubes lead to nearly frictionless flow of water through them, enabling transport rates that are orders of magnitude higher than transport in conventional pores. Finally, the structure of CNTs permit targeted specific modifications of the pore entrance without destroying the unique properties of the inner nanotube surface [127][128][129][130][131]. The combination of these three factors could enable a new generation of membranes whose transport efficiency, rejection properties, and lifetimes drastically exceed those of the current membranes.
Mostafavi et al. [132] fabricated a hollow cylindrical nanofilter with suitable mechanical strength from CNT with nanoscale porosity to remove MS 2 virus from water. Figure 20 shows that the CNTs are arranged so closely that ablock is formed and the length of the nanotube bundles is consistent to the thickness of the block. As shown in Figure  20(b), the nanofilter withsuitable porosity is prepared by closely arranging the nanotubes. Further more, Figure 20(c) shows that CNT bundle s are about 45 nm in size. The TEM image in Figure 21 shows that these entangled carbon nanotubes bundles are almost composed of MWNTs. The diameter of these carbon nanotubes among the bundles is about 30 nm. The fabricated nanofilter presents the largest pore size with a permeability of 30.7×10 -6 m.s -1 Mpa -1 while the NF90 seems to have the smallest pore size with a low permeability of 10.5×10 -6 m.s -1 Mpa -1 . Further, the permeability of the ESNA1-LF and NF270 are 20.3×10 -6 m.s -1 Mpa -1 and 25.8 m.s -1 Mpa -1 , respectively. The efficiency of the virus removal by using the fabricated nanofilter was expressed by a log reduction value (LRV). Figure 22 shows filter efficiency at different pressures. With increasing pressure, the velocity of water increases through nanofilter but mass transfer velocity of virus is low and hence the different mass transfer velocity causes that the virus concentration decreases in permeate with increasing pressure. These results indicate that fabricated nanofilter can remove virus and nanoparticles from water at 20ºC and pressure of about 11 bar.
It should be pointed out that some different approaches to producing an aligned CNT-polymer composite membrane (Polymeric/ CNT Membranes) have been investigated, such as a polypyrrole (PPy)-CNT-tyrosinase biocomposite film prepared by electrochemical polymerization [133], carbonized electrospun [134] and a MWCNT/ polyaniline (PANI) multilayer film formed by alternate casting of treated MWCNT-ethanol dispersion and electrodeposition of aniline [135].
In 2006, Choi et al. [136] prepared the multi-walled carbon nanotubes (MWNTs)/polysulfone (PSf) blend membranes by a phase inversion process, using N-methyl-2-pyrrolidinone (NMP) as a solvent and water as a coagulant. Before making the blend membranes, MWNTs were first treated with strong acid to make them well dispersed in organic solvents such as NMP for the preparation of homogeneous   MWNTs/PSf blend solutions. Where: polysulfone (PSf, Udel® P 3500, Amoco, Marietta, OH) was used as a membrane material. Carbon nanotubes (multi-walled carbon nanotubes, MWNTs) manufactured by CVD process whose purity of greater than 95% were used for the modification of a PSf membrane. The MWNTs in tubular shape, composed of six-membered carbon rings, were like a rolled graphite sheet with 10-20 nm of outer diameter, 10-50μm of the length, and 4.3±2.3 nm of inner diameter. The MWNTs were surface modified with strong acids, concentrated nitric (HNO 3 ) and sulfuric acids (H 2 SO 4 ) (1:3 invol. %), to make them easy to be dispersed in the organic solvents. Also, Polyvinyl pyrrolidone (PVP) with molecular weight of 55,000 g/ mol and poly(ethyleneoxide) (PEO) with molecular weight of 100,000 g/mol were used for permeation tests. Figures 23 and 24 show the FESEM photographs of the surfacesand the fractures of the MWNTs/PSf blend membranes, respectively. As one can see from Figure 23, by the MWNTs, the PSf membrane surface became rougher with increased pore size. The PSf membrane with 1.5 wt. % MWNTs looks to have the highest surface roughness and the largest pore size. These pictures suggest that the modified MWNTs with hydrophilic functional groups make the PSf membrane to have nodular structure, with increased pore size and rough surface. This result might be explained by the fast exchange of solvent and non-solvent in the phase inversion process due to the hydrophilic MWNTs.
However when the content of MWNTs increased further, especially when it was 4 wt. %, the surface structure started to become smooth again. This is maybe due to the increased viscosity of the MWNTs/ PSf blend solution. As explained above, the viscosity of the blend solutions increased along with the contents of MWNTs. The increased viscosity usually retards the exchange of solvent and non-solvent, making smooth membrane surface. In such case, two factors (increased hydrophilicity and increased viscosity by the added hydrophilic MWNTs) acted at same time for the formation of the microporous blend membranes. When the content of the added MWNTs was less than 4 wt.%, increased hydrophilicity of thesolution did major role to form a nodular structure, but when it was more than 4 wt.%, increased viscosity of the solution wasthe major factor to make a smooth membrane surface.
On the other hand, there was not a distinct difference in the structures of the fractures of the blend membranes, all showing finger like structure (Figure 24) except for the different numbers of the MWNTs positioned on the surface layer. With increasing contents of MWNTs, more MWNTs were found from the surface layer, making the membrane surface hydrophilic. Results show that the MWNTs turned out to be a good modifier for the formation of functional microporous PSf membranes, controlling the hydrophilicity of the membrane surface, adjusting the pore size and porosity. The permselective properties of the MWNTs/PSf blend membranes were appeared to be very dependent on the contents of the MWNTs used. By using proper amount of MWNTs, it was possible to increase the flux and the solute rejection at the same time for the PSf microporous membranes.
After that Peng et al. [137] described the novel nanocomposite membranes (PVA-CNT(CS)) prepared by incorporating chitosanwrapped multiwalled carbon nanotube (MWNT)into poly(vinyl alcohol) (PVA).To fully explore the potential of CNT, two issues have to be solved. One is the serious aggregation of CNTs leading to difficulties in their manipulation and incorporation into polymeric matrixes. Another is the strong hydrophobicity leading to significant decrease of membrane selectivity. Hydrophilic modification seems a promising and feasible solution. Therefore, chitosan (CS) with distinct hydrophicility due to the existence of high proportion amino and hydroxyl groups was introducedin the study. The dispersion and solubility behavior of CNT can be remarkably improved through substantial wrappingof chitosan utilizing the emulsifying capacity of chitosan, the unique solubility behavior of chitosan. In addition, as a rigid polymer, chitosan has relatively favorable free volume characteristics. In their previous study [138], they found t hat blending chitosan into PVA matrix could considerably enhance the pervaporation performance of benzene and cyclohexane mixtures. Authors firstly investigated the free volume characteristics of PVA-CNT nanocomposite membranes by molecular dynamics simulation (using SWNT instead of MWNT in MD simulationfor simplicity) in order to theoretically elucidate whether or not CNT is an effective inorganic component to improve the pervaporation properties of PVA-based membranes according to the obtained free volume data. Four amorphous cell models forpure PVA membrane, PVA-CNT nanocomposite membrane, PVA-CS blend membrane and PVA-CNT(CS) nanocomposite membrane were constructed by molecular dynamics simulations carried out on Materials Studio (Discover and Amorphous Cell modulus) developed by Accelrys Software Inc. and shown in Figure 25. Figure 26 show the cross-section of PVA-CNT(CS)nanocomposite membrane, where the CNT content was2.0 wt.%. The CNTs wrapped by chitosan were well dispersed within PVA matrix, which indicated that chitosan was a suitable polymer to alleviate the serious aggregation of CNTs.
Results demonstrate that novel nanocomposite pervaporation membranes composed of PVA and chitosan-wrapped CNT could exhibit simultaneous increase of permeation flux and separation factor compared with pure PVA membrane. It is supposed that the improvement originates from the appropriate physicochemical and structural characteristics of these nanocomposite membranes: (1) carbon nanotube showed preferential affinity toward benzene; (2) both carbons nanotube and chitosan could modify PVA polymer chain packing, f urther increased free volume and tailored free volumecavity size of PVA membrane.
In recent years, research on membranes with CNTs has been explored further [139][140][141]. However, up to now although CNTs have excellent separation, electrical and mechanical properties, and synthesis of mixed matrix membrane susing these materials have the following problems. First, during the CNTs synthesis, some impurities are produced and they must be purified by chemical or physical methods. Second, the synthesized CNTs are generally close ended and they must be cut during the purification processes or separately. Third, dispersion and dissolution of the as-grown CNTs are not usually sufficient in various organic solvents and different polymers and interaction of the interface between CNTs and polymer matrix is weak. Therefore, synthesis of highly pure, open ended and functionalized CNTs is important to form the membrane with the high quality. Nowadays, Shawky et al. [142] synthesize the multi-wall carbon nanotube (MWCNT)/aromatic polyamide (PA) nanocomposite membranes by a polymer grafting process. Polymerization was carried out in a 500 mL flask with a very high-speed magnetic stirrer. A solution of 4.326 gm (0.04 mol) of m-phenylene diamine and 8.48 gm (0.08 mol) of sodium carbonate in120 mL deionized water was placed in the flask. Vigorous stirring was begun and a solution of 8.12 gm (0.04 mol) of isophthaloyl chloride in150 mL of tetrahydrofuran was rapidly poured into the flask from a beaker. Stirring was continued for 5 min. The product thus obtained is a white fibrous precipitate. The polymer was separated by filtration; washed with excess deionized water and thus dried under vacuum at 80-90°C. The yield of polymer is about 100% of the theoretical value. For the preparation of MWCNT-PA composite membrane, MWCNTs were dispersed in DMAc solvent via ultrasonication for 2 h. DMAc solvent was prepared with lithium chloride salts (1% (wt./v)) as a casting solution for the PA membranes. This DMAcLiCl solution   was prepared by adding 0.1 g of LiClto 10 mL of DMAC and heating to complete dissolution. Multi-wall carbon nanotubes were supplied from NanoTech Labs Inc., USA, and characterized by the company as follows; purity of 95 wt.% an average diameter of 15 nm and lengths ranging from0.5 μm to 1 μm with most of the materials closer to 1 μm. The addition of benzoyl peroxide (BPO) initiator was done with the goal of forming of free-radicals on both CNTs and PA resulting in polymer-grafted nanotubes that would be better dispersed throughout the casting solution and forming morehomogenous MWCNTs-PA composite membranes. After the addition of benzoyl peroxide (BPO) initiator (0.25% (wt./v)), the mixture was heated with stirring for 3 h at 80 °C. Stirring was continued for 24 h, then the casting solution was evacuated to remove the dissolved gas, casted onto a dried clean glass Petri dish and spread with the aid of a glass rod to form a uniform thin film. A membrane thickness of 200 μm was obtained by controlling the amount of casting solution. Thus, the film was immediately placed in an oven at 90 °C for 30 min. When the solvent was completely evaporated, the Petri dish with the membrane was cooled and immersed in a deionized water bath for atleast 15 h at room temperature. Four different membranes were synthesized containing 2.5, 5, 10 and 15 mg MWCNTs/g PA,respectively with a constant PA concentration of 10% by weight inthe casting solution.
SEM imagery of the functionalized MWCNT in the resulting film showed MWCNTs to be well mixed and evenly distributed throughout the polymer matrix even at the highest concentration ( Figure 27) in contrast with earlier efforts to create CNT/polymer composite membranes where CNTs were non-uniformly distributed. No evidence was observed of CNT clustering, even as CNT concentration increased to 15 mg MWCNT/g PA. Results show that the addition of MWCNTs to the PA membranes to form a nanocomposite structure improved the mechanical properties of these membranes and their ability to reject key contaminants with little compromise in membrane permeability. However, there may be a reduction in longer-term membrane performance due to adsorptive fouling since the addition of MWCNTs increased membrane hydrophobicity. Additional factors to consider in the development of thesecomposites include the need to balance the dimensions of the membrane film cast with those of the reinforcing CNTs, and the effectof derivative CNTs on membrane casting and performance.
In order to investigate the protein fouling behavior of carbon nanotube/polyethersulfone composite membranes during water filtration, Celik et al. [143,144] prepared the multi-walled carbon nanotube/polyethersulfone (C/P) composite membranes via the phase inversionmethodwith bovine serum albumin (BSA) and ovalbumin (OVA) used as the model protein for assessing the protein fouling behavior.
To prepare the composite membranes, functionalized MWCNTs were dispersed homogenously in NMP prior to dissolving 20% PES in the blend solution while continuously stirring and heating at 60ºC until the solution became completely homogenous. The resultant polymer solution was ultrasonicated to allow a complete release of air bubbles. The blend solution was then casted on a glass plate using a casting knife at room temperature, and the glass plate was immersed in a coagulation bath of DI water. The pristine membranes were peeled off and subsequently rinsed with and stored in DI water until use. Note that membranes marked as C/P-0% refer to the original polyethersulfone membrane and C/P-4% refers to the membranes prepared in a casting solution in which the amount of MWNTswith respect to polyethersulfone was 4% by weight.
The adsorption of BSA onto C/P composite membranes is demonstrated in Figure 28. At pH 3, electrostatic attraction ccurred since the protein and all composite membranes had opposite charges. Moreover, at pH values below the IEP, the protein denaturation is relatively high. Hence, at pH 3 a higher amount of protein adsorption was observed on the membranes. At pH 7, both BSA and composite membrane surfaces were negatively charged, and electrostatic repulsion was dominant between the membrane surfaces and the protein. In this case, protein adsorption decreased with an increasing amount of MWCNTs in the blend solution, at both pH 3 and pH 7, due to the increased hydrophilicity of the composite membranes. And even though the protein adsorption decreased with increasing amounts of MWCNTs in the membrane structure, the protein adsorptions on the C/P-2% and C/P-4% membranes were similar at a 95% confidence level, probably due to the similar hydrophilicity of these two membranes.
Results show that the C/P composite membranes were fouled less compared to the barepolyethersulfone (PES) membrane at 4 h of static protein adsorption at neutral pH. Moreover, the irreversible fouling ratio of the C/P composite membranes was less than the bare PES membrane after 1 h of protein ultrafiltration, and the flux recovery ratio of the C/P composite membranes was higher than the bare PES membrane after 20 min of DI water filtration. Based on these results, C/P composite membranes were shown to have the potential to alleviate the effects of protein fouling, thereby enabling C/P composite membranes to be used for several runs of protein filtration after simple washing with water.
It should be realized that the most promising property of CNT membranes for water purification applications is their extremely high permeability [145,146]. This property should translate into more water per unit of applied pressure, more efficient, smaller purification units and ultimately into lower purification or desalinations costs. Rich possibilities for chemical functionalization, coupled with the rather unique ability to manipulate only the chemistry at the nanotube mouth open up the possibility of producing membranes tailored for specific applications (e.g., RO desalination or impurity purification) while maintaining the basic membrane structure and high permeability.
However, a true assessment of the potential impact of CNT membranes on water purification (and specifically on water desalination) applications requires a more comprehensive comparison of the membrane characteristics with the general requirements of the membrane purification process. At least in the case of RO desalination, the process efficiency comes from three main sources: capital costs, energy costs, and operation costs (which include costs for pretreatment, post treatment, and membrane cleaning and regeneration).
Meanwhile, the CNT technology is still in its infancy; therefore, most attempts at quantitative evaluation will face large uncertainties associated with predicting the future technological milestones, or the fact that some of the major membrane characteristics (e.g., fouling properties) have not been sufficiently evaluated. Another large source of uncertainty is the lack of availability and cost estimates for a manufacturing process that allows scale-up of membrane fabrication. However, some qualitative conclusions based even on the limited set of data that is available now can be still reached. The high flux of CNT membranes provides a clear advantage for both the energy costs and the capital costs, as the same amount of product water could be obtained with smaller driving pressures and less membrane area. But, some of the other important advantages of CNT membrane technology could come from the factors contributing to the third cost factor: the operation costs. The uniform pore size of CNT membranes could simplify or even eliminate the requirements for complicated multistage pretreatment efforts. The membrane pore surface is also rather chemically inert, which could increase the membrane lifetime against the harsh agents used for pretreating water before RO or other purification steps. Unlike most polymeric membrane surfaces, the CNT membrane surface is hydrophilic; therefore, it could offer an increased resistance to fouling, as well as easier cleanup by rinsing or backwashing. These factors could all contribute to an increased membrane lifespan and ultimately to operation cost savings.
If considering these factors, it becomes clear that the real impact of CNT membrane technology may lie in its potential to improve all of the major areas that contribute to the costs of water purification processes.
Clearly, much work needs to be done before these promises translate into field applications. Researchers need to develop approaches for fabricating CNTs with an even narrower distribution of the pore sizes, ideally with pores that are less than1 nm. Targeted chemical modification of the pore entrances should improve dramatically the rejection characteristics of the membrane. Further studies are necessary to quantify the membrane fouling resistance and useful lifespan. Finally, development of large-scale, low-cost manufacturing processes is imperative to ensure that CNT membrane technology can achieve significant penetration into the water purification market.

Issueson human health and environmental effects
To date, nanotechnology continues to evolve and develop at breakneck speed. Many nano-based applications are already available commercially but little is known about the human health and environmental effects of exposure. This rapid surge in worldwide research and development and the potential large-scale use of nanomaterials in consumer products point toward the need for more definitive information regarding environmental health and safety impacts from the manufacturing process, disposal (industrial or personal), remediation, and treatment applications. It is noted that the applications now show enormous promise for improving public health and the environment-specifically water quality. However, because of nanomaterials' exceedingly small size, the assumption is that they will be extremely mobile in porous media, thereby increasing the likelihood of human exposure due to dispersion and potential persistence in the environment. Yet, how reactive such materials could be in the environment, what residual compounds are formed during degradation, and where and how these materials partition to various environmental and biological media are crucial but largely unknown. How these materials move from one medium to another, from one organism or ecosystem to another, and from organisms to the environment and vice versa will be critical for understanding and implementing proper manufacturing, usage, and recycling/disposal options that are most protective of human health and the environment. In order to effectively assess these impacts, a full life-cycle perspective (impact of a product from the accumulation of starting materials to the development, manufacture, use, and eventual disposal or reuse of the item or portions thereof) of the various constituents and end products is an important component of a research framework. Although complete, robust life-cycle analyses are difficult, and potentially impossible, when data is limited, and this should not be an excuse for ignoring the full life cycle in any risk analysis.

Current&future developments
For the application in a desalination process, the distinguished membranes must exhibit a number of characteristics such as high water flux, high salt rejection, mechanical stability, resistance to fouling, and low cost. A number of polymer materials such as cellulose acetates, thin film composite (TFC) polyamides, cross-linked poly(furfuryl alcohol) and sulfonated polyethersulfone have been investigated for desalination. Although cellulose acetates and TFC polyamides have been the most successful, cellulose acetate membranes slowly hydrolyze overtime, are easily attacked by bacteria, and generally are not used above 35 o C. Aromatic polyamides, although without those problems, have low resistance to fouling, as well as a low chlorine tolerance due to the existence of secondary amides and electron-rich aromatic rings. The currently available substrates used for the preparation of TFC membranes include polysulfone and polyethersulfone. Most of these membrane substrates, however, have relatively poor chemical and thermal stabilities, and for ultrafiltration (UF) are susceptible to performance degradation caused by fouling. Therefore, the development of more thermally and chemically resistant selective top layers and substrates is critically required for water purification.
Up to now, mostpressure-driven membranes are in these two categories, that is, phase separation membranes and thin film composite (TFC) membranes. According to pore diameters, there are four categories of pressure-driven membranes-these include microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO) membranes. Nanofiltration, defined by many "as a process between Ultrafiltration (UF) and Reverse Osmosis (RO)," is a relatively recent technology, largely developed over the past decade. Typically, NF membranes have sodium chloride rejections between 20 and 80 percent and molecular weight cutoffs for dissolved organic solutes of 200-1000 Dalton. Their low-pressure operation (4-14 bars) provides increased energy savings with significantly lower installation and operating costs. A number of polymer materials have been investigated for NF applications. For instance: Cellulose acetate (CA) membrane and polyamide TFC membrane have been the most successful: 1. Cellulose acetate membranes. Cellulose is a linear, rod-like material that renders membranes mechanically robust. Cellulose acetate was the first high-performance NF/RO membrane material discovered. Today, CA membranes still maintain a small fraction of the market because they are easy to make, mechanically tough and resistant to degradation by chlorine up to 1 ppm. However, CA membranes slowly hydrolyze over time and are stable only in the pH range of 4-6. Salt rejection decreases as temperature increases and therefore CA membranes are generally not used above 35ºC. In addition, they are easily attacked by bacteria.
2. Polyamide TFC membranes. A more successful, commercially available NF/RO membrane is the TFC aromatic polyamide membrane. This type of membrane has higher salt rejections and fluxes than CA membranes. On the other hand, aromatic polyamides have several disadvantages including: a. Low resistance to fouling. Membrane fouling (scale, silt, biofouling, organic fouling, etc.) is the main cause of permeate flux decline and loss of water quality. Fouling can decrease total throughput more than 30 percent, and further increase capital costs by 15 percent resulting from membrane cleaning.
b. Limited oxidant tolerance due to the existence of secondaryamides and electron-rich aromatic rings. Selectivity is rapidlyand permanently lost once exposed to feed water containingmore than a few ppb levels of chlorine or hypochloritedisinfectants. It creates major problems in designing foreffective pretreatment of chlorine to destroy microorganisms.
c. Chemical and thermal instabilities. The substrate layer of TFC membranes, usually polysulfone, is attacked by many solvents,which limits deposition of a permselective, ultrathin membranefrom solvent systems, such as Dimethylformamide (DMF),and N-methyl-pyrrolidone (NMP). Most TFC membranescan only be used below 30-50ºC because of limitations of the polysulfone. On the other hand, hot wastewater from the food, chemical, and petroleum processing industries are discharged directly and the energy lost is estimated at higher percent of the total energy consumption. Developing chemically stable substrates, which operate at higher temperatures, will result in substantial energy savings. Therefore, the development of more thermally and chemically resistant selective top layers and substrates is critically required for water purification.
With new breakthroughs in membrane technology, constituent specific membranes are a possibility. Research in the use of embedded "nanoparticles" in reverse osmosis membranes has shown promise and may be applicable to nanofiltration membranes in the future. Meanwhile, the issues of nanomaterials leaching out from the membranes with its relevant contamination should be investigated in detail.

Conflict of interest
The author declares no conflict of interest.