WO2018091273A1 - New processes for treating water - Google Patents

Abstract

Process for treating water, wherein feed water is being treated with a membrane M, said membrane M being a reverse osmosis membrane, a forward osmosis membrane or a nanofiltration membrane, wherein said membrane M is contacted with an anionic surfactant A, wherein said anionic surfactant comprises a C₆ to C₁₀ aliphatic hydrocarbon chain.

Description

New Processes for treating water

The present invention is directed to processes for treating water, wherein feed water is being treated with a membrane M, said membrane M being a reverse osmosis membrane, a forward osmosis membrane or a nanofiltration membrane, wherein membrane M is contacted with an anionic surfactant A, wherein said anionic surfactant preferably comprises a C6 to C10 aliphatic hydrocarbon chain.

Different types of membranes play an increasingly important role in many fields of technology. In particular, methods for treating water rely more and more on membrane technology. In particular, membranes are used for the desalination of water like sea water of brackish water.

One problem of membranes, especially of reverse osmosis (RO) membranes, forward osmosis (FO) membranes and nanofiltration (NF) membranes is that defects in the separation layer of such membranes may impact the performance. Such defects can for example be formed during the manufacture of such membranes or during operation of such membranes, for example due to aging, chemical cleaning or mechanical strain.

There is a need for membranes with improved separation characteristics and with a long life- time.

US 8025159 B2 discloses an agent for increasing the rejection of inorganic electrolytes and organic substances soluble in water in the membrane separation using a selective permeable membrane such as a nano filtration membrane and a reverse osmosis membrane, process for increasing the rejection using the agent. Said agents comprise a first agent for increasing rejection with a permeable membrane which comprises an aqueous solution of a cationic macromol- ecule having a weight average molecular weight of 100,000 or greater, wherein said cationic macromolecule is a polyvinylamidine or derivative thereof. A second agent for increasing rejection with a permeable membrane which comprises an aqueous solution of an anionic macro- molecule having a weight average molecular weight of 100,000 or greater.

WO12121208A1 discloses enhancement of rejection of reverse osmosis membranes by use of certain organic compounds in a sequential manner without significantly reducing permeation flux. The method involves passing a first organic compound having a molecular weight of less than 200, a second organic compound having a molecular weight of 200 to less than 500, and a third organic compound having a molecular weight of at least 500 through the reverse osmosis membrane. The first organic compound is preferably an amino acid or an amino acid derivative.

US2010133172A discloses coatings being applied to and insolubilized upon an active surface of the selective membrane. The surface-active agent comprises an amine-containing surfactant such as polyethoxylated aliphatic polyamine. US2012168370AA describes a method of improving a rejection of a permeable membrane, more specifically, by allowing the low-molecular-weight amino compound to pass through the membrane, a degraded portion of the membrane can be restored without considerably lowering the permeation flux.

JP2008086945A discloses a method for recovering the performance of a permselective membrane by reducing a permeation flux to a proper range with regard to a nanofiltration membrane or reverse osmosis membrane. The nanofiltration membrane or reverse osmosis membrane with anion charge is treated by bringing its surface into contact with a nonion surfactant to re- duce the permeation flux to the range from +20 to -20 percent of the value at the beginning of the use. The performance of the permselective membrane is thus recovered.

A. E. Childress et al, Desalination 1 18 (1998) 167-174 discloses anionic surfactant (SDS) influence performance of RO membrane by forming hemimicelle over RO surface which in turn acts as secondary layer; hence causing decrase in flux but improved rejection.

The objective of the present invention was to provide processes for treating water that show excellent flux and rejection and that result in a long lifetime of the membranes used. This objective was achieved by processes for treating water, wherein feed water is being treated with a membrane M, said membrane M being a reverse osmosis membrane, a forward osmosis membrane or a nanofiltration membrane, wherein said membrane M is contacted with an anionic surfactant A, wherein said anionic surfactant comprises preferably a C6 to Cio aliphatic hydrocarbon chain.

Such processes for treating water can in principle be all processes in which a feed water is subjected to a membrane based treatment. Preferably such processes are desalination processes of sea water or brackish water. In the context of this application a membrane shall be understood to be a thin, semipermeable structure capable of separating two fluids or separating molecular and/or ionic components or particles from a liquid. A membrane acts as a selective barrier, allowing some particles, substances or chemicals to pass through, while retaining others. Membranes M can for example be reverse osmosis (RO) membranes, forward osmosis (FO) membranes or nanofiltration (NF) membranes. These membrane types are generally known in the art.

FO membranes are normally suitable for treatment of seawater, brackish water, sewage or sludge streams. Thereby pure water is removed from those streams through a FO membrane into a so-called draw solution on the backside of the membrane having a high osmotic pressure. Typically, FO type membranes, similar as RO membranes separate liquid mixtures via a solu- tion diffusion mechanism, where only water can pass the membrane whereas monovalent ions and larger components are rejected.

In a preferred embodiment, FO membranes M are thin film composite (TFC) FO membranes. Preparation methods and use of thin film composite membranes are principally known and, for example described by R. J. Petersen in Journal of Membrane Science 83 (1993) 81-150.

In a further preferred embodiment, FO membranes M comprise a support layer, a separation layer and optionally a protective layer. Said protective layer can be considered an additional coating to smoothen and/or hydrophilize the surface.

Said fabric layer can for example have a thickness of 10 to 500 μηη. Said fabric layer can for example be a woven or nonwoven, for example a polyester nonwoven. Said support layer of a TFC FO membrane M normally comprises pores with an average pore diameter of for example 0.5 to 100 nm, preferably 1 to 40 nm, more preferably 5 to 20 nm. Said support layer can for example have a thickness of 5 to 1000 μηη, preferably 10 to 200 μηη. Said support layer may for example comprise at least one polysulfone, polyethersulfone, polyphe- nylenesulfone (PPSU), PVDF, polyimide, polyimideurethane or cellulose acetate. Nano particles such as zeolites, may be comprised in said support membrane. This can for example be achieved by including such nano particles in the dope solution for the preparation of said support layer.

Said separation layer can for example have a thickness of 0.05 to 1 μηη, preferably 0.1 to 0.5 μηη, more preferably 0. 15 to 0.3 μηη. Preferably, said separation layer can for example comprise polyamide or cellulose acetate as the main component.

Optionally, TFC FO membranes M can comprise a protective layer with a thickness of 5 to 500 preferable 10 to 300 nm. Said protective layer can for example comprise polyvinylalcohol (PVA) as the main component. In one embodiment, the protective layer comprises a halamine like chloramine.

In one preferred embodiment, membranes M are TFC FO membranes comprising a support layer comprising polyethersulfone as main component and polymer P, a separation layer com- prising polyamide as main component and optionally a protective layer comprising polyvinylalcohol as the main component.

In a preferred embodiment FO membranes M comprise a separation layer obtained from the condensation of a polyamine and a polyfunctional acyl halide. Said separation layer can for example be obtained in an interfacial polymerization process.

RO membranes are normally suitable for removing molecules and ions, in particular monovalent ions. Typically, RO membranes separate mixtures based on a solution/diffusion mechanism. In a preferred embodiment, membranes M are thin film composite (TFC) RO membranes. Preparation methods and use of thin film composite membranes are principally known and, for example described by R. J. Petersen in Journal of Membrane Science 83 (1993) 81-150. In a further preferred embodiment, RO membranes M comprise a fabric layer, a support layer, a separation layer and optionally a protective layer. Said protective layer can be considered an additional coating to smoothen and/or hydrophilize the surface

Said fabric layer can for example have a thickness of 10 to 500 μηη. Said fabric layer can for example be a woven or nonwoven, for example a polyester nonwoven.

Said support layer of a TFC RO membrane normally comprises pores with an average pore diameter of for example 0.5 to 100 nm, preferably 1 to 40 nm, more preferably 5 to 20 nm. Said support layer can for example have a thickness of 5 to 1000 μηη, preferably 10 to 200 μηη. Said support layer may for example comprise at least one polysulfone, polyethersulfone, polyphe- nylenesulfone (PPSU), PVDF, polyimide, polyimideurethane or cellulose acetate. Nano particles such as zeolites, may be comprised in said support membrane. This can for example be achieved by including such nanoparticles in the dope solution for the preparation of said support layer.

Said separation layer can for example have a thickness of 0.02 to 1 μηη, preferably 0.03 to

0.5 μηη, more preferably 0.05 to 0.3 μηη. Preferably, said separation layer can for example comprise polyamide or cellulose acetate as the main component. Optionally, TFC RO membranes M can comprise a protective layer with a thickness of 5 to 500 preferable 10 to 300 nm. Said protective layer can for example comprise polyvinylalcohol (PVA) as the main component. In one embodiment, the protective layer comprises a halamine like chloramine. In one preferred embodiment, membranes M are TFC RO membranes comprising a nonwoven polyester fabric, a support layer comprising polyethersulfone as main component and polymer P, a separation layer comprising polyamide as main component and optionally a protective layer comprising polyvinylalcohol as the main component. In a preferred embodiment RO membranes M comprise a separation layer obtained from the condensation of a polyamine and a polyfunctional acyl halide. Said separation layer can for example be obtained in an interfacial polymerization process.

Suitable polyamine monomers can have primary or secondary amino groups and can be aro- matic (e. g. a diaminobenzene, a triaminobenzene, m-phenylenediamine, p-phenylenediamine,

1 , 3, 5-triaminobenzene, 1 ,3,4-triaminobenzene, 3, 5-diaminobenzoic acid, 2, 4-diaminotoluene,

2, 4-diaminoanisole, and xylylenediamine) or aliphatic (e. g. ethylenediamine, propylenedia- mine, piperazine, and tris(2-diaminoethyl)amine). Suitable polyfunctional acyl halides include trimesoyl chloride (TMC), trimellitic acid chloride, isophthaloyl chloride, terephthaloyl chloride and similar compounds or blends of suitable acyl halides. As a further example, the second monomer can be a phthaloyl halide. In one embodiment of the invention, a separation layer of polyamide is made from the reaction of an aqueous solution of meta-phenylene diamine (MPD) with a solution of trimesoyl chloride (TMC) in an apolar solvent.

NF membranes are normally especially suitable for removing separate multivalent ions and large monovalent ions. Typically, NF membranes function through a solution/diffusion or/and filtration-based mechanism.

NF membranes are normally used in cross filtration processes. NF membranes M can for example comprise as the main component polyarylene ether, polysul- fone, polyethersulfones (PES), polyphenylensulfone (PPSU) or mixtures thereof. In a preferred embodiment, said main components of NF membranes are positively or negatively charged.

Nanofiltration membranes M often comprise charged polymers comprising sulfonic acid groups, carboxylic acid groups and/or ammonium groups.

In an especially preferred embodiment, membrane M comprises a separation layer of polyamide. In one embodiment membrane M does not comprise any nanoparticles.

In one embodiment, membrane M comprises nanoparticles N. Nanoparticles N are preferably comprised in the separation layer of membrane M.

Nanoparticles N are normally comprised in amounts on 0.01 to 2 wt%, based on the membrane casting solution or based on the respective layer in which nanoparticles N are comprised. For example, in the case on TFC RO or FO membranes, nanoparticles can be comprised in amounts of 0.01 to 2 wt% in the casting solution for the preparation of membrane M.

Suitable nanoparticles N normally have median particle size (d50) as determined by Dynamic light scattering (DLS) method of 1 to 300 nm, preferably 2 to 100 nm

Example of suitable nanoparticles N include particles of silica, titania, zirconia, alumina, zinc oxide or zeolites. In one embodiment, nanoparticles N are nanoparticles with a core-shell morphology, wherein the core and the shell are made of different materials.

The core of core-shell nanoparticles N can for example be made of silica, titania, zirconia, alumina, zinc oxide of combinations thereof. The shell of core-shell nanoparticles N can for example be made of titania, silver, ceria, zirco- nia, alumina, zinc oxide or combinations thereof.

In one embodiment, nanoparticles N are included in membranes M upon the manufacture of membrane M and are not subjected to a posttreatment with an organic solvent.

In one embodiment, membranes M in which nanoparticles N have been included are subjected to a treatment with an organic solvent L after the manufacture of membrane M. Organic solvent L can in principle be any organic solvent capable of at least partly removing nanoparticles from said membrane M. Suitable solvents L are organic solvents and include Cs to C20 alkanes like pentane, hexane, heptane, aromatic solvents like benzene, toluene, xylene, alcohols like etha- nol, n/i-propanol, n/i/sec butanol.

Treatment of membranes M with organic solvent L is carried out at a temperature of 15 to 50°C preferably at room temperature by immersing the membrane M completely in solvent L for 1 to 60 min preferably from 5-15 minutes.

Through treatment with solvent L it is assumed that nanoparticles N which are not completely embedded or chemically linked in or with polyamide layer can dissociate from membrane M, thus forming defects or voids in the membrane.

According to the invention membranes M are contacted with an anionic surfactant A.

The term "contacted" in this context shall mean that membrane M is in some way treated with anionic surfactant A. In one preferred embodiment, membrane M is treated with an aqueous solution S of anionic surfactant A.

In one embodiment, anionic surfactant A is added to the feed water of a water treatment process. This way membrane M can be contacted with anionic surfactant A during operation of membrane M without having to interrupt the water treatment process.

Anionic surfactant A is a surfactant bearing at least one anionic group. Preferably anionic surfactant A comprises a Ce to C10 hydrocarbon chain and no hydrocarbon chain longer than that. Preferably anionic surfactant A is an alkyl sulfate, herein also preferred to alkylsulfate A.

An alkyl sulfate shall be understood to mean the salt of a monoester of an alkanol with sulfuric acid. In one embodiment alkylsulfate A is a Ce to C18 alkylsulfate-Preferably alkylsulfate A is an oc- tylsulfate, more preferably n-octylsulfate, also referred to as caprylsulfate or cprylylsulfate. Preferably alkylsulfate A is a sodium salt or a potassium salt. Especially preferably, alkylsulfate a is sodium n-octylsulfate or potassium n-octylsulfate.

Solution S normally comprises alkylsulfate A in a concentration of 0.1 to 5000 ppm by weight, based on solution S, preferably 100 to 1500 ppm.

Preferably, solution S has a pH of 4 to 12.

Normally membrane M is contacted with solution S for 5 minutes to 5 days per interval, preferably 10 min to 1 day.

In one embodiment membrane M is treated with solution S in intervals that differ between one day to five years. This means that from one treatment (contacting) of membrane M with solution S to the next treatment of membrane M with solution S a time from 1 day to five years passes without such treatment.

Preferably the intervals between two treatments of membrane M with solution S is 2 days to 365 days, more preferably 7 days to 30 days.

The intervals between treatments of membrane M with solution S may vary from time to time, depending on the performance and the condition of membrane M.

The temperature of the treatment of membrane M with solution S is not critical and is normally the operation temperature of the water treatment process conducted with membrane M, usually between 1 °C and 99 °C, preferably 15 °C to 50 °C.

Processes according to the invention are useful for treating water.

Processes according to the invention can for example be used for treating industrial waste water, municipal waste water, sea water, brackish water, fluvial water, surface water, drinking wa- ter, mining water, waste water from oil wells or power plants.

In one preferred embodiment of the invention, Processes according to the invention are used for the desalination of sea water or brackish water. Processes according to the invention can be used for the desalination of water with a particularly high salt content of for example 3 to 8 % by weight. For example, processes according to the invention are suitable for the desalination of water from mining and oil/gas production and frack- ing processes, to obtain a higher yield in these applications. Processes according to the invention, particularly with RO and/or FO membranes, can be used in food processing, for example for concentrating, desalting or dewatering food liquids (such as fruit juices), for the production of whey protein powders and for the concentration of milk, wine processing, providing water for car washing, making maple syrup, during electrochemical production of hydrogen to prevent formation of minerals on electrode surface, for supplying water to reef aquaria.

Processes according to the invention can be used for rehabilitation of mines, homogeneous catalyst recovery, desalting reaction processes.

Processes according to the invention, particularly using NF membranes, can be used for separating divalent ions or heavy and/or radioactive metal ions, for example in mining applications, homogeneous catalyst recovery, desalting reaction processes.

Processes according to the invention are easy and economical to carry out. They do not require suspension of the operation of the membrane, rather defects in the membrane can be repaired during operation. Processes according to the invention require the addition of only a small amount of anionic surfactant A. Through processes according to the invention, defects in membranes M can be repaired of reduced. Processes according to the invention extend the life time of membranes M Processes according to the invention allow for the treatment of water with high flux and high rejection.

Another aspect of the present invention is the use of C6 to Cio alkyl sulfates for healing defects of a membrane M, membrane M being a reverse osmosis membrane, a forward osmosis membrane or a nanofiltration membrane. In one embodiment, membranes can be stored in contact with a solution S.

Another aspect of the present invention are membranes obtained in a process comprising the following steps:

a) Providing a membrane comprising nanoparticles N,

b) Subjecting the membrane obtained in step a) to a treatment with an organic solvent L, c) Treating the membranes obtained in step b) with a solution S comprising anionic surfactant A.

Nanoparticles N and organic solvent L are those named above Step c) herein is carried out as a one-time treatment after steps a) and b).

In one embodiment, anionic surfactant A used in step c) in processes for making membranes according to the invention is an alkyl sulfate.

In one embodiment, anionic surfactant A used in step c) in processes for making membranes according to the invention is a C6 to Cis alkylsulfate.

In one embodiment, anionic surfactant A used in step c) in processes for making membranes according to the invention is sodium dodecylsulfate (SDS) or potassium dodcecylsulfate. In one embodiment, anionic surfactant A used in step c) in processes for making membranes according to the invention is a C6 to C10 alkylsulfate.

In one embodiment, anionic surfactant A used in step c) in processes for making membranes according to the invention is sodium or potassium octylsulfate.

In processes for making membranes, solution S comprises 0.1 to 5000 ppm of anionic surfactants A.

Solution S is applied in step c) for a period of 1 minute to 5 days preferably 10 min to 1 day by directly adding anionic surfactant A into the feed water. Normally this is done at a temperature of between 1 °C and 99 °C preferably at room temperature.

Membranes so prepared have are easy and economical to make, have a long lifetime, have excellent flux and rejections. Another aspect of the present invention are thin film composite membranes (TFC) with an ultra- thin polyamide layer formed through an interfacial polymerization process remain to be of paramount importance due to their commercial success in desalination applications. Incorporation of nanoparticles in the polyamide layer to produce thin film nanocomposite (TFN) membranes is one of the most promising approaches to improve water flux. However, the puzzle of permeabil- ity/selectivity trade-off remains unresolved. Additional challenges of defect formation with the introduction of nanoparticles need careful adjustments of the membrane properties. Here, we explored the influential effects of nanoparticles and feed additive (surfactant) on the desalination performance of polyamide membranes in an RO testing unit at 55 bar, 35000 ppm salt concentration and 25 °C. Compared to their TFC counterpart, the TFN membranes with varied loading of polyhedral oligomeric silsesquioxane (POSS) and Ti02-Si02 core-shell nanoparticles, respectively, exhibited enhanced water flux at the expense of salt rejection. Unexpectedly, a remarkable effect of the surfactant that seems to be able to heal defects in the membranes was observed. The tailored TFN membranes with combining the effects of nanoparticles, post treatment by a suitable solvent and followed by surfactant treatment was found to outperform most of the commercial TFN membranes under our test conditions. This work may provide useful insights about development of innovative approaches to improve membrane properties to overcome the usual trade-off between permeability and selectivity of the RO membrane by a targeted approach with a suitable surface active material. According to the United Nations World Water Development Report, today around 780 million people lack access to drinking water while the demand of process water is expected to increase by 400 percent between 2000 to 2050 globally [1]. Desalination provides a promising solution to secure water supply by producing fresh and clean water from seawater. Desalination processes are typically divided into thermal and membrane processes [2]. The thermal process usually comprises energy intensive multi-stage flash distillation (MSF) to treat water with high salinity [3]. Membrane processes based on reverse osmosis (RO) are usually less energy intensive and are expected to maintain their leadership in desalination for the near future [4, 5]. Over the years, the RO process has shown continuous improvements in terms of membrane materials, process modifications, module design and pre-treatment processes in order to improve its economics [6, 7]. Most of today's RO membranes, which are thin film composite (TFC) membranes, are based on Cadott's work in which the membrane performance is achieved by a thin selective polyamide layer [8-10]. Many novel approaches have been reported to improve the performance of polyamide-TFC membranes and to enhance the productivity and efficiency of the desalination process [2-7]. One of them is based on nanotechnology that offers completely new possibilities for the new generations of RO membranes.

The concept of thin film nanocomposite (TFN) membranes was firstly introduced by Hoek and his co-workers [1 1-13]. By incorporating NaA zeolite nanoparticles as inorganic fillers into the polyamide layer, they produced high performance TFN membranes with improved flux and comparable rejection to commercial RO membranes. The NaA zeolite nanoparticles not only improved the hydrophilicity, surface charge, porosity and antimicrobial properties, but also improved water permeability because of preferential water flow through zeolite pores. Meanwhile, a high solute rejection was maintained due to the combination of steric and Donnan exclusion [12, 13]. Similar hypotheses were assumed by Huang et.al. [14]. The authors observed that dispersing NaA zeolite particles in the organic phase resulted in TFN membranes with a higher degree of crosslinking, more homogeneous zeolite-polymer structure and better salt rejection, while the addition of zeolite particles in the aqueous phase produced TFN membrane with a lower flux and salt rejection because of micro-void formation.

Apart from zeolites, various nanoparticles including single walled carbon nanotubes, silica, tita- nia, silver, and metal organic frameworks have been used in TFN membranes [13, 15, 16]. Amongst them, T1O2 nanoparticles received much attention because of their unique physico- chemical properties to impart TFN membranes with superior hydrophilicity, antifouling properties and overall membrane performance [17-20]. The other one was the introduction of polyhedral oligomeric silsesquioxane (POSS) into the selective polyamide layer. Dalwani et. al. [21] observed that POSS particles did not react readily with the acyl chloride solution but required special conditions in order to form a highly crosslinked hybrid structure. Based on experimental and simulation work, Moon et. al. [22] concluded that the incorporation of POSS particles into the polyamide layer affected its hydrophilicity, porosity and charge characteristics, and thus improved membrane's performance than pure polyamide membranes. However, the presence of nanoparticles in the polyamide layer has been reported to reduce the degree of cross linking of the polyamide layer thereby, affecting salt rejection and mechanical properties [17, 23]. Lind et al. proposed two additional mechanisms for the flux enhancement in TFN membranes; namely, defect formation and changes in the cross-linked structure of polyamide films [24-26]. These findings suggested that defects or void formation in the polyamide layer might play a major role in determining the water permeability of TFN membranes. From the aforementioned literature, it appears that many factors need to be considered in order to overcome the trade-off relationship between permeability and selectivity. These include the degree of cross linking of the polyamide layer, surface charge, hydrophilicity, and the presence of nanovoids, defects or water channels. Apart from these, with the use of nanoparticles, addi- tional challenges are required to be addressed. The primary challenges include the poor dis- persibility of nanoparticles, particle agglomeration and potential release of nanoparticles from a membrane matrix [27]. The agglomeration of nanoparticles in the selective layer could lead to defect formation and reduce salt rejection [28]. To avoid agglomeration, mechanical treatments such as sonication and grinding, and chemical modifications of nanoparticles have been suggested [29]. All these challenges and influential factors must be controlled and solved collectively in order to optimize both productivity and selectivity of TFN membranes.

In this work, we aimed to design TFN membranes with desired performance by the use of na- noparticles, solvent post-treatment and surfactant addition. Two inorganic nanomaterials, T1O2- S1O2 core-shell nanoparticles, (2) commercially available water-soluble octammonium POSS (AM0285), would be used to fabricate TFN membranes. These nanoparticles could disperse easily in casting solutions and form highly cross-linked membrane structures. The selfsynthe- sized TiC"2-Si02 nanoparticles were employed because of their core-shell morphology with multi- functional characteristics of aforementioned individual metal oxides. Once TFN membranes with different particle chemistry and sizes were fabricated, they were evaluated by RO tests. Some selected TFN membranes were further treated by solvents and sodium dodecyl sulphate (SDS) in order to assess their influences on RO performance. This work may provide useful insights to design next-generation TFN membranes for seawater desalination.

Experimental Materials Titanium (IV) isopropoxide (97%, Ti[OCH(CH3)2]4, Sigma Aldrich), sodium dodecyl sulphate (>99%, Sigma Aldrich), sodium chloride (>99%, NaCI. Sigma Aldrich) triethylamine (>99%, (C₂H₅)3N , Sigma Aldrich), acetyl acetone (99%, CH3COCH2COCH3, Fluka) were used without purification. Nexsil silica suspensions and POSS particles (namely, AM0285) were procured from Nyacol Inc and Hybrid Plastics, respectively. Ethanol (analytical grade) was supplied by Fischer Scientific. Deionized (Dl) water was generated by a Millipore water purification system. m-Phenylenediamine (>98%, MPD, Tokyo Chemical Industry Co. Ltd, Japan), trimesoyi chloride ( >98%, TMC, Tokyo Chemical Industry Co. Ltd, Japan), and n-Hexane (Fisher Scientific) were utilized to synthesize the selective polyamide layer of TFC and TFN membranes. All reagents were used as received without further purification. Commercially available polysulfone ultrafiltra- tion flat sheet membranes (under the trade name of UP20) with an average pore radius of 5 nm was procured from Microdyn Nadir. Prior to use, the UP20 membrane was soaked in Dl water overnight.

Synthesis of titania coated silica nanoparticles

Figure 1 shows the schematic procedures of an in-house proprietary method to synthesize Ti0₂-

Si0₂ nanoparticles based on the procedures reported elsewhere [30]. Briefly, Titanium /sotri- propoxide of 0.005 mole was first dissolved in ethanol of 0.1 mole. Then acetyl acetone of 0.0025 mole was added into the mixture to form a deep red solution. In a typical synthesis, the prepared titania precursor was added into an aqueous solution containing colloidal silica (Nexsil 85) of 0.05 mole in Dl water of 2 moles, followed by stirring and forming a uniform solution. Subsequently, triethylamine of 0.02 mole was added dropwise under stirring overnight at room temperature. The resultant suspension was precipitated by the addition of 10 moles of acetone, centrifuged and washed three times with fresh acetone. The sediments were dried in a vacuum oven for 4 h at 60 °C. Afterwards, the solids were crushed by using a mortar and pestle to fine powders and then calcined at 600 °C for 4 h. The calcined core-shell type nanoparticles were consecutively nano-milled in Dl water at a concentration of -25 wt% in Dl water using zirconia beads of 2 mm in size for 48 h, sonicated prior to use in membrane casting solutions.

Preparation of thin film composite (TFC) and thin film nanocomposite (TFN) membranes The formation of a crosslinked polyamide layer on the UP20 membrane was conducted via in- terfacial polymerization, as described in the literature [8-10, 31 , 32]. Generally, the membrane substrate was firstly soaked overnight in Dl water at room temperature. After removing the ex- cess water, the substrate was immersed in an aqueous solution containing 2 wt% m- pheny- lenediamine (MPD) for 10 min. Afterwards, the membrane surface was mopped carefully by a paper towel and a rubber roller was employed to remove the excess MPD. A n-hexane solution containing 0.05-0.2 wt% trimesoyl chloride (TMC) was poured onto the membrane surface and allowed to react for a predefined time, followed by washing with n-hexane. The resulting TFC membrane was dried at room temperature for 15 min and then soaked in Dl water before performance testing.

The procedure to fabricate TFN membranes was similar except a predefined amount of nanoparticles was prepared in MPD solution. In the case of using titania-silica core-shell nanoparti- cles, they were uniformly dispersed by sonication before adding into the MPD solution. In the case of POSS, AM0285 particles were dissolved completely into MPD solution. The TFN membranes incorporated with AM0285 and titania-silica core-shell nanoparticle were referred to as TFN-P, and TFN-T. Sodium dodecyl sulphate (SDS) was used to enhance the RO performance in two different approaches. Method A: the TFC and TFN membranes were soaked in a 0.1wt% aqueous solution of SDS for 15 min. Method B: SDS was added directly into the feed tank at a concentration of 0.1 wt% after the first reading of salt rejection without interrupting the test run. The RO results were compared before and after the SDS addition.

Membrane performance evaluation

RO tests of TFN and TFC RO membranes were carried out using a model seawater (i.e., 35000 ppm in Dl water) as the feed at 55bar (-800 psi) and 25 °C in a continuous cross flow process, as illustrated in Figure 2. Water flux (J) was calculated by measuring the permeate flow at a specific interval of time

where M\s the permeate mass collected at the time interval Δ t, A is the effective membrane area. Salt rejection (R) was calculated by measuring the salt conductivities of the feed and permeate, respectively, and using the following formula

= (i - f¾ (%) = (ΐ -ψ *ιοο ...(2) where C_p and Cf are the conductivities of the permeate and feed, respectively. Permeate fluxes and rejection values were collected at an interval of 1 h (minimum 3 readings) after system equilibrium for 1 hr. In case of membrane evaluation with Method B, the following procedures were carried out for the conductivity measurements, as schematically shown in figure 3.

1 ) First salt reading: The procedure was the same as the described above.

2) Salt and surfactant: SDS of 0.1 wt% was added directly into the feed tank immediately after the first salt reading, followed by readings at an interval of 1 h each.

3) Second salt reading: After the above two sets of readings, the RO testing machine was de- pressurized, thoroughly washed with Dl water (5 washing cycles if not stated otherwise). The second salt reading was started with a freshly prepared NaCI solution under conditions identical to the first salt reading.

Figure 4 (a) to 4 (c) shows the morphology and size of the synthesized Ti0₂-Si0₂ nanoparticles and their elemental compositions determined by transmission electron microscopy (TEM, 200KV JEOL 201 OF microscope) and its EDX. The samples were prepared by placing a drop of well dispersed nanoparticle dispersion onto a copper grid with a carbon film, and then dried in a desiccator. The nanoparticle size was also confirmed by dynamic light scattering (DLS)

(Zetasizer Nano ZS, Malvern instruments, UK), as illustrated in Figure 4 (d). For the measurement, a stable dispersion of nanoparticles in Dl water was prepared.

Membrane characterizations

Prior to SEM observation by a field emission scanning electron microscope (FESEM JEOL JSM-6700F), the membrane samples were freeze dried and fractured in liquid nitrogen before platinum coating using a JEOL JFC- 1300 Platinum coater. X-ray Photoelectron Spectroscopy (XPS, Kratos AXIS UltraDLD, Kratos Analytical Ltd., England) was employed to investigate the nanoparticles in the polyamide layer using the procedures described elsewhere [33, 34]. Wide scans in the binding energy range of 0-1 100 eV and narrow scans of core-level Ti2p and Si2p were performed on the selective surface of the membranes. The TFC and TFN membranes were measured by Doppler Broadening Energy Spectroscopy (DBES) using a positron annihilation system coupled with a slow positron beam at the National University of Singapore. Details of the system and measurements have been described in the literature [35, 36]. A radioisotope ²²Na with an energy of 50 mCi was used as the positron source. The DBES spectra were recorded using an HP Ge detector at a counting rate of approximately 2000 cps and the total number of counts for each spectrum was 1 .0 million.

A SurPASS electrokinetic analyzer (Anton Paar GmbH, Austria) was utilized to measure the surface charge of the TFC and TFN membranes. The experiments were based on streaming potential measurements [37, 38]. The samples were die cast to fit into the measuring cell followed by circulation of a 0.01 M NaCI solution through the measuring cell to attain the zeta- potential of the membrane surfaces. Afterwards, 0.1 M HCI and 0.1 M NaOH solutions were used to auto-adjust pH values at the pre-specified range of 2-10 at 25 °C. Results and discussion

Analyses of nanoparticles

As shown in Figure 4 (a) and (b), the TEM images of the calcined nanoparticles exhibit a core- shell ("raspberry") morphology with silica as the core and titania as the shell material. The titania nanoparticles have a size of ~5 nm, they are uniformly deposited on a spherical silica core. The EDX analysis in Figure 4 (c) indicates the ratio of Si to Ti to be nearly the same as the ratio of the starting materials used in the synthesis (i.e., 0.005 mole of titanium isopropoxide and 0.05 mole of silica). The DLS data in Figure 4 (d) are consistent with the TEM observation, the T1O2- S1O2 core-shell particles have an average particle size of about 90 nm. The POSS particles could not be observed in SEM and TEM because their sizes are too small (around 1-3 nm) as reported previously [16, 39-41].

Characterization of the membranes

Figure 5 shows the surface and cross-sectional FESEM images of the membrane substrate (i.e., UP20), TFC and TFN membranes. The substrate layer has an asymmetrical structure that consists of a tighter skin, a sponge-like porous sublayer, and finger-like macrovoids underneath. In contrast to the UP20 membrane, the top surfaces of TFC and TFN membranes have a ridge- and-valley morphology with a thickness of around 150 nm resulting from interfacial polymerization of the polyamide layer [12, 31 , 32]. The polyamide layer of the TFN-T membrane has spherical globules in the ridge-and-valley morphology, which can be observed easily in the cross-section image (Figure 5 (d) bottom). This is due to the fact that the Ti02-Si02 nanoparticles are wrapped by the polyamide during the interfacial polymerization.

Figure 6 summarizes the XPS spectra of the TFC and TFN membranes. The inset (b) clearly shows the Si 2p peaks in the spectra of the two TFN membranes, as an indication for the suc- cessful incorporation of nanoparticles. In contrast, there is no Ti signal in all three membranes. The absence of the Ti signal in the TFN-T membrane may be attributed to its low concentration in the membrane. Nevertheless, a very low amount of Ti is detected by EDX, as shown in Figure 4 (c). The microstructure of the TFC, TFN-P and TFN-T membranes was investigated by PAS.

Figure 7 presents their S and R parameters as a function of the incident positron energy. The initial sharp increase in S parameter is due to the back diffusion and scattering of positroniums near the membrane surface [35, 36]. A smaller S parameter in the selective layer indicates ei- ther a smaller free volume size or a lower free volume content of the membrane. As a result, the smaller S parameter of the TFN-T membrane suggests its denser selective layer compared with the other two membranes. On the other hand, the R parameter indicates that TFN-T has a thicker polyamide layer than the other two membranes, which is consistent with the observation from FESEM.

Effects of particle concentration and surfactant on the TFN membranes

Figure 8 compares the flux and rejection between TFC and TFN-P membranes as a function of POSS content and testing procedures at 15.5 bar and 55 bar for feed concentrations of

1500 ppm and 35000 ppm, respectively, followed by the surfactant addition in the feed tank. The TFC membrane shows a rejection of 96% and a permeate flux of 20 kg/h.m² at 55 bar. The TFN-P membranes are also stable at both pressures. Figure 8 (b) indicates that the flux shows a maximum at 0.25 wt% POSS, while the rejection displays an opposite trend. Since the polyamide film is growing on the organic side of the interface between the two monomer solvents [9, 10, 31 , 32, 42], the water soluble nanoparticles may be either chemically or physically trapped in the polyamide layer. Hence, a reduction in selectivity could be explained by the presence of interfacial voids or defects between the polyamide layer and POSS particles because functionalized POSS particles cannot form a highly crosslinked hybrid structure unless under alkaline conditions [21 ]. Overall, the water transport seems to be influenced by defect formation at the loss of some selectivity. A similar concept can be found in the literature as described by Pendergast et.al. and Lind et.al. [15, 26].

Interestingly, solid improvements in rejection are observed for both TFC and TFN-P membranes once SDS is added into the feed tank, as shown in Figure 8 (a). This may be because SDS could cover defects of the polyamide layer and hence exhibits the "healing effect" by forming a surfactant layer. The surfactant healing effect is more pronounced for TFN-P membranes containing higher POSS concentrations. Once the surfactant is drained and a fresh salt solution of the same concentration (i.e., 35000 ppm) is introduced, the rejection appears to be almost retained after thorough washing prior to this second salt reading, as shown in Figure 8 (a). The improved rejection is found to accompany with a slight reduction in permeate flux as shown in figure 8 (b). The additional layer of the surfactant over the membrane surface could have affected the water transport through the membrane.

Figure 9 shows the performance of TFN-T membranes with different concentrations of T1O2- S1O2 particles; followed by the surfactant addition in the feed tank. The TFN-T membrane con- taining 0.125 wt% Ti02-Si02 has a rejection of 93% and a permeate flux of 23 kg/h.m² at 55 bar. However, the rejection decreases while the permeate flux increases as a function of nanoparti- cle content. Larger error bars were seen for TFN-T membranes with a higher nanoparticle concentration possibly because of greater non-uniform particle distribution and agglomeration in the polyamide layer as well as interference of inorganic nanoparticles during the interfacial polymerization [17, 21 , 23]. Interestingly, as shown in figure 9 (a) a similar "healing effect" by SDS is also observed. However, as shown in figure 9 (b) flux decreases with SDS in the feed tank. After draining the surfactant from the system, the TFN-T membranes show an improvement in rejection compared with those without the SDS treatment.

Comparison of surfactant treatment methods

In order to figure out the mechanism of the surfactant-assisted healing effect, two experiments were conducted as introduced in Section 2.4. As illustrated in Figure 10 (a), the membranes soaked in SDS according to Method A show better rejections than the pristine TFC and TFN membranes. However, the improvement is more significant when the membranes were treated by Method B. The latter also retain rejections higher than the former even after changing to fresh feed solutions. Consistent with literatures [43, 44], Zeta potential data (Figure 1 1 ) reveal that all pristine TFC and TFN membranes are negatively charged at the given pH range. There- fore, the post-treatment by the in-tank addition probably forms a uniform surfactant layer with the most thermodynamically stable structure on the negatively charged membrane surfaces.

As reported by Mai et. al. [45] the hydrophobic segments of SDS can be easily adsorbed on the membrane surface via hydrophobic-hydrophobic interaction, which limits the contact between the surfactant head and the membrane surface. Therefore, the anionic SDS surfactant may be aligned in a hydrophilic "head out" position and overlaps the defects by forming a secondary filter layer. Furthermore, the electrostatic repulsion between the SDS and negatively charged membrane surface also facilitates the alignment of the surfactant in "head out" position. The self-assembly characteristics of surfactants on membrane surface have been observed by Chil- dress et.al. [46]. As a result, the aforementioned surfactant assisted healing effect will improve rejection but reduce flux.

Combined effect of nanoparticles, solvent post treatment and surfactant-assisted healing effect on the overall membrane performance

Solvent post treatment has been demonstrated to play a very significant role in adjusting performance of TFN membranes [10, 31 , 32, 47]. The current study shows the addition of nanoparticles are prone to form defects or imperfections in the membrane and with the aid of an anionic surfactant additions these imperfections can be healed to recover the rejections. As a con- sequence, the permeability/selectivity trade-off can be effectively broken. Figure 12 demonstrates how those steps can be combined together to design TFC and TFN membranes with improved flux without sacrificing the rejection. Schematics for our hypotheses on the mechanism are illustrated in Figure 13. As seen in Figure 12, both TFC and TFN-T membranes show improvement in the flux upon ethanol treatment. This is probably due to an increase in the free volume in the case of TFC membranes and the removal of oligomeric parts from the polyamide layer. Increasing the free volume in the TFN-T membrane could also be attributed to formation of voids contributed from loosely bound nanoparticles from the membrane surface. Irrespective of the size of defects, the TFN membranes showed significant recovery in the selectivity in the presence of SDS. As afore-explained, the anionic SDS surfactant may be aligned in a hydro- philic "head out" position and exhibits "healing effect" by overlapping the defects. As a result, overall improvements of both rejection (through the surfactant self-assembly) and permeate flux takes place. Transport mechanism through RO membranes are commonly explained by a solu- tion diffusion model (dense polyamide layer without pores) and a pore flow model (dense polyamide layer with existence of nanometer size pores) [48, 49]. The current findings are suggesting that pore flow may be what is responsible for increased flux. It does not exclude the solution diffusion model as a transport mechanism that is happening in parallel, but the observed effects are easier explained by the pore flow model. The preferential flow path could be established through the selective layer of the membrane and accounted for increase in the permeate flow. This probably could be due to presence of extensive defects/imperfections inside the polyamide layer. The Maxwell mixing model used to describe performance of the TFN-T membranes in which it was proposed that impermeable nanoparticles/fillers could increase membrane permeability though the defect formation but at the cost of salt rejection [15]. The proposed mecha- nism differs from the Maxwell's mixing model in terms of improving flux through the defect formation without sacrificing the rejection.

Performance benchmarking

Commercial RO membranes namely NanoH₂0 ES (without PVA), Dow Filmtec (SW30XLE), Toray UTC-80E were tested at the identical conditions of 35000 ppm NaCI salt concentrations, 55bar pressure at 25°C and compared with the TFC and TFN membranes developed in this work. Figure 14 (a) shows their separation performance. It is important to note that the rejection of commercial membranes found in lab operations is lower than respective specification values (>99%). These variabilities could be from difference in processing steps and testing conditions [50]. Amongst all commercial membranes, the Nanoh O membrane exhibits the highest performance. The TFC membrane shows comparable rejection but the lowest flux. The TFN-T membrane by incorporating 0.125 wt% of Ti02-Si02 nanoparticles and with combinational effect of solvent post treatment and surfactant healing shows competitive performance compared to all commercial membranes. This example in accordance with the hypothesis explained in Figure 13 illustrates an effective way of tailoring RO membrane properties by overcoming trade-off effect between permeability and selectivity.

Conclusions and prospects

Nanoparticles irrespective of size and shape are prone to form defects in polyamide layer. The defect formation appeared explicitly with larger size and higher concentrations of nanoparticles; as in most of these cases lower salt rejections were observed.

Addition of an anionic surfactant (e.g. SDS) in the feed tank offered significant increase in solute rejection values for the TFN membranes while maintaining the high flux from the afore men- tioned defect formation. The imperfections appeared to be healed by surfactant adsorption over the membrane surface. Applying the surfactant during the RO process helps to selectively heal the defects without sacrificing flux, even after removing the SDS from the system and continuing without further SDS addition. The methodology demonstrated from current research work, TFN membrane exhibited comparable rejection to the TFC counterpart, and commercial membranes under identical conditions. The flux could be increased by 50% compared to the TFC membrane. The transport mechanism was influenced by existence of preferential flow channel through imperfections of polyam- ide layer and the presence of a surfactant acting as secondary filter layer. This concept could open the door to a new paradigm of hybrid RO membrane fabrication in which desired performance properties can be achieved by a clever design of a surfactant-induced surface active polymer in combination with an imperfect polyamide layer. By increasing the selectivity of the deposition sites of the surface active compounds, highly productive and selective membranes could be produced. Defective membranes could be repaired in situ. The concept may also help to extend the useful lifetime of RO membranes by continuous low dosage treatment.

Tables and figures

Figure 1 : Schematic procedures for the synthesis of Ti02-Si02 nanoparticles

Figure 2: Flow diagram of reverse osmosis testing unit

Figure 3: Procedures for the "surfactant in-tank" experiment

Figure 4: (a, b) TEM images, (c) EDX and (d) particle size distribution of Ti02-Si02 nanoparticles

Figure 5: FESEM images of top surfaces and cross-sections of (a) PES support, (b) TFC, (c)

TFN-P and (d) TFN-T membranes.

Figure 6: (a) XPS wide-scanned spectra, (b) Si 2p core-level spectra and (c) Ti 2p core-level spectra of TFC, TFN-P and TFN-T membranes

Figure 7: S and R parameters of TFC, TFN-P and TFN-T membranes as a function of incident positron energy (or mean depth).

Figure 8: Performance comparison between TFC and TFN-P membranes as a function of

POSS content and testing procedures: (1 ) First salt reading (no SDS), (2) Salt +

SDS in the feed tank, (3) Second salt reading (after the SDS removal)

Figure 9: Performance comparison between TFC and TFN-T membranes as a function of particle loading and testing procedures: (1 ) First salt reading (no SDS), (2) Salt + SDS in the feed tank, (3) Second salt reading (after SDS removal)

Figure 10: Performance comparison for two surfactant treatment approaches (Method A: SDS post treatment vs Method B: SDS in feed tank addition) for TFC, TFN-P, TFN-T membranes as per testing procedures of 1 ) First salt reading (no SDS) 2) Salt + SDS in feed tank 3) Second salt reading (after removal of SDS). For each TFN membrane nanoparticle concentration was 0.125 wt%

Figure 1 1 : Surface zeta potential of TFC, TFN-P and TFN-T membranes

Figure 12: Illustrative example of tailoring the membrane properties by combining influential effects of nanoparticles, ethanol post-treatment solvent and SDS in feed tank addition and comparing with standard TFC membrane

Figure 13: Schematics for healing membrane hypothesis showing surfactant assisted healing effect to overcome defect/imperfections in polyamide layer of (a) TFC and (b) TFN membranes Figure 14: Performance benchmarking of commercial and hand casted TFC, TFN-T membranes tested at 55 bar, -35000 ppm. Commercial Membranes: (1 ) NanoH20 ES (without PVA), (2) Dow Filmtec (SW30XLE), (3) Toray UTC-80E, TFN-T: TFN membrane with 0.125wt% Ti02-Si02 nanoparticles, with ethanol post treatment and SDS in feed tank.

In a second set of experiments, the use of anionic surfactants A has been tested during operation of a membrane M. Preparation of thin film composite (TFC) and thin film nanocomposite (TFN) membranes

A typical approach for the formation of a crosslinked polyamide layer via interfacial polymerization on top of a commercial ultrafiltration membrane as substrate was employed, as commonly described in the literature. Generally, the membrane substrate which had been soaked overnight in Dl water at room temperature; after removal of excess water was firstly immersed in an aqueous 2 wt% m- phenylenediamine (MPD) solution for 10 min. Afterwards, the membrane surface was treated carefully with a paper towel and a rubber roller to remove excess MPD. A 0.05-0.2 wt% trimesoyl chloride (TMC) in n-hexane solution was poured onto the membrane surface and allowed to react for a predefined exposure time, followed by washing with n-hexane. The resulting membrane was dried at room temperature for 15 min. Before performance testing, as standard procedure the membrane was soaked with Dl water for 15 min. The procedure to fabricate TFN membranes was similar, with an exception that predefined amount of BASF nanoparticles were uniformly dispersed by sonication before adding into the MPD solution and in the case of POSS materials namely, AM0285 and AM0265 particles were dissolved completely into MPD and TMC solution respectively prior to the casting process. Solvent post treatment methods were explored by soaking membrane into (1 ) ethanol only or (2) n-hexane only or (3) n-hexane followed by ethanol. The soaking time was either 15 min or 1 h. The TFN membranes incorporated with the AM0285, AM0265 and BASF nanoparticle were referred to as TFN-P1 , TFN-P2 and TFN-T.

Surfactant treatment

Anionic Surfactant A was added directly into feed tank at a concentration of 0.1 wt% after the first reading of salt rejection without interrupting the test run and comparing the results before / after anionic surfactant addition. Membrane performance evaluation

RO tests of TFN and TFC RO membranes were carried out at 55bar (-800 psi), 35000 ppm salt concentration (NaCI in Dl water) and 25 °C in a continuous cross flow process using proprietary testing systems. The schematics of the RO testing unit is shown in Figure 2. Prior to tests, all membrane samples were run and allowed to equilibrate for 1 h until a stable reading could be obtained. Water flux (J) was calculated by measuring the permeate flow at a specific interval of time

Act where M\s the permeate mass collected at the time interval Δ t, A is the effective membrane area. The salt rejection (R) was calculated by measuring the salt conductivities of feed and permeate, respectively, applying the following formula

* (%) = {1 - f^Cjxi00 ...(2) where C_p and CVare the conductivity of permeate and feed, respectively. Permeate flux and rejection values were noted at intervals of 1 h (minimum 3 readings) after allowing an equilibration time of 1 hr. In case of membrane evaluation with Method B, the following conductivity measurement procedures were followed as schematically shown in figure 3.

First salt reading: The procedure was the same as described above.

Salt and surfactant: Surfactant (0.1 wt%) was added directly into feed tank immediately after first salt readings, followed by readings at intervals of 1 h each.

Second salt reading: After the above two sets of readings, the RO testing machine was depres- surized, thoroughly washed with Dl water (5 washing cycles if not stated otherwise). The second salt reading was started with freshly prepared NaCI solution under conditions identical to the first salt reading.

Membrane evaluation

Surfactant healing effect: addressing current problem of the RO membrane

As seen from the below table, sodium octylsulfate performs better compared to SDS under identical conditions. TFN membranes (with nanoparticles induced defects) are compared with standard TFC membranes.

Both flux and membrane performance improved in case of the standard TFC membrane and TFN membranes. Both type membrane show no loss in flux. It appears that defects of the TFN membrane are healed by sodium octylsulfate addition. Surfactant healing effect: Potential use of anionic surfactant to improve permeability/selectivity

As seen from figures 15, 16 and 17 the nanoparticle addition with/without combination of solvent post treatment methods can induce defects or imperfections into the PA layer of the membrane. These carefully induced imperfections can offer preferential flow for water; as consequence multifold increase in the flux but at loss of selectivity. With aid of surfactant additions these imperfections can be healed to recover rejections. The anionic SDS surfactant may be aligned in a hydrophilic "head out" position and exhibits "healing effect" by overlapping the defects. However, with use of SDS, compared to the data without SDS addition compromise in the flux can be observed though imperfections of the membrane are repaired to the certain extent. Use of suit- able surfactant (e.g. o) gives opportunity (as seen from one of the example from figure 4a and b) to overcome compromise in flux hence effectively can break permeability/selectivity trade-off the RO membrane with overall improvement flux as well as rejection.

[I] UNESCO, The United Nations World Water Development Report 2015, Water for a Sus- tainable World (2015).

[2] L. F. Greenlee, D. F. Lawler, B. D. Freeman, B Marrot, P. Moulin, Reverse osmosis desalination: Water sources, technology, and today's challenges, Water Res. 43 (2009) 2317- 2348.

[3] C. Fritzmann, J. Lowenberg, T. Wintgens, T. Melin, State-of-the-art of reverse osmosis desalination, Desalination 216 (2007) 1 -76.

[4] K. P. Lee, T. C. Arnot, D. Mattia, A review of reverse osmosis membrane materials for desalination-Development to date and future potential, J. Membr. Sci. 370 (201 1 ) 1 -22.

[5] D. Li, H. Wang, Recent developments in reverse osmosis desalination membranes, J.

Mater. Chem. 20 (2010) 4551-4566.

[6] M. Wilt, C. Bartels, Optimization of seawater RO systems design, Desalination 173 (2005) 1-12.

[7] A.F. Ismail, M. Padaki, N. Hilal, T. Matsuura, W.J. Lau, Thin film composite membrane - Recent development and future potential, Desalination 356 (2015) 140-148.

[8] J. E. Cadotte. Reverse osmosis membrane. US 4,039,440 (1977).

[9] J. E. Cadotte, R. J. Petersen, R. E. Larson, E. E. Erickson. A new thin-film composite

seawater reverse osmosis membrane. Desalination 32 (1980) 25-31.

[10] S. Karan, Z. Jiang, A.G. Livingston, Sub-10 nm polyamide nanofilms with ultrafast solvent transport for molecular separation, Science, 348 (2015) 1347-1351 .

[I I ] E.M.V. Hoek, B.H. Jeong, Y. Yan, Nanocomposite membranes and methods of making and using same, U.S. Application 1 1/364,885, Priority: US 60/660,428, March 9 2005.

[12] B. H. Jeong, E.M.V. Hoek, Y. Yan, A. Subramani, X. Huang, G.Hurwitz, A.K. Ghosh, A.

Jawor, Interfacial polymerization of thin film nanocomposites: A new concept for reverse osmosis membranes, J. Membr. Sci. 294 (2007) 1-7. [13] M. Zargar, Y.Hartanto, B.Jin, S.Dai, Understanding functionalized silica nanoparticles incorporation in thin film composite membranes: Interactions and desalination performance, J. Membr. Sci. 521 (2017) 53-64.

[14] H. Huang, X. Qu, H. Dong, L. Zhang, H. Chen, Role of NaA zeolites in the interfacial polymerization process towards a polyamide nanocomposite reverse osmosis membrane,

RSC Adv. 3 (2013) 8203-8207.

[15] M. M. Pendergast, E.M.V. Hoek, A review of water treatment membrane nanotechnolo- gies, Energy Environ. Sci., 4 (201 1 ) 1946-1971 .

[16] J. Duan, E. Litwiller, I. Pinnau, Preparation and water desalination properties of POSS- polyamide nanocomposite reverse osmosis membranes, J. Membr. Sci. 473 (2015) 157-

164.

[17] H. S. Lee, S. J. Im, J. H. Kim, H. J. Kim, J. P. Kim, B. R. Min, Polyamide thin-film nanofil- tration membranes containing ΤΊ02 nanoparticles, Desalination 219 (2008) 48-56.

[18] S. H. Kim, S.Y. Kwak, B. H. Sohn, T. H. Park, Design of ΤΊ02 nanoparticle self-assembled aromatic polyamide thin-film-composite (TFC) membrane as an approach to solve biofoul- ing problem, J. Membr. Sci. 21 1 (2003) 157-165.

[19] A.M.A. El-Aassar, Improvement of reverse osmosis performance of polyamide thin-film composite membranes using ΤΊ02 nanoparticles, Desalin. Water Treat. 55 (2015) 2939-

2950.

[20] E. Bet-moushoul, Y. Mansourpanah, Kh. Farhadi, M. Tabatabaei, ΤΊ02 nanocomposite based polymeric membranes: A review on performance improvement for various applications in chemical engineering processes, Chem. Eng. J. 283 (2016) 29-46.

[21] M. Dalwani, J. Zheng, M. Hempenius, M. J. T. Raaijmakers, C. M. Doherty, A.J. Hill, M.

Wessling, N. E. Benes, Ultra-thin hybrid polyhedral silsesquioxane-polyamide films with potentially unlimited 2D dimensions, J. Mater. Chem. 22 (2015) 14835-14838.

[22] J. H. Moon, A. R. Katha, S. Pandian, S. M. Kolake, S. Han, Polyamide-POSS hybrid

membranes for sea water desalination: effect of POSS inclusion on membrane properties, J. Membr. Sci. 461 (2014) 89-95.

[23] J. Yin, B. Deng, Polymer-matrix nanocomposite membranes for water treatment, J.

Membr. Sci. 479 (2015) 256-275.

[24] M. L. Lind, A. K. Ghosh, A. Jawor, X. Huang, W. Hou, Y. Yang, E. M. V. Hoek, Influence of zeolite crystal size on zeolite-polyamide thin film nanocomposite membranes, Langmuir 25 (2009) 10139-10145.

[25] M. L. Lind, B.H. Jeong, A. Subramani, X. Huang, E. M.V. Hoek, Effect of mobile cation on zeolite-polyamide thin film nanocomposite membranes. J. Mater. Res. 24 (2009) 1624-

1631.

[26] M. L. Lind, D. E. Suk, T. Nguyen, E. M. V. Hoek, Tailoring the structure of thin film nanocomposite membranes to achieve seawater RO membrane performance, Environ. Sci. Technol. 44 (2010) 8230-8235.

[27] J. Kim, B. Van der Bruggen, The use of nanoparticles in polymeric and ceramic membrane structures: Review of manufacturing procedures and performance improvement for water treatment, Environ. Pollut. 158 (2010) 2335-2349. [28] G-R. Xu, J-N. Wang, C-J. Li, Strategies for improving the performance of the polyamide thin film composite (PA-TFC) reverse osmosis (RO) membranes: Surface modifications and nanoparticles incorporations, Desalination 328 (2013) 83-100.

[29] A. Razmjou, J. Mansouri, V. Chen, The effects of mechanical and chemical modification of Ti02 nanoparticles on the surface chemistry, structure and fouling performance of PES ultrafiltration membranes, J. Membr. Sci. 378 (201 1 ) 73-84.

[30] Z. Zhang, V. Dalvi, B. Mehta, A. Fechtenkoetter, Y. Li, M. Lauter, Raspberry-type metal oxide nanostructures coated with Ce02 nanoparticles for chemical mechanical planariza- tion, US20120077419 A1

[31] R. Wei, S. Zhang, Y. Cui, R. C. Ong, T. S. Chung, B. J. Helmer, J. S. de Wit, Highly permeable forward osmosis (FO) membranes for high osmotic pressure but viscous draw solutes, J. Membr. Sci. 496 (2015) 132-141.

[32] Y. Cui, X.Y. Liu, T. S. Chung, M. Weber, C. Staudt, C. Maletzko, Removal of organic micro-pollutants (phenol, aniline and nitrobenzene) via forward osmosis (FO) process: Eval- uation of FO as an alternative method to reverse osmosis (RO), Water Res. 91 (2016)

104-1 14.

[33] Y. P. Tang, J. X. Chan, T.S. Chung, M. Weber, C. Staudt, C. Maletzko, Simultaneously covalent and ionic bridging towards antifouling of GO-imbedded nanocomposite hollow fiber membranes, J. Mater. Chem. A. 3 (2015) 10573-10584.

[34] Y.P. Tang, H. Wang, T. S. Chung, Towards high water permeability in triazine-framework- based microporous membranes for dehydration of ethanol, Chemsuschem 8 (2015) 138- 147.

[35] H. M. Chen, W. S. Hung, C. H. Lo, S. H. Huang, M. L. Cheng, G. Liu, K.R. Lee, J.Y. Lai, Y. M. Sun, C. C. Hu, R. Suzuki, T. Ohdaira, N. Oshima, Y. C. Jean. Free-volume depth profile of polymeric membranes studied by positron annihilation spectroscopy: layer structure from interfacial polymerization. Macromolecules, 40 (2007) 7542-7557.

[36] Y. P. Tang, N. Widjojo, T. S. Chung, M. Weber, C. Maletzko, Nanometric thin skinned dual-layer hollow fiber membranes for dehydration of isopropanol, AIChE J. 59 (2013) 2943-2956.

[37] Y. P. Tang, S. Yuwen, T. S. Chung, M. Weber, C. Staudt, C. Maletzko, Synthesis of hy- perbranched polymers towards efficient boron reclamation via hybrid ultrafiltration process, J. Membr. Sci. 510 (2016) 1 12-121 .

[38] J. Gao, S. P. Sun, W. P. Zhu, T. S. Chung, Green modification of outer selective P84

nanofiltration (NF) hollow fiber membranes for cadmium removal, J. Membr. Sci. 499 (2016) 361-369.

[39] V. Freger, Nanoscale heterogeneity of polyamide membranes formed by interfacial

polymerization, Langmuir. 19 (2003) 4791-4797.

[40] E. Ayandele, B. Sarkar, P. Alexandridis, Polyhedral ooligomeric silsesquioxane (POSS)- containing polymer nanocomposites, Nanomaterials. 2 (2012) 445-475.

[41] P. Iyer, G. Iyer, M. Coleman, Gas transport properties of polyimide-POSS nanocomposites, J. Membr. Sci. 358 (2010) 26-32.

[42] G.-Y. Chai, W.B. Krantz, Formation and characterization of polyamide membranes via interfacial polymerization, J. Membr. Sci. 93 (1994) 175-1 92. [43] J. S. Louie, I. Pinnau, I. Ciobanu, K. P. Ishida, A. Ng, M. Reinhard, Effects of polyether- polyamide block copolymer coating on performance and fouling of reverse osmosis membranes, J. Membr. Sci. 280 (2006) 762-770.

[44] Y. Zhou, S. Yu, C. Gao, X. Feng, Surface modification of thin film composite polyamide membranes by electrostatic self deposition of polycations for improved fouling resistance,

Sep. Purif. Technol. 66 (2009) 287-294.

[45] Z. Mai, V. Butin, M. Rakib, H. Zhu, M. Rabiller-Baudry, E. Couallier, Influence of bulk concentration on the organization of molecules at a Membrane surface and flux decline during reverse osmosis of an anionic surfactant, J. Membr. Sci. 499 (2016) 257-268.

[46] A. E. Childress, S. S. Deshmukh, Effect of humic substances and anionic surfactants on the surface charge and performance of reverse osmosis membranes, Desalination 1 18 (1998) 167-174.

[47] S. P. Sun, T. S. Chung, K. J. Lu, S. Y. Chan, Enhancement of flux and solvent stability of Matrimid® thin film composite membranes for organic solvent nanofiltration, AIChE J. 60 (2014) 3623-3633.

[48] L. Lin, R. Lopez, G. Z. Ramon, O. Coronell, Investigating the void structure of the polyamide active layers of thin-film composite membranes, J. Membr. Sci. 497 (2016) 365-376.

[49] S. M. J. Zaidi, F. Fadhillah, Z. Khan, A.F. Ismail, Salt and water transport in reverse osmosis thin film composite seawater desalination membranes, Desalination 368 (2015) 202-213.

[50] W. Xie, G. M. Geise, B. D. Freeman, H-S. Lee, G. Byun, J. E. McGrath, Polyamide inter- facial composite membranes prepared from m-phenylene diamine, trimesoyi chloride and a new disulfonated diamine, J. Membr. Sci. 403-404 (2012) 152-161 .

Claims

Claims

1. Process for treating water, wherein feed water is being treated with a membrane M, said membrane M being a reverse osmosis membrane, a forward osmosis membrane or a nanofiltration membrane, wherein said membrane M is contacted with an anionic surfactant A, wherein said anionic surfactant preferably comprises a C6 to Cio aliphatic hydrocarbon chain.

2. Process according to claim 1 , wherein said alkyl sulfate A is a C6 to Cio alkyl sulfate.

3. Process according to claims 1 to 2 wherein said alkyl sulfate A is an octyl sulfate.

4. Process according to claim 3, wherein said alkyl sulfate a is selected from sodium octyl sulfate or potassium octyl sulfate.

5. Process according to any of claims 1 to 4, wherein said membrane M is treated with anionic surfactant A in intervals varying in the range of one day to five years.

6. Process according to any of claims 1 to 5, wherein said alkyl sulfate A is being applied to membrane M as an aqueous solution S comprising 0.1 to 5000 ppm by weight of alkyl sulfate A.

7. Process according to any of claims 1 to 6, wherein membrane M is treated with aqueous solution S for a period of 5 minutes to 5 days.

8. Process according to any of claims 1 to 7, wherein said aqueous solution S has a pH of 4 to 12.

9. Process according to any of claims 1 to 8, wherein said membrane M comprises a separa- tion layer of polyamide.

10. Process according to any of claims 1 to 9, wherein said process is a desalination process.

1 1. Process according to any of claims 1 to 10, wherein said alkyl sulfate A is being admixed to the feed water during the operation of said process.

12. Process according to any of claims 1 to 1 1 , wherein said membrane M comprises nano- particles.

13. Process according to claim 12, wherein said nanoparticles have a core-shell morphology.

14. Process according to claim 13, wherein the core of said nanoparticles is made of silica, titania, zirconia, alumina, zinc oxide or combinations thereof. Process according to claim 13, wherein the shell of said nanoparticles is made of titania, silver, ceria, zirconia, alumina, zinc oxide or combinations thereof.

Use of C6 to Cio alkyl sulfates for healing defects of a membrane M, membrane M being a reverse osmosis membrane, a forward osmosis membrane or a nanofiltration membrane.

Membranes obtained in a process comprising the following steps:

a) Providing a membrane comprising nanoparticles N, preferably core-shell nanoparticles,

b) Subjecting the membrane obtained in step a) to a treatment with an organic solvent L,

c) Treating the membranes obtained in step b) with a solution S comprising anionic surfactant A, preferably a dodecyl sulfate or octylsulfate.