High-flux reverse osmosis membranes incorporated with NaY zeolite nanoparticles for brackish water desalination

Journal of Membrane Science 476 (2015) 373–383

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

High-flux reverse osmosis membranes incorporated with NaY zeolite nanoparticles for brackish water desalination Hang Dong a,b, Lin Zhao b, Lin Zhang a, Huanlin Chen a, Congjie Gao a, W.S. Winston Ho b,n a Key Laboratory of Biomass Chemical Engineering, Ministry of Education, College of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, PR China b William G. Lowrie Department of Chemical and Biomolecular Engineering, Department of Materials Science and Engineering, The Ohio State University, 2041 College Road, Columbus, OH 43210-1178, USA

art ic l e i nf o

a b s t r a c t

Article history: Received 3 July 2014 Received in revised form 1 November 2014 Accepted 30 November 2014 Available online 9 December 2014

Thin film nanocomposite (TFN) membranes incorporated with NaY zeolite nanoparticles were prepared via interfacial polymerization (IP) of trimesoyl chloride and m-phenylenediamine on nanoporous polysulfone supports. The isolated zeolite-filled polyamide layer was observed by transmission electron microscopy, and the nanoparticles dispersed in the dense nodular polyamide on the polysulfone side. The effects of IP reaction time and zeolite loading on membrane separation performance were investigated. The results showed that a longer IP reaction time was necessary to form a denser zeolite–polyamide layer for higher salt rejection, and the optimum zeolite loading was determined to be 0.15 wt%. Under the optimum conditions, the water flux increased from 0.95 to 1.78 m3/m2/day (23.3 to 43.7 gal/ft2/day (gfd)) with the incorporation of the zeolite nanoparticles, while providing a high salt rejection of 98.8% (2000 ppm NaCl solution, 225 psi (1.55 MPa), 25 1C). The TFN membranes were then post-treated with aqueous solutions containing glycerol, camphorsulfonic acid-triethylamine salt, and sodium lauryl sulfate to further improve the water flux. By optimizing the post-treatment solution composition, an improved brackish water desalination performance was achieved with 2.06 m3/m2/day (50.6 gfd) water flux, which was more than double compared to that of the TFC membrane without the zeolite nanoparticles, along with 98.4% salt rejection. & 2014 Elsevier B.V. All rights reserved.

Keywords: NaY zeolite nanoparticle Thin film nanocomposite (TFN) Reverse osmosis Brackish water desalination

1. Introduction The shortage of fresh water is an increasingly crucial challenge around the world due to the rapid growth in population and the expanding needs of agriculture and energy consumption along with pollution and deficiencies in water management [1]. Seawater and brackish water desalination, which can supply fresh water for drinking, irrigation and industrial development, has emerged as the most sustainable approach to alleviate the water scarcity [2]. Among various desalination technologies, the membrane-based reverse osmosis (RO) process is simple to design and has excellent scale-up capability, and it currently dominates the global desalination market [3]. In the RO desalination process, high performance semi-permeable membranes play the most important role because they determine the product water quality and affect the overall energy consumption [4]. Therefore, high efficiency RO membranes with improved productivity along with good salt rejection are highly desired.

n

Corresponding author. Tel.: þ 1 614 292 9970; fax: þ 1 614 292 3769. E-mail address: [email protected] (W.S. Winston Ho).

http://dx.doi.org/10.1016/j.memsci.2014.11.054 0376-7388/& 2014 Elsevier B.V. All rights reserved.

Polyamide (PA) thin film composite (TFC) membranes were developed in 1980s and are now the most widely used desalination membranes due to the high intrinsic water permeability and good stability over a wide range of pH values [5]. A typical TFC membrane consists of a thick, porous, nonselective layer, along with an ultrathin barrier layer on its top surface. This asymmetric structure brings some key advantages to the membranes. Specifically, each individual layer of TFC membranes can be optimized for its particular function, i.e., the thin film layer can be designed for the desired combination of water flux and salt rejection, while the porous support layer can be prepared for good mechanical properties combined with minimum resistance to permeate flow [6]. A number of recent studies have been carried out in terms of the selection and modification of both suitable thin barrier layers and porous substrates towards the goal of enhancing separation performance, fouling resistance and chlorine tolerance [7]. In particular, the modification of the thin film layers can be achieved by the incorporation of hydrophilic additives or by monomer substitution [8–11], while the substrates can be modified with the addition of organic solvents or replaced by other polymers to assist in the preparation of high-performance TFC membranes with improved mechanical and thermal properties [12,13].

374

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

However, the structure of the polymer membrane is difficult to control, and some studies have suggested that it is hard to further increase the water permeability of polymer-based membranes without sacrificing selectivity [14]. Nanotechnology may offer new avenues to develop novel water treatment membranes that can exceed state-of-the-art performance [15]. A number of recent studies have focused on the incorporation of inorganic materials into both the thin film layer and the support layer of organic RO TFC membranes [16,17]. Notably, a new class of thin-film nanocomposite (TFN) membranes with dramatically improved water permeability was developed by incorporating nanoparticles within polyamide films [18]. This concept offers the potential for the next generation of high performance RO membranes and has attracted considerable interest. Different types of nanomaterials have been investigated for such applications including zeolite, silica, carbon nanotubes, mesoporous carbon, pure metal and nanometal oxides [19–23]. The unique functionalities of these materials have introduced new degrees of freedom in the RO membrane design, which combines the essential properties of conventional membrane polymers with the advantages of nanomaterials. Among various nanoparticles, zeolite nanoparticles are now the most frequently used nano-fillers in TFN membranes that lead to more permeable polyamide active layers. The physical and chemical properties of zeolite are crucially important to the TFN membrane performance and have been widely studied in recent years, such as particle size, adding method, chemical resistance and surface modification [24–28]. Generally, it is hypothesized that zeolite nanoparticles with a tight pore distribution less than the diameter of a hydrated salt ion can exclude the solute to pass through, while also simultaneously providing preferential flow paths for water transport [18]. Besides the molecular-sieving mechanism, it is also believed that the presence of zeolite nanoparticles may change the structure of the polyamide thin film layer by the formation of nano-gaps at the organic–inorganic interfaces, which can reduce the crosslinking density of the polyamide layer [19]. Notably, the nanostructure of polyamide active layers is believed to have two distinct regions: a dense phase of nodular polyamide on the support side and a more open structure of loose polyamide on the surface side. Many studies have suggested that the polyamide dense phase presumably acts as the true separation barrier [29–31]. Therefore, as for zeolite-filled TFN membranes, the distribution characteristics of zeolite nanoparticles in different polyamide regions is especially crucial to bring the function of zeolite nanoparticles into full play. It is reasonable to hypothesize that the zeolite nanoparticles will play a greater role in affecting the membrane properties when the nanoparticles are mainly dispersed in the dense core sub-layer of polyamide. As a result, the study of specific zeolite distribution state is essential to achieve a better understanding of the transport mechanism of TFN membranes. Zeolite-Y possesses the Faujasite (FAU) framework with a 3-dimensional pore structure similar to Linde Type A (LTA). The pore diameter is relatively large, i.e., 7.4 Å, which is formed by a 12-member oxygen and connected to the inner cavity with a diameter of 12 Å [32]. FAU type zeolite has been shown as a good candidate for the application in seawater desalination by molecular dynamics simulations [33]. Therefore, the incorporation of zeolite-Y nanoparticles in TFN membranes may create more preferential water pathways than zeolite-A (4.2 Å pore diameter) and result in a better membrane separation performance. The effects of adding FAU type zeolites with a NaX composition into polyamide thin layers on membrane separation performance have been reported [34]. However, the increased water flux was achieved at the cost of salt rejection due to the void defects formed in the structure of the thin film layers. Thus, it is important to

observe and study the detailed dispersibility and distribution of zeolite nanoparticles in polyamide layers to avoid the non-selective defects. Besides, the post-treatment of TFC membranes by glycerol solutions is an effective way to enhance the membrane separation performance [35]. As a result, the effects of post-treatment process on the properties of zeolite-filled TFN membranes is also worth studying. In this study, NaY zeolite nanoparticles were synthesized via the hydrothermal method. The framework structure, particle diameter, and elemental composition of the synthesized zeolite nanoparticles were characterized by X-ray powder diffraction (XRD), dynamic light scattering (DLS), and X-ray photoelectron spectroscopy (XPS), respectively. The TFN membranes were prepared by adding NaY zeolite nanoparticles in amine solution in the interfacial polymerization (IP) process. Scanning electron microscopy (SEM) and attenuated total reflectance infra-red (ATR-IR) were employed to investigate the membrane morphology and composition. Transmission electron microscope (TEM) was used to observe the isolated polyamide thin film layer and determine the dispersion and distribution of zeolite nanoparticles in polyamide. Contact angles of deionized water were measured on air-dried samples of synthesized TFN membranes with different zeolite loadings to verify the hydrophilicity of the membranes. The effects of IP reaction time and zeolite loading were evaluated by using brackish water desalination tests with 2000 ppm NaCl solution at 225 psi (1.55 MPa) and 25 1C. After the membrane preparation conditions were optimized, the post-treatment using glycerol solutions containing camphorsulfonic acid-triethylamine salt was investigated to further improve the water flux of TFN membranes. For the post-treatment, the effects of heat-treatment time and solution composition on TFN membrane separation performance were studied.

2. Experimental 2.1. Materials The polysulfone support (ultrafiltration membrane) with nonwoven polyester fabric backing was purchased from TriSep Corporation. Isopropanol (IPA, 499.9%, Fisher Scientific) was used for the pretreatment of the polysulfone supports. Chemicals used for the synthesis of the polyamide thin film layer in the aqueous phase included mphenylenediamine (MPD, 499%, Sigma-Aldrich), triethylamine (TEA, 499.5%, Sigma-Aldrich), (þ)-10-camphorsulfonic acid (CSA, 499%, Advanced Asymmetrics) and sodium lauryl sulfate (SLS, 485%, Fisher Scientific). Trimesoyl chloride (TMC, 498%, TCI America) and Isopar Gs (ExxonMobil Chemical) were used as monomer and solvent in the organic phase, respectively. Chemicals used for the post-treatment process included sodium sulfate anhydrous (Na2SO4, 499.4%, Fisher Scientific), sodium carbonate (Na2CO3, 499.5%, Sigma-Aldrich), glycerol (499.5%, Sigma-Aldrich), TEA, CSA, and SLS. The membrane separation performance was tested using sodium chloride (NaCl, 499%, Sigma-Aldrich). All the chemicals were used as received without further purification. Deionized water was used for the solution preparation and the desalination study. 2.2. Zeolite synthesis and characterization Zeolite-Y crystals were synthesized hydrothermally from the starting molar composition of 0.37 Na2O: 1.0 Al2O3: 3.13 (TMA)2O: 4.29 SiO2: 497 H2O according to previously reported procedures [36]. Typically, the clear synthesis solution was placed in a polypropylene bottle and stirred at 98 1C for 4 days until white crystals were observed. The product was collected by centrifugation and washed to remove excess reactants. The clean product was sodium ion exchanged to remove tetramethylammonium

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

(TMA þ ) cations occluded within the framework supercages, leaving only completely trapped TMA þ ions entrapped within the primary building units of zeolite. The pore-opened NaY zeolite nanoparticles obtained from the replacement of the TMA þ counter-ions with Na þ were used for all subsequent experiments. The zeolite framework structure and crystallinity were characterized by X-ray powder diffraction (XRD, X'Pert PRO diffractometer using Cu Kα radiation). The zeolite samples were dried overnight at 80 1C and ground into powder form before XRD testing. The particle diameter was determined by dynamic light scattering (DLS, Zetasizer 3000HSA, Malvern Instrument). The measurements were carried out using dilute aqueous colloidal suspensions of the synthesized zeolite product. The weight loading of the suspension was on the order of 0.01 wt% and was clear to the naked eye. Morphological characterization of zeolite nanoparticles was carried out using scanning electron microscopy (SEM, ULTRA 55, Carl Zeiss). Samples were prepared on silicon wafers and sputter-coated with gold. X-ray photoelectron spectroscopy (XPS, Escalab 250Xi, Thermo Scientific) was used to determine the elemental composition and the Si/Al ratio of zeolite nanoparticles. 2.3. Membrane preparation Both TFC and TFN membranes were synthesized on the polysulfone support through interfacial polymerization. The polysulfone support was first soaked in an IPA/water (1:1 by weight) solution overnight and rinsed with deionized water for 5 min. Then, the support was soaked in deionized water for another 2 h before it was taped onto a 5 in.  5 in.  0.2 in. glass plate. The excess water on the support surface was removed and dried at room temperature upon standing vertically. The polysulfone support together with the glass plate was then firmly clamped by a custom-fabricated 5 in.  5 in.  0.6 in. Teflon frame (inner cavity: 4.2 in.  4.2 in.) with eight long tail clips. Two separate monomer solutions were prepared and stirred at room temperature for 3 h prior to use. The amine solution contained 2 wt% MPD, 5 wt% CSA–TEA salt, and 0.2 wt% SLS in water was poured onto the surface of the polysulfone support that clamped by the frame and allowed to soak for 6 s. The frame was removed and a squeegee roller was employed to gently drain off the excess amine solution on the top surface. Then, the support was dried upon standing vertically in the air for about 2.5 min until no droplet can be seen on the membrane surface. The frame was clamped again and a solution of 0.1 wt% TMC in Isopar Gs was slowly poured on the amine saturated support. After 7 s of interfacial polymerization reaction, the TMC solution was poured off and the membrane was dried in the oven at 90 1C for 5 min for hydrocarbon removal. Finally, the membrane was peeled off from the glass plates for further treatments and soaked in deionized water before testing. During this membrane preparation process, the pouring of aqueous and organic phase solutions was gently carried out from the frame corner, and a customized air convection with the flow rate of 45 L/min was applied in an oven to remove the hydrocarbon evenly. It should be noted that these optimal membrane preparation conditions for TFC membranes were used according to our previous studies [8–10]. The preparation of zeolite-filled TFN membranes was essentially identical to the TFC membranes. The organic phase was prepared by adding 0.1 wt% TMC in Isopar Gs. The composition of amine solution was 2 wt% MPD, 5 wt% CSA–TEA salt, and 0.2 wt% SLS in water. Since the minute quantity of zeolite powder was difficult to weigh, the synthesized zeolite was first pre-dispersed in water to prepare a 0.75 wt% concentrated solution under ultrasonication. Then, a predetermined amount of this concentrated zeolite solution was dispersed in the aqueous amine

375

solution to make the mass fraction of zeolite in the aqueous phase ranging from 0.02 wt% to 0.2 wt%. The zeolite was dispersed in the aqueous phase due to its better dispersibility in water than Isopar Gs. Homogeneous zeolite dispersion in the amine solution could be obtained by ultrasonication for 1 h at room temperature immediately prior to use in the interfacial polymerization reaction.

2.4. Post-treatment process after membrane formation After the formation of the polyamide thin film layer according to the procedures described in Section 2.3, the membranes were soaked in a neutralization solution containing 0.2 wt% Na2CO3 and 3.3 wt% Na2SO4 for 20 s. The membrane samples were then washed by dipping in deionized water at about 47 1C four times each for 4 min. After rinsing, the post-treatment solution containing various amounts of glycerol, CSA–TEA salt and 0.3 wt% SLS was used to soak the membrane for 2 min. An air knife was employed to remove the extra post-treatment solution on the membrane surface, followed by a second step drying in the oven at 90 1C for a different length of time. This procedure of the post-treatment process was employed based on our previous work [8,9].

2.5. Membrane characterization Attenuated total reflectance infra-red (ATR-IR, Renishaw inVia system) was used to determine the functional groups on the membrane surface. The membrane samples were dried in a desiccator overnight before ATR-IR measurements. SEM (as described in Section 2.2) was used to characterize membrane surface and cross-section morphologies. Membrane cross-section samples were prepared by peeling away the polyester fabric layer, remaining the polyamide–polysulfone composite layers together. Then, the samples were gently freeze fractured in liquid nitrogen to prepare the membrane sections. The morphologies of thin film surface and cross-section were also characterized by transmission electron microscope (TEM, JEM-1230, Jeol Ltd.) at an accelerating voltage of 120 kV. For the TEM characterization of membrane surface, N,N-dimethylacetamide was used to dissolve the polysulfone support and the isolated polyamide thin layer was loaded on the copper grid for TEM observation. For the TEM characterization of membrane cross-section, small pieces of the fabric free membrane samples were embedded in the Epon resin and cut by an ultramicrotome (Ultracut UC7, Leica). Approximately 60–80 nm thick membrane sections were obtained and placed on copper grids for TEM observation. Sessile drop contact angles of TFC and TFN membranes were evaluated by a contact angle goniometer (Digidrop, GBX Instruments). The equilibrium value was the steady-state average of left and right angles. The data reported were the average of six measurements for each membrane sample. The desalination performance of the synthesized TFC and TFN membranes was evaluated in terms of water flux and salt rejection by a cross-flow RO water permeation apparatus. The membrane sample was loaded in the stainless cell with an effect area of 28.6 cm2 (4.43 in.2). The feed solution was prepared using 2000 ppm NaCl to simulate the brackish water. The transmembrane pressure was kept at 225 psi and the operating temperature was maintained at 25 1C by a cooling system. After the membranes were compacted for 2 h to reach the steady state, the water flux can be calculated by measuring the weight of the water penetrated through the membranes per effective area per unit time, and its unit was converted into m3/m2/day or gallon/ft2/day (gfd) (1 gfd ¼0.0407 m3/m2/day). The salt rejection (R) was achieved

376

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

according to the following equation: R ¼ 1

Cp Cf

where Cp and Cf are the permeate and feed salt concentrations, respectively, which were determined by a conductivity meter. The reported RO experimental results were the average values obtained from at least three membrane samples prepared at different times, and the error bars showed the standard deviation.

3. Results and discussion 3.1. Characterization of zeolite nanoparticles The morphology and size of the synthesized zeolite nanoparticles are illustrated in Fig. 1. It can be seen from the SEM image in Fig. 1(a) that the nanoparticles had cubic-like shapes and most of them exhibited particle sizes ranging from 100 nm to 200 nm. The particle diameter distribution was characterized by DLS as shown in Fig. 1(b). The mean hydrodynamic diameter of the synthesized nanoparticles could be confirmed as around 150 nm from the peak analysis by number. Then, the crystal structure of nanoparticles was characterized by XRD. As shown in Fig. 2(a), the synthesized nanoparticles could be determined to be pure Faujasite (FAU) type zeolite by comparison with the standard XRD powder patterns [37]. It is wellknown that a typical FAU framework consists of cubo-octahedral “sodalite cages” and the pores are arranged perpendicular to each other. The three-dimensional channel system is formed by a 12-

Fig. 2. Structure and composition of the synthesized zeolite nanoparticles characterized by (a) XRD (the standard XRD powder pattern of FAU type zeolite is included as the inset for comparison [37]) and (b) XPS.

membered ring with the diameter of 7.4 Å, which is believed to allow water molecules (2.7 Å kinetic diameter) to pass through, while þ rejecting dissolved hydrated salt ions (e.g., Na(aq) : 7.2 Å kinetic diameter) [34]. The chemical compositions of FAU zeolite were determined by XPS analysis. Fig. 2(b) shows the Si 2s, A1 2s, Si 2p, Al 2p and Na 2s XPS spectra of zeolite. Since the photoionization cross sections for the 2p subshells of Si and Al are significantly larger than those for the 2s subshells [38], it is reasonable to measure the 2p spectra for the calculation of the silicon to aluminum ratio. Accordingly, the Si 2p/Al 2p atomic concentration ratio was calculated to be 1.7, which was a typical zeolite-Y composition. Nevertheless, to determine the Na concentration, the intensity of the 2s spectrum had been used instead of that of 2p due to the Na 2p spectrum was interfered by the O 2s spectrum. The Al 2p/Na 2s ratio was found to be around 1.0 that resulted from the sodium ion exchange process. Thus, the NaY zeolite nanoparticles were successfully synthesized with the mean particle diameter smaller than 200 nm, and these nanoparticles were proper to incorporate into the thin film polyamide layer of TFC membranes. 3.2. Characterization of thin film nanocomposite membranes

Fig. 1. Morphology of the synthesized zeolite nanoparticles. (a) SEM image for the zeolite nanoparticles under 10,000  magnification and (b) particle diameter distribution characterized by DLS.

3.2.1. Attenuated total reflectance infra-red (ATR-IR) spectroscopy characterization ATR-IR spectroscopy can provide insight into the chemical functionality as well as the type of chemical bonds present in

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

TFC and zeolite-filled TFN membranes. Fig. 3(a) shows the ATR-IR spectra from 2000 to 800 cm  1 of the pristine TFC and the 0.1 wt% zeolite incorporated TFN membranes. However, there were no obvious differences between the two membranes, especially for the absence of the expected broad peak at 1050–950 cm  1 (Si–O and Al–O functionality of NaY nanoparticles) for the TFN membrane. The depth of penetration of the reflected IR beam in the ATR technique is typically below 1 μm and can exceed the thickness of

Fig. 3. ATR-IR spectra from (a) 2000 to 800 cm  1 and (b) 4000 to 2600 cm  1 for TFC and TFN membranes with 0.1 wt% zeolite loading.

377

polyamide skin layer. Thus, the bands ascribable to the polysulfone sub-layer could also be revealed, which might cover the Si–O and Al–O bands, especially in the condition with a relatively low zeolite loading in the polyamide layer. As shown in Fig. 3(b), it can be seen that though the peak positions of the two membranes were almost identical, the peak intensity from 3500 to 3200 cm  1 differed a lot. For the TFN membrane, the intensity ratio of the peak at 3749 cm  1 to the centered peak (3361 cm  1) of the broadened shoulder was around 1.2, while for the TFC membrane, this ratio was 2.6. These results indicated the existence of the slightly broadened and enhanced shoulder around 3400 cm  1. This phenomenon was also observed in another study and it corresponded to the interaction between the hydroxyl groups of the zeolite and the functional groups of the polyamide [19]. The results indicated that the NaY zeolite nanoparticles were present in the polyamide thin film layer.

3.2.2. Morphology characterization The surface and cross-section morphology of both TFC and TFN membranes were analyzed using SEM to investigate the effects of zeolite incorporation. The SEM images of the top surface of TFC and TFN (0.1 wt% zeolite loading) membranes are shown in Fig. 4. A familiar “hill and valley” structure can be seen in the SEM images, which is the typical polyamide morphology formed through interfacial polymerization of MPD and TMC. The incorporation of zeolite nanoparticles did not affect the overall morphology of the polyamide thin film layer. Nevertheless, a few cubic-like structures (featured by white circles) could be observed on the surface of TFN membrane, which were in consistent with the morphology of zeolite nanoparticles. However, since most of the zeolite nanoparticles should be embedded in the polyamide layer rather than attached on the membrane surface, it was very difficult to conclude from SEM images that zeolite nanoparticles were observed on the TFN membrane surface. TEM should be more reliable to study the real distribution of zeolite nanoparticles in TFN membranes. It is worth noting that several studies have shown that adding zeolite in the organic phase could affect the interfacial polymerization reaction and might result in a more “leaf-like” membrane surface structure [24,26]. This could be explained by the fact that zeolite in the organic phase would hydrate and release heat when they encountered hydrated MPD from the aqueous phase, which in turn locally enhanced the miscibility of the two phases and increased the breadth of the reaction zone. However, the zeolite nanoparticles were added in the aqueous phase in our study and they should be already hydrated before the interfacial polymerization, so the reaction zone was mainly unaltered and the pristine surface morphology of polyamide remained almost the same.

Fig. 4. SEM images for the surfaces of (a) TFC membrane and (b) TFN membrane with 0.1 wt% zeolite loading under 50,000  magnification.

378

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

Fig. 5. SEM images for the cross-section of (a) TFC membrane and (b) TFN membrane with 0.1 wt% zeolite loading under 50,000  magnification.

The SEM images of membrane cross-section are shown in Fig. 5. While both the active layers of TFC and TFN (0.1 wt% zeolite loading) membranes had the thickness ranging from 100 to 300 nm, the polyamide layer of TFN membrane showed a more flourish structure than that of the TFC membrane, which indicated the existence of the zeolite nanoparticles. TEM is considered as a more effective imaging technique than SEM for studying the surface and cross-section nanostructures of TFN membranes because the zeolite nanoparticles can be easily discerned in the polyamide layer due to their dark appearance under the electron beam. Fig. 6(a) and (b) shows the surface TEM images of the TFN membrane with 0.1 wt% zeolite loading. As shown in this figure, the cubic-like zeolite nanoparticle could be clearly observed among the leaf-like polyamide structure. Fig. 6 (c) and (d) shows the cross-section TEM images of the same membrane. It should be noted that the polyamide layer was detached from the polysulfone support during the sample preparation, which could be observed in low magnification images (images not shown). As shown in Fig. 6(c) and (d), the isolated polyamide layer had a thickness ranging from 100 to 300 nm, which was in accordance with the SEM images, and the zeolite nanoparticles with the mean particle size smaller than 200 nm were well dispersed in the polyamide thin film. Many studies have reported that the thin film layer of TFC membrane actually includes two distinct polyamide regions, which are a dense phase of nodular polyamide that forms a relatively smooth interface with the polysulfone support and a more open “ridge-and-valley” structure of loose polyamide extending outward from the dense phase. This double-layer structure is formed due to different kinetic regimes, and it is commonly believed that the dense polyamide sub-layer presumably constitutes the true separation barrier [29–31]. The higher magnification TEM image in Fig. 6(d) indicated that the zeolite nanoparticles were embedded in the nodular polyamide dense phase on the polysulfone side rather than in the “ridge-and-valley” loose polyamide on the surface side. This result could be mainly ascribed to the following three reasons. First, during the membrane preparation process, the zeolite nanoparticles were dispersed in the aqueous solution and the nanoparticles would stay on the polysulfone support side after the solution was drained off and kept their position when the interfacial polymerization occurred. Second, the zeolite used in this study was hydrophilic, which resulted in lower dispersibility in Isopar Gs, and the zeolite nanoparticles thus stayed close to the surface of polysulfone during the fast interfacial polymerization between MPD and TMC. Third, the mean hydrodynamic diameter of the synthesized zeolite nanoparticles was around 150 nm, which was

well suited to the thickness of the nodular polyamide dense layer. For these reasons, the zeolite nanoparticles tended to disperse in the bottom of the polyamide thin layer on the polysulfone side. Since the dense polyamide sub-layer was indeed the active barrier that determined the water flux and salt rejection of TFC membranes, the orientated distribution of zeolite in TFN membranes would let the nanoparticles play a greater role in ameliorating the membrane separation performance. Fig. 7 shows TEM cross-section images of the TFN membrane prepared with higher zeolite loading. It can be seen that some of the zeolite nanoparticles tended to aggregate and formed many defects between the nanoparticles as well as at the interface between the zeolite and the polyamide matrix. These nonselective defects were believed to decrease the membrane separation performance, especially the salt rejection. The results indicated that the zeolite loading should have a great influence on their dispersibility, thereby affecting the membrane structure and performance, which will be discussed in detail in the following sections. 3.2.3. Contact angle characterization Fig.8 shows the contact angles of the TFC membrane and the TFN membranes with different zeolite loadings. The TFC membrane without the zeolite nanoparticles showed a relatively high contact angle of 68.41. After the incorporation of zeolite, the contact angle of the TFN membranes decreased from 63.41 to 46.21 with increasing zeolite loading from 0.02 wt% to 0.2 wt%. The decrease in the contact angle normally indicates the increase in hydrophilicity, and a more hydrophilic membrane surface typically produces better water permeability. The enhanced hydrophilicity of the TFN membranes could be due to the hydrophilic property of NaY zeolite nanoparticles. As discussed in membrane morphology characterization, some of the nanoparticles tended to aggregate and could be partially exposed on the membrane surface at higher zeolite concentration. Thus, the membrane surface hydrophilicity might increase because of the sorption of water through the hydrophilic pores of NaY zeolite via capillary effects. 3.3. Effects of IP reaction time on TFN membrane separation performance The polyamide thin film layers of TFC membranes are formed by interfacial polymerization reaction of the two monomers, which occur at the interface between aqueous and organic phases. Specifically, both MPD in water and TMC in Isopar Gs diffuse across the liquid–liquid interface and react to create a thin polymer film. However, the IP

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

379

Fig. 6. TEM images for the surface of TFN membrane with 0.1 wt% zeolite loading under (a) 30,000  magnification and (b) 100,000  magnification, and the cross-section images under (c) 20,000  magnification and (d) 50,000  magnification.

Fig. 7. TEM images for the cross-section of TFN membrane with 0.2 wt% zeolite loading under (a) 20,000  magnification and (b) 50,000  magnification.

380

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

when adding zeolite nanoparticles in the aqueous phase, the results showed that a longer IP reaction time was necessary, especially for better salt rejection. It suggested that when the zeolite nanoparticles existed during the IP reaction, the nanoparticles might provide a mass transfer resistance to the monomer (MPD) diffusion across the interface and affect the reaction of – NH2 and –COCl groups, which could result in a relatively looser structure of the thin film layer. Therefore, a longer IP reaction time was necessary to form a denser polyamide layer with higher salt rejection in the presence of zeolite nanoparticles. The results (Fig. 9) showed that 15 s was the optimal IP reaction time for the preparation of TFN membranes and could be applied in the subsequent experiments. 3.4. Effects of zeolite loading on TFN membrane separation performance

Fig. 8. Effects of zeolite loading on pure water contact angles of TFC and TFN membranes.

Fig. 9. Effects of interfacial polymerization reaction time on water flux and salt rejection (adding 0.1 wt% NaY zeolite nanoparticles in aqueous phase, 2000 ppm NaCl solution, 225 psi).

reactions predominantly happen in the organic phase due to the relatively low solubility and stability of the acid chlorides in water [39]. Therefore, the solubility and diffusivity of the amine monomer in the organic phase play an important role in the structure of the resulting polyamide layer and ultimately affect the separation performance. Generally, a dense polyamide layer can be formed within only a few seconds on the polysulfone supports. A relatively longer IP reaction time can further increase the membrane thickness and cross-linking degree, leading to a lower water flux and a higher salt rejection. Fig. 9 presents the effect of interfacial polymerization reaction time on the water flux and salt rejection of the TFN membranes by adding 0.1 wt% NaY zeolite nanoparticles in the aqueous phase. The results showed that the water flux of TFN membranes decreased with longer IP reaction time, while the salt rejection continuously increased. With the IP reaction time shorter than 12 s, the separation performance showed no obvious changes, which were around 1.75 m3/m2/day (43 gfd) water flux and 98.5% salt rejection. When employing 15 s reaction time, the water flux slightly declined to 1.69 m3/m2/day (41.4 gfd), while the salt rejection increased to 98.7%. With further increase in the reaction time, the salt rejection enhanced to 98.9%, but the water flux dropped to 1.51 m3/m2/day (37.0 gfd). In the preparation of TFC membranes without adding zeolite nanoparticles, 7 s of IP reaction time was found to be the optimum condition to obtain the best separation performance [8]. However,

Fig. 10 shows the effects of zeolite loading in the aqueous phase ranging from 0.02 wt% to 0.2 wt% on membrane desalination performance. The water flux of TFN membranes increased continuously from 1.35 m3/m2/day (33.0 gfd) at the lowest zeolite loading (0.02 wt%), up to 1.92 m3/m2/day (47.0 gfd) at the highest zeolite loading (0.2 wt%), which was more than double compared to that of the TFC membranes (0.95 m3/m2/day or 23.3 gfd) [8,9]. On the other hand, the salt rejection was higher than 98.5% and kept nearly constant at the zeolite loading from 0.02 wt% to 0.15 wt%, but then showed a marked decrease to lower than 97.9% at the highest loading with 0.2 wt% zeolite nanoparticles. At the 0.15 wt% zeolite loading, the TFN membrane showed the best separation performance in consideration of both the water flux (1.78 m3/m2/day or 43.7 gfd) and salt rejection (98.8%). The results indicated that the zeolite nanoparticles made a great contribution to the significant increase of the water flux, while maintaining the salt rejection. It was suggested that zeolite nanoparticles might change the bulk polyamide film structure possibly by the formation of nano-gaps at the organic–inorganic interface, which could be responsible for the enhanced water flux. Besides, the NaY zeolite nanoparticles with the uniform porous structure of approximately 0.74 nm pore size might show a molecular sieve effect, which allowed the water molecules preferentially to go through while the solute molecules were rejected. In addition, as shown by the contact angle measurements, the hydrophilicity of the TFN membranes was improved by the incorporation of hydrophilic zeolite nanoparticles, and thus the water permeability was enhanced. However, the zeolite nanoparticles were easily to aggregate at the high concentration, which could lead to the creation of

Fig. 10. Effects of zeolite loading on water flux and salt rejection (adding NaY zeolite nanoparticles in amine solution; 2000 ppm NaCl solution, 225 psi).

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

interstitial defects during the membrane synthesis process. Thus, perfect salt rejection could not be achieved. The aggregation of zeolite nanoparticles was observed by TEM as discussed in the membrane morphology characterization. The interstitial nonselective defects formed at the interfaces could displace the favorable nano-gaps and molecular sieve effects. Therefore, the optimal zeolite loading was about 0.15 wt%, and the proper range of NaY zeolite in the aqueous phase during the TFN membranes preparation was from 0.05 wt% to 0.15 wt%. In this proper range, the TFN membranes showed the water flux higher than 1.62 m3/m2/day (40 gfd) and the salt rejection greater than 98.5%. 3.5. Effects of post-treatment on TFN membrane separation performance In the preparation of the conventional TFC membranes, it is a common practice to treat the membrane after its formation with post-treatment solution containing glycerol, surfactant and organic salt to further enhance the separation performance. In our previous study, the post-treatment solution of 5 wt% glycerol, 6 wt% CSA–TEA salt and 0.3 wt% SLS was demonstrated to be effective on improving the desalination performance of TFC membranes [8,9]. Glycerol is believed to plasticize the polyamide thin film, while the CSA–TEA salt is considered to protect the polysulfone support from pore collapse during heat treatment [35]. The addition of the surfactant may help to promote rewetting of the membrane during the treating process. To evaluate the post-treatment effects on TFN membranes, different heattreatment times and various post-treatment solutions were applied to the TFN membranes, and 0.1 wt% zeolite was employed in the aqueous phase as the typical zeolite concentration in the proper range of zeolite loading. 3.5.1. Effects of heat-treatment time on TFN membrane separation performance The heat-treatment process after dipping membranes into the post-treatment solution can rearrange the polyamide structure that is swelled by the glycerol. The TFN membranes with 0.1 wt% zeolite loading were treated by the solution containing 5 wt% glycerol, 6 wt% CSA–TEA salt and 0.3 wt% SLS, followed by drying in the oven at 90 1C for different lengths of time. The results in Fig. 11 exhibited that by using 5 and 10 min of heattreatment time, the TFN membranes showed a water flux at around 2.32 m3/m2/day (57 gfd), which was much higher than that before post-treatment. When the heat treatment time was increased to 15 min, the water flux was further enhanced to 2.47 m3/m2/day (60.5 gfd) but the salt rejection reduced considerably to only 97.2%.

Fig. 11. Effects of heat-treatment time on TFN membrane separation performance.

381

The CSA–TEA salt in the post-treatment solution enabled the membrane to sustain the second step of heat-treatment by avoiding the damage of the nanoporous structure in the polysulfone support and protecting the PA structure rearrangement process. However, with intensive heat-treatment, the polysulfone support might still be damaged even though CSA–TEA salt was employed. Thus, 10 min could be determined as the optimal heat-treatment time for TFN membranes. 3.5.2. Effects of solution composition on TFN membrane separation performance Glycerol and CSA–TEA salt were considered as the fluxenhancing additives in the membrane post-treatment process while the membrane salt rejection should be maintained. However, the treatment with 5 wt% glycerol, 6 wt% CSA–TEA salt and 0.3 wt% SLS combined with a 10-min drying step caused the salt rejection decrease from higher than 98.5% to only 98.0%. Therefore, different compositions of the post-treatment solution were investigated to optimize the membrane separation performance, especially for preventing the loss of the salt rejection. As depicted in Fig. 12, the effects of glycerol and CSA–TEA salt were first studied by using the post-treatment solution containing 2 wt% CSA–TEA salt without glycerol and 5 wt% glycerol without CSA–TEA salt, respectively. Under these conditions, the water flux was around the similar range with that before post-treatment whereas the salt rejection was low, only 97.5%. Then, a combination of the glycerol concentration and the CSA–TEA salt concentration at the ratio of 1:1.2 by weight, i.e., 1 wt% glycerol and 1.2 wt% glycerol, in the posttreatment solution was employed. The TFN membrane showed much better separation performance, i.e., 2.06 m3/m2/day or 50.6 gfd water flux, 98.4% salt rejection. But, when further increasing CSA–TEA salt to the glycerol:CSA–TEA salt ratio at 1:3 by weight, the salt rejection again dropped down to lower than 98.0%. The results showed that for TFN membranes, the posttreatment process significantly increased the water flux but reduced the salt rejection for 5 out of the 6 glycerol:CSA–TEA salt ratios investigated and depicted in Fig. 12. These results suggested that the nano-gaps formed between polyamide and zeolite nanoparticles were easily enlarged by the post-treatment solution due to the plasticization and swelling effects of glycerol and CSA–TEA salt. The enlarged nano-gaps could be seen as the non-selective defects of the membrane, which were difficult to recover even after the second step of heat treatment. However, by the combination of glycerol and CSA–TEA salt at a proper concentration ratio of 1:1.2 as described earlier, improved membrane desalination performance, especially for higher water flux with good salt rejection

Fig. 12. Effects of different glycerol and CSA–TEA salt compositions with 0.3 wt% SLS on TFN membrane separation performance.

382

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

(2.06 m3/m2/day or 50.6 gfd water flux with 98.4% salt rejection), could still be obtained by the post-treatment. This flux was more than double compared to that of the TFC membranes without the zeolite nanoparticles. Although the mechanism of the synergistic effect by glycerol and CSA–TEA salt at a proper concentration ratio is not clear, one possible explanation could be offered through the tendency of the TFN membrane separation performance. On one hand, the plasticizing effect of glycerol and the swelling effect of CSA–TEA salt tended to increase the water flux of the TFN membrane. But it might cause undesired defects and rearrangement of the thin film nanocomposite layer, leading to the reduction of salt rejection. On the other hand, CSA–TEA salt could also protect the membrane structure during the second step of the heat-treatment process to minimize the loss of salt rejection. As a result, the plasticizing and swelling effects were predominant under the high post-treatment solution concentration, such as 5 wt% glycerol and 6 wt% CSA–TEA salt, while these effects might be partially inhibited by the protection of CSA–TEA salt to maintain the salt rejection under a relatively lower solution concentration with a proper glycerol to CSA–TEA salt ratio. Therefore, the optimum composition of the post-treatment solution was found to be 1 wt% glycerol and 1.2 wt% CSA–TEA salt.

4. Conclusions The NaY zeolite nanoparticles with the mean particle size smaller than 200 nm were embedded into the polyamide layer via interfacial polymerization to form a novel zeolite–polyamide TFN membrane. Both ATR-IR spectrometry and electron microscopy confirmed the existence of zeolite nanoparticles in the polyamide thin film layer. The zeolite nanoparticles dispersed in the dense nodular polyamide on the polysulfone side rather than in the loose polyamide on the surface side. Under the optimized interfacial polymerization reaction time and zeolite loading, TFN membranes showed a significantly improved separation performance for brackish water desalination. The water flux increased from 0.95 m3/m2/day (23.3 gfd) without zeolite nanoparticles to 1.78 m3/m2/day (43.7 gfd) with the optimal 0.15 wt% zeolite nanoparticles in the diamine solution, and the salt rejection was maintained at a high value of 98.8%. By employing proper posttreatment, the water flux of TFN membranes could be further increased but the salt rejection was reduced. The unrecoverable decline of the salt rejection indicated the existence of nano-gaps on the zeolite–polyamide interface, which were presumably enlarged after the post-treatment allowing the passage of NaCl. By optimizing the composition of the post-treatment solution with a proper glycerol: CSA–TEA salt ratio of 1:1.2 by weight, an improved brackish water desalination performance was achieved with 2.06 m3/m2/day (50.6 gfd) water flux, which was more than double compared to that of the TFC membrane without the zeolite nanoparticles, along with 98.4% salt rejection.

Acknowledgments The authors would like to thank Dr. Michael Severance and Dr. Prabir K. Dutta at The Ohio State University for providing the zeolite nanoparticles and also thank Ying Xu and Hua Wang at the Center of Analysis & Measurement of Zhejiang University for the TEM analyses. The first author would like to gratefully acknowledge the financial support by the China Scholarship Council (201206320080). The authors would also like to gratefully acknowledge the Office of Naval Research (N00014-03-1-0994, N00014-05-1-00800 and N00014-10-1-00147), the National Basic Research Program of China (2015CB655300), and the Research Fund for the Doctoral Program

of Higher Education of China (20130101110064) for the financial support of this work.

References [1] R. Connor, Managing Water under Uncertainty and Risk: The United Nations World Water Development Report, 4th ed., UNESCO, Paris, 2012. [2] M. Elimelech, W.A. Phillip, The future of seawater desalination: energy, technology, and the environment, Science 333 (2011) 712–717. [3] B. Penate, L. Garcia-Rodriguez, Current trends and future prospects in the design of seawater reverse osmosis desalination technology, Desalination 284 (2012) 1–8. [4] M.A. Shannon, P.W. Bohn, M. Elimelech, J.G. Georgiadis, B.J. Marinas, A.M. Mayes, Science and technology for water purification in the coming decades, Nature 452 (2008) 301–310. [5] R.E. Larson, J.E. Cadotte, R.J. Petersen, The FT-30 seawater reverse osmosis membrane-element test results, Desalination 38 (1981) 473–483. [6] J.E. Cadotte. Interfacially synthesized reverse osmosis membrane, U. S. Patent 4,277,344; 1981. [7] D. Li, H. Wang, Recent developments in reverse osmosis desalination membranes, J. Mater. Chem. 20 (2010) 4551–4566. [8] L. Zhao, P.C.-Y. Chang, W.S.W. Ho, High-flux reverse osmosis membranes incorporated with hydrophilic additives for brackish water desalination, Desalination 308 (2013) 225–232. [9] L. Zhao, P.C.-Y. Chang, C. Yen, W.S.W. Ho, High-flux and fouling-resistant membranes for brackish water desalination, J. Membr. Sci. 425 (2013) 1–10. [10] L. Zhao, W.S.W. Ho, Novel reverse osmosis membranes incorporated with a hydrophilic additive for seawater desalination, J. Membr. Sci. 455 (2014) 44–54. [11] J. Qin, S. Lin, S. Song, L. Zhang, H. Chen, 4-Dimethylaminopyridine promoted interfacial polymerization between hyperbranched polyesteramide and trimesoyl chloride for preparing ultralow-pressure reverse osmosis composite membrane, ACS Appl. Mater. Interfaces 5 (2013) 6649–6656. [12] A.K. Ghosh, E.M.V. Hoek, Impacts of support membrane structure and chemistry on polyamide–polysulfone interfacial composite membranes, J. Membr. Sci. 336 (2009) 140–148. [13] J. Wei, X.G. Jian, C.R. Wu, S.H. Zhang, C. Yan, Influence of polymer structure on thermal stability of composite membranes, J. Membr. Sci. 256 (2005) 116–121. [14] G.M. Geise, H.B. Park, A.C. Sagle, B.D. Freeman, J.E. McGrath, Water permeability and water/salt selectivity tradeoff in polymers for desalination, J. Membr. Sci. 369 (2011) 130–138. [15] X.L. Qu, P.J.J. Alvarez, Q.L. Li, Applications of nanotechnology in water and wastewater treatment, Water Res. 47 (2013) 3931–3946. [16] M.T.M. Pendergast, E.M.V. Hoek, A review of water treatment membrane nanotechnologies, Energy Environ. Sci. 4 (2011) 1946–1971. [17] M.T.M. Pendergast, A.K. Ghosh, E.M.V. Hoek, Separation performance and interfacial properties of nanocomposite reverse osmosis membranes, Desalination 308 (2013) 180–185. [18] B.H. Jeong, E.M.V. Hoek, Y.S. Yan, A. Subramani, X.F. 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. [19] M.L. Lind, D.E. Suk, T.-V. Nguyen, E.M.V. Hoek, Tailoring the structure of thin film nanocomposite membranes to achieve seawater RO membrane performance, Environ. Sci. Technol. 44 (21) (2010) 8230–8235. [20] G.L. Jadav, P.S. Singh, Synthesis of novel silica–polyamide nanocomposite membrane with enhanced properties, J. Membr. Sci. 328 (2009) 257–267. [21] E.S. Kim, B.L. Deng, Fabrication of polyamide thin-film nano-composite (PATFN) membrane with hydrophilized ordered mesoporous carbon (H-OMC) for water purifications, J. Membr. Sci. 375 (2011) 46–54. [22] E.S. Kim, G. Hwang, M.G. El-Din, Y. Liu, Development of nanosilver and multiwalled carbon nanotubes thin-film nanocomposite membrane for enhanced water treatment, J. Membr. Sci. 394 (2012) 37–48. [23] H. Zhao, S. Qiu, L. Wu, L. Zhang, H. Chen, C. Gao, Improving the performance of polyamide reverse osmosis membrane by incorporation of modified multiwalled carbon nanotubes, J. Membr. Sci. 450 (2014) 249–256. [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] C.L. Kong, T. Shintani, T. Tsuru, Pre-seeding-assisted synthesis of a high performance polyamide–zeolite nanocomposite membrane for water purification, New J. Chem. 34 (2010) 2101–2104. [26] 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. [27] H. Huang, X. Qu, X. Ji, X. Gao, L. Zhang, H. Chen, L. Hou, Acid and multivalent ion resistance of thin film nanocomposite RO membranes loaded with silicalite-1 nanozeolites, J. Mater. Chem. A 1 (2013) 11343–11349. [28] H. Dong, X.Y. Qu, L. Zhang, L.H. Cheng, H.L. Chen, C.J. Gao, Preparation and characterization of surface-modified zeolite–polyamide thin film nanocomposite membranes for desalination, Desalin. Water. Treat. 34 (2011) 6–12. [29] V. Freger, Nanoscale heterogeneity of polyamide membranes formed by interfacial polymerization, Langmuir 19 (2003) 4791–4797. [30] V. Freger, Kinetics of film formation by interfacial polycondensation, Langmuir 21 (2005) 1884–1894.

H. Dong et al. / Journal of Membrane Science 476 (2015) 373–383

[31] F.A. Pacheco, I. Pinnau, M. Reinhard, J.O. Leckie, Characterization of isolated polyamide thin films of RO and NF membranes using novel TEM techniques, J. Membr. Sci. 358 (2010) 51–59. [32] C.h. Baerlocher, L.B. McCusker, D.H. Olson, Atlas of Zeolite Structure Types, 6th ed., Elsevier, Amsterdam, 2007. [33] Y. Liu, X. Chen, High permeability and salt rejection reverse osmosis by a zeolite nano-membrane, Phys. Chem. Chem. Phys. 15 (2013) 6817–6824. [34] M. Fathizadeh, A. Aroujalian, A. Raisi, Effect of added NaX nano-zeolite into polyamide as a top thin layer of membrane on water flux and salt rejection in a reverse osmosis process, J. Membr. Sci. 375 (2011) 88–95. [35] M.A. Kuehne, R.Q. Song, N.N. Li, R.J. Petersen, Flux enhancement in TFC RO membranes, Environ. Prog. 20 (2001) 23–26.

383

[36] W.S.W. Ho, P.K. Dutta, Novel inorganic/polymer composite membranes for CO2 capture (U.S. Department of Energy Annual Report of Budget Period 1, No. DEFE0007632), , 2013. [37] M.M.J. Treacy, J.B. Higgins, Collection of Simulated XRD Powder Patterns for Zeolites, 5th ed., Elsevier, Amsterdam, 2007. [38] J.J. Yeh, I. Lindau, Atomic subshell photoionization cross sections and asymmetry parameters: 1 ⩽ Z ⩽ 103, At. Data Nucl. Data Tables 32 (1985) 1–155. [39] P.W. Morgan, S.L. Kwolek, Interfacial polycondensation 2. Fundamentals of polymer formation at liquid interfaces, J. Polym. Sci. A: Polym. Chem. 34 (1996) 531–559.