Tailoring the Separation Behavior of Polymer-Supported Organosilica

Research Article www.acsami.org

Tailoring the Separation Behavior of Polymer-Supported Organosilica Layered-Hybrid Membranes via Facile Post-Treatment Using HCl and HN3 Vapors Genghao Gong, Hiroki Nagasawa, Masakoto Kanezashi, and Toshinori Tsuru* Department of Chemical Engineering, Hiroshima University, Higashi-Hiroshima 739-8527, Japan S Supporting Information *

ABSTRACT: A promising layered-hybrid membrane consisting of a microporous organosilica active layer deposited onto a porous polymer support was prepared via a facile sol−gel spin-coating process. Subsequently, the pore sizes and structures of the organosilica top layers on the membrane surface were tuned at mild temperature combined with vapor treatment from either hydrochloric acid (HVT) or ammonia (AVT), thereby tailoring the desalination performance of the membranes during reverse osmosis (RO) processing. The effects of HVT and AVT on the pore size, structure, and morphology of organosilica layers and on the separation performances of membranes were investigated in detail. We confirmed that both HVT and AVT processes accelerated the condensation of silanol (SiOH) in the organosilica layer, which led to dense silica networks. The layeredhybrid membranes after HVT showed an improved salt rejection and reduced water flux, while membranes after AVT exhibited a decrease in both salt rejection and water permeability. We found that HVT gave rise to smoother and denser organosilica layers, while AVT produced large voids and formed pinholes due to Ostwald ripening. These conclusions were supported by a comparative analysis of the results obtained via FTIR, TG-MS, SPM, and RO desalination. KEYWORDS: organosilica, layered-hybrid membrane, hydrochloric acid vapor treatment, ammonia vapor treatment, reverse osmosis

■

acids.8,9 Recent research in organosilica membrane fabrication, however, has been dominated by ceramic membranes. These organosilica membranes are often prepared on flat or tubular ceramic supports,4,10,11 and their applications continue to be limited by poor reproducibility and a high level of difficulty in the preparation of inorganic membranes, as well as by the high cost of the ceramic supports.12 A new trend in the development of advanced separation membranes involves the fabrication of layered-hybrid membranes consisting of a thin and dense organosilica film layer deposited onto a porous flexible polymeric support. Compared with ceramic membranes, these new types of layered-hybrid organosilica membranes have shown excellent reproducibility at a lower cost in membrane fabrication as well as good separation performance, and these membranes have been applied to several membrane separation processes.13−15 For example,

INTRODUCTION In recent years, organically bridged silica membranes have attracted an ever-increasing amount of attention for use in separation applications, mainly due to their excellent molecular sieving abilities, adjustable pore sizes, and their superior levels of thermal and chemical resistance.1−3 Several research groups successfully prepared popular organosilica membranes using a bridged precursor, 1,2-bis (triethoxysilyl)ethane (BTESE), via a sol−gel process.4 This BTESE-derived silica membrane showed amazing hydrothermal stability for more than 1000 days during the continuous pervaporation dehydration of n-butanol at 150 °C.5 Also, we fabricated a BTESE membrane via a “spacer” technique to allow tuning of the silica networks, which also improved the hydrothermal stability and high hydrogen permeability in gas separations.6 Our research group was the first to apply this BTESE membrane to the reverse osmosis (RO) desalination of a NaCl aqueous solution, confirming a superior chlorine tolerance over a wide range of chlorine concentrations (35 000 ppm·h),7 and good stability was also reported for both inorganic (nitric) and organic (acetic) © 2016 American Chemical Society

Received: February 16, 2016 Accepted: April 12, 2016 Published: April 12, 2016 11060

DOI: 10.1021/acsami.6b01986 ACS Appl. Mater. Interfaces 2016, 8, 11060−11069

ACS Applied Materials & Interfaces

■

Ngamou et al. successfully deposited a BTESE-derived silica layer onto porous polyamide-imide substrates via an expanding thermal plasma chemical vapor deposition (ETP-CVD) approach.16 The obtained polymer-supported organosilica membrane showed separation performance comparable to that of ceramic membranes for the pervaporative dehydration of an n-butanol water mixture. Although this novel approach avoids the calcination step, the key ethane bridges (CH2 CH2) in the silica networks would be partially decomposed by the abundant reactive oxygen species during the ETP-CVD process. The ethane bridges played a vital role in enhancing the hydrothermal stability of the silica membranes, but they were only retained at a rate of about 30%. An alternative approach to fabricating organosilica layers on porous polymeric supports was first proposed by our research group. Using a facile, sol−gel spin coating process, a thin and uniform BTESE-derived silica layer was successfully deposited onto a porous polysulfone support.17,18 One of the advantages of this approach is that the ethane bridges can be preserved completely. These layeredhybrid organosilica membranes were applied to the vapor permeation dehydration of an isopropanol−water solution and to the reverse osmosis (RO) desalination of a NaCl solution, and showed good stability and reproducibility. Nevertheless, polymeric supports generally require lower calcination temperatures, and both the ceramic and polymer supported organosilica membranes described above decreased the separation selectivity.18,19 Mild methods that can be used to solidify and modify sol− gel silica-based films have been successfully applied to antireflective (AR) and scratch-resistant coatings. Boudot et al. utilized an ammonia vapor treatment (AVT) process to prepare AR coatings on poly(methyl methacrylate) (PMMA) at room temperature.20 Belleville and Li et al. reported a similar AVT process that causes surface silanol condensation for silicabased AR coatings.21,22 A hydrochloric acid vapor treatment (HVT) process has also been developed to prepare silica colloidal AR coatings on polymer substrates at a mild temperature.23 Both HVT and AVT processes are known to effectively enhance mechanical robustness and the densification of these silica films or coatings,23,24 which suggested that both HCl and NH3 vapors could be used to change the silica network structure in silica-based films or membranes. More importantly, these processes avoided thermal treatments, or at least required only mild curing temperatures that polymers could generally withstand. This suggested that both acid and base vapor treatments at low temperatures could probably be used to tune the pore sizes of organosilica active layers to form a denser silica network structure, thereby improving the separation performance. As far as we could ascertain, there has been no report concerning the effects of HVT or AVT on the separation behavior of layered-hybrid membranes for a RO desalination process. This was the first fabrication of a polymer-supported organosilica layered-hybrid membrane via a facile, sol−gel spin coating process with heat treatment at a mild temperature. We also explored the possibility of tuning the pore sizes of organosilica networks via HVT and AVT processing at a mild temperature to improve the separation selectivity of the layered-hybrid membranes. Herein, we provide a detailed comparison of the effects of HVT and AVT processes on the morphology, structure, and characterization of layered-hybrid organosilica membranes, along with separation performances during the RO desalination of NaCl aqueous solutions.

Research Article

EXPERIMENTAL SECTION

Sol Synthesis. A polymeric sol, 1,2-bis (triethoxysilyl)ethane (BTESE: (EtO)3SiH2C−CH2Si(OEt)3, Gelest, Inc.), was synthesized via hydrolysis and polymerization in a mixture of water, HCl and 1propanol. First, a given mass of BTESE was mixed with 1-propanol, and then distilled water and HCl were added dropwise to the mixture under continuous stirring. The molar ratio of the BTESE/H2O/HCl mixture was 1:60:0.1 with 5.0 wt % of BTESE. This mixed solution was stirred for 1.5 h in a closed glass bottle at 60 °C. After that, the obtained BTESE sol was maintained at 2 °C before use. To prepare a BTESE-derived gel powder, the BTESE polymeric sol was dried at 100 °C under air, and then a mortar was used to grind it into powder. Preparation of a Layered-Hybrid Membrane. A nanofiltration membrane (NTR 7450, Nitto-Denko, Japan; see Figure S1 in the Supporting Information, SI-1) was used as a porous polymeric support. Fabrication of the NTR-supported organosilica layered-hybrid membrane was carried out via a sol−gel, spin-coating process that is described elsewhere.15 Briefly, an NTR support with a 2.5 cm diameter was connected to a macroporous stainless substrate (Pore diameter: 100 μm, porosity: 50%) and was then placed onto a spin coater. Approximately 200 μL of BTESE sol was subsequently dispensed onto the NTR support. The rotation speed was accelerated from 0 to 5000 rpm in 5 s and then held there for 30 s. The spin-coated samples were dried for 10 min at room temperature. This spin coating process was repeated twice. Finally, the obtained samples were dried at 100 °C for 20 min. BTESE-derived silica films on glass substrates were also prepared using the same steps and conditions. Hydrochloric Acid and Ammonia Vapor Treatment. As shown in Figure 1, the layered-hybrid organosilica membranes were placed

Figure 1. Schematic diagrams of the hydrochloric acid and ammonia vapor treatments. horizontally onto a sealed glass cell. Vapor treatments with hydrochloric acid (HVT) and ammonia (AVT) were performed using the vapors that were naturally generated from a 25% HCl aqueous solution and a 28% aqueous ammonia solution, respectively. All treatment equipment was contained in an oven in order to control the treatment temperature (60 °C). The HVT and AVT processes were also applied to BTESE-derived silica gel powders and films. Characterization. The static water contact angle (CA) of the BTESE films on glass substrates was measured using a Dropmaster (DM-300, Kyowa Co., Japan). Thermogravimetric (TG)-mass spectrometric (MS) analysis was carried out at a ramping rate of 10 °C min−1 up to 120 °C, and was then held for 1 h under He flow (300 cm3 min−1) to remove the adsorbed water. The heating was then increased to 1000 °C with 300 cm3 min−1 He gas at a ramping rate of 10 °C min−1. Attenuated total reflection Fourier transform infrared spectroscopy (ATR-FT/IR-4100, JASCO, Japan) was applied to determine the chemical structure of the BTESE films before and after HVT and AVT. The structures and morphologies of the layeredhybrid membranes were examined by field-emission scanning electron microscopy (FESEM, Hitach S-4800, Japan) with an acceleration voltage of 5 kV. The surface topography of the BTESE films on glass substrates before and after HVT and AVT was characterized using a scanning probe microscope (SPM, Nanocute Sii, Japan) and a 3D laser microscope (VK-9700, Keyence, Japan). 11061


Research Article


Figure 2. Changes in the water contact angles of the BTESE-derived silica films during the HVT (left) and AVT (right) processes as a function of time and treatment temperatures (25 °C vs 60 °C).

Figure 3. TG curves of BTESE gel powders under different treatment conditions for temperatures ranging from 25 to 1000 °C (a) and from 120 to 1000 °C (b). For b, the TG data were adopted after the samples were heated to 120 °C and held for 1 h. Before measurement, all the samples were placed in a drybox for 10 h at room temperature. Membrane Performance Test. Reverse osmosis (RO) desalination experiments using a 2000 ppm sodium chloride (NaCl) aqueous solution were performed using a typical RO test installation that is described elsewhere.18 Molecular weight cutoff (MWCO) measurements were carried out using 500 ppm neutral solutes: ethanol, isopropanol, glucose, and maltose. The concentrations of solutions in both the feed and permeate were measured using a conductivity meter (HORIBA, ES-51) for NaCl and a total organic carbon analyzer (Shimadzu, TOC-VE) for neutral solutes. Unless otherwise specified, all the RO tests were conducted at 25 °C. The RO performances included water permeability, Lp, and observed rejection, Robs, and were evaluated using eqs 1 and 2, respectively:

Lp = Jv /(ΔP − Δπ )

reduction in the hydrophilic groups thereby showing a more hydrophobic surface. In addition, a much higher contact angle was achieved at 60 °C compared with the one at 25 °C. This indicated that a higher temperature for the HVT process might have enhanced the dehydration caused by condensation of the silanol groups. A similar phenomenon was also observed in the AVT process, whereby the water contact angle of all the samples was increased with treatment time. This might also have been a result of the catalytic action of NH3·H2O vapor promoting the dehydration of the SiOH groups on the BTESE film surface. Meanwhile, in the AVT process, the contact angles of the samples were also higher at 60 °C than at 25 °C. This suggested that the catalytic action of the NH3·H2O vapor on the BTESE films might have been more effective at a higher AVT temperature. In addition, a higher water contact angle was obtained in the HVT process compared with the AVT process. The main reason might be because the condensation degree of silanol groups in the HVT process was greater than in the AVT process. Another possible reason might be the difference of roughness of the membrane surface, which will be explained in subsequent sections. Therefore, the following HVT and AVT processes were performed at 60 °C for 2 h. In order to confirm the above conclusions, thermogravimetric (TG)-mass spectrometric (MS) analysis was conducted to detect the decomposition behavior of BTESE-derived silica gel powders following HVT and AVT. For comparison, BTESE gel powder without HVT and AVT was also heated at 60 °C under air for 2 h. Figure 3a shows the weight losses of BTESE gel powders under treatment conditions that ranged from 25 to 1000 °C (Before measurement, all the samples were placed in the drybox for 10 h at room temperature). Under these conditions, TG weight residue decreased rapidly at 25−120 °C

(1)

Where ΔP (=P1 − P2) is the difference in applied pressure, Δπ (=π1 − π2) is the difference in osmotic pressure. Jv is the volume flux.

R obs = (1 − C p/Cf ) × 100%

(2)

where Cp and Cf are the permeate and feed concentrations, respectively. The effect of concentration polarization was not considered in this work due to the low permeate flux. To confirm a steady state, each of the feed and permeate solutions were collected after the RO system had operated for at least 3 h.

■

RESULTS AND DISCUSSION First, the water contact angles of BTESE-derived silica films on glass substrates were investigated as a function of treatment time and different treatment conditions (HVT vs AVT, 25 °C vs 60 °C), and the results are shown in Figure 2. During the HVT process, the contact angles of all the BTESE films increased with treatment time and tended to be constant after about 2 h. One explanation for this phenomenon could have been that during HVT process HCl acted as a catalyst to further promote the condensation of residual silanol groups (Si−OH) on the BTESE film surface, which would have resulted in a 11062


Research Article

ACS Applied Materials & Interfaces probably due to the elimination of adsorbed water in the samples. Following the HVT and AVT processes, however, the weight losses of the samples were 12.5 and 15%, respectively, which were lower than that of the untreated sample (weight loss: 18.7%). This showed that both the HVT and AVT processes induced a decrease in the content of adsorbed water in the samples. In addition, to avoid the effect of adsorbed water on the weight loss of samples and to further explore the decomposition behavior of these samples, the TG curves of samples were plotted by normalizing the weight at 120 °C, as shown in Figure 3b. The second weight loss step was observed at 200− 500 °C, and was caused mainly by the dehydration of the silanol groups. This was confirmed when a corresponding peak of m/z 18 for H2O appeared at this temperature range (The MS-peaks of all samples are given in Figure S2). As the amount of silanol groups was reduced at about 500 °C, the dehydroxylation reaction of silanol groups became less frequent, leading to a decrease in the rate of weight loss. The third weight loss step occurred above 500 °C, and originated mainly from decomposition of the organic moiety. The release of methane (CH3 or CH4) and hydrogen (H2) was observed in the MS peaks (see Figure S2), indicating that pyrolysis of the ethane bridges took place via the cleavage of the carbon− carbon bond. It was reported that an organic moiety (CH2 CH2) will be retained at temperatures of 300 and 500 °C under air and N2 atmospheres, respectively. Hence, the difference in weight loss for the samples was attributed to the different amounts of silanol groups participating in the dehydroxylation reaction in samples rather than to the decomposition of the organic groups. It was obvious that the greater the amount of silanol groups in the samples, the greater the weight loss of the samples in a TG experiment. Figure 3b shows that the weight losses at 500 °C, which could be an index of the amount of dehydroxylation, were 3.12 and 3.86% for the HVT and AVT post-treated samples, respectively, and were less than that (4.42%) of the untreated sample. Finally, weight losses of the HVT and AVT post-treated samples at 1000 °C were 8.62% and 8.64%, respectively, both of them were similar but less than that (9.43%) of the untreated sample. This result suggests that HVT and AVT could accelerate the dehydroxylation of a portion of the silanol groups in BTESE-derived silica networks.” Figure 4 shows the FTIR spectra of BTESE thin films before and after HVT and AVT processes. All samples showed the typical characteristic absorption bands of pure BTESE films.25

For example, the intense peaks of a SiOSi stretch appeared at 1020−1200 cm−1; the characteristic absorptions bands of silanol bonds (SiOH) were located at 910 and 3750 cm−1; and, a board absorption band between 3000 and 3500 cm−1 could be assigned to hydroxyl groups (OH) and adsorbed water. Moreover, the peaks that appeared in the region from 2850 to 2930 cm−1 were ascribed to the CH2 bonds from the bridging ethane groups. For HVT and AVT post-treated BTESE films, the peak intensity for silanol (SiOH) groups was reduced while the absorbance of the SiOSi groups was apparently increased. This indicates the occurrence of condensation reactions for the silanol groups in both HVT and AVT processes, which led to the formation of siloxane bonds (SiOSi). Meanwhile, peak intensities between 3000 and 3500 cm−1 for BTESE films after both HVT and AVT processes was also lower than that for untreated films. This showed that both HVT and AVT at least partially compelled the removal of adsorbed water in the films, and that this effect was more significant compared with thermal treatment at 100 °C. These results also agreed well with those of the previous results of TG-MS (see SI-2). To quantitatively analyze the condensation degree of silanol group for BTESE films in HVT and AVT processes, a deconvolution of peaks was applied to identify the overlapping peaks in the 1200−850 cm−1 region. As shown in Figure 5a, The SiOSi asymmetric stretching band could be deconvoluted into four peaks, and the absorption band around 900 cm−1 was ascribed to SiOH (920 cm−1) and SiO− (890 cm−1), as reported elsewhere.16,26,27 The peak at 1160 cm−1 overlapped the SiOC adsorption band from the BTESE precursor, and the remaining peaks centered at 1115; 1068; and 1020 cm−1 were assigned solely to the SiOSi stretching modes, respectively.28,29 Therefore, the ratio of the peak area of condensed silica species (peaks II, III, and IV) and uncondensed silicon species (peaks V and VI) can be used as indicators of the condensation degree of the SiOH groups to SiOSi groups in BTESE films. For comparison, the heat treatment temperature was also investigated for BTESE films (BTESE-100, BTESE-200, and BTESE-300 films, which were prepared by heat-treatment at 100, 200, and 300 °C for 20 min, respectively). Figure 5b shows the calculated peak area ratio of SiOSi/SiO(H) for BTESE films under different treatment conditions. It was clear that the ratio of the peak area increased with increases in the heat-treatment temperature because of increases in the condensation reaction degree of silanol groups to SiOSi groups in the BTESE films. More importantly, the area ratios for HVT and AVT post-treated BTESE films were similar, and both were apparently higher than that of BTESE-100 film and actually approximated that of BTESE-200. This indicates that the condensation degree of BTESE films in both HVT and AVT process at 60 °C were obviously increased compared with the virgin BTESE film (BTESE-100), and were similar to that of BTESE film heattreated at 200 °C. This finding would be beneficial for the fabrication of Si-based membranes at lowered temperatures. The condensation of silanol groups in the BTESE gel layer, however, led to a change in thickness. Therefore, the contraction ratio of film could also be considered a quantitative indicator that could be used to indirectly evaluate the condensation degree of BTESE-derived silica films.30 The thickness of BTESE films was measured for samples under different treatment conditions according to the indentation method (a detailed description in SI-3). Table 1 summarizes

Figure 4. FTIR spectra of BTESE films before and after AVT (a) and HVT (b) processes. 11063


Research Article


Figure 5. (a) FTIR spectrum (small circles) and peak deconvolution of HVT post-treated BTESE film. The red curves are the summation (red) of the fitted peaks (gray); (b) FTIR peak area ratios of SiOSi to SiO(H) as a function of different treatment conditions (BTESE-100, BTESE200 and BTESE-300 films, which were prepared by heat-treatment at 100, 200, and 300 °C for 20 min, respectively).

Table 1. Contact Angle, Thickness Contraction, Condensation Degree and Total SiOH Ratio of BTESE-Derived Silica Film under Different Treatment Conditions BTESE samples

treatment conditions

BTESE-100 (Virgin) BTESE-HVT BTESE-AVT BTESE-200 BTESE-300

100 °C for 20 min in air HCl·H2O vapor at 60 °C for 2 h NH3·H2O vapor at 60 °C for 2 h 200 °C for 20 min in N2 300 °C for 20 min in N2

contact angle (deg) 40.5 90.3 73.3 66.9 84.7

(±1.0) (±2.3) (±0.7) (±1.9) (±2.1)

thickness contractiona (%)

condensation degreeb ()

total SiOH ratioc ()

− 13.3 11.6 17.7 26.8

− 3.4 3.19 3.51 5.47

0.56 0.45 0.46 − −

Contraction ratios = (Tuntreated − Ttreated)/Tuntreated × 100%; (Tuntreated and Ttreated are the film thicknesses before and after treatment, respectively). Condensation degree = ARtreated/ARuntreated; (ARtreated and ARuntreated are the ratios of the peak areas for BTESE films before and after treatment, respectively, according to Figure 5b) cSee SI-4. a b

A typical SEM image of the cross-section of the BTESE/ NTR layered-hybrid membrane is shown in Figure 6. A dense,

the contact angles, condensation degree, thickness contractions of BTESE films, and total SiOH ratio for different applied conditions. The thickness of BTESE films after HVT and AVT processes contracted by 13.3 and 11.6%, respectively, because of the condensation of SiOH groups to SiOSi groups in the films. The order of the thickness contraction was consistent with that of the condensation degree calculated using the peak area ratio shown in Figure 5b. Meanwhile, the changes in the amount of silanol groups in the organosilica networks before and after treatments was also an important factor, which was able to be quantitatively analyzed from the differences in the condensation degrees of the silanol groups for BTESE-derived silica networks under different treatment conditions. Therefore, the amount of silanol groups of the samples was evaluated via the 29Si MAS solid-state NMR measurement. The results showed the total SiOH ratio of BTESE samples after both HVT and AVT processes decreased compared with the untreated BTESE sample (a detailed description in SI-4). Moreover, according to the results of the FTIR spectra, the elimination of partially adsorbed water in the film during the HVT and AVT processes might be another possible reason for the thickness contraction of BTESE films. It is worth noting that although the condensation degrees of HVT and AVT posttreated BTESE films were similar to those of BTESE-200 film, their thickness contraction ratios were lower than 17.7% after heat-treatment at 200 °C. This might have been caused by the presence of H2O vapor in both the HVT and AVT processes, which suppressed the further removal of absorbed water in the film because of a dynamic equilibrium with the atmosphere. As expected, however, both the contraction ratio and the condensation degree for BTESE films following HVT and AVT processes were still lower than that of the film following thermal treatment at 300 °C.

Figure 6. SEM image of the BTESE/NTR layered-hybrid membrane. The insert is a cross-section of the membrane that shows the NTR support and a BTESE-derived silica top layer.

uniform and approximately 200 nm thick BTESE-derived silica top layer can be clearly observed (see insert Figure 6). The separation performance of the membrane was evaluated by the reverse osmosis (RO) desalination of a 2000 ppm of NaCl solution. Figure 7 shows the time course of the RO performance of the original BTESE/NTR membrane, and HVT and AVT post-treated BTESE/NTR membranes. All the membranes exhibited a basically stable level of water permeability and salt rejection over time. Compared with the 11064


Research Article


Figure 7. Time courses for the RO performances of BTESE/NTR membranes before and after HVT and AVT during the desalination of a 2000 ppm of NaCl solution (The deviation in water permeability and NaCl rejection was less than 3 and 0.5%, respectively).

Figure 8. (a) The trade-offs between NaCl rejection and water permeability for NTR supports and BTESE/NTR membranes before and after HVT and AVT and (b) the trade-offs of the desalination performances for HVT post-treated BTESE/NTR membranes and other inorganic membranes (organosilica/ceramic,7,25,32 silicalite,33 and ZSM-534).

original BTESE/NTR membrane, the HVT post-treated membrane showed higher salt rejection and low water permeability. This was probably because the HVT process intensified the condensation of uncondensed silicon (SiOH and SiO−) species in the BTESE top layer due to acid catalysis, resulting in the formation of siloxane bond (SiO Si). This further cross-linked the silica network structure and probably reduced the pore sizes, which led to a denser BTESE top layer.31 It is interesting that the opposite result was obtained for a BTESE/NTR membrane treated with the AVT process. Although AVT post-treated membranes showed lower levels of water permeability compared with the original version, the salt rejection of this membrane was also decreased from approximately 93 to 89%. In order to confirm this phenomenon, a series of identical experiments for RO desalination were performed for NTR supports and layered-hybrid membranes before and after HVT and AVT processes. Figure 8a shows the trade-offs between NaCl rejection and water permeability for different membrane samples under different treatment conditions. It was apparent that the water permeability and salt rejection of NTR supports had not changed either before or after HVT and AVT processes, indicating that both HVT and AVT would not cause a change in the RO performance or damage the structure of NTR supports. However, all the HVT post-treated BTESE/ NTR membranes not only showed an improved salt rejection compared with original membranes, but also exhibited basically reproducible desalination performances. This indicated that the

HVT process improved the salt rejection of the membrane, which also showed good reproducibility and reliability. Moreover, as shown in Figure 8b, the desalination performance of the BTESE-NTR membrane after HVT was similar to, or better than, that of many inorganic membranes. Hence, the HVT post-treated BTESE-NTR membrane showed a competitive desalination performance in both water permeability and NaCl rejection compared with ceramic membranes. Unexpectedly, however, almost all the AVT post-treated membranes showed lower values for water permeability and salt rejection compared with the original ones. In addition, a larger fluctuation in both water permeability (0.15−1.0 × 10−12 m3/(m2·s·Pa)) and salt rejection (80−90%) was also observed for these membranes. These results suggested that the AVT process might induce a change in the surface morphology or in the network structure of the BTESE top layer, which would result in inconsistent separation performance. Therefore, the values for the molecular weight cutoff (MWCO) of membranes before and after treatment were investigated by measuring the level of rejections for a series of neutral solutes with different molecular weights (MW), which showed the effective pore size change of membranes after the HVT and AVT processes. Figure 9 shows the MWCO curves of membranes for different treatment conditions, using a series of 500 ppm neutral solutes: ethanol (MW: 46), isopropanol (MW: 60), glucose (MW: 180), and maltose (MW: 342). The MWCO, defined at 90% rejection of the original BTESE/NTR membrane was approximately 160 Da. Clearly, following the 11065


Research Article


which was caused by reduced pore sizes, and a slight decrease in rejections of the large ones such as glucose (0.73 nm) and maltose (1.0 nm), which was caused by the formation of large pores.25 This could have been because the AVT process induced further condensation between adjacent SiOH groups, leading to a smaller pore size but simultaneously resulting in the formation of large pores such as pinholes in the BTESE top layer. Instability of the silica-based networks (Si OSi) under alkaline conditions could have been one of the reasons. In addition to this, the results of nitrogen sorption experiments (see SI-5) also suggested that BTESE-derived silica networks became denser after the HVT process due to a decrease in both BET specific surface and pore volume, whereas after the AVT process the BTESE sample showed an extremely low BET specific surface and pore volume, indicating the formation of dead-end pores and macropores, or even a nonporous structure after the AVT process. However, it was difficult to observe subtle structure and/or morphology changes in the BTESE top layer after HVA or AVT processes based on the SEM images of the membrane surface. Therefore, the surface topographies of the BTESE top layer before and after the HVT and AVT processes were measured using SPM. Figure 10 shows the SPM threedimensional images of the BTESE film on glass substrates before and after HVT and AVT processes. After the HVT process (Figure 10b), the surface roughness of the BTESE film was decreased slightly, and the size of the particles on the film surface also trended to become smaller compared with that before the HVT process (Figure 10a). This could have been caused by the further condensation of the uncondensed silicon species in the BTESE film due to catalysis of the HCl vapor, as mentioned before, which would have led to a denser and

Figure 9. Rejection and volume flux as a function of the MWs of neutral solutes for BTESE/NTR membranes before and after HVT/ AVT.

HVT process, the MWCO was decreased to 80 Da, indicating a smaller pore size was obtained after HCl vapor treatment. On the contrary, after the AVT process, the MWCO of the BTESE/NTR membrane was increased to approximately 200 Da, which might suggest that a larger pore size was formed following the NH3 vapor treatment. Although the MWCO was increased after the AVT process, it is worth noting that the volume of permeate flux for the AVT post-treated membrane remained lower than that of the virgin membrane. Moreover, compared with the virgin membrane, after the AVT process, the membrane showed increased rejection for small solutes such as ethanol (Stokes diameter: 0.4 nm) and isopropanol (0.48 nm),

Figure 10. SPM three-dimensional images of BTESE film surfaces before (a) and after (b) HVT, and the BTESE film before (c) and after (d) AVT. 11066


Research Article


Figure 11. Schematic illustrations of the mechanisms of HVT and AVT in BTESE-derived silica networks.

thereby improving the salt rejection of layered-hybrid membranes for the RO desalination of NaCl solutions. By contrast, although the condensation reaction of SiOH also occurred in the silica networks during the AVT process, the BTESE membrane surface was prone to the production of large voids that formed pinholes, which resulted in decreased salt rejection during the RO process. Moreover, Ostwald ripening occurred randomly in the BTESE layer during the AVT process, which caused the formation of nonuniform pores in the BTESE-derived silica networks, and led to fluctuations in the membrane performance.

smoother surface. Moreover, the results of nitrogen sorption experiments (see SI-5) also suggested that organosilica networks become denser after the HVT process due to a decrease in both BET specific surface and pore volume, and possess a narrower pore size distribution compared with untreated BTESE sample in the microporous region. On the basis of the catalysis mechanism in the sol−gel processing of silica, the acidic catalysis generally promoted a dense and smooth surface due to the formation of a linear chain network consisting of small primary particles.23,35,36 This agreed with our SPM results. It was interesting that an opposite phenomenon was observed for the BTESE film following the AVT process, as shown in Figure 10c, d. Following the AVT process, the BETESE film surface became rough and showed a morphology that was characterized by larger-sized particles. Meanwhile, some pinholes were also observed on the surface of the AVT post-treated BTESE film (indicated by red arrows), which could have been the reason for a decrease in the salt rejection of layered-hybrid membranes following the AVT process. This could have been caused by a base-catalyzed mechanism under the presence of NH3·H2O vapor, which usually leads to a porous structure and a rough surface with coarse particle morphology during sol−gel silica coating.36,37 This could have been why the water contact angle of the AVT post-treated BTESE film was smaller than that of the HVT post-treated version, because although the AVT accelerated the condensation of the SiOH groups in BTESE film, this condensation reaction tends to occur between adjacent silica particles, as shown in Figure 11, which leads to a formation of large silica particles.22,36,38 Simultaneously, Ostwald ripening occurred within the BTESE films during the AVT process,20,39 so that the silica particles grew in size and were reduced in number. The resultant larger interparticle voids would produce pinholes in the BTESE film. Moreover, this process is generated randomly in the BTESE layer, leading to a nonuniform void distribution and a roughened surface. This could also be why the RO performances of AVT post-treated layered-hybrid membranes have always tended to fluctuate. The HVT process tended to make the top layer of a BTESE layered-hybrid membrane surface more dense and smooth due to the further condensation of uncondensed silicon moieties via the catalysis of HCl vapor. As shown in Figure 11, this process led to smaller pore sizes in the BTESE-derived silica networks,

■

CONCLUSIONS A promising layered-hybrid membrane consisting of a microporous organosilica active layer deposited onto a porous polymer support was prepared via a facile sol−gel, spin-coating process. The membranes received either a hydrochloric acid vapor treatment (HVT) or an ammonia vapor treatment (AVT). Both the HVT and AVT processes promoted the condensation of silanol (SiOH) in the BTESE top layer via acid/base catalysis. Following the HVT process, the BTESE top layer became more dense and uniform, and the pore sizes in the silica networks were smaller. Thus, the membranes showed improvements in salt rejection of from 93 to 98% and a reduced water flux for RO desalination of NaCl solutions. In contrast, the condensation reaction of SiOH that also occurred in the AVT process produced large voids that formed pinholes in the top layer of the BTESE membrane surface, which resulted in a decrease in both salt rejection and water permeability during the RO process. Therefore, the HVT process improved membrane performance, while the AVT process had an adverse influence on the separation behavior of the layered-hybrid membranes during the RO desalination process.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b01986. SI-1: Surface and cross-section SEM images of NTR support membranes; SI-2: TG-MS characterization of samples before and after treatment; SI-3: indentation method for measuring the thickness of BTESE films; SI4: 29Si MAS NMR characterization of samples before and 11067


Research Article


■

the Application to Vapor Permeation. J. Membr. Sci. 2014, 464, 140− 148. (16) Ngamou, P. H.; Overbeek, J. P.; Kreiter, R.; van Veen, H. M.; Vente, J. F.; Wienk, I. M.; Cuperusc, P. F.; Creatore, M. PlasmaDeposited Hybrid Silica Membranes with a Controlled Retention of Organic Bridges. J. Mater. Chem. A 2013, 1, 5567−5576. (17) Gong, G. H.; Wang, J.; Nagasawa, H.; Kanezashi, M.; Yoshioka, T.; Tsuru, T. Synthesis and Characterization of a Layered-Hybrid Membrane Consisting of an Organosilica Separation Layer on a Polymeric Nanofiltration Membrane. J. Membr. Sci. 2014, 472, 19−28. (18) Gong, G. H.; Nagasawa, H.; Kanezashi, M.; Tsuru, T. Reverse Osmosis Performance of Layered-Hybrid Membranes Consisting of an Organosilica Separation Layer on Polymer Supports. J. Membr. Sci. 2015, 494, 104−112. (19) Wang, J.; Kanezashi, M.; Yoshioka, T.; Tsuru, T. Effect of Calcination Temperature on the PV Dehydration Performance of Alcohol Aqueous Solutions through BTESE-derived Silica Membranes. J. Membr. Sci. 2012, 415, 810−815. (20) Boudot, M.; Gaud, V.; Louarn, M.; Selmane, M.; Grosso, D. Sol−Gel Based Hydrophobic Antireflective Coatings on Organic Substrates: A Detailed Investigation of Ammonia Vapor Treatment (AVT). Chem. Mater. 2014, 26, 1822−1833. (21) Belleville, P. F.; Floch, H. G. Ammonia Hardening of Porous Silica Antireflective Coatings. SPIE’s 1994 International Symposium on Optics, Imaging, and Instrumentation; International Society for Optics and Photonics, 1994, pp 25−32. (22) Li, X.; Gross, M.; Oreb, B.; Shen, J. Increased Laser-Damage Resistance of Sol−Gel Silica Coating by Structure Modification. J. Phys. Chem. C 2012, 116, 18367−18371. (23) Li, T.; He, J.; Yao, L.; Geng, Z. Robust Antifogging Antireflective Coatings on Polymer Substrates by Hydrochloric Acid Vapor Treatment. J. Colloid Interface Sci. 2015, 444, 67−73. (24) Wu, G.; Wang, J.; Shen, J.; Yang, T.; Zhang, Q.; Zhou, B.; Deng, Z.; Fan, B.; Zhou, D.; Zhang, F. Properties of Sol−Gel Derived Scratch-Resistant Nano-Porous Silica Films by a Mixed Atmosphere Treatment. J. Non-Cryst. Solids 2000, 275, 169−174. (25) Xu, R.; Kanezashi, M.; Yoshioka, T.; Okuda, T.; Ohshita, J.; Tsuru, T. Tailoring the Affinity of Organosilica Membranes by Introducing Polarizable Ethenylene Bridges and Aqueous Ozone Modification. ACS Appl. Mater. Interfaces 2013, 5, 6147−6154. (26) Deshmukh, S. C.; Aydil, E. S. Investigation of Low Temperature SiO2 Plasma Enhanced Chemical Vapor Deposition. J. Vac. Sci. Technol., B: Microelectron. Process. Phenom. 1996, 14, 738−743. (27) Wang, S.; Wang, D. K.; Smart, S.; Diniz da Costa, J. C. Ternary Phase-Separation Investigation of Sol-Gel Derived Silica from Ethyl Silicate 40. Sci. Rep. 2015, 5, 14560. (28) Wang, C. Y.; Shen, Z. X.; Zheng, J. Z. Thermal Cure Study of a Low-k Methyl Silsesquioxane for Intermetal Dielectric Application by FT-IR Spectroscopy. Appl. Spectrosc. 2000, 54, 209−213. (29) Tejedor-Tejedor, M. I.; Paredes, L.; Anderson, M. A. Evaluation of ATR-FTIR Spectroscopy as an ″In Situ″ Tool for Following the Hydrolysis and Condesation of Alkoxysilanes under Rich H2O Conditions. Chem. Mater. 1998, 10, 3410−3421. (30) Falcaro, P.; Grosso, D.; Amenitsch, H.; Innocenzi, P.; Pado, V.; Marzolo, V. J. Silica Orthorhombic Mesostructured Films with Low Refractive Index and High Thermal Stability. J. Phys. Chem. B 2004, 108, 10942−10948. (31) Wang, J.; Gong, G. H.; Kanezashi, M.; Yoshioka, T.; Ito, K.; Tsuru, T. Pervaporation Performance and Characterization of Organosilica Membranes with a Tuned Pore Size by Solid-phase HCl Post-Treatment. J. Membr. Sci. 2013, 441, 120−128. (32) Ibrahim, S. M.; Xu, R.; Nagasawa, H.; Naka, A.; Ohshita, J.; Yoshioka, T.; Kanezashi, M.; Tsuru, T. Insight into the Pore Tuning of Triazine-based Nitrogen-rich Organoalkoxysilane Membranes for Use in Water Desalination. RSC Adv. 2014, 4, 23759−23769. (33) Liu, N.; Li, L.; McPherson, B.; Lee, R. Removal of Organics from Produced Water by Reverse Osmosis using MFI-type Zeolite Membranes. J. Membr. Sci. 2008, 325, 357−361.

after treatment; and SI-5: nitrogen sorption characterization of samples before and after treatment (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (T.T.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Core Research for Evolutional Science and Technology (CREST) Program of the Japan Science and Technology Agency (JST).

■

REFERENCES

(1) Castricum, H. L.; Paradis, G. G.; Mittelmeijer-Hazeleger, M. C.; Kreiter, R.; Vente, J. F.; Ten Elshof, J. E. Tailoring the Separation Behavior of Hybrid Organosilica Membranes by Adjusting the Structure of the Organic Bridging Group. Adv. Funct. Mater. 2011, 21, 2319−2329. (2) Castricum, H. L.; Sah, A.; Kreiter, R.; Blank, D. H.; Vente, J. F.; Johan, E. Hybrid Ceramic Nanosieves: Stabilizing Nanopores with Organic Links. Chem. Commun. 2008, 9, 1103−1105. (3) Xu, R.; Ibrahim, S. M.; Kanezashi, M.; Yoshioka, T.; Ito, K.; Ohshita, J.; Tsuru, T. New Insights into the Microstructure-Separation Properties of Organosilica Membranes with Ethane, Ethylene, and Acetylene Bridges. ACS Appl. Mater. Interfaces 2014, 6, 9357−9364. (4) Agirre, I.; Arias, P. L.; Castricum, H. L.; Creatore, M.; Johan, E.; Paradis, G. G.; Ngamou, P. H. T.; Van Veen, H. M.; Vente, J. F. Hybrid Organosilica Membranes and Processes: Status and Outlook. Sep. Purif. Technol. 2014, 121, 2−12. (5) Van Veen, H. M.; Rietkerk, M. D.; Shanahan, D. P.; Van Tuel, M. M.; Kreiter, R.; Castricum, H. L.; Vente, J. F. Pushing Membrane Stability Boundaries with HybSi Pervaporation Membranes. J. Membr. Sci. 2011, 380, 124−131. (6) Kanezashi, M.; Yada, K.; Yoshioka, T.; Tsuru, T. Design of Silica Networks for Development of Highly Permeable Hydrogen Separation Membranes with Hydrothermal Stability. J. Am. Chem. Soc. 2009, 131, 414−415. (7) Xu, R.; Wang, J.; Kanezashi, M.; Yoshioka, T.; Tsuru, T. Development of Robust Organosilica Membranes for Reverse Osmosis. Langmuir 2011, 27, 13996−13999. (8) Tsuru, T.; Shibata, T.; Wang, J.; Lee, H. R.; Kanezashi, M.; Yoshioka, T. Pervaporation of Acetic Acid Aqueous Solutions by Organosilica Membranes. J. Membr. Sci. 2012, 421, 25−31. (9) Castricum, H. L.; Kreiter, R.; van Veen, H. M.; Blank, D. H.; Vente, J. F.; Johan, E. High-Performance Hybrid Pervaporation Membranes with Superior Hydrothermal and Acid Stability. J. Membr. Sci. 2008, 324, 111−118. (10) Qi, H.; Han, J.; Xu, N.; Bouwmeester, H. J. Inorganic Microporous Membranes with High Hydrothermal Stability for the Separation of Carbon Dioxide. ChemSusChem 2010, 3, 1375−1260. (11) Qureshi, H. F.; Nijmeijer, A.; Winnubst, L. Influence of Sol−Gel Process Parameters on the Micro-Structure and Performance of Hybrid Silica Membranes. J. Membr. Sci. 2013, 446, 19−25. (12) Burgraaf, A.; Cot, L. Fundamentals of Inorganic Membrane Science; Elsevier Science: New York, 1996; Vol. 4. (13) Jang, K. S.; Kim, H. J.; Johnson, J. R.; Kim, W. G.; Koros, W. J.; Jones, C. W.; Nair, S. Modified Mesoporous Silica Gas Separation Membranes on Polymeric Hollow Fibers. Chem. Mater. 2011, 23, 3025−3028. (14) Patrick, H. T.; ávan Veen, H. M. On the Enhancement of Pervaporation Properties of Plasma-Deposited Hybrid Silica Membranes. RSC Adv. 2013, 3, 14241−14244. (15) Gong, G. H.; Wang, J.; Nagasawa, H.; Kanezashi, M.; Yoshioka, T.; Tsuru, T. Fabrication of a Layered Hybrid Membrane Using an Organosilica Sseparation Layer on a Porous Polysulfone Support, and 11068


Research Article

ACS Applied Materials & Interfaces (34) Li, L. X.; Liu, N.; McPherson, B.; Lee, R. Enhanced Water Permeation of Reverse Osmosis through MFI-type Zeolite Membranes with High Aluminum Contents. Ind. Eng. Chem. Res. 2007, 46, 1584− 1589. (35) Wu, G.; Wang, J.; Shen, J.; Yang, T.; Zhang, Q.; Zhou, B.; Deng, Z.; Fan, B.; Zhou, D.; Zhang, F. A New Method to Control NanoPorous Structure of Sol-Gel-Derived Silica Films and Their Properties. Mater. Res. Bull. 2001, 36, 2127−2139. (36) Vincent, A.; Babu, S.; Brinley, E.; Karakoti, A.; Deshpande, S.; Seal, S. Role of Catalyst on Refractive Index Tunability of Porous Silica Antireflective Coatings by Sol-Gel Technique. J. Phys. Chem. C 2007, 111, 8291−8298. (37) Guo, Y. J.; Zu, X. T.; Jiang, X. D.; Yuan, X. D.; Zheng, W. G.; Xu, S. Z.; Wang, B. Y. Laser-Induced Damage Mechanism of the Sol− Gel Single-Layer SiO2 Acid and Base Thin Films. Nucl. Instrum. Methods Phys. Res., Sect. B 2008, 266, 3190−3194. (38) Guo, Y. J.; Zu, X. T.; Jiang, X. D.; Yuan, X. D.; Xu, S. Z.; Lv, H. B.; Wang, B. Y. Effect of Ammonia Treatment on Laser-Induced Damage of Nano-Porous Silica Film. Optik 2009, 120, 437−441. (39) Ciriminna, R.; Fidalgo, A.; Pandarus, V.; Béland, F.; Ilharco, L. M.; Pagliaro, M. The Sol−Gel Route to Advanced Silica-Based Materials and Recent Applications. Chem. Rev. 2013, 113, 6592−6620.

11069


Tailoring the Separation Behavior of Polymer-Supported Organosilica

Recommend Documents