English

• 0.   Introduction

  Dehydration of acidic organic aqueous solutions (such as high concentrations of acetic acid) is of extremely importance in chemical industries^[1]. Zeolite membrane pervaporation technology has received widespread attention for this purpose due to the lower energy consumption and non-pollution^[2-7]. Mordenite membranes with a medium Si/Al ratio (n_Si/n_Al) and excellent pore structure are widely regarded as promising potential applications for dehydration of acidic organic-aqueous solutions due to their good hydrophilicity and superior acid resistance^[8-12].

  The synthesis of high-performance zeolite mem- branes obviously depends on the control and optimization of membrane microstructure such as membrane thickness and grain boundary defects^[13-19]. Several researchers have used microwave synthesis methods to shorten the synthesis time and reduce the membrane thickness^[16-19]. Li et al.^[16] prepared high-quality mordenite membranes by microwave synthesis. Under the optimal conditions, the thickness of the as-synthesized membrane was only 1.5 μm and the membrane exhibited a flux of 1.48 kg·m^-2·h^-1 for dehydration of mass fraction of 90% HAc/water mixtures at 75 ℃. Alternatively, many related studies have shown that adding mineralizer fluoride ions into synthesis gel would significantly improve the PV performance of mordenite membranes^[20-27]. Chen et al. ^[20] reported that fluoride ions could optimize the distribution of aluminum atoms in mordenite membrane layer and reduce the grain boundary defect of zeolite crystals, thus the assynthesized membrane showed a long-term acid stability for dehydration of high-concentration acetic acid mixtures. In our previous study^[27], the fluoridecontaining synthesis gel was used to synthesize a com- pact and high-quality mordenite membrane with a flux of 1.36 kg·m^-2·h^-1 for separation of a 90% HAc/H₂O mixture at 75 ℃.

  Generally, fluoride-containing systems can improve the PV performance of mordenite membranes^[20-27]. However, different kinds of fluorides have different effects on the synthesis and quality of zeolite membranes, mainly due to the different cations contained in synthesis gel^[26]. The structure-directing role of alkali-metal cations in the synthesis of zeolites with a low Si/Al ratio has been widely confirmed^[28-29]. In the high silica zeolite and membrane, alkalimetal cations also had a significant influence on the synthesis process^[30-38]. Liu et al.^[35] investigated the influences of the addition of alkali-metal cations on the synthesis of ZSM -5 zeolite, and the results showed that Na⁺ and K⁺ had a significant acceleration on the crystallization of zeolite. Therefore, alkali-metal cations played an important role in the zeolite framework structure, which might also show a remarkable influence on the PV performance of zeolite membranes. Xu et al.^[36] reported that small amounts of sodium ions could improve the quality of pure-silica MFI zeolite membranes, and the membrane exhibited a lower PV performance in the presence of large amounts of sodium. It was because large amounts of sodium ions would decelerate the nucleation and result in the formation of gel particles. Similarly, Fu et al. ^[37] showed that Na⁺ as the mineralizer could promote lateral crystal growth and eliminate intercrystallite defects. To our knowledge, the role of alkali-metal cations in the synthesis of mordenite membranes has not yet been discussed in detail. Consequently, in order to further improve the PV performance of mordenite membrane, it is necessary to systematically study the role of alkali-metal cations in the synthesis of mordenite membranes.

  So, different alkali-metal cations including Li⁺, Na⁺, K⁺ and Cs⁺ as well as the mixed Na⁺ - Li⁺, Na⁺ - K⁺ and Na⁺-Cs⁺ were added into synthesis gel and the influences of alkali-metal cations on the morphology and PV performance of mordenite membrane in the fluoride-containing system were systematically investigated by means of X-ray diffraction (XRD), field emission scanning electron microscopy (FE-SEM), energy dispersive X-ray spectroscopy (EDX), mapping and water contact angle measurement. Moreover, the effect of Na⁺/K⁺ ratio (n_Na⁺/n_K⁺) on the hydrophilicity, PV performance, morphology and thickness of mordenite membranes was discussed in detail. And the role of alkalimetal cations on the synthesis of mordenite membrane was discussed. It is found that Na⁺-K⁺ simultaneously existing in synthesis gel could improve the hydrophilicity and reduce the thickness of mordenite membranes, thereby enhancing PV performances of these membranes. In addition, the long-term acid stability of mordenite membranes was investigated.

  1.   Experimental

  1.1   Materials

  The reagents and chemicals used for mordenite membrane preparation included mordenite seed crystals (HS-642, Si/Al ratio=9, Wako), colloidal silica (TM - 40, 40%, Aldrich), aluminum hydroxide (Al(OH)₃, 99%, Wako), lithium hydroxide (LiOH, 99%, Aladdin), sodium hydroxide (NaOH, 96%, Sinopharm Chemical Reagent), potassium hydroxide (KOH, 82%, Sinopharm Chemical Reagent), cesium hydroxide (CsOH, 99%, Aladdin), lithium fluoride (LiF, 99%, Aladdin), sodium fluoride (NaF, 99%, Wako), potassium fluoride (KF, 99%, Wako), cesium fluoride (CsF, 99%, Aladdin) and deionized (DI) water. Porous mullite tubes (Noritake, inner diameter=9 mm, out diameter=12 mm, pore diameter=1.3 μm) were used as supports.

  1.2   Preparation of mordenite membranes

  Mordenite membranes were prepared on mullite supports with different alkali-metal cations in synthesis gel by hydrothermal secondary growth method. Prior to the synthesis, the outer surfaces of mullite supports were manually rubbed with mordenite seed crystals and the entire seeding process was following the proce- dure as described in our previous studies^{[24, 26]}. The molar ratio of various alkali - metal cations containing synthesis gel was n_SiO2∶n_Al2O3∶(n_Na2O +n_M2O)∶(n_NaF+n_MF)∶ n_H2O =1∶0.08∶0.2∶0.1∶35 (M=Li, K, Cs). And the detailed preparation process of synthesis gel was also according to our previous studies^{[24, 26]}. The final synthe- sis gel was aged for 6 h at a room temperature and then poured into a stainless-steel autoclave. Then two seed- ed mullite supports were vertically immersed into syn- thesis gel. Finally, the autoclave was placed in an oven at 170 ℃ for 5 h. After synthesis, the as-synthesized membranes were repeatedly washed with boiling DI water until the solution became neutral and then dried overnight in a 60 ℃ oven.

  1.3   Characterization of mordenite membranes

  The crystal phase and structures of as-synthesized mordenite membranes were characterized by XRD (Ultima Ⅳ, Rigaku) using a Cu Kα radiation (λ=0.154 06 nm) in the 2θ range of 5°~45° at a scanning speed of 4 (°)·min^-1. The tube voltage was 40 kV and the tube current was 40 mA. The surface and cross-sectional morphologies of mordenite membranes were characterized by cold FE-SEM (Hitachi SU8020) with the acceleration voltage of 5 kV. All samples were sprayed with platinum. The water contact angle (JC-2000CD) was calculated to determine the hydrophilicity of the mordenite membrane surface. The elemental analysis (Na, K, Al, Si) and Si/Al ratios of the mordenite membrane surface were obtained by EDX (Q200, Bruker) equipped in the SU8020 machine.

  1.4   PV performance of mordenite membranes

  PV experiments were carried out using these membranes to separate water-acetic acid mixtures. PV performance of the as-synthesized membrane was evaluated using a PV experimental apparatus described in our previous work^[27]. The feed solution was pumped into a heater to be heated to a given temperature, and then sent to the outside of the membrane modules. The permeate side was kept below 200 Pa by a vacuum pump. The permeate vapor was collected into the feed tank by liquid nitrogen. The acetic acid and water mixture compositions of the feed and permeate were ana- lyzed by a gas chromatograph (GC-14C, Shimadzu) equipped with a thermal conductivity detector (TCD) and a 3 m packed column. Two parameters were used to evaluate membrane separation performance, the permeation flux (J) and the separation factor (α), defined by equations as follows: J=m/(At), α_H2O/HAc= (Y_w/Y_a)/(X_w/X_a), where m is the weight of the permeation (kg); A is the effective membrane area (m²); t is the testing time (h); X_w, X_a, Y_w and Y_a correspond to the mass fractions of components w (water) and a (acetic acid) in the feed and permeate, respectively.

  2.   Results and discussion

  2.1   Effect of single alkali-metal cation

  In order to study the influence of single alkali-metal cation on PV performance and growth of mordenite membranes, the membranes were prepared by adding Li⁺, Na⁺, K⁺ and Cs⁺ into synthesis gel, respectively. Fig. 1 shows XRD patterns of mordenite membranes synthesized with different single alkali-metal cations. As shown in Fig. 1c, the membrane synthesized in the presence of Na⁺ had the typical characteristic peaks of pure mordenite zeolite. But when Li⁺, K⁺ and Cs⁺ were separately added into synthesis gel, except for the characteristic peaks of mullite support, the weak mordenite characteristic peaks were found (Fig. 1b, 1d and 1e). Hence, it could be inferred that the addition of Na⁺ can significantly increase the crystallization rate of mordenite crystals in a certain extent, while Li⁺, K⁺ and Cs⁺ show a slower crystallization rate.

  Figure 1

  

  Figure 1.  XRD patterns of (a) mullite support and mordenite membranes prepared with different single alkali-metal cations: (b) Li⁺, (c) Na⁺, (d) K⁺, (e) Cs⁺

  下载: 全尺寸图片 幻灯片

  Through the surface and cross - sectional SEM images of mordenite membranes (Fig. 2), it could be clearly seen the influence of single alkali-metal cation on the crystallization process of mordenite membranes. When the synthesis gel contained Li⁺, mordenite crystals could not sufficiently grow and plenty of amorphous substances were loosely scattered on the membrane surface. A lot of apparent defects occurred and no obvious membrane layer was formed (Fig. 2a and 2b). Once Na⁺ was added into synthesis gel, plenty of homogeneous elliptical mordenite crystals fully covered on the mullite support, but the membrane surface was rough and the membrane thickness was approximately 5 μm (Fig. 2c and 2d). However, when K⁺ and Cs⁺ were added into synthesis gel, the size of mordenite crystals was greatly reduced, and the elliptical crystal morphology changed to small cylinder (K⁺) and small sphere crystals (Cs⁺). These small crystals were unevenly scattered on membrane surface and did not stack to form a dense membrane layer, even a lot of exposed support appeared (Fig. 2e and 2g). As shown in the cross-sectional images of Fig. 2f and 2h, some small crystals entered into the support pore and no obvious dense membrane layers occurred on the support surface. These SEM images also verify that the alkali-metal cations play a structure-directing role for constructing mordenite framework. Compared with Li⁺, K⁺ and Cs⁺, Na⁺ can obviously promote the growth of mordenite crystals in the equal crystallization time.

  Figure 2

  

  Figure 2.  Surface and cross-sectional SEM images of mordenite membranes prepared with different single alkali-metal cations: (a, b) Li⁺, (c, d) Na⁺, (e, f) K⁺, (g, h) Cs⁺

  下载: 全尺寸图片 幻灯片

  Table 1 shows PV performances of mordenite membranes with different alkali-metal cations in synthesis gel. Combined with Fig. 1 and 2, it could be concluded from Table 1 that when keeping the content of F^- and OH^-/Si ratio (n_OH^-/n_Si) unchanged, the alkali-metal cations have obvious influences on PV performances. The membrane (M2) with synthesis gel containing Na⁺ exhibited excellent separation performance. The permeation flux was 1.91 kg·m^-2·h^-1 and the separation factor was more than 3 500. But when Li⁺, K⁺ and Cs⁺ were separately added into synthesis gel, the as-synthesized membranes (M1, M3 and M4) showed no separation performance. The difference for PV performances of these membranes was consistent with the previous XRD results in Fig. 1 and the corresponding growth of crystal difference in the surface and cross-sectional SEM images of membranes in Fig. 2.

  Table 1

  

  Table 1.  PV performances of mordenite membranes prepared with different alkali-metal cations for a 90% HAc/H₂O mixture at 90 ℃

  下载: 导出CSV

  Membrane*	Alkali-metal cation	J/(kg·m-2·h-1)	αH2O/HAc
  M1	Li+	—	—
  M2	Na+	1.91	3 500
  M3	K+	—	—
  M4	Cs+	—	—
  M5	Na+-Li+	4.48	70
  M6	Na+-K+	2.67	4 000
  M7	Na+-Cs+	2.49	1 200
  * Membrane gel composition: nSiO2∶nAl2O3∶nM2O∶nMF∶nH2O=1∶0.08∶0.2∶0.1∶35 (M=Li+, Na+, K+, Cs+) for M1~M4; nSiO2∶nAl2O3∶(nNa2O+nM2O)∶(nNaF+nMF)∶nH2O=1∶0.08∶0.2∶0.1∶35 (M=Li, K, Cs; nNa+/nM+=2) for M5~M7; Synthesis temperature: 170 ℃; Synthesis time: 5 h.

  In order to further increase the permeation flux of the membrane, the reduce of Na⁺-content in the synthesis gel to reduce the crystallization rate was investigated. Table 2 presents PV performances of mordenite membranes prepared with different Na⁺-contents in synthesis gel. As the Na⁺-content decreased, the permeation flux of the membranes rapidly increased, but separation selectivity rapidly decreased. Lower OH^- and F^- were not conducive to synthesis excellent PV performance membranes.

  Table 2

  

  Table 2.  PV performances of mordenite membranes prepared with different Na⁺-contents for a 90% HAc/H₂O mixture at 90 ℃

  下载: 导出CSV

  Membrane*	Na+-content	J/(kg·m-2·h-1)	αH2O/HAc
  M2	0.5	1.91	3 500
  M8	0.3	2.84	200
  M9	0.1	—	—
  * Membrane gel composition: nSiO2∶nAl2O3∶nNa2O∶nNaF∶nH2O =1∶0.08∶x∶0.1∶35 (x=0.2, 0.1, 0); Synthe- sis temperature: 170 ℃; Synthesis time: 5 h.

  Generally, Na⁺ was added together with alkali source or fluorine source to the synthesis gel. Replacement of a small amount Na⁺ with Li⁺, K⁺ or Cs⁺ reduced the amount of Na⁺ but did not reduce the OH^-/Si ratio and F^-/Si ratio(n_F^-/n_Si), thus slightly reducing the crystallization rate and increasing the permeability flux of as-synthesized mordenite membranes.

  2.2   Effect of binary alkali-metal cations

  Fig. 3 shows the XRD patterns of mordenite mem- branes prepared with different binary alkali-metal cat- ions (M5: Na⁺-Li⁺, M6: Na⁺ -K⁺, M7: Na⁺ -Cs⁺) and their corresponding surface and cross-sectional SEM images are shown in Fig. 4. These membranes synthesized with different binary alkalimetal cations all had the characteristic peaks of pure mordenite zeolite in addition to the characteristic peaks of mullite support (Fig. 3). When Na⁺-Li⁺ were added into synthesis gel, the membrane surface with many small particles exhibited poor stack- ing and no obvious membrane layer was found (Fig. 4a and 4b). Once the synthesis gel containing Na⁺-K⁺ and Na⁺-Cs⁺, the surface of membrane began to form inter- grown mordenite zeolite layers (Fig. 4c~4f). XRD results in Fig. 3b~3d confirmed that the peak intensities of membranes with synthesis gel containing Na⁺-K⁺ and Na⁺-Cs⁺ were stronger than that of the membrane with synthesis gel containing Na⁺-Li⁺, which are consistent with the results in SEM observations. When Na⁺-K⁺ were added into synthesis gel, the square crystals of membrane surface were well-intergrown and dense without obvious pinhole defects (Fig. 4c). The thickness of the membrane layer was about 4 μm (Fig. 4d), which was thinner than that of the membrane with Na⁺. When Na⁺-Cs⁺ were added into synthesis gel, the mordenite crystal morphology changed into small ellipsoidal crystals and the surface was less intergrown, with several pinhole defects (Fig. 4e). As shown in the cross-sectional SEM image (Fig. 4f), the thickness of the membrane slightly increased to ~4.5 μm. These characterization results are consistent with our previous expectations: the replacement of a small amount of Na⁺ with Li⁺, K⁺ and Cs⁺ in synthesis gel can slow down the crystallization rate of the crystal to a certain extent, resulting a thinner membrane layer.

  Figure 3

  

  Figure 3.  XRD patterns of (a) mullite support and mordenite membranes prepared with different binary alkali-metal cations: (b) Na⁺-Li⁺, (c) Na⁺-K⁺, (d) Na⁺-Cs⁺

  下载: 全尺寸图片 幻灯片

  Figure 4

  

  Figure 4.  Surface and cross-sectional SEM images of mordenite membranes prepared with different binary alkali-metal cations: (a, b) Na⁺-Li⁺, (c, d) Na⁺-K⁺, (e, f) Na⁺-Cs⁺

  下载: 全尺寸图片 幻灯片

  PV performances of these membranes are listed in Table 1. The membrane M5 prepared with Na⁺-Li⁺ displayed a high permeation flux of 4.48 kg·m^-2·h^-1 but a poor separation factor of 70, which is attributed to the poorly intergrown membrane layer with large defects and a large number of crystals entering the support channel (Fig. 4a and 4b). The membrane M6 prepared with Na⁺-K⁺ showed the best PV performance: the permeation flux was up to 2.67 kg·m^-2·h^-1 and the separation factor was about 4 000. When Na⁺-Cs⁺ were added into synthesis gel, the permeation flux and separation factor of membrane M7 decreased to 2.49 kg·m^-2·h^-1 and 1 200, respectively. This is related to the less intergrown and thickened membrane layer. Aiello et al. ^[38] showed that in the synthesis of ZSM-5 zeolite, K⁺ was more capable of incorporation of aluminum into the zeolite framework. In this study, K⁺ might have the similar influence in synthesis of mordenite membranes. The Al content of the membrane prepared with Na⁺-K⁺ might be higher and the membrane surface might be more hydrophilic, thus the membrane M6 showed a high permeation flux. Therefore, it was necessary to further study the content of K⁺ in synthesis gel and the role on the synthesis of mordenite membranes.

  2.3   Effect of Na⁺/K⁺ ratio

  To investigate the influence of Na⁺/K⁺ ratio on PV performance and crystal growth of mordenite mem- branes, the membranes were prepared with different Na⁺/K⁺ ratios (0.5~3). Fig. 5 and 6 present the XRD pat- terns and morphologies of mordenite membranes prepared with different Na⁺/K⁺ ratios. As shown in Fig. 5, all these membranes had the characteristic peaks of typical mordenite crystals. Additionally, the peak intensities of mordenite membrane gradually increased with the Na⁺/K⁺ ratio increasing from 0.5 to 3, which indicates that the crystallization rate increases with the increase of Na⁺ content. As seen in surface morphologies of the membranes (Fig. 6a, 6c, 6e and 6g), with the increase of Na⁺ content, the crystal grains in membrane layer gradually became larger, the compactness of membrane surface gradually increased and the intergranular voids gradually decreased. When the Na⁺/K⁺ ratio was 2, a compact and highly intergrown mordenite membrane layer with a thickness of approximately 4 μm occurred on the support surface (Fig. 6e and 6f). As observed from cross - sectional morphologies (Fig. 6b, 6d, 6f and 6h), the thicknesses of membrane layers were found to increase from about 2 μm to ca. 5 μm with the increase of Na⁺/K⁺ ratio from 0.5 to 3. It also indicates that Na⁺ can significantly increase the crystal- lization rate of mordenite crystals, which is accordant with XRD characterization.

  Figure 5

  

  Figure 5.  XRD patterns of (a) mullite support and mordenite membranes prepared with different Na⁺/K⁺ ratios: (b) 0.5, (c) 1, (d) 2, (e) 3

  下载: 全尺寸图片 幻灯片

  Figure 6

  

  Figure 6.  Surface and cross-sectional SEM images of mordenite membranes prepared with different Na⁺/K⁺ ratios: (a, b) 0.5, (c, d) 1, (e, f) 2, (g, h) 3

  下载: 全尺寸图片 幻灯片

  Table 3 presents separation performances of mordenite membranes with different Na⁺/K⁺ ratios in synthesis gel. As Na⁺/K⁺ ratio increased, the separation selectivity of the membranes gradually increased, but the permeation flux gradually decreased. When the Na⁺/K⁺ ratio was 2, the membrane M6 exhibited a high permeation flux and a high separation factor, which was consistent with the XRD and SEM characterization results. Therefore, when there are equal contents of F^- and OH^- as mineral reagents in synthesis gel, an appropriate Na⁺/K⁺ ratio is the key to the preparation of high-performance mordenite membranes. This is due to that a certain content of Na⁺ can ensure the proper crystalliza- tion rate of mordenite membrane, thereby forming a dense, thin and defect-free membrane layer. In this study, the Na⁺/K⁺ ratio of 2 is the optimal synthesis condition of mordenite membrane.

  Table 3

  

  Table 3.  PV performances of mordenite membranes prepared with different Na⁺/K⁺ ratios in synthesis gel for a 90% HAc/H₂O mixture at 90 ℃

  下载: 导出CSV

  Membrane*	Na+/K+ ratio	J/(kg·m-2·h-1)	αH2O/HAc	Si/Al ratio in crystal
  M10	0.5	4.05	350	2.6
  M11	1	3.12	620	3.6
  M6	2	2.67	4 000	4.2
  M13	3	2.15	4 200	4.7
  * Membrane gel composition: nSiO2∶nAl2O3∶(nNa2O+nK2O)∶(nNaF+nKF)∶nH2O=1∶0.08∶0.2∶0.1∶35 (Na+/K+ ratio=0.5, 1, 2 and 3); Synthesis temperature: 170 ℃; Synthesis time: 5 h.

  2.4   Discussion on role of alkali - metal cations on synthesis of mordenite membrane

  The role of alkali-metal cations in zeolite and membrane synthesis has been widely confirmed^{[28-31, 36, 39-40]}. In this study, during the mordenite membrane synthesis process, the presence of alkali-metal cations can accelerate the cleavage of Si—O bonds in silica and then the silicates are easier to react with aluminates to form aluminosilicates. Moreover, alkali-metal cations play the structure-directing role, and can promote the rearrangement or connection of the primary structure unites, thus increasing the nucleation and crystallization rate of mordenite crystals. Na⁺ has a significant acceleration effect on the mordenite crystal- lization while Li⁺, K⁺ and Cs⁺ do not significantly promote mordenite crystal growth. The possible reason is that the alkalinity of Li⁺-containing synthesis gel is lower than Na⁺-containing synthesis gel under the same OH-/Si ratio, which can not effectively promote the dissolution of silicon source in synthesis gel. As shown in Fig. 2a, the membrane surface was coated with agglomerates of amorphous substances similar to gel precipitation. K⁺ and Cs⁺ are similar with Na⁺, but due to the structure breaking role of K⁺ and Cs⁺, the formation of structure unites slow down and thus K⁺ and Cs⁺ can not accelerate crystallization in the equal crystallization time (a short synthesis time of 5 h). It could be appar- ently seen from SEM images that plenty of small cylin- der crystal aggregates were formed by K⁺ (Fig. 2e) and several smaller sphere crystals were formed by Cs⁺ (Fig. 2g) on the surface of mordenite membranes. This mechanism of action is also consistent with effect of K⁺ in the synthesis of heulandite-type zeolite^[41].

  In addition, Aiello et al.^[38] showed that in the syn- thesis of ZSM-5 zeolite, K⁺ was more capable of incorporation of aluminum into the zeolite framework. To further investigate the influence of Na⁺-K⁺ on the synthesis of mordenite membranes, the EDX mapping of the optimal membrane (M6) was carried out. The surface mapping by EDX displayed relatively homoge- neously intense red and green colours (Fig. 7b and 7c), indicating that Na and K are relatively uniform distributed on the mordenite membrane layer. This illustrates that Na⁺ and K⁺ are participated in the crystallization process of mordenite membrane. And also, the surface mapping by EDX displayed rather homogeneously intense sapphire blue and peacock blue colours (Fig. 7d and 7e), indicating the membrane has a rather uniform distribution of Al and Si atoms. Moreover, the Si/Al ratios of mordenite membranes prepared with different Na⁺/K⁺ ratios are listed in Table 3. It can be found that with the increase of Na⁺ and K⁺ contents, the Si/Al ratio of the membrane surface increased, demonstrating that the hydrophilicity of the membrane decreased. This is attributed to the fact that K⁺ can promote the incorporation of aluminum into the zeolite framework and too much amount of Na⁺ can increase the hydrophobicity of the membrane. Furthermore, Fig. 8 presents the water contact angles of the membranes prepared with Na⁺ and Na⁺/K⁺ ratio of 2. As shown in Fig. 8, the contact angle of the membrane prepared with Na⁺/K⁺ ratio of 2 (θ = 48°) was smaller than that of the membrane prepared with single Na⁺ (θ =67°). It also demonstrates that the membrane prepared with Na⁺-K⁺ displayed better hy- drophilicity. Various characterization results show that the replacement of an amount of Na⁺ with K⁺ can increase the hydrophilic of membrane surface, and thus greatly improving the permeation flux of mordenite membranes in this study.

  Figure 7

  

  Figure 7.  (a) Surface SEM images of mordenite membrane prepared with Na⁺/K⁺ ratio=2 (M6); EDX mappings of (b) Na, (c) K, (d) Al and (e) Si atoms of membrane M6 surface

  下载: 全尺寸图片 幻灯片

  Figure 8

  

  Figure 8.  Water contact angle of mordenite membranes prepared with (a) Na⁺ and (b) Na⁺/K⁺ ratio=2

  下载: 全尺寸图片 幻灯片

  In summary, alkali-metal cations have a significant influence on the rate of crystallization, the crystal size and morphology, and the Si/Al ratio of the mordenite membrane. The addition of other alkali-metal cations into the Na⁺-containing synthesis gel can slow down the rate of crystallization. But when K⁺ is added into the Na⁺-containing synthesis gel, the hydrophilicity of membrane surface is improved.

  2.5   Long-term acid stability of mordenite membrane

  Additionally, a long-term acid stability was required for industrial applications of mordenite membranes, the membrane (M6) was evaluated by long-term dehydration of a 90% HAc/H₂O mixture at 90 ℃ for 240 h. As shown in Fig. 9, the permeation flux of membrane M6 slightly decreased from 2.67 to 2.42 kg·m^-2· h^-1 at the first 48 h, and ultimately kept stable at approximately 2.42 kg·m^-2·h^-1. However, the separation factor gradually increased from about 4 000 to 5 000 and then ultimately kept constant at ca. 5 000. The slight decrease of permeation flux and the increased separation factor are due to healing of few intercrystalline pores by some impurities in wateracetic acid mixtures during PV separation process. Furthermore, the membrane surface adsorbed a small amount of acetic acid molecules which blocked some membrane pores, thus reducing the effective pores of the membrane and resulting in a decrease in the permeation flux during the initial period of 48 h. Fig. 10 shows the XRD patterns and surface SEM images of mordenite membrane M6 before and after long-term PV performances test. Even though the long-term PV test time was 240 h, the membrane still kept the typical and high intensity mordenite structural diffraction peaks (Fig. 10a), and the dense mordenite membrane layer was fully covered on the support surface (Fig. 10b). These results were consistent with PV performances, and also consistent with our previous study^[27]. Apparently, the assynthesized mordenite membranes exhibited a good longterm acid stability in this study, which will be promising candidates of industrial applications for dehydration of acetic acid solutions and acidic aqueous mixtures.

  Figure 9

  

  Figure 9.  Long-term PV performances of mordenite membrane M6 for a 90% HAc/H₂O mixture at 90 ℃

  下载: 全尺寸图片 幻灯片

  Figure 10

  

  Figure 10.  XRD patterns (a) and surface SEM images (b) of mordenite membrane M6 before and after long-term PV performance test

  下载: 全尺寸图片 幻灯片

  2.6   Comparison with other reported mordenite membranes

  Table 4 summarizes PV performances of mordenite membranes in literatures and this work for separat- ing HAc/H₂O mixtures. As demonstrated, the as- synthesized membranes in this study further greatly improved the permeation flux compared to other refer- ences. This may be because mordenite membrane M6 has better hydrophilicity and the membrane layers are thinner and denser. The membrane also showed a high permeation flux of 1.89 kg·m^-2·h^-1 even at 75 ℃. To our knowledge, the as-synthesized mordenite mem- brane in this study exhibited a highest permeation flux currently reported in the literature. The membrane with superior PV performance at 90 ℃ shows an excellent promising prospect of industrial applications at high temperatures.

  Table 4

  

  Table 4.  Comparison of PV performances of mordenite membranes

  下载: 导出CSV

  Membrane	Feed (Mass fraction of HAc/%)	Separation temperature/℃	PV performance		Ref.
  			J/(kg·m-2·h-1)	αH2O/HAc
  M6	HAc/H2O(90)	90	2.67	4 000	This work
  M6	HAc/H2O(90)	75	1.89	6 500	This work
  M14	HAc/H2O(90)	75	1.48	> 10 000	[16]
  M3	HAc/H2O(90)	75	1.36	2 300	[27]
  M1004	HAc/H2O(50)	80	0.61	299	[8]
  No name	HAc/H2O(50)	80	2.00	600	[9]
  No name	HAc/H2O(90)	100	0.18	83	[42]

  3.   Conclusions

  The role of alkali-metal cations on the synthesis of mordenite membranes was systematically discussed. It is found that the hydrothermal crystallization process and morphology of mordenite membranes are significantly influenced by alkali - metal cations. Na⁺ has a meaningful promotion effect on the crystallization of mordenite membrane and K⁺ can promote the incorporation of Al into the mordenite framework. An appropriate Na⁺/K⁺ ratio in synthesis gel can improve the hydrophilicity and PV performances of mordenite membranes. The membrane prepared with the Na⁺/K⁺ ratio of 2 in synthesis gel displayed a high permeation flux of 2.67 kg·m^-2·h^-1 and a high separation factor of 4 000 for separating a 90% HAc/H₂O mixture at 90 ℃. And this high-performance mordenite membrane remained stable in the long-term acid stability test for up to 240 h. Therefore, this work provides a feasible method for fabrication of high-flux mordenite membranes.

  
  Acknowledgements: This work was supported by the National Natural Science Foundation of China (Grants No. 21968009, 21766010, 21868012), the Jiangxi Provincial Department of Science and Technology (Grants No. 20171BCB24005, 20181ACH80003, 20192ACB80003, 20192BBH80024), the Science and Technology Project of the Education Department of Jiangxi Province (Grant No. GJJ200321), the Sponsored Program for Cultivating Youths of Outstanding Ability in Jiangxi Normal University and the Gradu- ate Innovation Fund in Jiangxi Normal University (Grant No. YC2020-S155). ^#共同第一作者。
  
• 1. [1]

     Yang J H, Li L Q, Li W Z, Wang J Q, Chen Z, Yin D H, Lu J M, Zhang Y, Guo H C. Chem. Commun. , 2014, 50(93): 14654-14657 doi: 10.1039/C4CC04747H

  2. [2]

     Alomair A A, Al-Jubouri S M, Holmes S M. J. Mater. Chem. A, 2015, 3: 9799-9806 doi: 10.1039/C5TA00124B

  3. [3]

     Bowen T C, Noble R D, Falconer J L. J. Membr. Sci. , 2004, 245: 1-33 doi: 10.1016/j.memsci.2004.06.059

  4. [4]

     Jiang J, Wang L, Peng L, Cai C, Zhang C, Wang X R, Gu X H. J. Membr. Sci. , 2017, 527: 51-59 doi: 10.1016/j.memsci.2017.01.005

  5. [5]

     Wang X R, Chen Y Y, Zhang C, Gu X H, Xu N P. J. Membr. Sci. , 2014, 455: 294-304 doi: 10.1016/j.memsci.2014.01.006

  6. [6]

     Sebastian V, Mallada R, Coronas J, Julbe A, Terpstra R A, Dirrix R W J. J. Membr. Sci. , 2010, 355: 28-35 doi: 10.1016/j.memsci.2010.02.073

  7. [7]

     Morigami Y, Kondo M, Abe J, Kita H, Okamoto K. Sep. Purif. Technol. , 2001, 25: 251-260 doi: 10.1016/S1383-5866(01)00109-5

  8. [8]

     Li G, Kikuchi E, Matsukata M. Sep. Purif. Technol. , 2003, 32: 199-206 doi: 10.1016/S1383-5866(03)00035-2

  9. [9]

     Sato K, Sugimoto K, Kyotani T, Shimotsuma N, Kurata T. J. Membr. Sci. , 2011, 385(1/2): 20-29

  10. [10]

      Navajas A, Mallada R, Téllez C, Coronas J, Menéndez M, Santamaría J. Desalination, 2002, 148: 25-29 doi: 10.1016/S0011-9164(02)00648-3

  11. [11]

      Casado L, Mallada R, Téllez C, Coronas J, Menéndez M, Santamaría J. J. Membr. Sci. , 2003, 216: 135-147 doi: 10.1016/S0376-7388(03)00065-6

  12. [12]

      Zhang Y F, Xu Z Q, Chen Q L. J. Membr. Sci. , 2002, 210: 361-368 doi: 10.1016/S0376-7388(02)00414-3

  13. [13]

      Lee P S, Zhang X Y, Stoeger J A, Malek A, Fan W, Kumar S, Yoo W C, Hashimi S A, Penn R L, Stein A, Tsapatsis M. J. Am. Chem. Soc. , 2011, 133: 493-502 doi: 10.1021/ja107942n

  14. [14]

      Yoo W C, Stoeger J A, Lee P S, Tsapatsis M, Stein A. Angew. Chem. Int. Ed. , 2010, 49: 8699-8703 doi: 10.1002/anie.201004029

  15. [15]

      Choi J, Jeong H K, Snyder M A, Stoeger J A, Masel R I, Tsapatsis M. Science, 2009, 325: 590-593 doi: 10.1126/science.1176095

  16. [16]

      Li L Q, Li J J, Cheng L J, Wang J X, Yang J H. J. Membr. Sci. , 2020, 612: 118479 doi: 10.1016/j.memsci.2020.118479

  17. [17]

      Hu N, Li Y Q, Zhong S L, Wang B, Zhang F, Wu T, Yang Z, Zhou R F, Chen X S. Microporous Mesoporous Mater. , 2016, 228: 22-29 doi: 10.1016/j.micromeso.2016.03.018

  18. [18]

      Wang M Q, Bai L, Li M, Gao L W Y, Wang M X, Rao P H, Zhang Y F. J. Membr. Sci. , 2019, 572: 567-579 doi: 10.1016/j.memsci.2018.11.049

  19. [19]

      Wang L, Yang J H, Wang J Q, Raza W, Liu G R, Lu J M, Zhang Y. Microporous Mesoporous Mater. , 2020, 306: 110360 doi: 10.1016/j.micromeso.2020.110360

  20. [20]

      Chen Z, Li Y H, Yin D H, Song Y M, Ren X X, Lu J M, Yang J H, Wang J Q. J. Membr. Sci. , 2012, 411-412: 182-192 doi: 10.1016/j.memsci.2012.04.030

  21. [21]

      Li L L, Yang J H, Li J J, Han P, Wang J X, Zhao Y, Wang J Q, Lu J M, Yin D H, Zhang Y. J. Membr. Sci. , 2016, 512: 83-92 doi: 10.1016/j.memsci.2016.03.056

  22. [22]

      Zhou R F, Hu Z L, Hu N, Duan L Q, Chen X S, Kita H. Microporous Mesoporous Mater. , 2012, 156: 166-170 doi: 10.1016/j.micromeso.2012.02.023

  23. [23]

      Sasaki H, Oumi Y, Itabashi K, Lu B W, Teranishi T, Sano T. J. Mater. Chem. , 2003, 13: 1173-1179 doi: 10.1039/b212519f

  24. [24]

      Zhu M H, Xia S L, Hua X M, Feng Z J, Hu N, Zhang F, Kumakiri I, Lu Z H, Chen X S, Kita H. Ind. Eng. Chem. Res. , 2014, 53: 19168-19174 doi: 10.1021/ie501248y

  25. [25]

      Zhu M H, Feng Z J, Hua X M, Hu H L, Xia S L, Hu N, Yang Z, Kumakiri I, Chen X S, Kita H. Microporous Mesoporous Mater. , 2016, 233: 171-176 doi: 10.1016/j.micromeso.2016.01.038

  26. [26]

      Zhu M H, Hua X M, Liu Y S, Hu H L, Li Y Q, Hu N, Kumakiri I, Chen X S, Kita H. Ind. Eng. Chem. Res. , 2016, 55: 12268-12275 doi: 10.1021/acs.iecr.6b02125

  27. [27]

      Li Y Q, Zhu M H, Hu N, Zhang F, Wu T, Chen X S, Kita H. J. Membr. Sci. , 2018, 564: 174-183 doi: 10.1016/j.memsci.2018.07.024

  28. [28]

      Breck D W. Zeolite Molecular Sieves: Structure, Chemistry, and Use. New York: John Wiley & Sons Inc. , 1974.

  29. [29]

      Kosanović C, Čižmek A, Subotić B, Šmit I, Stubičar M, Tonejc A. Zeolites, 1995, 15: 632-636 doi: 10.1016/0144-2449(95)00036-6

  30. [30]

      Bhat Y S, Das J, Rao K V, Halgeri A B. J. Catal. , 1996, 159: 368-374 doi: 10.1006/jcat.1996.0099

  31. [31]

      Kang N Y, Song B S, Lee C W, Choi W C, Yoon K B, Park Y K. Microporous Mesoporous Mater. , 2009, 118: 361-372 doi: 10.1016/j.micromeso.2008.09.016

  32. [32]

      Shiralkar V P, Joshi P N, Eapen M J, Rao B S. Zeolites, 1991, 11: 511-516 doi: 10.1016/S0144-2449(05)80127-7

  33. [33]

      Althoff A, Schulz-Dobrick B, Schüth F, Unger K. Microporous Mater. , 1993, 1: 207-218 doi: 10.1016/0927-6513(93)80079-A

  34. [34]

      Reding G, Maurer T, Kraushaar-Czarnetzki B. Microporous Mesoporous Mater. , 2003, 57: 83-92 doi: 10.1016/S1387-1811(02)00558-9

  35. [35]

      Liu C Y, Gu W Y, Kong D J, Guo H C. Microporous Mesoporous Mater. , 2014, 183: 30-36 doi: 10.1016/j.micromeso.2013.08.037

  36. [36]

      Xu C C, Lu X F, Wang Z B. J. Membr. Sci. , 2017, 524: 124-131 doi: 10.1016/j.memsci.2016.11.014

  37. [37]

      Fu D L, Schmidt J E, Pletcher P, Karakilic P, Ye X W, Vis C M, Bruijnincx P C A, Filez M, Mandemaker L D B, Winnubst L, Weckhuysen B M. Angew. Chem. Int. Ed. , 2018, 57: 12458-12462 doi: 10.1002/anie.201806361

  38. [38]

      Aiello R, Crea F, Nigro E, Testa F, Mostowicz R, Fonseca A, Nagy J B. Microporous Mesoporous Mater. , 1999, 28: 241-259 doi: 10.1016/S1387-1811(98)00241-8

  39. [39]

      Goepper M, Li H X, Davis M E. J. Chem. Soc. Chem. Commun. , 1992(22): 1665-1666 doi: 10.1039/c39920001665

  40. [40]

      Burkett S L, Davis M E. Chem. Mater. , 1995, 7: 1453-1463 doi: 10.1021/cm00056a009

  41. [41]

      Zhao D Y, Szostak R, Kevan L. J. Mater. Chem. , 1998, 8: 233-239 doi: 10.1039/a705026g

  42. [42]

      Zhang Y Q, Nakasaka Y, Tago T, Hirata A, Sato Y, Masuda T. Microporous Mesoporous Mater. , 2015, 207: 39-45

• 

  Figure 1  XRD patterns of (a) mullite support and mordenite membranes prepared with different single alkali-metal cations: (b) Li⁺, (c) Na⁺, (d) K⁺, (e) Cs⁺

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Surface and cross-sectional SEM images of mordenite membranes prepared with different single alkali-metal cations: (a, b) Li⁺, (c, d) Na⁺, (e, f) K⁺, (g, h) Cs⁺

  下载: 全尺寸图片 幻灯片
  

  Figure 3  XRD patterns of (a) mullite support and mordenite membranes prepared with different binary alkali-metal cations: (b) Na⁺-Li⁺, (c) Na⁺-K⁺, (d) Na⁺-Cs⁺

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Surface and cross-sectional SEM images of mordenite membranes prepared with different binary alkali-metal cations: (a, b) Na⁺-Li⁺, (c, d) Na⁺-K⁺, (e, f) Na⁺-Cs⁺

  下载: 全尺寸图片 幻灯片
  

  Figure 5  XRD patterns of (a) mullite support and mordenite membranes prepared with different Na⁺/K⁺ ratios: (b) 0.5, (c) 1, (d) 2, (e) 3

  下载: 全尺寸图片 幻灯片
  

  Figure 6  Surface and cross-sectional SEM images of mordenite membranes prepared with different Na⁺/K⁺ ratios: (a, b) 0.5, (c, d) 1, (e, f) 2, (g, h) 3

  下载: 全尺寸图片 幻灯片
  

  Figure 7  (a) Surface SEM images of mordenite membrane prepared with Na⁺/K⁺ ratio=2 (M6); EDX mappings of (b) Na, (c) K, (d) Al and (e) Si atoms of membrane M6 surface

  下载: 全尺寸图片 幻灯片
  

  Figure 8  Water contact angle of mordenite membranes prepared with (a) Na⁺ and (b) Na⁺/K⁺ ratio=2

  下载: 全尺寸图片 幻灯片
  

  Figure 9  Long-term PV performances of mordenite membrane M6 for a 90% HAc/H₂O mixture at 90 ℃

  下载: 全尺寸图片 幻灯片
  

  Figure 10  XRD patterns (a) and surface SEM images (b) of mordenite membrane M6 before and after long-term PV performance test

  下载: 全尺寸图片 幻灯片

  Table 1.  PV performances of mordenite membranes prepared with different alkali-metal cations for a 90% HAc/H₂O mixture at 90 ℃

  Membrane*	Alkali-metal cation	J/(kg·m-2·h-1)	αH2O/HAc
  M1	Li+	—	—
  M2	Na+	1.91	3 500
  M3	K+	—	—
  M4	Cs+	—	—
  M5	Na+-Li+	4.48	70
  M6	Na+-K+	2.67	4 000
  M7	Na+-Cs+	2.49	1 200
  * Membrane gel composition: nSiO2∶nAl2O3∶nM2O∶nMF∶nH2O=1∶0.08∶0.2∶0.1∶35 (M=Li+, Na+, K+, Cs+) for M1~M4; nSiO2∶nAl2O3∶(nNa2O+nM2O)∶(nNaF+nMF)∶nH2O=1∶0.08∶0.2∶0.1∶35 (M=Li, K, Cs; nNa+/nM+=2) for M5~M7; Synthesis temperature: 170 ℃; Synthesis time: 5 h.

  下载: 导出CSV

  Table 2.  PV performances of mordenite membranes prepared with different Na⁺-contents for a 90% HAc/H₂O mixture at 90 ℃

  Membrane*	Na+-content	J/(kg·m-2·h-1)	αH2O/HAc
  M2	0.5	1.91	3 500
  M8	0.3	2.84	200
  M9	0.1	—	—
  * Membrane gel composition: nSiO2∶nAl2O3∶nNa2O∶nNaF∶nH2O =1∶0.08∶x∶0.1∶35 (x=0.2, 0.1, 0); Synthe- sis temperature: 170 ℃; Synthesis time: 5 h.

  下载: 导出CSV

  Table 3.  PV performances of mordenite membranes prepared with different Na⁺/K⁺ ratios in synthesis gel for a 90% HAc/H₂O mixture at 90 ℃

  Membrane*	Na+/K+ ratio	J/(kg·m-2·h-1)	αH2O/HAc	Si/Al ratio in crystal
  M10	0.5	4.05	350	2.6
  M11	1	3.12	620	3.6
  M6	2	2.67	4 000	4.2
  M13	3	2.15	4 200	4.7
  * Membrane gel composition: nSiO2∶nAl2O3∶(nNa2O+nK2O)∶(nNaF+nKF)∶nH2O=1∶0.08∶0.2∶0.1∶35 (Na+/K+ ratio=0.5, 1, 2 and 3); Synthesis temperature: 170 ℃; Synthesis time: 5 h.

  下载: 导出CSV

  Table 4.  Comparison of PV performances of mordenite membranes

  Membrane	Feed (Mass fraction of HAc/%)	Separation temperature/℃	PV performance		Ref.
  			J/(kg·m-2·h-1)	αH2O/HAc
  M6	HAc/H2O(90)	90	2.67	4 000	This work
  M6	HAc/H2O(90)	75	1.89	6 500	This work
  M14	HAc/H2O(90)	75	1.48	> 10 000	[16]
  M3	HAc/H2O(90)	75	1.36	2 300	[27]
  M1004	HAc/H2O(50)	80	0.61	299	[8]
  No name	HAc/H2O(50)	80	2.00	600	[9]
  No name	HAc/H2O(90)	100	0.18	83	[42]

  下载: 导出CSV
•