Healing of Microdefects in SSZ-13 Membranes via Filling with Dye

3 hours ago - It is quite challenging to avoid microdefect formation during hydrothermal growths and/or calcination processes, while manufacturing hig...
0 downloads 3 Views 5MB Size
Download PDF
Article Cite This: Chem. Mater. XXXX, XXX, XXX−XXX

pubs.acs.org/cm

Healing of Microdefects in SSZ-13 Membranes via Filling with Dye Molecules and Its Effect on Dry and Wet CO2 Separations Sungwon Hong,† Dongjae Kim,‡ Yanghwan Jeong,† Eunjoo Kim,† Jae Chil Jung,§ Nakwon Choi,∥ Jaewook Nam,⊥ Alex C. K. Yip,*,# and Jungkyu Choi*,† †

Department of Chemical and Biological Engineering, College of Engineering, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, Republic of Korea ‡ School of Chemical Engineering, Sungkyunkwan University, Suwon 16419, Republic of Korea § FINETECH. Co., Ltd., 53-60 Jinwisandan-ro, Jinwi-myeon, Pyeongtaek-si, Gyeonggi-do 17709, Republic of Korea ∥ Center for BioMicrosystems, Brain Science Institute, Korea Institute of Science and Technology (KIST), Hwarang-ro 14-gil, Seongbuk-gu, Seoul 02792, Republic of Korea ⊥ Institute of Chemical Process, School of Chemical and Biological Engineering, Seoul National University, Seoul 08826, Republic of Korea # Chemical and Process Engineering, University of Canterbury, Private Bag 4800, Christchurch 8041, New Zealand S Supporting Information *

ABSTRACT: It is quite challenging to avoid microdefect formation during hydrothermal growths and/or calcination processes, while manufacturing high-quality zeolite membranes in a reproducible manner. Even less than 1% of defects, which generally provide nonselective pathways, will considerably worsen the intrinsic, high molecular sieving-based separation performance of a continuous zeolite membrane. Herein, we propose a simple and reliable method for blocking defects using water-soluble dye molecules, which were originally used for the visualization of nonzeolitic, defective structures in a zeolite membrane. Because the dye molecules are ∼1 nm in size, they cannot diffuse into the zeolitic pores and selectively access the defects. For the demonstration of dye-based defect healing, we chose a siliceous chabazite type SSZ-13 zeolite membrane (pore size = 0.37 × 0.42 nm2) with some degree of defects and investigated the effect of defect healing on the final CO2 separation performance. Because the defects were gradually filled by the dye molecules, both CO2/N2 and CO2/CH4 separation performances were concomitantly increased. Intriguingly, the CO2 permselectivity test with ternary mixtures including H2O vapor (the third largest component in the flue and natural/shale/bio gas streams) in the feed diminished CO2 separation performance. This could be ascribed to inhibited transport of the fast permeating species, here CO2, from the adsorbed H2O molecules on the dye-treated and water-friendly (relatively hydrophilic) membrane surface. On the contrary, the intact, siliceous (water-repelling or hydrophobic) SSZ-13 membranes showed improved CO2 permselectivities in the presence of H2O vapor, seemingly due to defect blocking by the physisorbed H2O molecules.

1. INTRODUCTION Zeolite membranes have been widely studied for the separation of gas mixtures1−4 and liquid mixtures5−10 due to their rigid molecular-sized pore structure and/or preferred adsorption/ diffusion properties.11−14 The molecular sieving ability of the zeolite membrane can be further complemented with a reaction as a means to overcome equilibrium-limited reactions.15,16 Thus, most studies have focused on the fabrication of defectfree zeolite membranes, as it is believed to be essential for realizing intrinsic molecular sieving.17−20 Defects in zeolite films are usually (1) grain boundaries that exist among the polycrystalline membrane constituents or grains and (2) cracks that interconnect grain boundaries and/or cross membrane grains.21,22 Both types of defects are known to occur primarily © XXXX American Chemical Society

during the removal of organic templates at high temperatures (>500 °C).23,24 Such defects, although few in number, hamper the intrinsic molecular sieving of zeolite membranes,25,26 as they often provide nonselective pathways for all permeating species. Although attempts have been undertaken to develop alternative approaches for template removal,19,27 many efforts have been made to repair such defects via using a suitable posttreatment (e.g., chemical vapor deposition (CVD),28−30 chemical liquid deposition (CLD),31,32 silica coating,33−35 Received: February 20, 2018 Revised: April 30, 2018

A

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

nonzeolitic defects. Dye molecules (here, fluorescein sodium salt; ∼1 nm19) are small enough to enter defects (i.e., grainboundary defects and cracks with sizes less than ∼2 nm), but too large to penetrate the zeolite pores. As a proof-of-concept, we applied the dye-based post-treatment to a SSZ-13 (chabazite type, CHA) membrane, resulting in the improvement of CO2/ N 2 and CO2/CH 4 separation performances under dry conditions. Importantly, we could change the effective defect size by simply varying the dye concentration, and thus control the effect of post-treated defects on the permeation rate and separation performance. Along with performance investigation, fluorescent confocal optical microscopy (FCOM) analysis was performed to qualitatively visualize the amount and distribution of dyes in the defects. Furthermore, we processed the FCOM images to obtain quantitative information such as the porosity and tortuosity of the defective structures. Accordingly, the degree of defect reduction could be quantified and correlated with the CO2 molar flux through the defects. Finally, we attempted to comprehend the effect of H2O vapor in the feed stream by changing the degree of defect healing, from intact, defective SSZ-13 membranes to post-treated, less-defective SSZ-13 membranes. The results suggested that the H2O molecules, adsorbed onto the defects, contributed to an improvement in the separation performance over that under dry conditions. Therefore, defective zeolite membranes were beneficial for the effective separation of CO2 molecules in the presence of H2O vapor in the feed.

catalytic cracking deposition (CCD),36,37 coke formation,38,39 and so on). Although the concomitant decrease in molar flux across a post-treated membrane is unavoidable because of the diminished transport through defects, several post-treatment approaches have successfully improved selectivity.17 For instance, silica precursors such as tetramethyl orthosilicate (TMOS; 0.89 nm),40 tetraethyl orthosilicate (TEOS; 0.95 nm),29,40−42 and methyldiethoxysilane (MDES; 0.4 nm)36,43 are often used to form silica on zeolite films in post-treatment processes. Such processes are expected to control the effective size of the zeolite pore and/or pore mouth via the appropriate choice of the precursors while simultaneously blocking the defects. The CVD and CLD approaches have been shown to be effective for improving the separation performance. For example, the CO2/N2 separation factors (SF) were increased from 2.5 to 7.5 for CHA membranes29 and from 1 to 15 for MFI membranes,31 while the H2/CO2 perm-selectivity was increased from 2.6 to 33 for DDR membranes.30 In addition, the selective blocking of defects, and thus the improvement in CO2 perm-selectivity over CH4, has been reported with the use of organic molecules of α-cyclodextrin and β-cyclodextrin in SAPO-34 membranes.44 The sizes of α-cyclodextrin (1.46 nm) and β-cyclodextrin (1.54 nm)45 are larger than the pore size (0.38 nm) of SAPO-34 zeotypes so that they will selectively penetrate the defects. Silica dip-coating was reported for improving the p-/o-xylene separation performance of a MFI membrane via the selective repairing of cracks.33 Coke formation via a methanol-to-olefins reaction on SAPO-34 membranes was reported to selectively reduce defects in the membranes, thus improving the H2O/2-propanol separation performance. 39 As revealed by these studies, selective deposition of fillers in the zeolitic pores and nonzeolitic, defective pores in a zeolite film is highly desirable for achieving high separation performance. A SSZ-13 zeolite (Standard Oil Synthetic Zeolite-13; an aluminosilicate CHA type zeolite with a Si/Al ratio from 2.5 to 175)46 with a pore size of 0.37 × 0.42 nm2 is a good candidate for separating CO2 (0.33 nm) from slightly larger molecules, N2 (0.364 nm) and CH4 (0.38 nm), by its molecular sieving ability. Highly siliceous SSZ-13 zeolite membranes showed high CO2 separation performance,47−50 which is suitable for selective CO2 capture from the flue gas emitted by fossil fuel-based power plants or in the biogas upgrading process. However, despite the continuity under SEM resolution, CHA membranes sometimes do not provide good CO2 perm-selectivity, mainly due to undesired nanosized defects resulting from the calcination process.51 Furthermore, a recent study50 revealed that defects in CHA membranes were beneficial for achieving CO2 perm-selectivity in the presence of H2O vapor in the feed. That is, H2O molecules were adsorbed on the defects existing in the CHA membranes and contributed to retrieval of the intrinsic high CO2 separation performance, albeit with some flux reduction. For improving the CO2 separation performance and realizing industrial applications, it is necessary to understand and control the defects in CHA membranes. In this study, we proposed an effective method to improve the membrane separation performance via a simple but reliable dye-based post-treatment. To the best of our knowledge, this is the first report on the dye-based defect healing for a zeolite membrane. Specifically, we adopted a fluorescent dyeing process, which was originally used for the visualization of defects in a MFI zeolite film,19,22,52 to plug, block, or heal

2. EXPERIMENTAL SECTION 2.1. Synthesis of SSZ-13 Seed Particles. SSZ-13 zeolite particles were synthesized by modifying a literature procedure.47 A synthesis gel was prepared with a molar composition of 20TMAdaOH:100SiO 2 :20NaOH:5Al(OH) 3 :1600H 2 O, with the minor difference being the amount of H2O (vs 4000 H2O in ref 47). The resulting SSZ-13 particles were smaller than those reported in the literature, and their size was estimated to be 250 nm. For gel preparation, specified amounts of N,N,N-trimethyl-1-adamantanammonium hydroxide (TMAdaOH, 25 wt % aqueous solution, SACHEM Inc.), NaOH (98%, Sigma-Aldrich), silica source (LUDOX HS-40; 40 wt % suspension in H2O, Sigma-Aldrich), and Al(OH)3 (reagent grade, Sigma-Aldrich) were added to deionized (DI) water in sequence. The final synthesis gel was homogenized by stirring for 2 d at room temperature, and then transferred to a Teflon-lined autoclave. This was followed by a hydrothermal reaction at 160 °C for 7 d in a forced convection oven, which had been preheated to 160 °C. After completion of the hydrothermal reaction, the solid particles were recovered by repeated centrifuging and washing with fresh DI water. The recovered particles were further calcined at 550 °C at a ramp rate of 1 °C min−1 under an air flow of 200 mL min−1. 2.2. Synthesis of SSZ-13 Seed Layers and Membranes. We used porous α-alumina discs with a thickness of ∼2 mm and diameter of ∼20 mm as supports. The α-alumina discs were lab-prepared by following a method reported elsewhere.53 For the formation of the SSZ-13 seed layers, the calcined SSZ-13 particles were deposited on top of α-alumina discs using dip-coating. For the dip coating, a seed suspension was prepared by adding ∼0.05 g of the SSZ-13 particles to ∼40 mL of ethanol and dispersing them by sonication for ∼20 min (UC-10P, JEIO TECH, South Korea). One side of the α-alumina disc was smoothened by using a polisher (GLP-AP105, GLP Korea, South Korea) with sand paper and then contacted with the seed suspension for 30 s. The disc was withdrawn from the seed suspension and dried for 30 s under ambient conditions. This dip-coating procedure was repeated four times to uniformly cover the disc surface. Finally, SSZ-13 seed layers were calcined at 450 °C at a ramp rate of 1 °C min−1 under an air flow of 100 mL min−1. B

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials Finally, SSZ-13 membranes were synthesized by hydrothermal growth of the prepared SSZ-13 seed layer based on a literature method described elsewhere.54 Specifically, a secondary growth gel was prepared with a molar composition of 15TMAdaOH:5tetraethylammonium hydroxide (TEAOH):100SiO2:20NaOH:1Al(OH)3:8000H2O. The specified amounts of TMAdaOH, TEAOH (Alfa Aesar), NaOH, Al(OH)3, and LUDOX HS-40 were sequentially added to DI water. The gel was further mixed on the shaking machine for 1 d at room temperature. The seeded α-alumina disc was placed with the seeded side facing down, in a tilted position inside a Teflon liner, and the well-mixed secondary growth gel (∼30 g) was added to the Teflon liner (40 mL) in an autoclave. A hydrothermal reaction was conducted under static conditions for 2 d in an oven preheated to 160 °C. After completion of the hydrothermal reaction, the autoclave was quenched with tap water, and then the as-synthesized SSZ-13 membranes were immersed in DI water to eliminate any impurities and dried at 100 °C overnight. The dried SSZ-13 membrane was calcined at 550 °C at a ramp rate of 0.5 °C min−1 under an air flow of 200 mL min−1 to remove the organic templates. 2.3. Dye-Based Post-Treatment on SSZ-13 Zeolite Membranes. Dye solutions with various concentrations (1, 10, and 50 mM) were prepared. For solution preparation, a dye, fluorescein sodium salt (CA index name: spiro[isobenzofuran-1(3H),9′-[9H]xanthen]-3-one, 3′,6′-dihydroxy-, disodium salt, empirical formula C20H10Na2O5, CAS no. 518-47-8, and purchased from Sigma-Aldrich (product no.: F6377)), was dissolved in DI water and stirred for 30 min. Scheme 1 shows the chemical structure of the fluorescein sodium

Association (IZA; http://www.iza-online.org). In addition, FCOM images of the dyed SSZ-13 membranes were recorded along the membrane thickness by using a ZEISS LSM 700 confocal microscope with a solid-state laser (488 nm wavelength). The FCOM images were obtained as described elsewhere,56 with a minor difference, that is, by using an oil immersion objective lens instead of a water immersion objective lens. Variables such as source intensity and gain intensity for data acquisition were kept identical for all of the samples to evaluate the defective structures in the membrane samples in a reliable manner. Separation performance tests for CO2/N2 and CO2/CH4 mixtures were conducted using a homemade permeation system in the WickeKallenbach mode, with the total pressures of both the feed and the permeate sides held at ∼1 atm. Detailed information related to the permeation test can be found elsewhere.29 Under dry conditions, the partial pressures of CO2 and N2 (or CH4) for the CO2/N2 and CO2/ CH4 mixture feeds were 50.5 and 50.5 kPa, respectively. The partial pressures of CO2, N2 (or CH4), and H2O for the CO2/N2 and CO2/ CH4 mixture feeds under wet feed conditions were maintained at 49, 49, and 3 kPa, respectively. To investigate the CO2/N2 separation performance of the membrane samples under different humidity conditions (0%, 26%, 60%, and 100%), the partial pressure of H2O vapor was increased from 0 kPa through ∼3 kPa and ∼7 kPa to ∼12 kPa. To include the H2O vapor in the feed, equimolar CO2/N2 and CO2/CH4 mixtures were allowed to pass through a water-containing gas bubbler at different temperatures (25, 40, and 50 °C were used to generate 3, 7, and 12 kPa of H2O vapor, respectively). The flow rates of the feed binary mixture (dry basis) and the He sweep were maintained at ∼100 mL min−1. For the internal standard required for reliable gas chromatographic analysis, ∼5 mL min−1 of CH4 for CO2/ N2 mixtures and ∼5 mL min−1 of H2 for CO2/CH4 mixtures were added to the permeate stream carried by the He sweep gas toward the gas chromatograph (GC) column. A GC (YL 6100 GC system, YL Instruments, South Korea) installed with a packed column (6 ft × 1/8” Porapak T) and a thermal conductivity detector (TCD) was used for the online analysis of the CO2/N2 permeates, while a GC (YL 6500 GC system, YL Instruments, South Korea) installed with a capillary column (30 m × 0.320 mm GS-GasPro) and a pulsed discharge ionization detector (PDD) was used for the online analysis of the CO2/CH4 permeates.

Scheme 1. Chemical Structure of Fluorescein Sodium Salt Used for Dyeing the Intact SSZ-13 Membranes

salt used as the dye reagent. The compound is an orange-red, odorless powder that is highly soluble in water (henceforth, the fluorescein sodium salt will be referred to as the dye). The conjugated system of xanthene in the chemical form of the dye allows for absorption of light and causes light emission at a specific wavelength.55 The molecular size of the dye is ∼1 nm, which is too large for the zeolite pores, but small enough to selectively block microdefects (< ∼2 nm). Prior to the dyeing process, the calcined SSZ-13 membranes were dried at 100 °C at least overnight. The dried SSZ-13 membrane was placed in the middle of a beaker with the help of a Teflon tube, while its membrane side was faced downward. The dye solution then was poured into the beaker until the membrane sample was fully immersed. Subsequently, the beaker was sealed first with parafilm and fully wrapped with aluminum foil. After dyeing for 1 d, the SSZ-13 membranes were taken out of the beaker and dried under ambient conditions overnight. The SSZ-13 membranes were further dried at 160 °C at least overnight before the permeation test. For convenience, the resulting dyed membrane samples are referred to as M_xmM, where M represents the membrane sample and x indicates the concentration of the dye solution in mM (x = 1, 10, and 50); accordingly, an intact SSZ membrane is denoted as M_0mM. 2.4. Characterizations. Scanning electron microscopy (SEM) images were obtained with a Hitachi S-4300 instrument. The surfaces of the particle and membrane samples were Pt-coated at a configuration of 15 mA for 100 s (Hitachi, E-1030). X-ray diffraction (XRD) patterns were obtained on a Rigaku model D/Max-2500 V/PC diffractometer (Japan) with Cu Kα radiation (λ = 0.154 nm). The simulated XRD pattern of CHA zeolite was duplicated by using Mercury software (available from the Cambridge Crystallographic Data Centre) with a crystallographic information file (CIF) of all-silica CHA zeolite. The CIF file was acquired from the International Zeolite

3. RESULTS AND DISCUSSION 3.1. Dye-Treated SSZ-13 Membranes. SEM and XRD results (Figure S1) show that a uniform SSZ-13 seed layer, which is further intergrown toward a continuous SSZ-13 membrane, was obtained through compact packing of ∼250 nm-sized SSZ-13 particles. Figure 1a shows that the secondary growth of the SSZ-13 seed layer allowed for the continuous SSZ-13 membrane, comparable to previously reported ones.47,48,50 In this study, we manufactured highly siliceous (thus, hydrophobic) SSZ-13 zeolite membranes to minimize the physical adsorption of dye molecules onto the membrane outer surface via their favored electrostatic interactions. In addition, the CO2/CH4 separation performance of a SSZ-13 zeolite membrane was improved monotonically with increasing Si/Al ratio, and the resulting siliceous SSZ-13 zeolite membrane was beneficial for achieving high CO2 separation under wet conditions.48,50 In this dye-based post-treatment, we decoupled the role of the defects from the combined effects of the zeolitic and nonzeolitic (i.e., defective) parts on the final CO2 separation performance. In this aspect, the siliceous SSZ13 membrane can serve as a good model for understanding the effect of defects on the perm-selectivity of a certain target molecule (here, CO2) under both dry and wet feed conditions. Dyeing of the SSZ-13 membranes was performed at varying concentrations of the dye solution; here, the concentrations of 1, 10, and 50 mM were adopted. The color of an intact SSZ-13 membrane was white (inset of Figure 1a). After the dyeing C

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 1. Top view SEM images of M_xmM (x = (a) 0, (b) 1, (c) 10, and (d) 50). The photo of each sample is shown in the upper right inset.

process was completed, the dyed membranes became yellowish, and the yellowish color deepened monotonically with increasing concentration of the dye solution (insets of Figure 1b−d). The surface morphology of the intact SSZ-13 membrane (i.e., M_0mM), equivalent to that of other reported SSZ-13 membranes,47,48,50 was not noticeably changed after post-treatment with dye solutions of concentrations 1, 10, and 50 mM (Figure 1). Considering the molecular size (∼1 nm) of the dye, it was challenging to find any difference after the dyeing post-treatment under SEM resolution. In addition, the XRD pattern of M_0mM (Figure 2) confirms that it comprised pure CHA zeolite. The XRD patterns of the other dyed SSZ-13 membranes (i.e., M_xmM; x

= 1, 10, and 50) were almost identical to that of M_0mM, which confirms their nondistinguishable morphologies as observed under SEM resolution (Figure 1). Although the color difference between the intact and dyed SSZ-13 membranes was visible to the naked eye (insets of Figure 1), neither SEM (Figures 1 and S1a1−d1) nor XRD (Figure 2) results could reveal any considerable difference between the intact and dyed SSZ-13 membranes, indicating that the dye post-treatment induced no structural change in the bulk scale and the dye molecules were highly dispersed on the top of the membrane surface. We further acquired the EDX results via line scanning to compare the Si and Al atom profiles and to evaluate the ratio of Si to Al atoms along the membrane thickness (Figure S2a2− d2). The intact SSZ-13 zeolite membranes were fabricated using a synthetic precursor with a nominal Si/Al ratio of 100; the resulting SSZ-13 membrane was highly siliceous, with the corresponding Si/Al ratio being as high as ∼70 (Table 1). The dyeing process did not change the chemical composition of the intact CHA membrane throughout the sample surface (Table 1) and thickness (Figure S2a2−d2). The reason for the actual Si/Al ratio in the membrane samples being lower than the nominal Si/Al ratio in the synthetic precursor is the undesired Table 1. Si/Al and Na/Al Ratios Measured by EDX Analysis on the Surface of M_xmM (x = 0, 1, 10, and 50)a sample M_0mM M_1mM M_10mM M_50mM

Figure 2. XRD patterns of M_xmM (x = 0, 1, 10, and 50) along with a simulated XRD pattern of all-silica CHA zeolite, which is included for comparison. The asterisks (*) indicate the XRD peaks from the αAl2O3 disc.

Si/Al 68 70 65 71

± ± ± ±

2.5 4.2 3.7 3.5

Na/Al 0.8 0.5 0.6 0.7

± ± ± ±

0.4 0.2 0.3 0.3

a EDX data were obtained at three different regions on the surface of each membrane sample.

D

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 3. (a1−c1) Top view and (a2−c2) cross-sectional view of FCOM images of M_xmM (x = 1, 10, and 50). The yellow lines in (a2)−(c2) indicate the location where the top view FCOM images (a1)−(c1) were obtained. (a3−c3) Magnified cross-sectional FCOM images (2-fold higher magnification with respect to the FCOM images shown in (a2)−(c2)).

molecules, if any, inside the SSZ-13 membrane. Similar to the SEM and XRD results (Figures 1 and 2, respectively), the EDX mapping approach could not distinguish between intact and dyed SSZ-13 membranes. In addition, we investigated the defective structures of SSZ13 membranes by visualizing them using a nondestructive FCOM analysis,19,22,52 because the overall separation performance of a zeolite membrane is a highly sensitive function of a minor nonzeolitic portion.25,35 Figure 3 shows the resulting top view and cross-sectional view FCOM images of M_xmM (x = 1, 10, and 50). The crack densities in all of the dyed membranes (M_xmM; x = 1, 10, and 50) were comparable (Figure 3a1−c1), supporting the reliable synthesis of SSZ-13 membranes. Although the crack densities were similar in the three membranes, both brightness (i.e., fluorescent intensity) and width of cracks were monotonically increased with increasing concentration of the dye solution. Considering that the actual crack sizes in the intact SSZ-13 membranes were technically identical, this phenomenon indicates that the amount of the defect-trapped dye molecules was monotonically increased with increasing concentration of the dye solution. At this point, we would like to mention that, although the dyeing was originally attempted to visualize the defective structure,22 we adopted the same approach for improving membrane performance via the effective blocking of nonzeolitic or extrinsic defects. Although the FCOM images in Figure 3 provided qualitative information relevant to the defective structure, their appropriate image analysis will allow for quantifying the corresponding structural properties (mainly, tortousity and porosity). These properties can be used to estimate the molar flux of a permeating species through the defects in a membrane, and to elucidate the effect of defects on the final separation performance and extract a correlation between them. A detailed discussion will be given later in section 3.3. 3.2. CO2/N2 and CO2/CH4 Separation Performances under Dry and Wet Conditions. Figures 4 and 5 show the CO2/N2 and CO2/CH4 separation performances of all of the membrane samples studied herein (M_xmM; x = 0, 1, 10, and 50) under both dry and wet conditions. Figure 4 shows that the CO2/N2 SF and CO2 permeance of M_1mM under dry conditions (Figure 4a1) were almost identical to those of

incorporation of additional Al atoms (plausibly, leached from the disc support) into the CHA framework of the membrane.57 To reveal any possible change in the hydrophobicity of membrane samples due to the dyeing process, we also measured the contact angle of a water droplet on each membrane sample. Given the inevitable adsorption of water droplets onto both zeolitic and nonzeolitic pores in a membrane, we monitored the water contact angles of all membrane samples (M_xmM; x = 0, 1, 10, and 50) for up to 5 min (Figure S3). Considering that the contact angle of a water droplet in a hydrophobic zeolitic imidazolate framework-8 (ZIF-8) was ∼70°,58,59 the contact angles on M_xmM (x = 0, 1, and 10) higher than 70° (Figure S3) supported the fact that the surfaces of these films were considerably hydrophobic. In contrast, M_50mM, which was supposed to retain more dyes, was less hydrophobic than M_xmM (x = 0, 1, and 10), suggesting that the water-soluble dye molecules accommodated the preferential wetting on membrane surface. In fact, the contact angles on membranes M_xmM (x = 0, 1, and 10) did not change significantly with monitoring time (Figure S3a−c), although it was considerably decreased with measurement time in M_50mM. This gradual, pronounced decrease in contact angle with time in membrane M_50mM (Figure S3d) also indicates continuous wetting associated with the dissolution of dye molecules. We obtained the mappings of C, Si, and Al atoms through EDX measurements on the cross-sectional membrane samples (Figure S2a3−d3, S2a4−d4, and S2a5−d5). In particular, we tried to detect any differences in carbon profiles due to the dye molecules trapped in the defects. However, the carbon profile or distribution obtained from the EDX mapping could not provide any meaningful clue as to the location and existence of the dye molecules for the dyed membranes. This can be ascribed to the difficulty in preserving a carbon-free ambient environment because ubiquitous carbon sources can be easily deposited on all samples. Considering that a SEM instrument equipped with an EDX function is valid for quantifying chemical compounds of content over 0.1 wt %,60 the identification of carbon atoms, which were apparently present in very minor quantities among the majority of the Si and Al atoms in the membrane samples, was quite challenging. Instead, this result indicates the presence of a trace amount of the dye E

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 4. CO2/N2 separation performances of M_xmM (x = (a1−a2) 0, (b1−b2) 1, (c1−c2) 10, and (d1−d2) 50)) as a function of temperature up to 200 °C under dry (left) and wet (right) conditions. In all graphs, the red dashed lines represent the CO2/N2 SF of 10, which was obtained from the product of the diffusion and sorption selectivities.29

CO2/N2 perm-selectivity of ∼10 obtained using highly siliceous CHA type zeolite membranes.29 Despite the monotonic increase in CO2/N2 separation performance with increasing dye concentration up to 50 mM (Figure S4), we found that such a trend did not hold for that in the presence of H2O vapor in the feed. Under wet conditions, the max CO2/N2 SFs of all four samples were in the comparable range of ∼10−12 at 30−50 °C. In particular, the CO2/N2 SFs (the max value of ∼10) of the intact SSZ-13 membrane (i.e., M_0mM) under wet conditions were increased as compared to those under dry conditions (max value of ∼4.6) measured at temperatures up to 200 °C. This trend was in good agreement with the permeation behavior of reported SSZ-13 membranes.50 The dyed membranes (M_xmM; x = 1 and 10) showed a similar tendency (Figure 4b1,b2 and c1,c2). However, the CO2/N2 separation performance of M_50mM was distinctly different between dry and wet conditions (Figure 4d1,d2). First, the CO2/N2 SF at 30 °C was decreased from 12.3 under dry conditions (Figure 4d1) to 9.0 under wet

M_0mM (Figure 4b1), suggesting that the dye concentration of 1 mM was not sufficient to block the defects. For M_10mM, the maximum (max) CO2/N2 SF (5.8) under dry conditions was increased 1.3-fold and the corresponding CO2 permeance was decreased by ∼25% as compared to the corresponding values for M_0mM. With further increase in dye concentration to 50 mM, the max CO2/N2 SF was increased ∼3-fold to become as high as 12.3 and the CO2 permeance at 30 °C was decreased by approximately one-half of those of M_0mM. Such an improvement in CO2/N2 SF resulted from the marked decrease in the permeance of the bulkier N2 molecule, due to the effective reduction of defects, which originally provided a nonselective pathway. Here, we further assumed that M_50mM was a defect-free SSZ-13 zeolite membrane because its max CO2/N2 SF of ∼12 at 30 °C through alumina-supported SSZ13 membranes could be corrected to be ∼15 through pure SSZ13 membranes. This value of ∼15 was comparable to the simulated max CO2/N2 SF of ∼2061 and the expected ideal F

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 5. CO2/CH4 separation performances of M_xmM (x = (a1,a2) 0, (b1,b2) 1, (c1,c2) 10, and (d1,d2) 50) as a function of temperature up to 200 °C under dry (left) and wet (right) conditions. In all graphs, the red dashed lines that represent the CO2/CH4 SF of 50 were included for easy comparison.

top of the zeolitic parts of the relatively less hydrophobic surface of M_50mM (as reflected by the lower water contact angle in Figure S3). Surprisingly, the formation of a water droplet on hydrophobic M_0mM was well preserved despite some degree of defects, while water molecules on the less hydrophobic surface of M_50mM were well spread throughout the surface with time (Figure S3), supporting prevailing wetting of the top surface of M_50mM. In addition to the CO2/N2 separation performance, we measured the CO2/CH4 separation performances of the four samples under both dry and wet conditions (Figure 5). As observed for the CO2/N2 separation performance (Figure 4), the max CO2/CH4 SF under dry conditions was observed at ∼30 °C; its value was increased from 50 for M_0mM through 44 for M_1mM and 75 for M_10mM to 100 for M_50mM (Figure 5a1−d1). Notably, the dyeing post-treatment imparted the max CO2/CH4 SF of ∼100 at 30 °C, which was almost double that (50) for M_0mM. In addition, the CO2/CH4 separation performance under wet conditions (Figure 5a2−d2)

conditions (Figure 4d2), while the corresponding CO2 permeance was decreased approximately by one-half. Second, max CO2/N2 SFs under wet conditions (∼10.4−10.5) were observed at 50−75 °C, not at 30−50 °C, where max SF values were observed for the other three samples (Figure 4a2−c2). The effective repairing of the defects in M_50mM (max SF of 12.3 in Figure 4d1 vs 4.6 of the intact membrane in Figure 4a1) appeared beneficial for achieving the CO2/N2 SF at a higher temperature of 75 °C. The permeation tests of the dyed and intact SSZ-13 membranes reveal that the nonzeolitic defects were beneficial for achieving high molecular sieving for CO2 perm-selectivity in the presence of H2O vapor. For M_0mM, the much-increased CO2/N2 SF (∼10) under wet conditions (against 4.6 under dry conditions) suggests that because defects in a hydrophobic M_0mM (Figure S3) were selectively occupied by water molecules, the resulting membrane could achieve high CO2/N2 separation performance. On the contrary, because the defects in M_50mM were already filled with the dye molecules, water molecules were likely to be adsorbed on G

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 6. (a,b) CO2/N2 and (c,d) CO2/CH4 separation performances of M_xmM (x = 0, 1, 10, and 50) at 50 °C under dry (left) and wet (right) conditions as a function of the concentration of dye solution. The points corresponding to the separation performance of M_0mM were placed at the x-axis position of a virtual concentration of 0.1 mM for demonstration on the logarithmic scale.

Figure 7. (a) Effect of relative humidity (∼26−100%) on the CO2/N2 separation performance of M_0mM at 50 °C and (b) results of long-term stability test of M_0mM with respect to the equimolar CO2 and N2 mixture under wet conditions (∼12 kPa H2O vapor). The used temperature profile is shown in (b) as well. For better readability, the averaged and standard deviation values of CO2 and N2 permeances and the corresponding CO2/N2 SFs, continuously measured at 50 °C for ∼3 d, are added in the first and last sections of the temperature profile.

helping to recover an intrinsic molecular sieving ability through the zeolitic parts. For better comparison, in Figure 6, we summarized the CO2/ N2 and CO2/CH4 separation performances at 50 °C under both dry and wet conditions as a function of the dye solution concentration (0, 1, 10, and 50 mM). Apparently, the dry CO2 perm-selectivity through dyed SSZ-13 membranes was increased monotonically with increasing dye concentration. As discussed above, the intact SSZ-13 membranes showed higher CO2 perm-selectivities over both N2 and CH4 under wet conditions. However, M_50mM, which was assumed to be defect-free due to defect blockage by the dye molecules, provided comparable CO2/N2 and CO2/CH4 SFs under both dry and wet conditions. Interestingly, the permeances of all of the permeating components through M_0mM were decreased to a similar degree with respect to those under dry conditions. Along with this similar permeance reduction pattern in M_50mM, its wet CO2/N2 and CO2/CH4 SFs were close to those in M_0mM, with the only difference being the lowered permeances (Figure 6b and d). These combined features reveal that the adsorbed water molecules spread on the membrane

showed a pattern similar to that found for the CO2/N2 separation performance (Figure 4a2−d2). The only minor difference for M_xmM (x = 0, 1, and 10) was that the max CO2/CH4 SF appeared at 30 °C, while the max CO2/N2 SF was observed at 30−50 °C. Such a trend can be attributed to the stronger adsorption of water molecules at the lower temperature of 30 °C, where the CH4 molecules (0.38 nm), larger than N2 molecules (0.364 nm), were unlikely to penetrate into the SSZ-13 zeolite pores. This suggests that the transport of CH4 molecules was disfavored by the physisorbed water molecules on the membrane surface. Interestingly, we also found that the CO2 separation performance of M_50mM under wet conditions was lower than those of M_xmM (x = 0, 1, and 10). Among them, the intact SSZ-13 membrane (i.e., M_0mM) showed the highest separation performance with the max CO2/CH4 SF as high as ∼195 at 30 °C. This strongly supports that the nonzeolitic defects play a beneficial role in achieving the high CO2 perm-selectivity in the presence of H2O vapor in the feed. This occurred seemingly because the physisorbed water in the zeolitic parts blocked the defects and retarded the molar flux of the slowly permeating species (here, CH4) through the nonselective defects, thus H

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 8. Porosity of nonzeolitic parts (i.e., defects) and zeolitic parts, and their corresponding contribution to the total CO2 molar fluxes for M_xmM (x = 0, 1, 10, and 50). For convenience, M_0mM and M_50mM were assumed to be a 100% defective membrane and a 0% defect-free membrane, respectively. Left and right columns at each concentration of dye solution represent the porosity and CO2 molar flux, respectively. In every column, the parts, designated by no pattern and upward diagonal patterns, indicate the contributions from the zeolite and defect, respectively. For better understanding, estimated defect sizes are illustrated on the top.

of 200 °C for ∼2 d to simulate long-term use. Figure 7b reveals that its separation performance was remarkably well preserved without any recognizable degradation, supporting the superior performance of the intact, siliceous, and hydrophobic SSZ-13 membranes. It should be emphasized that the intact SSZ-13 membrane also showed high CO2/N2 performance with respect to the simulated flue gas conditions (partial pressures of CO2, N2, and H2O vapor being 13.4, 75.6, and 12 kPa, respectively, at 50 °C) (Figure S5). Desirably, the resulting CO2/N2 SF of 13.7 ± 1.8 at 50 °C (Figure S5) was slightly higher than the counterpart of 10.5 ± 0.7 with respect to an equimolar CO2/N2 mixture (Figure 7b). These similar CO2/N2 SFs can be ascribed to the linear adsorption behavior of both CO2 and N2 molecules inside the CHA zeolites.29 The long-term stability results demonstrated that hydrophobic zeolite membranes with some defects were conceptually desirable for securing marked CO2 capture ability under water-containing feed conditions. 3.3. Correlation of the Properties of Defective Structures with the CO 2 Molar Flux Across the Membranes. Figure 3 shows that the original FCOM images were mainly composed of black and white spots. Figure S6 shows the schematic of the defective structure, obtained by processing the FCOM images. Such image processing provided unique quantitative features of the defective pore structures (mainly in terms of tortuosity and porosity), which could not be acquired otherwise. Detailed information relevant to the tortuosity and porosity is summarized in Table S1. Considering the tortuosity along the z-axis (i.e., the direction of membrane thickness), we tried to account for the CO2 molar flux across a SSZ-13 membrane using a Maxwell−Stefan equation.64−66 Accordingly, we could estimate the diffusivity of CO2 that passed through a defect-free SSZ-13 zeolite membrane (here, M_50mM was assumed to be a defect-free SSZ-13 membrane in section 3.2) to be 1.26 × 10−11 m2 s−1. The detailed procedure is described in the Supporting Information. In addition, flux assessment with an assumption that the fluxes through zeolitic and nonzeolitic parts can be linearly combined

surface of M_50mM hindered the transport of all permeating components without improving perm-selectivity. We also attempted to figure out the effect of H2O vapor on the CO2/N2 separation performance of M_0mM at the representative flue gas temperature of 50 °C (Figure 7). For this purpose, the relative humidity (RH) was gradually increased from 26% through 60% to 100% at 50 °C. The intact SSZ-13 membrane (M_0mM) was chosen because it showed the highest CO2 perm-selectivity under wet conditions (Figure 6). Because the flue gas stream usually contained saturated H2O vapor of ∼12 kPa, the variation of its partial pressure from 3 kPa (RH 26%) through 7 kPa (RH 60%) up to 12 kPa (RH 100%) helped to elucidate the permeation behavior of M_0mM. Figure 7a shows that M_0mM attained high CO2/N2 separation performance under wet conditions and well preserved its performance regardless of the relative humidity at 50 °C. The fact that the effect of H2O vapor on the performance M_0mM was rather insensitive supports the effective, reliable use of hydrophobic SSZ-13 membranes for practical carbon capture processes. To the best of our knowledge, we reported, for the first time, such high CO2/N2 separation performance of a zeolite membrane with respect to more realistic flue gas streams. Despite the saturated pressure of ∼12 kPa at the representative flue gas temperature of 50 °C, many studies have reported good CO2 separation performance of zeolite membranes at ∼3 kPa of H2O vapor, apparently due to easier handling of H2O vapor on the lab scale.50,62 In contrast, a hydrophilic NaY (FAU type; Si/Al ratio of ∼1.7− 1.8) zeolite membrane showed very poor separation performances,18 because its powder form zeolite (Si/Al ratio of ∼2.4) had extremely high affinity for H2O adsorption even under very low partial pressure (as low as 0.01 kPa).63 Furthermore, a long-term stability test of the performance using M_0mM was conducted at 50 °C and the saturated H2O vapor of ∼12 kPa for up to 6 d (Figure 7b). This is highly desirable for ensuring its availability for practical purposes. In particular, we had exposed M_0mM to a harsh feed condition I

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 9. (a) CO2/N2 SFs versus CO2 permeances and (b) CO2/CH4 SFs versus CO2 permeances for M_0mM and M_50mM (this work) along with those of other zeolite membranes reported in previous studies. For CO2/N2 mixtures, the performances of SSZ-13,50 CVD-treated CHA,29 NaY,18 and DDR56 membranes were included, while for CO2/CH4 mixtures, the performances of SAPO-34,73 DDR,62 Si-CHA,74 and SSZ-1350,75 membranes were included. The red dashed lines, which represent the CO2/N2 SF of 10 and CO2/CH4 SF of 100, are included as a guide to the eye. For fair comparison, information on the total feed pressure is added next to each membrane in both graphs.

reveals that the size, ddefect, and porosity of the defects were estimated to be 3.1 nm and 0.27%, respectively (Figure 8). Surprisingly, this result shows that the molar flux through the defects (Nd) with porosity of 0.27% accounted for up to 51% of the total molar flux (Nm) (Figure 8). In other words, the defects, although present in a very minor portion (here, ≤1%), could provide facile, nonselective pathways to all permeating components, thereby worsening the membrane separation performance. The same calculation protocol was extended to other membranes (M_xmM; x = 1 and 10), resulting in the estimation of the defect size and CO2 molar flux through the defects. For M_1mM, we obtained results similar to those of M_0mM, as expected from almost the identical separation performance shown in Figure 4a1−b1. This indicates that the defects in M_1mM were hardly blocked by the dye molecules. However, dyeing with the 10 mM dye solution resulted in decreasing both the defect size from 3.1 to 2.5 nm and the porosity from 0.27% to 0.22%. Concomitantly, the portion of Nz in the total flux decreased from 51% to 34% (Figure 8). Quantitative flux assessment suggested that the defects, present in the range of 0.2−0.3%, accounted for 30−50% of the total flux in M_xmM (x = 0, 1, and 10) (Figure 8). This finding was in good agreement with previous studies.25,35 For example, Korelskiy and co-workers reported that 2% of defects in a MFI membrane accounted for ∼30% of the total flux.25 Karimi et al. stated that 10% of the total He flux through a MFI-type zeolite membrane passed through the defects (content of ∼0.2%).35 Therefore, the removal of nonzeolitic defects is essential for realizing the intrinsic properties of zeolite membrane constituents in a membrane form and apparently for ensuring its molecular sieving-based separation performance. Herein, the post-treatment based on the dyeing process provided different degrees of healing or blocking defects and accommodated a systematic variation in the dependence of permeation rates on the defects. This approach helps provide a platform to elucidate the effect of the defects on the perm-selectivity of a certain molecule. 3.4. Comparison of CO2/N2 and CO2/CH4 Separation Performances with Other Literature. As a final attempt to evaluate membrane performance, we compared the CO2/N2 and CO2/CH4 separation performance of M_0mM (with some amount of defects) and M_50mM (plausibly with no defects) with those of other reported membranes. The resulting comparison with respect to the CO2/N2 mixtures (Figure 9a)

indicates that some defects in M_0mM (mainly cracks as shown in Figure 3) were desirable for achieving the marked CO2/N2 SF as high as ∼10 with modest CO2 permeance. Considering the representative temperature of the flue gas (∼50 °C)67,68 from the coal-fired power plant, the CO2/N2 separation performance measured in the similar temperature range (40−50 °C) was collected and illustrated in Figure 9a. In particular, we also compared the separation performances between dry and wet conditions to comprehend the effect of water on the performance. Figure 9a shows that most membranes had higher CO2/N2 SFs under dry conditions compared to wet conditions, except for the hydrophilic NaY membranes.18 The hydrophilic NaY zeolite membranes showed very poor (actually, almost null) separation performances under wet conditions, mainly due to the inhibition of CO2 permeation by strongly and copiously adsorbed water.63,69 On the contrary, siliceous SSZ-13,50 DDR,56 and CVD-treated CHA29 membranes had a different trend in that they showed high CO2/N2 SFs under wet conditions. Among them, the CVD-treated CHA membrane showed a permeating behavior similar to that of M_0mM in this study, describable by the much-pronounced reduction in N2 permeance than CO2 permeance when dry feed was changed to wet feed. This could result mainly from defect blocking through the strong adsorption of water molecules. Accordingly, a molecular sieving effect in favor of CO2 permeance was then realized in the presence of H2O vapor, which played a beneficial role in CO2 perm-selectivity. In contrast, SSZ-13 and DDR membranes showed high CO2/N2 SFs under both dry and wet conditions. Specifically, the CO2/ N2 SFs under wet conditions were slightly higher than those under dry conditions, while the corresponding CO2 permeance was considerably decreased (∼2−3 times). Although a possibility due to the above-mentioned defect healing on the SSZ-13 and DDR membranes could not be excluded completely,50,56 the influence of the defects on the performance was not remarkable as compared to M_0mM. The permeation behaviors of these two membranes were rather close to that of M_50mM, also supporting the less significant effect of the defects. For practical uses, securing a very high molar flux through the fabrication of a very thin SSZ-13 membrane70 on high flux supports71,72 is desirable as a follow-up task. In addition, the CO2/N2 separation performances of other zeolite membranes with respect to dry CO2/N2 feeds were added to J

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

H2O vapor in the feed stream, we conducted a CO2/N2 separation performance test at 50 °C by increasing the partial pressure of H2O vapor up to ∼12 kPa. Because of the hydrophobicity of the SSZ-13 membranes, the CO2/N2 separation performance at 50 °C under relative humidities up to 100% was well maintained for up to 72 h. We found that some degree of defects in SSZ-13 membranes was rather helpful for attaining high CO2 separation performance under wet conditions, while the significant reduction of defects allowed for improving the CO2 perm-selectivity over both N2 or CH4 under dry conditions. In summary, we verified that a simple and reliable defect healing with dye molecules was quite effective for decreasing the effect of undesired defects. It is obvious that the energyefficient technology that allows for reducing nonintrinsic defects in zeolite membranes is critical for accelerating and securing their commercialization. Because this dye-based posttreatment reported in this study can be directly applied to various zeolites and even other inorganic polycrystalline membranes, this approach will be a great contributor to realizing membrane-based separation processes.

those of the membrane samples used for Figure 9a, now Figure S7, where the main temperature range was ∼20−30 °C. Along with the CO2/N2 separation performance, we compared the CO 2 /CH 4 separation performances of M_0mM and M_50mM at 40−50 °C with those of SAPO34,73 DDR,62 SSZ-13,50,75 and Si-CHA74 zeolite membranes (Figure 9b). In particular, the CO2/CH4 separation performances were compared at 40−50 °C, which are the representative temperatures of biogas streams.76,77 Under dry conditions, M_50mM showed CO2/CH4 SF superior to that of the high-quality SSZ-13 membrane.50 Indeed, both eightmembered ring CHA and DDR-type zeolite membranes, which are highly expected to recognize the size difference between the CO2 and CH4 molecules, were desirable for performing the separation of CO2 from the bulkier CH4 than from the slightly smaller N2. The CO2/CH4 separation performance of zeolite membranes has been primarily reported near room temperature. Despite the lab-scale test, investigation of the CO2/CH4 separation performance under more realistic conditions is required. In this aspect, the intact SSZ-13 membranes (i.e., M_0mM) showed very high separation performance especially under wet conditions at ∼50 °C, apparently due to the abovementioned defect blocking by the physisorbed water molecules. The comparison of the separation performances of M_0mM and M_50mM with other reported membranes (Figure 9) provided an insight into the effect of the defects on the final separation performance. First, M_50mM, which can be considered as a defect-free membrane, had excellent CO2 perm-selectivities over both N2 and CH4 under dry conditions, but not under wet conditions. In contrast, M_0mM, albeit containing some degree of defects, showed separation performance superior to the defect-free membranes under wet conditions, due to defect healing or blocking by the physisorbed water molecules.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemmater.8b00757. Detailed information on the quantitative analysis of defective structures in the SSZ-13 membrane; SEM and XRD results of SSZ-13 seed particles and a SSZ-13 seed layer; EDX results and CO2/N2 separation performances of the intact and dyed SSZ-13 membranes; the contact angle of a water droplet on the intact and dyed SSZ-13 membranes; long-term stability test of the intact SSZ-13 membrane with respect to a realistic feed; schematic of the defective structure based on the image analysis of the FCOM images; and comparison of CO2/N2 separation performances of zeolite membranes (PDF)

4. CONCLUSIONS We attempted to reduce the defects in SSZ-13 zeolite membranes via a simple and reliable dye-based post-treatment with fluorescein sodium salt (denoted by “dye molecule” throughout this Article). The effective width of defects, originally presented in an intact SSZ-13, was reduced via post-treatment with increasing dye solution concentrations. Accordingly, the resulting post-treated SSZ-13 membranes showed much improved CO2/N2 and CO2/CH4 separation performances under dry conditions; the respective max CO2/ N2 and CO2/CH4 SFs were as high as ∼12.3 and ∼100, which were 2−3 times higher than those of the intact SSZ-13 membranes. With increasing dye solution concentration, the corresponding CO2 perm-selectivity was monotonically increased, indicating the presence of nonzeolite blocking toward the realization of intrinsic molecular sieving through SSZ-13 zeolites. Indeed, the quantitative analysis of defects presented for the intact SSZ-13 membranes revealed that a small portion of defects (less than 1%) accounted for more than 50% of the final CO2 molar flux. However, under wet conditions, the defect-healed M_50mM did not show any improvement in separation performance. Instead, the intact membrane (i.e., M_0mM), although it contained some degree of defects, showed higher separation performances. This indicates that defective zeolite membranes are suitable for CO2 separation under wet conditions, because the defects were apparently blocked by the condensed and/or physisorbed water molecules. To comprehend the effect of

■

AUTHOR INFORMATION

Corresponding Authors

*Phone: +64-3-369-4086. E-mail: [email protected]. *Phone: +82-2-3290-4854. Fax: +82-2-926-6102. E-mail: [email protected]. ORCID

Nakwon Choi: 0000-0003-2993-9233 Jungkyu Choi: 0000-0003-1137-4799 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Korea CCS R&D Center (KCRC) (2014M1A8A1049309) and by the International Research & Development Program (2016K1A3A1A48954031) through the National Research Foundation (NRF) of Korea. These grants were funded by the Korea government (Ministry of Science and ICT). In addition, this research was supported by a Korea University Future Research Grant. Part of the SEM characterizations were carried out at the Korea Basic Science Institute (KBSI). K

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

■

(21) Dong, J. H.; Lin, Y. S.; Hu, M. Z. C.; Peascoe, R. A.; Payzant, E. A. Template-Removal-Associated Microstructural Development of Porous-Ceramic-Supported MFI Zeolite Membranes. Microporous Mesoporous Mater. 2000, 34, 241−253. (22) Bonilla, G.; Tsapatsis, M.; Vlachos, D. G.; Xomeritakis, G. Fluorescence Confocal Optical Microscopy Imaging of the Grain Boundary Structure of Zeolite MFI Membranes Made by Secondary (Seeded) Growth. J. Membr. Sci. 2001, 182, 103−109. (23) Hedlund, J.; Noack, M.; Kolsch, P.; Creaser, D.; Caro, J.; Sterte, J. ZSM-5 Membranes Synthesized without Organic Templates Using a Seeding Technique. J. Membr. Sci. 1999, 159, 263−273. (24) Mintova, S.; Hedlund, J.; Valtchev, V.; Schoeman, B. J.; Sterte, J. Preparation of ZSM-5 Films from Template Free Precursors. J. Mater. Chem. 1997, 7, 2341−2342. (25) Korelskiy, D.; Ye, P.; Nabavi, M. S.; Hedlund, J. Selective Blocking of Grain Boundary Defects in High-Flux Zeolite Membranes by Coking. J. Mater. Chem. A 2017, 5, 7295−7299. (26) Yu, M.; Falconer, J. L.; Noble, R. D. Characterizing Nonzeolitic Pores in MFI Membranes. Ind. Eng. Chem. Res. 2008, 47, 3943−3948. (27) Heng, S.; Lau, P. P. S.; Yeung, K. L.; Djafer, M.; Schrotter, J. C. Low-Temperature Ozone Treatment for Organic Template Removal from Zeolite Membrane. J. Membr. Sci. 2004, 243, 69−78. (28) Gu, X. H.; Tang, Z.; Dong, J. H. On-Stream Modification of MFI Zeolite Membranes for Enhancing Hydrogen Separation at High Temperature. Microporous Mesoporous Mater. 2008, 111, 441−448. (29) Kim, E.; Lee, T.; Kim, H.; Jung, W. J.; Han, D. Y.; Baik, H.; Choi, N.; Choi, J. Chemical Vapor Deposition on Chabazite (CHA) Zeolite Membranes for Effective Post-Combustion CO2 Capture. Environ. Sci. Technol. 2014, 48, 14828−14836. (30) Zheng, Z.; Hall, A. S.; Guliants, V. V. Synthesis, Characterization and Modification of DDR Membranes Grown on α-Alumina Supports. J. Mater. Sci. 2008, 43, 2499−2502. (31) Zhang, B. Q.; Wang, C.; Lang, L.; Cui, R. L.; Liu, X. F. Selective Defect-Patching of Zeolite Membranes Using Chemical Liquid Deposition at Organic/Aqueous Interfaces. Adv. Funct. Mater. 2008, 18, 3434−3443. (32) Hong, Z.; Zhang, C.; Gu, X. H.; Jin, W. Q.; Xu, N. P. A Simple Method for Healing Nonzeolitic Pores of MFI Membranes by Hydrolysis of Silanes. J. Membr. Sci. 2011, 366, 427−435. (33) Xomeritakis, G.; Lai, Z. P.; Tsapatsis, M. Separation of Xylene Isomer Vapors with Oriented MFI Membranes Made by Seeded Growth. Ind. Eng. Chem. Res. 2001, 40, 544−552. (34) Shin, D. W.; Hyun, S. H.; Cho, C. H.; Han, M. H. Synthesis and CO2/N2 Gas Permeation Characteristics of ZSM-5 Zeolite Membranes. Microporous Mesoporous Mater. 2005, 85, 313−323. (35) Karimi, S.; Korelskiy, D.; Yu, L.; Mouzon, J.; Khodadadi, A. A.; Mortazavi, Y.; Esmaeili, M.; Hedlund, J. A Simple Method for Blocking Defects in Zeolite Membranes. J. Membr. Sci. 2015, 489, 270−274. (36) Hong, Z.; Wu, Z. J.; Zhang, Y. T.; Gu, X. H. Catalytic Cracking Deposition of Methyldiethoxysilane for Modification of Zeolitic Pores in MFI/α-Al2O3 Zeolite Membrane with H+ Ion Exchange Pretreatment. Ind. Eng. Chem. Res. 2013, 52, 13113−13119. (37) Hong, M.; Falconer, J. L.; Noble, R. D. Modification of Zeolite Membranes for H2 Separation by Catalytic Cracking of Methyldiethoxysilane. Ind. Eng. Chem. Res. 2005, 44, 4035−4041. (38) Yan, Y. S.; Davis, M. E.; Gavalas, G. R. Preparation of Highly Selective Zeolite ZSM-5 Membranes by a Post-Synthetic Coking Treatment. J. Membr. Sci. 1997, 123, 95−103. (39) Hirota, Y.; Watanabe, K.; Uchida, Y.; Egashira, Y.; Yoshida, K.; Sasaki, Y.; Nishiyama, N. Coke Deposition in the SAPO-34 Membranes for Examining the Effects of Zeolitic and Non-Zeolitic Pathways on the Permeation and Separation Properties in Gas and Vapor Permeations. J. Membr. Sci. 2012, 415, 176−180. (40) Nomura, M.; Yamaguchi, T.; Nakao, S. Silicalite membranes modified by counterdiffusion CVD technique. Ind. Eng. Chem. Res. 1997, 36, 4217−4223. (41) Kanezashi, M.; Lin, Y. S. Gas Permeation and Diffusion Characteristics of MFI-Type Zeolite Membranes at High Temperatures. J. Phys. Chem. C 2009, 113, 3767−3774.

REFERENCES

(1) Huang, Y.; Wang, L.; Song, Z. N.; Li, S. G.; Yu, M. Growth of High-Quality, Thickness-Reduced Zeolite Membranes towards N2/ CH4 Separation Using High-Aspect-Ratio Seeds. Angew. Chem., Int. Ed. 2015, 54, 10843−10847. (2) Kusakabe, K.; Kuroda, T.; Murata, A.; Morooka, S. Formation of a Y-Type Zeolite Membrane on a Porous α-Alumina Tube for Gas Separation. Ind. Eng. Chem. Res. 1997, 36, 649−655. (3) Carreon, M. A.; Li, S. G.; Falconer, J. L.; Noble, R. D. Aluminasupported SAPO-34 membranes for CO2/CH4 separation. J. Am. Chem. Soc. 2008, 130, 5412−5413. (4) Tomita, T.; Nakayama, K.; Sakai, H. Gas Separation Characteristics of DDR Type Zeolite Membrane. Microporous Mesoporous Mater. 2004, 68, 71−75. (5) Yuan, W. H.; Lin, Y. S.; Yang, W. S. Molecular sieving MFI-type zeolite membranes for pervaporation separation of xylene isomers. J. Am. Chem. Soc. 2004, 126, 4776−4777. (6) Liu, Q.; Noble, R. D.; Falconer, J. L.; Funke, H. H. Organics/ Water Separation by Pervaporation with a Zeolite Membrane. J. Membr. Sci. 1996, 117, 163−174. (7) Wang, Z. B.; Ge, Q. Q.; Shao, J.; Yan, Y. S. High Performance Zeolite LTA Pervaporation Membranes on Ceramic Hollow Fibers by Dipcoating-Wiping Seed Deposition. J. Am. Chem. Soc. 2009, 131, 6910−6911. (8) Wee, S. L.; Tye, C. T.; Bhatia, S. Membrane Separation ProcessPervaporation through Zeolite Membrane. Sep. Purif. Technol. 2008, 63, 500−516. (9) Morigami, Y.; Kondo, M.; Abe, J.; Kita, H.; Okamoto, K. The first Large-Scale Pervaporation Plant Using Tubular-Type Module with Zeolite NaA Membrane. Sep. Purif. Technol. 2001, 25, 251−260. (10) Okamoto, K.; Kita, H.; Horii, K.; Tanaka, K.; Kondo, M. Zeolite NaA Membrane: Preparation, Single-Gas Permeation, and Pervaporation and Vapor Permeation of Water/Organic Liquid Mixtures. Ind. Eng. Chem. Res. 2001, 40, 163−175. (11) Caro, J.; Noack, M.; Kolsch, P. Zeolite Membranes: From the Laboratory Scale to Technical Applications. Adsorption 2005, 11, 215− 227. (12) Lai, Z. P.; Bonilla, G.; Diaz, I.; Nery, J. G.; Sujaoti, K.; Amat, M. A.; Kokkoli, E.; Terasaki, O.; Thompson, R. W.; Tsapatsis, M.; Vlachos, D. G. Microstructural Optimization of a Zeolite Membrane for Organic Vapor Separation. Science 2003, 300, 456−460. (13) Davis, M. E. Ordered Porous Materials for Emerging Applications. Nature 2002, 417, 813−821. (14) Gascon, J.; Kapteijn, F.; Zornoza, B.; Sebastian, V.; Casado, C.; Coronas, J. Practical Approach to Zeolitic Membranes and Coatings: State of the Art, Opportunities, Barriers, and Future Perspectives. Chem. Mater. 2012, 24, 2829−2844. (15) Kim, S. J.; Liu, Y. J.; Moore, J. S.; Dixit, R. S.; Pendergast, J. G.; Sholl, D.; Jones, C. W.; Nair, S. Thin Hydrogen-Selective SAPO-34 Zeolite Membranes for Enhanced Conversion and Selectivity in Propane Dehydrogenation Membrane Reactors. Chem. Mater. 2016, 28, 4397−4402. (16) Zhou, C.; Wang, N. Y.; Qian, Y. N.; Liu, X. X.; Caro, J.; Huang, A. S. Efficient Synthesis of Dimethyl Ether from Methanol in a Bifunctional Zeolite Membrane Reactor. Angew. Chem., Int. Ed. 2016, 55, 12678−12682. (17) Maghsoudi, H. Defects of Zeolite Membranes: Characterization, Modification and Post-treatment Techniques. Sep. Purif. Rev. 2016, 45, 169−192. (18) Gu, X. H.; Dong, J. H.; Nenoff, T. M. Synthesis of Defect-Free FAU-Type Zeolite Membranes and Separation for Dry and Moist CO2/N2 Mixtures. Ind. Eng. Chem. Res. 2005, 44, 937−944. (19) Choi, J.; Jeong, H. K.; Snyder, M. A.; Stoeger, J. A.; Masel, R. I.; Tsapatsis, M. Grain Boundary Defect Elimination in a Zeolite Membrane by Rapid Thermal Processing. Science 2009, 325, 590−593. (20) Zhou, M.; Korelskiy, D.; Ye, P. C.; Grahn, M.; Hedlund, J. A Uniformly Oriented MFI Membrane for Improved CO2 Separation. Angew. Chem., Int. Ed. 2014, 53, 3492−3495. L

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials (42) Wang, H. B.; Lin, Y. S. Synthesis and Modification of ZSM-5/ Silicalite Bilayer Membrane with Improved Hydrogen Separation Performance. J. Membr. Sci. 2012, 396, 128−137. (43) Zhu, X. F.; Wang, H. B.; Lin, Y. S. Effect of the Membrane Quality on Gas Permeation and Chemical Vapor Deposition Modification of MFI-Type Zeolite Membranes. Ind. Eng. Chem. Res. 2010, 49, 10026−10033. (44) Zhang, Y. F.; Avila, A. M.; Tokay, B.; Funke, H. H.; Falconer, J. L.; Noble, R. D. Blocking Defects in SAPO-34 Membranes with Cyclodextrin. J. Membr. Sci. 2010, 358, 7−12. (45) Szejtli, J. Introduction and General Overview of Cyclodextrin Chemistry. Chem. Rev. 1998, 98, 1743−1753. (46) Zones, S. I. Zeolite SSZ-13 and Its Method of Preparation. US4544538A, 1985. (47) Kalipcilar, H.; Bowen, T. C.; Noble, R. D.; Falconer, J. L. Synthesis and Separation Performance of SSZ-13 Zeolite Membranes on Tubular Supports. Chem. Mater. 2002, 14, 3458−3464. (48) Kosinov, N.; Auffret, C.; Borghuis, G. J.; Sripathi, V. G. P.; Hensen, E. J. M. Influence of the Si/Al Ratio on the Separation Properties of SSZ-13 Zeolite Membranes. J. Membr. Sci. 2015, 484, 140−145. (49) Hudson, M. R.; Queen, W. L.; Mason, J. A.; Fickel, D. W.; Lobo, R. F.; Brown, C. M. Unconventional, Highly Selective CO2 Adsorption in Zeolite SSZ-13. J. Am. Chem. Soc. 2012, 134, 1970−1973. (50) Kosinov, N.; Auffret, C.; Gucuyener, C.; Szyja, B. M.; Gascon, J.; Kapteijn, F.; Hensen, E. J. M. High Flux High-Silica SSZ-13 Membrane for CO2 Separation. J. Mater. Chem. A 2014, 2, 13083− 13092. (51) Zhang, Y. F.; Tokay, B.; Funke, H. H.; Falconer, J. L.; Noble, R. D. Template Removal from SAPO-34 Crystals and Membranes. J. Membr. Sci. 2010, 363, 29−35. (52) Snyder, M. A.; Vlachos, D. G.; Nikolakis, V. Quantitative Analysis of Membrane Morphology, Microstructure, and Polycrystallinity via Laser Scanning Confocal Microscopy: Application to NaX Zeolite Membranes. J. Membr. Sci. 2007, 290, 1−18. (53) Choi, J.; Ghosh, S.; King, L.; Tsapatsis, M. MFI Zeolite Membranes from a- and Randomly Oriented Monolayers. Adsorption 2006, 12, 339−360. (54) Zheng, Y. H.; Hu, N.; Wang, H. M.; Bu, N.; Zhang, F.; Zhou, R. F. Preparation of Steam-Stable High-Silica CHA (SSZ-13) Membranes for CO2/CH4 and C2H4/C2H6 Separation. J. Membr. Sci. 2015, 475, 303−310. (55) Jampol, L. M.; Cunha-Vaz, J. Diagnostic Agents in Ophthalmology: Sodium Fluorescein and Other Dyes. In Pharmacology of the Eye; Sears, M. L., Ed.; Springer: Berlin, Heidelberg, 1984; pp 699−714. (56) Kim, E.; Hong, S.; Jang, E.; Lee, J. H.; Kim, J. C.; Choi, N.; Cho, C. H.; Nam, J.; Kwak, S. K.; Yip, A. C. K.; Choi, J. An Oriented, Siliceous Deca-Dodecasil 3R (DDR) Zeolite Film for Effective Carbon Capture: Insight into its Hydrophobic Effect. J. Mater. Chem. A 2017, 5, 11246−11254. (57) Hedlund, J.; Jareman, F.; Bons, A. J.; Anthonis, M. A Masking Technique for High Quality MFI Membranes. J. Membr. Sci. 2003, 222, 163−179. (58) Liu, X. L.; Li, Y. S.; Ban, Y. J.; Peng, Y.; Jin, H.; Bux, H.; Xu, L. Y.; Caro, J.; Yang, W. S. Improvement of Hydrothermal Stability of Zeolitic Imidazolate Frameworks. Chem. Commun. 2013, 49, 9140− 9142. (59) Wu, Y. N.; Zhou, M. M.; Zhang, B. R.; Wu, B. Z.; Li, J.; Qiao, J. L.; Guan, X. H.; Li, F. T. Amino Acid Assisted Templating Synthesis of Hierarchical Zeolitic Imidazolate Framework-8 for Efficient Arsenate Removal. Nanoscale 2014, 6, 1105−1112. (60) Zhang, Y.; Shipway, P. H.; Boue-Bigne, F. Characterisation of Creep-Weak Zones (White Bands) in Grade 91 Weld Metal. Sci. Technol. Weld. Joining 2009, 14, 542−548. (61) Krishna, R.; van Baten, J. M. Segregation Effects in Adsorption of CO2-Containing Mixtures and Their Consequences for Separation Selectivities in Cage-Type Zeolites. Sep. Purif. Technol. 2008, 61, 414− 423.

(62) Himeno, S.; Tomita, T.; Suzuki, K.; Nakayama, K.; Yajima, K.; Yoshida, S. Synthesis and Permeation Properties of a DDR-type Zeolite Membrane for Separation of CO2/CH4 Gaseous Mixtures. Ind. Eng. Chem. Res. 2007, 46, 6989−6997. (63) Boddenberg, B.; Rakhmatkariev, G. U.; Hufnagel, S.; Salimov, Z. A calorimetric and statistical mechanics study of water adsorption in zeolite NaY. Phys. Chem. Chem. Phys. 2002, 4, 4172−4180. (64) Krishna, R.; Wesselingh, J. A. Review article number 50 - The Maxwell-Stefan approach to mass transfer. Chem. Eng. Sci. 1997, 52, 861−911. (65) Krishna, R. Diffusion in porous crystalline materials. Chem. Soc. Rev. 2012, 41, 3099−3118. (66) Kapteijn, F.; Moulijn, J. A.; Krishna, R. The generalized Maxwell-Stefan model for diffusion in zeolites: sorbate molecules with different saturation loadings. Chem. Eng. Sci. 2000, 55, 2923−2930. (67) Merkel, T. C.; Lin, H. Q.; Wei, X. T.; Baker, R. Power Plant Post-Combustion Carbon Dioxide Capture: An Opportunity for Membranes. J. Membr. Sci. 2010, 359, 126−139. (68) D’Alessandro, D. M.; Smit, B.; Long, J. R. Carbon Dioxide Capture: Prospects for New Materials. Angew. Chem., Int. Ed. 2010, 49, 6058−6082. (69) Wang, Y.; Levan, M. D. Adsorption Equilibrium of Binary Mixtures of Carbon Dioxide and Water Vapor on Zeolites 5A and 13X. J. Chem. Eng. Data 2010, 55, 3189−3195. (70) Agrawal, K. V.; Topuz, B.; Pham, T. C. T.; Nguyen, T. H.; Sauer, N.; Rangnekar, N.; Zhang, H.; Narasimharao, K.; Basahel, S. N.; Francis, L. F.; Macosko, C. W.; Al-Thabaiti, S.; Tsapatsis, M.; Yoon, K. B. Oriented MFI Membranes by Gel-Less Secondary Growth of Sub100 nm MFI-Nanosheet Seed Layers. Adv. Mater. 2015, 27, 3243− 3249. (71) Shan, L. J.; Shao, J.; Wang, Z. B.; Yan, Y. S. Preparation of Zeolite MFI Membranes on Alumina Hollow Fibers with High Flux for Pervaporation. J. Membr. Sci. 2011, 378, 319−329. (72) Shu, X. J.; Wang, X. R.; Kong, Q. Q.; Gu, X. H.; Xu, N. P. HighFlux MFI Zeolite Membrane Supported on YSZ Hollow Fiber for Separation of Ethanol/Water. Ind. Eng. Chem. Res. 2012, 51, 12073− 12080. (73) Li, S. G.; Martinek, J. G.; Falconer, J. L.; Noble, R. D.; Gardner, T. Q. High-Pressure CO2/CH4 Separation Using SAPO-34 Membranes. Ind. Eng. Chem. Res. 2005, 44, 3220−3228. (74) Kida, K.; Maeta, Y.; Yogo, K. Preparation and Gas Permeation Properties on Pure Silica CHA-Type Zeolite Membranes. J. Membr. Sci. 2017, 522, 363−370. (75) Wu, T.; Diaz, M. C.; Zheng, Y. H.; Zhou, R. F.; Funke, H. H.; Falconer, J. L.; Noble, R. D. Influence of propane on CO2/CH4 and N2/CH4 separations in CHA zeolite membranes. J. Membr. Sci. 2015, 473, 201−209. (76) Turco, M.; Ausiello, A.; Micoli, L. Treatment of Biogas for Feeding High Temperature Fuel Cells: Removal of Harmful Compounds by Adsorption Processes; Springer: New York, 2016. (77) Favre, E.; Bounaceur, R.; Roizard, D. Biogas, Membranes and Carbon Dioxide Capture. J. Membr. Sci. 2009, 328, 11−14.

M

DOI: 10.1021/acs.chemmater.8b00757 Chem. Mater. XXXX, XXX, XXX−XXX