Template Removal and Surface Modification of an SSZ-13 Membrane with Heated Sodium Chloride for CO2/CH4 Gas Separation.

Hydrothermal synthesis with an organic template of N,N,N trimethyl-1-adamantammonium hydroxide (TMAdaOH) is the most commonly used method to prepare an SSZ-13 zeolite membrane. In this paper, the synthesized membrane was treated in heated sodium chloride to remove TMAdaOH instead of calcination in air. The surface of the membrane was modified by the heated NaCl and resulted in an improved CO2/CH4 gas separation selectivity. TMAda+ in the channels of SSZ-13 zeolite decomposed completely, and the treatment time was shortened significantly compared with calcination in air. The recrystallization of zeolite reacting with heated NaCl was the possible reason for the improved gas separation performance of the membrane.

Introduction

Membrane technology has attracted increasing attention for biogas purification as an energy-saving and green method. Several types of zeolite membranes, including MFI,[1] DDR,[2] CHA,[3,4] AEI,[5] and ERI,[6] were reported to separate CO2/CH4 mixtures for molecular-sized pores and high mechanical, thermal, and chemical stabilities.

The CHA topological zeolite SSZ-13 has a three-dimensional structure with 8-membered pores of 0.38 nm and was reported to have high adsorption selectivity of CO2 over N2 and CH4.[7] SSZ-13 zeolite membranes were prepared by in situ crystallization,[8] hydrothermal secondary growth,[9−14] inter-zeolite conversion,[15−17] and so forth, among which hydrothermal secondary growth was the most widely used method for controlling nucleation and crystal growth easily.

Organic template of N,N,N-trimethyl-1-adamantammonium hydroxide (TMAdaOH) was essential for the preparation of an SSZ-13 zeolite membrane by hydrothermal secondary growth. Chen et al.[12,18] investigated the TMAdaOH/SiO2 ratio in gel and optimized the synthesis conditions. TMAdaOH must be removed before gas separation test for blocking the channels of SSZ-13 zeolite. The traditional way to remove TMAdaOH from the membrane was calcination at 550 °C in air for about 10 h,[10,13,14] which resulted in defects for mismatch of the thermal expansion coefficient between the membrane and the support. An improved way to reduce defects was calcination in oxygen[19,20] or pretreatment with ozonication,[21,22] but it needed more investment in equipment. Rapid thermal processing has been reported recently to avoid the formation of defects.[23] The membranes were placed in a quartz tube under vacuum and moved fast into a preheated oven at 1000 °C,[24] but the operation process was complex.

The defect generated in the preparation and calcination of the membrane was typically nonselective for gas separation, and several defect-patching methods were suggested. By deposition of methyldiethoxysilane,[25] tetraethylorthosilicate,[26] siloxane,[27] amorphous silica,[28] dye molecules,[29] or ionic liquids[30] on the surface of the membrane, the gas separation selectivity was improved successfully. However, the deposition layer was not a porous material, which resulted in the decrease in permeance notably, and the defects in the membrane remained. Another strategy is prompting the recrystallization of zeolite on the surface of the membrane to heal the defects. Molten salt method was widely used in the metal-processing field as a surface treatment technology with the advantages of heating evenly and no overheating.[31,32] However, the over-high temperature above the melting point of salt threatens the zeolite framework. In this paper, SSZ-13 zeolite membrane was treated in heated NaCl at temperatures varying from 500 to 550 °C to remove TMAdaOH, and the surface of the membrane was modified, too. The gas separation test showed that the membrane had an improved gas separation performance.

Results and Discussion

Template Removal from the Membrane with Heated NaCl

SSZ-13 zeolite membranes were prepared with the initial gel composition of 1.0 SiO2/0.005 Al2O3/0.10 Na2O/0.10 TMAdaOH/80 H2O and crystallized at 160 °C for 48 h. The membrane after hydrothermal synthesis had no gas separation performance because TMAda+ occupied the channels of zeolite.[24] The traditional way to remove TMAda+ was calcination in air at 550 °C for 10 h with heating and cooling rates of 0.25 °C/min, which takes a long time of about 3 days to avoid cracks.

In this paper, the synthesized SSZ-13 zeolite membrane was calcined in NaCl to remove TMAdaOH with the heating rate speeded up to reduce the time. The calcination time was shortened by using heated NaCl due to the difference in the specific heat capacity (SHC) between salt and air. The SHC of NaCl is 4.03 kJ/(kg °C) and that of air is 1.005 kJ/(kg °C) at 27 °C. Thus, salt could endure the temperature fluctuation better than air even at a high heating rate. As the decomposition temperature of TMAdaOH ranged from 420 to 600 °C, the temperature of NaCl was varied from 500 to 550 °C. At this temperature, NaCl was kept in the solid state. The membrane layer came in contact with air through the space between salt particles.

Figure shows the X-ray diffraction (XRD) patterns of fresh SSZ-13 zeolite membranes treated in air and in NaCl. All membranes correspond to CHA structure reported by zones.[33] The membranes treated with salt (Figure b–d) had typical CHA peaks and higher crystallinity than the one (Figure a) calcined in air. However, the membranes treated with 500 °C/6 h and 550 °C/4 h were pale yellow, showing that the organic template was removed incompletely.

Figure 1

XRD patterns of SSZ-13 zeolite membranes calcined (a) in air at 550 °C/10 h and in heated NaCl at (b) 500 °C/6 h, (c) 550 °C/6 h, and (d) 550 °C/4 h.

To investigate the mechanism of template removal with heated NaCl, uncalcined SSZ-13 zeolites, those calcined in air at 550 °C for 10 h, and those calcined in heated NaCl at 550 °C for 6 h were characterized with XRD, as shown in Figure . All the samples retained the CHA structure. The positions of the strongest peak in XRD patterns changed from 20.87 to 9.62 of 2θ after calcination in air or in salt compared to the fresh membrane, which resulted from the preferred orientation of planes in a lattice change from [201̅] to [100].[33,34] It was noticed that a weak peak appeared at 2θ of the 6.52 position in the XRD pattern of uncalcined SSZ-13 zeolite and those calcined in NaCl but disappeared after calcination in air, which was due to the flexibility of the rhombohedral lattice of SSZ-13 zeolite.

Figure 2

XRD patterns of (a) uncalcined SSZ-13 zeolite, (b) SSZ-13 zeolite calcined in air at 550 °C for 10 h, and (c) SSZ-13 zeolite calcined in heated NaCl at 550 °C for 6 h.

Thermogravimetric analysis (TGA) was applied to investigate the weight loss of fresh SSZ-13 zeolite and the one calcined in sodium chloride at various temperatures. Figure a shows the TGA curve of fresh SSZ-13 zeolite at a heating rate of 5 °C/min without an intermediate dwell time, indicating that the weight loss mainly occurred from 420 to 600 °C. TMAdaOH was removed effectively at 480,[5] 500 °C[12] in air, or 450 °C in pure oxygen,[6] and SSZ-13 zeolite membranes were reported with excellent gas separation performance. Figure b shows that the weight of SSZ-13 zeolite remained stable after calcination in heated NaCl at 550 °C for 6 h, indicating that the organic template in channels decomposed completely.

Figure 3

TGA of (a) uncalcined SSZ-13 zeolite and (b) SSZ-13 zeolite calcined with NaCl at 550 °C for 6 h.

Adsorption isotherms for N2 on SSZ-13 zeolite calcined in air at 550 °C for 10 h and in heated NaCl at 550 °C for 6 h are shown as Figure . The absorption amount for N2 on SSZ-13 zeolite calcined in NaCl was higher than that on zeolite calcined in air. The Brunauer–Emmett–Teller (BET) SSA of SSZ-13 zeolite calcined in air is 492 m2/g, which is higher than the BET SSA of SSZ-13 zeolite calcined in NaCl (420 m2/g). The pore size distributions of SSZ-13 zeolite calcined in air and in heated NaCl are shown in Figure S1. The pore size distributions of SSZ-13 zeolites were close between the two calcinations ways.

Figure 4

Adsorption isotherms for N2 on SSZ-13 zeolite calcined (a) in air at 550 °C for 10 h and (b) in heated NaCl at 550 °C for 6 h.

TMAdaOH reacted with oxygen and decomposed to N2, H2O, and CO2 when SSZ-13 zeolite was calcined in air at 550 °C for 10 h, which escaped from the channels easily. During the calcination with NaCl at 550 °C for 6 h, oxygen entered from the spaces between salt particles and reacted with TMAdaOH, which might decompose to carbon, similar to the process of carbon film preparation.[35−37] The CO2/CH4 ideal selectivity of the membrane treated with NaCl was improved maybe due to the blocking of pinhole defects with the produced carbon. In order to verify the hypothesis and investigate the change in functional groups after heated salt treatment, Fourier transform infrared spectrometry (FTIR) was carried out for SSZ-13 zeolite, as shown in Figure . The IR spectrum peaks corresponding to the functional groups are shown in Table .[38] After calcination in air or in heated salt, the peaks corresponding to functional groups of −CH3, −CH2, and N–H disappeared, indicating that TMAdaOH decomposed completely and no carbon was produced.

Figure 5

FTIR of (a) uncalcined SSZ-13 zeolite, (b) SSZ-13 zeolite calcined in air at 550 °C for 10 h, and (c) SSZ-13 zeolite calcined in heated NaCl at 550 °C for 6 h.

Table 1

FTIR Spectra Peaks Corresponding to Functional Groups

no.	wavenumber/cm–1	the corresponding functional groups
1	3440	N–H stretching vibration
2	2900–2850	–CH2 bending vibration
3	2850–2750	–CH3 stretching vibration
4	1630	N–H bending vibration
5	1450–1350	–CH3 bending vibration
6	1300–1000	C–N vibration
7	1240–1060	O–Si–O or O–Al–O asymmetric bending vibration
8	790	Al–O symmetric vibration
9	650–580	double 6-member vibration
10	420	Si–O or Al–O tetrahedral bending vibration

To verify whether carbon existed in the channels of SSZ-13 zeolite produced from the decomposition of TMAdaOH, Raman spectrum characterization was performed. Figure shows the Raman spectrum of SSZ-13 zeolite calcined in air at 550 °C for 10 h and in heated NaCl at 550 °C for 6 h. The Raman shift corresponds to natural graphite at 1580 cm–1 as G-line; at 1345 cm–1 as D-line; low-order structures at 1180, 1500, and 1620 cm–1; and so forth.[39] However, no typical peaks corresponding to carbon appeared. In addition, energy-dispersive X-ray spectroscopy (EDS) of SSZ-13 zeolite was carried out, and the results showed that almost no carbon existed on the surface of zeolite after the treatment with heated salt. As a result, TMAdaOH decomposed completely and no carbon was produced in the channels of SSZ-13 zeolite after treatment with heated NaCl.

Figure 6

Raman spectrum of SSZ-13 zeolite (a) calcined in air at 550 °C for 10 h and (b) treated with heated NaCl at 550 °C for 6 h.

Surface Modification of the Membrane with Heated NaCl

Figure shows the surface and cross-sectional SEM images of the SSZ-13 zeolite membrane treated with heated NaCl at 550 °C for 6 h. Compared to the membrane calcined in air at 550 °C for 10 h (Figure b,c), the surface of the SSZ-13 membrane treated with heated salt shows a dense and good intergrowth layer (Figure d), and the membrane layer integrated with the support (Figure e). Although some salts infected the surface of the membrane, an improved CO2/CH4 selectivity was obtained. The membrane treated with heated sodium chloride was characterized by EDS, and the sodium chloride had only 3.88 wt %.

Figure 7

SEM images of the (a) surface of mullite support and the surface and cross section of the SSZ-13 zeolite membrane calcined in air (b,c) at 550 °C for 10 h and in heated NaCl (d,e) at 550 °C for 6 h.

To investigate the surface modification with heated salt, SSZ-13 zeolite was used. The conventional morphology of the synthesized SSZ-13 zeolite was rhombohedral and spherical, as shown in Figure a. A crystal with complete crystallization showed rhombohedral morphology, and the spherical morphology came from the stacking of crystals. The rhombohedral and spherical morphologies coexisted when SSZ-13 zeolite was calcined in air, which was reported in our previous work.[10] However, only rhombohedral crystals were observed when SSZ-13 zeolite was treated with heated salt, as shown in Figure b, and the spherical crystals were not observe maybe due to recrystallization.

Figure 8

SEM images of (a) uncalcined SSZ-13 zeolite and (b) SSZ-13 zeolite treated with heated NaCl at 550 °C for 6 h.

No recrystallization behavior of SSZ-13 zeolite crystals was observed when calcined in air, why it occurred in heated salt? Recrystallization of conventional coarse-grained metals was investigated, and Hibbard et al.[40] reported nanocrystalline nickle recrystallized and grown by isothermal annealing. Reber et al.[41] reported coarse-grained silicon films on SiAlON ceramics recrystallized by zone melting. Guzmán-Castillo et al.[42] reported that Y zeolite was depolymerized with glycerol at 200 and 250 °C and partially recrystallized in the presence of either a structure-directing agent (CTAB) or ammonium ions (NH4NO3/NH4OH). However, there is no report on the recrystallization of zeolite with heated salt, but it is the most possible explanation for the improved gas separation performance of the SSZ-13 zeolite membrane. SSZ-13 zeolite was recrystallized in the treatment of heated salt, and the spherical morphology transformed to a rhombohedral morphology after salt treatment, as shown in Figure b, was an obvious proof. This supplied a novel way to control the quality of SSZ-13 zeolite and its membrane with no defect.

Gas Separation Performance of the Membrane Treated in Heated Salt

SSZ-13 membranes were prepared repeatedly with secondary hydrothermal growth, and the high performance of the membrane calcined in air had a CO2 permeance of 3.21 × 10–7 mol/(m2·s·Pa) and CO2/CH4 selectivity of 34. The gas separation performances of SSZ-13 zeolite membranes treated in air and with heated salt are shown as Table ; the membranes were tested at 25 °C and the feed gas pressure was 0.4 MPa. TMAdaOH was removed effectively from the SSZ-13 zeolite membrane with heated salt treatment at 550 °C/6 h, and the membrane had a CO2 permeance of 2.01 × 10–7 mol/(m2·s·Pa) and CO2/CH4 selectivity of 46. TMAdaOH occupying the channels of the SSZ-13 zeolite membrane was removed incompletely at 500 °C/6 h and 550 °C/4 h, which induced lower CO2 permeance but higher CO2/CH4 selectivity. All the membranes treated with heated salt had higher CO2/CH4 selectivities compared to the membranes calcined in air, and the membrane treated at 500 °C/6 h had an improved CO2/CH4 selectivity of 71% than the one calcined in air.

Table 2

Single Gas Separation Performance of SSZ-13 Zeolite Membranes Treated in Air and in Heated NaCl

no.	treatment methods	conditions	CO2 permeance [×10–7 mol/(m2·s·Pa)]	ideal selectivity of CO2/CH4
1	calcination in air	550 °C/10 h	3.21	34
2	in heated NaCl	500 °C/6 h	1.82	58
3	in heated NaCl	550 °C/6 h	2.01	46
4	in heated NaCl	550 °C/4 h	1.80	49

Conclusions

A novel method of heated NaCl treatment was used to remove the organic template of TMAdaOH from the channels of the SSZ-13 zeolite membrane instead of conventional calcination in air. The membrane generated less cracks by heated salt treatment, and the required time for template removal was reduced significantly. Besides, the treatment could be carried out in a normal muffle furnace. TMAda+ in the channels of SSZ-13 zeolite decomposed completely, which reacted with oxygen that entered from the space between salt particles. The SSZ-13 zeolite membrane treated with heated NaCl had an improved gas separation performance, and the possible reason was the recrystallization of zeolite crystals after the heated salt treatment, which changed the morphology of crystals from spherical to rhombohedral. The optimized condition for the heated salt treatment was 550 °C for 6 h, and the membrane had a CO2 permeance of 2.01 × 10–7 mol/(m2·s·Pa) and a CO2/CH4 selectivity of 46.

Experimental Section

Membrane Fabrication and Treatment

SSZ-13 zeolite and membranes were synthesized and introduced in our previous work.[10,12,18] The SSZ-13 zeolite membrane was synthesized on porous mullite tubes (length: 10 cm, outside diameter: 12 mm, inner diameter: 9 mm, thickness: 1.5 mm, average pore size: 1.3 μm, porosity: 30–40%, Nikkato Corp., Tokyo, Japan) by secondary hydrothermal synthesis. The synthesis gel with a molar ratio of 1.0 SiO2/0.005 Al2O3/0.1 Na2O/0.1 TMAdaOH/80 H2O was mixed with NaOH (96%, Junzheng, Wuhai, China), TMAdaOH (25%, Ankai, Tokyo, Japan), Al(OH)3 (99%, Wako, Tokyo, Japan), deionized (DI) water (homemade), and colloidal silica (Ludox TM-40, Sigma-Aldrich, St. Louis, MO, USA). After synthesis, the membranes were washed with running tap water, soaked in DI water until neutral, and dried in a conventional oven. Sodium chloride (AR, 99.5%, Sinopharm, Shanghai, China) was filled in a corundum crucible with two vertically placed SSZ-13 membranes at 550 °C for 6 h without controlling the heating rate for template removal and surface modification, as shown in Figure . There was no pretreatment of sodium chloride. The particle size of NaCl was less than 0.15 mm and the purity was ≥99.5%. As a comparison, the traditional way to remove TMAdaOH was calcined in air at 550 °C for 10 h with heating and cooling rate of 0.25 °C/min. After treatment, the membranes were stored in a conventional oven at 100 °C before characterization and gas separation test.

Figure 9

Schematic diagram for the procedure of template removal and surface modification of the SSZ-13 zeolite membrane by heated NaCl.

Characterization

The thermal weight loss of samples was analyzed using TGA (NETZSCH STA 449F3, USA), and the carrier gas was a mixture of N2 (80%) and O2 (20%), with a heating rate of 5 °C/min, initial temperature of 40 °C, and the finial temperature of 1000 °C. The sample weight was 5 ± 0.5 mg.

SSZ-13 zeolites and membranes were identified by XRD (Rigaku Ultima IV, Japan) using Cu Kα radiation (λ = 0.15406 nm), and the operating conditions were 40 kV and 40 mA in air with 2θ at 5–45° and a step size of 0.02°. Relative crystallinity was estimated by comparing the sum of typical peak intensities (at 2θ = 9.6, 12.9, and 20.6°) of the samples, with the membrane having the highest crystallinity as identified by XRD.

The morphologies of samples were observed with a field emission scanning electron microscope (Hitachi SU8020, Japan) at acceleration voltages from 3 to 10 kV. To improve the conductivity and obtain clear SEM images, the samples of zeolite and membranes were coated with Pt before SEM characterization. EDS (Bruker Q200, USA) was used to identify the elemental composition.

FTIR (Bruker Tensor II, USA) was used in the wavenumber range of 400–4000 cm–1, and the concentration of samples in the KBr pellets was 1%, which was accumulated with 256 scans to obtain the IR spectra.

Raman spectrum measurements were taken with a laser Raman spectrometer (LabRAM HR Jobin Yvon, France), and the excitation wavelength was 632.8 nm. The focal length of the spectrometer was 800 mm, and the scanning frequency shift range was 200–2000 cm–1.

Nitrogen adsorption isotherms were measured at −196 °C using a volumetric adsorption analyzer (Micromeritics ASAP2460, USA), with samples being degassed at 300 °C under vacuum. The BET model surface area (SBET), total pore volume (Vtotal), and micropore volume (Vmicro) were calculated from the nitrogen isotherms.

Single-gas permeation of SSZ-13 membranes was tested with a dead-end and retentate stream blocked system without sweep gas according to our previous study.[12,18] The permeate flow rates of CO2 and CH4 were measured with a soap bubble flowmeter, and the ideal selectivity of the membrane was the ratio of two single-gas permeances with each data point obtained at the stabilization status.

The permeance (P) is calculated by eq where N (mol/s) is the membrane flux of component i, A (m2) is the effective surface area of the SSZ-13 zeolite membrane, and ΔP (Pa) is the pressure drop across the membrane. The ideal selectivity (S) for CO2/CH4 is calculated by eq .