Synthesis and characterization of a novel type of mixed matrix membrane: surface sieving membrane

Abstract

A novel surface sieving membrane was prepared by attaching zeolite crystals on the polymer membrane surface. The top zeolite layer serves as a pre-screen layer which only allows the permeation of small molecules based on a molecular sieving mechanism. The separation factor of the new membrane was substantially increased by 300% for the pervaporation separation of a methanol–dimethyl carbonate mixture.

---

The application of membranes for gas separation, filtration, and pervaporation, is growing, since membrane separation offers many advantages, like energy savings, low capital investment, and ease of operation in comparison to other traditional separations.^1–4 There is always great interest in membranes with both high selectivity and permeability, which can lead to lower membrane cost and more energy savings. However, the trade-off between permeability and selectivity limits the application of polymer membranes.^5–7 On the other hand, inorganic membranes, such as zeolite membrane and Pd membrane, have great separation performance. But high manufacturing cost and poor reproducibility hindered their commercial application.^8,9 Mixed-matrix membranes (MMMs), first reported by UOP,^10,11 may surpass the trade-off limit of permeability verse selectivity, as defined in a Robeson plot,⁵ by combining the easy processability and low cost of polymer materials with the excellent gas separation properties of inorganic molecular sieves. Some of the common additives in mixed-matrix membranes include nanoporous molecular sieving materials such as carbon molecular sieves (CMS), zeolites and metal–organic frameworks.^12–14 The addition of these nanoparticles might improve the transport of target gases and enhance the permeability and/or selectivity of a polymer membrane. Although great progress has been made, the synthesis of defect-free mixed matrix membranes is still difficult and the improvement of separation performance is not significant. In this communication, a novel membrane, surface sieving membrane, was prepared and pervaporation separation of methanol–dimethyl carbonate was used to characterize the membrane.

Fig. 1 illustrated the configuration of surface sieving membrane, as well as pure polymer membrane and mixed matrix membrane. By definition, mixed matrix membrane consists of a continuous polymer phase containing dispersed particles, as shown in Fig. 1. Surface sieving membrane is composed of a continuous polymer phase and non-continuous particles covering the membrane surface. The particles used in surface sieving membranes, usually nanoporous materials, like zeolites, MOF, ZIF, CMS etc, should have the ability to pre-screen or separate molecules based the difference of diffusion and/or adsorption. In this surface sieving membrane configuration, the zeolite crystals on the polymer membrane surface serves as a pre-screen layer which only allow the diffusion of molecules smaller than the pore diameter of zeolite crystals. Of course, it is almost impossible to cover the surface of a polymer membrane completely. So two diffusion paths co-exist: diffusion through polymer phase directly (same selectivity as the pure polymer membrane) and highly selective diffusion through zeolite crystal first then polymer phase (shown in Fig. 2). Therefore, the overall separation factor is the weighted average of path A and path B. If the zeolite coverage is low and/or the diffusion resistance through zeolite crystals is high, then the surface sieving membrane should behave like pure polymer membrane. On the contrary, it is more like zeolite membrane. The key is to increase the zeolite coverage and reduce the diffusion resistance in zeolite crystals.

Fig. 1 Illustration of surface sieving membrane.

Fig. 2 Diffusion across the surface sieving membrane.

In this communication, we covered the surface of polyvinyl alcohol (PVA) membrane with SAPO-34 zeolite crystals to make surface sieving membrane. SAPO-34 crystals were dispersed in ethanol first and then coated on a glass substrate by dip coating method. PVA membrane was hand cast on the glass substrate coated with SAPO-34 crystals. After drying the membrane, the composite membrane was peeled off from the glass substrate and ready for use (detailed procedure see ESI Fig. S1†).

DMC is mainly used as solvent and methylating agent. The separation of methanol and DMC is an important separation process.^15–19 SAPO-34 zeolite is a small pore silicoaluminophosphate with CHA structure and its pore size is 0.38 nm. The kinetic molecular diameters of methanol and DMC are 0.36 nm and ∼0.6 nm, respectively. Therefore, the top layer SAPO-34 crystals only allow the diffusion of methanol molecules and block the permeation of DMC molecules, based on molecular sieving mechanism. Since polymer membrane surface was not covered completely by zeolite crystals, both molecules may permeate across the polymer phase directly, as shown in Fig. 2. Therefore, the overall separation factor is the weighted average of path A (extremely selective) and path B (polymer selective).

Fig. 3 showed the SEM image of SAPO-34 crystals used in this study. The SAPO-34 crystals exhibited rectangular plate morphology with dimensions of 300 × 300 × 100 nm. These thin plate crystals were chosen because the crystals may lie down on the glass substrate easily and minimize the diffusion resistance in the crystal. The thickness of PVA membranes used was ∼30 μm, significantly thicker than SAPO-34 crystals, which indicated the diffusion resistance in zeolite crystal might be very small. The XRD patterns and SEM images of glass substrates dip coated in zeolite suspensions with different concentrations can be found in the ESI.†

Fig. 3 SEM image of SAPO-34 crystals.

Fig. 4 showed the top and cross sectional view SEM images of surface sieving membranes with different SAPO-34 loadings. The pure PVA membrane exhibited smooth surface and cross section, as shown in Fig. 4a. From the top view SEM images, the rectangular SAPO-34 crystals can be seen easily on the membrane surface and the crystal coverage increases as the loading increases. In some area, the SAPO-34 crystals fell off and the rectangular concave dents were observed. From the cross sectional view, multiple layers of SAPO-34 crystals were observed. The adhesion between the polymer phase and zeolite crystals was good since no voids and cracks were present. The interfacial contact is critical since molecules may bypass zeolite crystals through the interfacial defects, which leads to lower selectivity.

Fig. 4 Top and cross sectional view SEM images of surface sieving membranes with different loadings, (a) 0.0 wt%, (b) 0.5 wt%, (c) 0.8 wt% and (d) 1.0 wt%.

Table 1 listed the performance of surface sieving membranes with different zeolite loading. Pure PVA membrane had separation factor of 4.2 and flux of 0.19 kg (m² h)⁻¹, which is consistent with literature data.¹⁵ As the zeolite coverage on the membrane surface (or loading) increased, the separation factor increased and peaked at 0.5 wt% loading. At higher loading, separation factor remained flat, even with double dip coating in 1 wt% suspensions. The highest separation factor obtained was 12.6, which is a 300% increase over pure PVA membrane. This is the result of more permeation through zeolite crystals. The glass transition temperature (T_g) of the membranes increased with the loading of zeolite crystals, from 353.9 K (pure PVA) to 360.7 K (1.0 wt% loading). This might be the result of strong interation between zeolite crystals and polymer chains.²⁰

Table 1 The performance of surface sieving membranes for pervaporation separation of methanol–DMC mixtures^a

SAPO-34 suspensions conc (wt%)	Flux kg (m^2 h)^−1	Separation factor	Tg (K)
a (feed: 90CH3OH:10DMC wt%, T = 333 K, downstream pressure ∼200 Pa).
0	0.19	4.2 ± 0.3	353.9
0.01	0.16	4.0	NA
0.1	0.17	4.4 ± 0.3	355.6
0.3	0.21	6.0 ± 0.4	356.4
0.5	0.33	12.6 ± 1.3	357.9
0.8	0.36	12.6 ± 2.2	358.9
1.0	0.36	11.7 ± 0.5	360.7
1.0	0.40	12.2	NA

---

Patricia et al., compared the permeance/selectivity verse flux/separation factor and preferred using permeance/selectivity since the effect of the driving force and operating conditions was eliminated.²¹ So, the comparison of permeance/selectivity and flux/separation factor was shown in Table 2. The selectivities of these membranes are much lower than the corresponding separation factors. The selectivity of pure PVA membrane is only 1.8, which indicated that PVA shows weak affinity for methanol. As the loading/coverage of zeolite crystals increases, the selectvities increase almost the same way as the separation factor. The methanol permeance is in the range of 10⁻⁸ mol (m² Pa s)⁻¹.

Table 2 Comparison of permeance/selectivity and flux/separation factor^a

SAPO-34 conc (wt%)	Flux kg (m^2 h)^−1	SF	Permeance 10^−8 mol (m^2 Pa s)^−1	Selectivity
a (feed: 90CH3OH:10DMCwt%, T = 333 K, downstream pressure ∼200 Pa, SF: separation factor).
0	0.19	4.2 ± 0.3	2.1	1.8
0.01	0.16	4.0	1.7	1.7
0.1	0.17	4.4 ± 0.3	1.8	1.9
0.3	0.21	6.0 ± 0.4	2.3	2.5
0.5	0.33	12.6 ± 1.3	3.6	5.3
0.8	0.36	12.6 ± 2.2	3.9	5.3
1.0	0.36	11.7 ± 0.5	3.9	4.9
1.0	0.40	12.2	4.3	5.2

---

Fig. 5 showed the long term performance of a surface sieving membrane for the pervaporation separation of CH₃OH–DMC mixture. Both flux and separation factor remained stable in an eight-day run.

Fig. 5 The effect of time on the pervaporation separation performance of a surface sieving membrane (feed: 90CH₃OH:10DMCwt%, T = 333 K, downstream pressure ∼200 Pa).

Conclusions

In conclusion, we demonstrated a new membrane configuration, surface sieving membrane, based on a polymer membrane covered with zeolite crystals. Such a membrane configuration showed enhanced separation factor for pervaporation separation of methanol and DMC. The SAPO-34 zeolite crystal layer attached on the surface of polymer membrane served as a pre-screen layer which only allow the permeation of smaller CH₃OH molecules and reject the diffusion of bigger DMC molecules. The highly selective permeation through zeolite crystals resulted into higher overall separation factor on condition that the zeolite coverage is high and diffusion resistance in zeolite crystals is low. Many factors affect the membrane quality, such as zeolite coverage, interfacial contact between polymer phase and zeolite crystals. More work is needed to address these issues.

Acknowledgements

The authors acknowledge the financial supports from National 863 Project (no: 2012AA050104) and Advanced Coal Processing Project of Chinese Academic Sciences (no: XDA07040401).

Notes and references

1. S. A. Stern, J. Membr. Sci., 1994, 94, 1.
2. P. Bernardo, E. Drioli and G. Golemme, Ind. Eng. Chem. Res., 2009, 48, 4638.
3. W. J. Koros and G. K. Fleming, J. Membr. Sci., 1993, 83, 1.
4. W. J. Koros, M. R. Coleman and D. R. B. Walker, Annu. Rev. Mater. Sci., 1992, 22, 47.
5. L. M. Robeson, J. Membr. Sci., 2008, 320(1), 390.
6. W. J. Koros and R. Mahajan, J. Membr. Sci., 2000, 175, 181.
7. D. Q. Vu, W. J. Koros and S. J. Miller, J. Membr. Sci., 2003, 211, 311.
8. J. Gascon, F. Kapteijn, B. Zornoza, V. Sebastián, C. Casado and J. Coronas, Chem. Mater., 2012, 24, 2829.
9. G. Soler-Illia, C. Sanchez, B. Lebeau and J. Patarin, Chem. Rev., 2002, 102, 4093–4138.
10. S. Kulprathipanja, R. W. Neuzil and N. N. Li, Separation of fluids by means of mixed matrix membranes, US Pat. No. 4,740,219, 1988.
11. S. Kulprathipanja, R. W. Neuzil and N. N. Li, Separation of gases by means of mixed matrix membranes, US Pat. No. 5,127,925, 1992.
12. T. S. Chung, L. Y. Jiang, Y. Li and S. Kulprathipanja, Prog. Polym. Sci., 2007, 32, 483.
13. H. Cong, M. Radosz, B. Towler and Y. Shen, Sep. Purif. Technol., 2007, 55, 281.
14. P. S. Goh, A. F. Ismail, S. M. Sanip, B. C. Ng and M. Aziz, Sep. Purif. Technol., 2011, 81, 243.
15. L. Wang, J. Li, Y. Lin and C. Chen, Chem. Eng. J., 2009, 146, 71.
16. L. Wang, X. Han, J. Li, X. Zhan and J. Chen, Chem. Eng. J., 2011, 171, 1035.
17. J. Chen, Q. Liu, J. Fang, A. Zhu and Q. Zhang, J. Colloid Interface Sci., 2007, 316, 580.
18. J. Chen, Q. Liu, A. Zhu, Q. Zhang and J. Fang, J. Membr. Sci., 2008, 315, 74.
19. L. Wang, J. Li, Y. Lin and C. Chen, J. Membr. Sci., 2007, 305, 238.
20. Y. Yua, C. Lina, J. Yeha and W. Lin, Polymer, 2003, 44, 3553.
21. P. Luis, J. Degreve and B. Bruggen, J. Membr. Sci., 2013, 429, 1.

---

Footnote

† Electronic supplementary information (ESI) available: Detailed information about the synthesis and SAPO-34 crystals, membrane casting and pervaporation test conditions. See DOI: 10.1039/c3ra46659k

---