Fig. 5: Effect of different different-sized seeding methods on (a) permeance and (b) ideal selectivity of tubular supported ZSM-5 zeolite membranes
The effects of different-sized seeding method on the performance of ZSM-5 zeolite membranes were evaluated by N2 and SF6 pure gas permeation. The N2/SF6 permeance and ideal selectivity are widely used as an indicator of membrane quality due to their significant difference in the kinetic diameter (0.36 nm and 0.55 nm for N2 and SF6, respectively) [12]. The permeance and ideal selectivity of the synthesized membranes are shown in Fig. 5. The results in Tables 1-3 and Fig. 5b indicate that the single-layer seeded membranes have low N2/SF6 ideal selectivities due to the poor intergrowth between zeolite crystals which results in creating more defects and non-zeolitic pores. Membranes with dual-layer seeds (DLS series) and triple-layer seeds (TLS series) show significantly higher selectivities confirming the positive effect of using different-sized seeding method in which the small seeds can effectively penetrate from the top seed layer into the gaps of the bottom layer, grow faster regarding both mass and diameter, and fill the defects during the hydrothermal synthesis process [14].
The results in Fig. 5b reveals that adding the second and third seed layers to the top of the primary seed layer significantly affects the membrane permeance. For example, N2/SF6selectivities increase by 2100 % and 230 % by adding the second and third seed layers, respectively, while the average of N2permeances diminishes by only 90% and 65% at the same condition (Fig. 5a). These results confirm that using the different-sized seeding method significantly enhances the membrane performance if both permeance and selectivity took into consideration.
A closer examining the changing trend of permeance versus seed size at the bottom layer (Fig. 5a) discloses that increasing the seed size in the bottom layer from 300 nm to 1.5 μm, if the size of smaller seeds at the top layers are kept constant, significantly increases the permeance by an average of 480%, while the selectivity decreases by only less than 30%. It is due to the fact that using larger seeds at the bottom layer (preferably larger than the pore size of the support), significantly decreases the penetration of seeds into the pores resulting in a reduction in effective membrane thickness and mass transfer resistance. These results also reveal that the seed size at the bottom layer mainly controls the membrane permeance. On the contrary, results in Fig. 5a and b indicate that the N2/SF6 ideal selectivity increases by 230 % and 160 % versus seed size change from 300 nm to 60 nm at the top seed layer for DLS and TLS series, respectively. These results confirm that the seed size at the top layer mainly determines the membrane selectivity.
The effect of using a mid-size intermediate seed layer between the large seed layer at the bottom and the small seed layer at the top can be identified by comparison between performances of M5 and M9 membranes. Where using a 300 nm intermediate seed layer increases selectivity up to 260 %, while decreases the N2 permeance by 40 % showing that a proper intermediate seed layer can improve membrane permeance, if the difference between seed sizes in the bottom and top layers is considerable.
During the different-sized seeding process, small seeds from the top layer penetrate into the inter-crystalline gaps/defects between large seeds in the bottom layer. Afterward, because of their higher growth rate compare to large seeds, small seeds effectively fill the gaps/defects during the hydrothermal synthesis, resulting in higher selectivities for membranes in series TLS and DLS than that of series SLS. Consequently, the reduction in seed size in the top layer, reduces the inter-crystalline gaps and makes the zeolite layer denser on the one hand, and increase the membrane thickness via increasing the penetration of seeds into the support pores, on the other hand. The former enhances selectivity and the latter decreases permeance. So, we need to find an optimum combination of seed sizes in the bottom and top layers to maximize the perm-selectivity of the membranes.
Fig. 6 plots N2/SF6 ideal selectivity versus N2 permeance for series SLS, DLS, and TLS membranes. It can be observed in Fig. 6 that series TLS membranes locate in high selectivity/low permeance region, while series DLS membranes locate in moderate selectivity/permeance region. Considering the results in Fig. 6, we selected the M9 membrane (from series TLS) for further modification using the variable-temperature/time method. The advantages of the new different-sized seeding method can be summarized in Table 4.