[83–86]. This has been explained by size exclusion, owing to the slight difference
between the effective pore size of Na-A and the critical molecular diameter of C 3
¼
and C 3 . Ca-A was reported to show a higher C 3
¼ uptake (2.5 mmol g
À1 at 323 K and
1.0 bar) but a much lower C 3
¼
/C 3 selectivity (ca. 1) compared to Na-A zeolite
because of its larger pore size [87]. The same conclusion can be drawn from the
large-pore zeolite Na-X with higher C 3
¼ and C 3 uptakes (2.7 and 2.1 mmol g
À1 at
303 K and 1.0 bar, respectively) [79, 88–90].
The C 3
¼ and C 3 uptake and selectivity on commercial zeolites described above
can be further modified via ion exchange. For example, in Na-X, although the uptake
of C 3
¼ and C 3 decreased after Li
+ ion exchange, the selectivity increased since there
is a larger decrease in C 3 uptake [91]. Similarly, Padin et al. reported that exchange
of most of the Na
+ ions in Na-A with smaller Li
+ (5% Na
+ - 95% Li
+
) could optimize
the C 3
¼ uptake (2.3 mmol g
À1 at 393 K and 1.0 bar) and C 3
¼ /C 3 selectivity (>15)
[92]. Such an increase in selectivity has been attributed to an increased C 3
¼ diffusion
rate, whereas C 3 adsorption is still hindered. To confirm this speculation, Hedin et al.
measured the self-diffusion coefficient rates of C 3
¼ and C 3 , using pulsed field
gradient (PFG) NMR spectroscopy, in LTA zeolites with different Si/Al ratios and
extraframework compositions. They found that a series of NaCa-A with different
Ca/Na ratios have higher diffusion rates than cation-free ITQ-29 with Si/Al ¼ 1.
While a slower olefin diffusion rate should be observed for cation-containing
zeolites due to interactions between the target molecule and acidic sites, it has
been postulated that the lower diffusion rate in ITQ-29 is mainly caused by its
smaller pore size. Olson et al. have determined the diffusion rates of C 1 -C 4 hydrocarbons on pure-silica DDR, CHA, and LTA zeolites using PFG NMR spectroscopy
and found that the rate increases with larger 8-ring pore size and with smaller kinetic
diameter of adsorbates [94]. This means that to kinetically separate olefin/paraffin
mixtures, one needs to not only control the pore size but also consider the
diffusion rate.
Another way to tune hydrocarbon separation is to change the zeolite framework
composition. In fact, significant differences in the diffusion rate of C 3
¼ and C 3 were
observed for pure-silica chabazite and AlPO 4 -34 with the same framework topology
(CHA) but different compositions [94, 95]. Pure-silica chabazite showed a high C 3
¼
selectivity due to size exclusion, while AlPO 4 -34 exhibited similar C 3
¼ and C 3
adsorption capacities (2.5 and 2.8 mmol g
À1 , respectively) at 298 K and 1.0 bar.
However, the latter material showed faster C 3
¼ diffusion rates than the former
material because of its larger 8-ring window size [93]. As already noted in C 2
¼ /C 2
separation, in general, adsorbent regeneration is one of the main issues in the
separation of olefin/paraffin with same number of carbon atoms. In general, C 3
¼
uptake gradually decreases in aluminosilicate zeolites since oligomerization takes
place due to the inherent acidity and/or framework polarity. Therefore, to avoid this
difficulty, many studies have investigated C 3
¼ /C 3 separation on pure-silica or AlPO 4
molecular sieves. However, in the case of AlPO 4 materials, a polarity still exists,
owing to the relatively large electronegativity difference between Al (1.61) and P
(2.19). In fact, Padin et al. reported that although AlPO 4 -14 (AFN) has a large
propylene working capacity of ca. 0.6 mmol g
À1 between 0.1 and 1.0 bar at 393 K
18
K. C. Kemp et al.
between the effective pore size of Na-A and the critical molecular diameter of C 3
¼
and C 3 . Ca-A was reported to show a higher C 3
¼ uptake (2.5 mmol g
À1 at 323 K and
1.0 bar) but a much lower C 3
¼
/C 3 selectivity (ca. 1) compared to Na-A zeolite
because of its larger pore size [87]. The same conclusion can be drawn from the
large-pore zeolite Na-X with higher C 3
¼ and C 3 uptakes (2.7 and 2.1 mmol g
À1 at
303 K and 1.0 bar, respectively) [79, 88–90].
The C 3
¼ and C 3 uptake and selectivity on commercial zeolites described above
can be further modified via ion exchange. For example, in Na-X, although the uptake
of C 3
¼ and C 3 decreased after Li
+ ion exchange, the selectivity increased since there
is a larger decrease in C 3 uptake [91]. Similarly, Padin et al. reported that exchange
of most of the Na
+ ions in Na-A with smaller Li
+ (5% Na
+ - 95% Li
+
) could optimize
the C 3
¼ uptake (2.3 mmol g
À1 at 393 K and 1.0 bar) and C 3
¼ /C 3 selectivity (>15)
[92]. Such an increase in selectivity has been attributed to an increased C 3
¼ diffusion
rate, whereas C 3 adsorption is still hindered. To confirm this speculation, Hedin et al.
measured the self-diffusion coefficient rates of C 3
¼ and C 3 , using pulsed field
gradient (PFG) NMR spectroscopy, in LTA zeolites with different Si/Al ratios and
extraframework compositions. They found that a series of NaCa-A with different
Ca/Na ratios have higher diffusion rates than cation-free ITQ-29 with Si/Al ¼ 1.
While a slower olefin diffusion rate should be observed for cation-containing
zeolites due to interactions between the target molecule and acidic sites, it has
been postulated that the lower diffusion rate in ITQ-29 is mainly caused by its
smaller pore size. Olson et al. have determined the diffusion rates of C 1 -C 4 hydrocarbons on pure-silica DDR, CHA, and LTA zeolites using PFG NMR spectroscopy
and found that the rate increases with larger 8-ring pore size and with smaller kinetic
diameter of adsorbates [94]. This means that to kinetically separate olefin/paraffin
mixtures, one needs to not only control the pore size but also consider the
diffusion rate.
Another way to tune hydrocarbon separation is to change the zeolite framework
composition. In fact, significant differences in the diffusion rate of C 3
¼ and C 3 were
observed for pure-silica chabazite and AlPO 4 -34 with the same framework topology
(CHA) but different compositions [94, 95]. Pure-silica chabazite showed a high C 3
¼
selectivity due to size exclusion, while AlPO 4 -34 exhibited similar C 3
¼ and C 3
adsorption capacities (2.5 and 2.8 mmol g
À1 , respectively) at 298 K and 1.0 bar.
However, the latter material showed faster C 3
¼ diffusion rates than the former
material because of its larger 8-ring window size [93]. As already noted in C 2
¼ /C 2
separation, in general, adsorbent regeneration is one of the main issues in the
separation of olefin/paraffin with same number of carbon atoms. In general, C 3
¼
uptake gradually decreases in aluminosilicate zeolites since oligomerization takes
place due to the inherent acidity and/or framework polarity. Therefore, to avoid this
difficulty, many studies have investigated C 3
¼ /C 3 separation on pure-silica or AlPO 4
molecular sieves. However, in the case of AlPO 4 materials, a polarity still exists,
owing to the relatively large electronegativity difference between Al (1.61) and P
(2.19). In fact, Padin et al. reported that although AlPO 4 -14 (AFN) has a large
propylene working capacity of ca. 0.6 mmol g
À1 between 0.1 and 1.0 bar at 393 K
18
K. C. Kemp et al.