Communications
À1
ethanol dehydration at 2008C and 0.64 h WHSV for 10 h is
Similar initial yields to those observed at a lower WHSV of
À1
about 2.5 times larger than that on the latter catalyst (Table 2).
The flushing experiments on the used H-levyne catalyst at the
same temperature show that ethene is the only major species
0.64 h (Figure 4) were obtained. However, unlike H-SSZ-13(I),
H-SSZ-39, and H-RTH(I), which are all microsized, two nanocrys-
talline zeolites H-SSZ-13(II) and H-RTH(II) exhibit no significant
decrease in ethene yield during 50 h on stream. This indicates
that H-RTH(II) is the most active and stable among the cata-
lysts studied here: its ethene yield (70%) becomes significantly
higher than the yields of H-SSZ-13(I) and H-mordenite (48 and
31%, respectively) after 50 h on stream. As shown in Figure 5,
in addition, both H-SSZ-39 and H-RTH(I) become almost com-
pletely deactivated. Therefore, it is clear that crystal size is the
key to governing the durability of cage-based small-pore zeo-
lites in ethanol dehydration. Most likely, nanocrystallinity short-
ens the diffusion length and thus slows the intrazeolitic accu-
mulation of the bulky byproduct (i.e., DEE) that can block the
acid sites. Despite the microcrystallinity, however, H-mordenite
exhibits no significant decrease in ethene yield over the period
of TOS. This can in our view be attributed to its 12-ring chan-
nels where the diffusion of reactant and (by)product molecules
must be faster than that in cage-based small-pore materials,
rendering H-mordenite more resistant to the accumulation of
organics, probably of DEE (Figure S3). It is worth noting that
when the used H-RTH(I) zeolite, which were completely deacti-
vated during the long-term durability test (Figure 5), was re-
generated by calcination in air at 5508C for 8 h and then react-
ed with ethanol under the same conditions as those described
above, it regained the initial activity (Figure S5). This suggests
that the organics deposited are far from hard coke, revealing
the high regenerability of our zeolite catalysts.
detected (Figure S3). However, even after flushing with N2
À1
(
30 mLmin ) for 10 h, the amount of organics in the resulting
catalyst [determined by thermogravimetric and differential
thermal analyses (TG/DTA)] is still high (9.6 wt%). It thus ap-
pears that most, if not all, of residual organics in flushed H-
levyne may be DEE molecules that may not be prone to de-
compose to ethene under flushing conditions. This suggests
that DEE production over this zeolite (Figure 4) is mainly cata-
lyzed by the acid sites located on its crystal surface.
Given that the size of 8-ring windows is larger in H-levyne
than in the other cage-based small-pore materials, the ob-
served decrease in ethene yield of the former catalyst can be
attributed to its lev cages that are considerably small com-
pared with the cages in the other zeolite structure types
(
Table 1). We speculate that this may lead to a faster pore
blockage of acid sites responsible for ethene formation, most
likely by DEE. If such is the case, the cage volume of cage-
based small-pore materials would then be a critical factor af-
fecting the selectivity of the low-temperature dehydration of
ethanol.
We have also examined the initial activities in ethanol dehy-
dration over H-SSZ-13(I), H-SSZ-13(II), H-SSZ-39, H-RTH(I), and
H-RTH(II) under wet conditions at temperatures lower than
2
008C, whose ethene yields are, to a certain degree, lower or
higher than the ethene yield over H-mordenite at 2008C. There
is a thermodynamic limitation not only in the intramolecular
dehydration of ethanol to ethene. Because of its endothermic
nature, in addition, the decomposition of DEE into ethanol and
ethene may not be easier at a lower temperature, as shown in
Figure S4. Thus, it is clear that the optimal temperature for the
efficient ethene formation from ethanol cannot be lower than
On the other hand, it is not difficult to infer that both 8-ring
windows in the RTH framework (Table 1) are too narrow to
[7b]
allow the diffusion of DEE with a kinetic diameter of 5.4 ꢀ,
without causing serious steric hindrance, like the case of H-
levyne. Therefore, the selective ethene formation over an RTH-
type zeolite cannot be rationalized in a way similar to that in
H-mordenite. Transition-state shape selectivity is the only credi-
ble hypothesis to explain the suppression of the formation of
2008C.
Figure 5 shows the long-term performance of five cage-
based small-pore zeolites studied, as well as H-mordenite, at
ethanol dimers within the 8-ring side pockets in this large-pore
À1
[5b,11]
2
008C and H O/EtOH=0.2, but at a higher WHSV (1.92 h ).
zeolite.
Apparently, the cages in H-RTH would be consider-
2
ably larger than even the 12-ring channels in H-mordenite
Table 1), easily allowing the formation of ethanol dimers and
(
thus DEE molecules in this small-pore zeolite. This led us to
conclude that its selective behavior for ethene formation can
be rationalized by product shape selectivity. However, from a
structural point of view, there is no reason only the RTH-type
zeolite should be active and selective for low-temperature eth-
anol dehydration, explaining the high ethene yields observed
for H-SSZ-13(I) and H-SSZ-39 (Figure 4).
In summary, we have demonstrated that nanocrystalline H-
RTH is considerably more active and selective than any of the
already known zeolitic catalysts for the low-temperature
(
2008C) dehydration of ethanol to ethene in the presence of
water vapor (H O/EtOH=0.2). The overall results of our work
2
suggest that its superior performance originates from a combi-
nation of the following factors: 1) cage volume and acid
strength, which should be large and strong, respectively,
enough to effectively catalyze the formation of DEE that subse-
Figure 5. Long-term performance of H-SSZ-13(I)(&), H-SSZ-13(II)(&), H-SSZ-
9(*), H-RTH(I)(~), H-RTH(II)(~), and H-mordenite(^) for ethanol dehydra-
3
tion at 2008C under wet conditions (H
2
O/EtOH=0.2). The feed contains
À1
17.7 kPa ethanol and 3.6 kPa water vapor with N
2
at 1.92 h WHSV (EtOH).
&
ChemSusChem 2018, 11, 1 – 6
4
ꢁ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!