146
M.-L. Gou et al. / Catalysis Communications 56 (2014) 143–147
2 2 3
within 3 h on stream. All the HZSM-5 zeolites (SiO /Al O = 25–360),
containing both weak and strong acid sites, could almost completely
catalyzed this reaction with initial conversion (TOS = 1.0 h) of over
9
9% and maintained much longer time before deactivation relative to
NaZ-25. So it can be concluded that the isomerization of styrene oxide
is mainly catalyzed by the strong acid sites of HZSM-5 zeolites.
To obtain a quantitative description of catalytic stability, catalyst life-
time was defined as the time at which the initial conversion dropped by
2
% and denoted by the dashed line in Fig. 4. The catalyst lifetimes of
HZSM-5 first decreased and then increased with increasing Si/Al ratio,
e.g., the lifetimes of low Si/Al HZSM-5 decreased in the order HZ-25
(
12.5 h) N HZ-38 (11.0 h) N HZ-50 (5.8 h), whereas the lifetimes of
high Si/Al HZSM-5 increased in the order HZ-135 (7.2 h) b HZ-150
9.1 h) b HZ-360 (15.7 h). It has been reported that stronger acid sites
(
favor the formation of condensation products which are responsible
for the fast deactivation of catalyst [13–15], thus the catalyst with the
strongest acid sites would exhibit the fastest deactivation. As deter-
3
mined by NH -TPD, the strong acid strength on HZSM-5 zeolites mono-
tonically decreased with the increase of Si/Al ratio, so their catalyst
lifetimes should monotonically increase according to the above litera-
tures. However, the catalyst lifetimes of HZSM-5 zeolites presented
here first decreased and then increased with increasing of Si/Al ratio.
Furthermore, the similar BET surface area, micropore volume and size
of HZSM-5 zeolites did not cause different catalytic performances.
Therefore, it is assumed that the catalyst lifetimes are affected by both
acid strength and concentration. With catalysts of similar acid strength,
conversion of styrene oxide begins to drop until the number of acid sites
deactivates to some extent. For example, the strong acid concentrations
of HZ-25, HZ-38 and HZ-50 decreased from 0.606 to 0.416 and
0
4
.336 mmol/g, but their acid strengths showed slight changes (from
38 to 417 °C), thus their lifetimes were mainly influenced by the acid
concentration and decreased with increasing Si/Al ratio. HZ-135,
HZ-150 and HZ-360 have much weaker acid sites (from 397 to
3
42 °C) and the acid strength was a more significant contributor in
this case, which caused the catalyst lifetimes being prolonged with
increasing Si/Al ratio even though their acid concentrations gradually
decreased. Therefore, increasing the acid concentration while decreas-
ing the acid strength can improve the catalyst lifetimes for this reaction.
The main by-products in this reaction were dimer (2,4-diphenyl-2-
butenal) and trimer (2,4,6-tribenzyl-s-trioxane, mp 155–156 °C) of
phenylacetaldehyde. The dimer was formed through aldol condensa-
tion of phenylacetaldehyde, while the trimer was obtained from
trimerization of phenylacetaldehyde by an acid-catalyzed process [30].
Some other by-products, such as phenylethanol, phenylethanediol,
styrene and so on, were also identified by GC–MS with total selectivity
of less than 1%.
Fig. 5 shows the main product distribution with time on stream over
the HZSM-5 samples. The initial selectivities of phenylacetaldehyde
(
9
TOS = 1.0 h) over HZ-25, HZ-38 and HZ-50 were only 56%, 76% and
2% respectively, which are caused by the formation of the trimer. How-
ever, the trimer selectivities over HZ-25, HZ-38 and HZ-50 gradually
decreased from 42%, 22% and 6% to zero after 11, 9 and 4 h on stream,
and correspondingly the phenylacetaldehyde selectivities increased up
to 96%. HZ-135, HZ-150 and HZ-360 had no trimer production and the
phenylacetaldehyde selectivities were all over 96% from beginning to
end of the reaction. As determined by FT-IR collidine adsorption, the ex-
ternal acid sites on HZ-25, HZ-38 and HZ-50 were significantly more in
number than those on HZ-135, HZ-150 and HZ-360 and decreased in
the order HZ-25 N HZ-38 N HZ-50. Combining the above reaction results
with those of the FT-IR collidine adsorption, it is evident that the
trimerization of phenylacetaldehyde probably occurs at the external
acid sites which are readily accessed without spatial constraints. For
the trimer formation, three adsorbed precursor species should be in ad-
jacent acid sites on the external surface. While the external acid sites are
preferentially deactivated than the acid sites in pores [25], thus the tri-
mer decreases with the reaction goes on and only occurs at the early
Fig. 5. Main product distribution with time on stream over the HZSM-5 zeolites. Reaction
conditions: T = 200 °C, P = 1 atm, catalyst loading = 0.5 g, flow-rate of N = 120 ml/min,
2
WHSV = 3.0 h−
1
.
stage of the reaction on HZ-25, HZ-38 and HZ-50. However, the
trimerization could not occur on HZ-135, HZ-150 and HZ-360, since
they contained only a trace amount of external acid sites.
As shown in Fig. 5(b), the dimer selectivities were maintained at the
same level (1–3%) from beginning to end of the reaction on all the
samples. Therefore, the aldol condensation of phenylacetaldehyde is