62
S.A. Regenhardt et al. / Catalysis Communications 35 (2013) 59–63
0.06
acidity increased. The deactivation of metal sites on Pt/Al2O3 and
Pt/SiO2–Al2O3 became more important when W/F0MA =24 g h mol−1
PA
GBL
CH4
.
It is probably that some compounds were slowly formed on the support
acid sites, at the metal-support interface, which then migrated to the
metal surface deactivating Pt sites. Thus, the deactivation of metal Pt
sites on Pt/Al2O3 and Pt/SiO2–Al2O3 became more important as the
space time increase. This last deactivation mechanism is not probably
taking place in the case of Pt/SiO2.
0.04
The only products detected from SA hydrogenolysis were PA, GBL
and CH4, in agreement with reaction scheme shown in Fig. 1. It is
worth to notice that in all of the cases, metal Pt resulted highly selective
to PA (Table 2), in contrast with other non-noble metal catalysts tested
0.02
in previous works [12–18]. Selectivity to PA at t=10 min, for W/F0
=
MA
12 g h mol−1, followed the pattern Pt/Al2O3 (96%)>Pt/SiO2–Al2O3
(85%)≅Pt/SiO2 (85%). Instead, for a W/F0MA=24 g h mol−1, the selec-
tivity pattern was Pt/Al2O3 (96%)>Pt/SiO2 (86%)>Pt/SiO2–Al2O3
(76%). For both space times, the selectivity to PA remained almost
constant during the 3 h run with both Pt/SiO2 and Pt/Al2O3, while it
increased slightly on Pt/SiO2–Al2O3. Thus, the highest selectivity to
PA was reached when the metal Pt phase was interacting with the
γ-Al2O3 surface.
0.009
0.006
0.003
For W/F0MA=12 g h mol−1, the PA, GBL and CH4 production rates
and yields with Pt/SiO2 diminished monotonically with time, until a
steady state was reached (Fig. 4). These trends were similar to that ob-
served for SA conversion (Fig. 3). This is in agreement with the fact that
metal sites, active for SA hydrogenolysis reactions, are deactivated with
time. This deactivation is probably due to strong adsorption of reactant
and/or product molecules on metal Pt sites, as it was suggested above
and in previous works [13,19]. In the case of Pt/Al2O3, PA and GBL
production rates followed similar trends to that determined for SA con-
version, i.e. a linear slight decrease with time indicating a very slow
deactivation of metal hydrogenolytic sites. However, CH4 production
rate showed exactly the opposite trend. Finally, with Pt/SiO2–Al2O3,
PA production rate kept practically constant. Instead, GBL production
rate decreased slowly with time, with a similar linear trend to that
observed for SA conversion, while CH4 production rate diminished
exponentially to reach a steady state after 1 h. The product evolutions
observed with Pt/Al2O3 and Pt/SiO2–Al2O3, are indicating that PA and
GBL are most probably coming from parallel SA hydrogenolysis on dif-
ferent types of metal Pt sites (Fig. 1). Instead, CH4 was not only a GBL
hydrogenolysis product but was also coming from a side reaction no
depicted in Fig. 1.
0.004
0.002
0.000
0
50
100
150
200
Time (min)
When W/F0MA=24 g h mol−1, with Pt/SiO2, PA production rate re-
mains constant at 100% SA conversion (Fig. 5). However, GBL production
rate decreased with time, indicating a deactivation of hydrogenolytic
metal sites similar to that observed at W/F0MA =12 g h mol−1. With
Pt/Al2O3, PA, GBL and CH4 production rates and yields diminished with
time (Table 2, Fig. 5). Instead, with Pt/SiO2–Al2O3, PA production rate
remains almost constant during the 3 h run, while GBL diminished
with time. This is again indicating that SA hydrogenolysis is taking
place on two different types of metallic hydrogenolytic sites, similarly
to that observed in previous works [19,20]. The relative surface concen-
tration of metal sites that are selective for SA hydrogenolysis into PA
respect to those selective to GBL increased as the metal Pt dispersion
was higher (Tables 1 and 2). On the other hand, the high CH4 production
rate with Pt/SiO2–Al2O3 and its rapid decrease with time are again indi-
cating that CH4 is mainly coming from a side reaction not shown in
Fig. 1. It is very likely that at the beginning SA was adsorbed on both
metal Pt and strong acid sites of the support. The SA adsorbed on metal
Pt sites reacted following the parallel and series reactions represented
in Fig. 1. Instead, SA adsorbed on the strongest Lewis and Brönsted acid
sites was converted into CH4 and other carbonaceous fragments. Then,
CH4 desorbed while the carbonaceous fragments remained strongly
adsorbed on the SiO2–Al2O3 surface, blocking the strong Lewis acid
sites. Thus, total SA conversion was due to hydrogenolysis reactions on
metal sites (Fig. 1) and some hydrocracking reactions on strong acid
Fig. 4. Production rates of propionic acid (PA), γ-butyrolactone (GBL) and CH4 at
240 °C, 1 bar and W/F0MA=12 g h mol−1. □ Pt/SiO2, Δ Pt/Al2O3, ○ Pt/SiO2–Al2O3.
sites present on SiO2–Al2O3 surface. As a consequence of the blockage of
strong acid sites, total SA conversion and CH4 production rate diminished
with time while PA production rate remained almost constant. Mean-
while, GBL production rate diminished slowly due to strong adsorption
of reactant and/or product molecules on metal Pt hydrogenolytic sites,
similarly to what happened for Pt/SiO2, with low or none acidity.
4. Conclusions
Propionic acid can be synthesized via gas-phase hydrogenation of
maleic anhydride to succinic anhydride on Pt-supported catalysts using
a single reactor system, at atmospheric pressure and low space times.
High activity in the selective hydrogenolysis of succinic anhydride is
obtained when the support has low acidity and Pt-support interaction
is weak. Increasing acidity and metal-support interaction lead to a dimi-
nution in metal activity and to an increase in methane yield. Thus, the
highest yield in propionic acid is obtained with Pt/SiO2, having the lowest
support acidity and the weakest Pt-support interaction of the catalyst
series used in this work.