116
W. Oberhauser et al. / Journal of Catalysis 325 (2015) 111–117
Table 3
spectrometry, respectively. Importantly, the carbon mass balance
was in all cases >99%. The results of the catalytic screening are
compiled in Table 2.
Average Pt-NPs’ size of supported catalysts before and after GLY hydrogenolysis at
453 and 433 K.
Catalyst
Pt-NP diameter in nm
As-synthesized
All catalytic hydrogenolysis reactions were carried out in the
presence of base, since in its absence, only a very low GLY conver-
sion was obtained (Table 1, entry 4) [6]. Regardless of the reaction
temperature, the obtained liquid-phase reaction products were
1,2-PD, ethylene glycol (EG), and MeOH (not reported in Table 2)
which were formed in a 1:1 molar ratio, due to a retro-aldol reac-
tion [30] and the sodium salt of lactic acid (Na(LA)) [31,32]
(Scheme 2). The gas-phase products consisted of CO, CO2, and
traces of CH4.
After catalysis at
453 K
After catalysis
at 433 K
4 h
8 h
4 h
8 h
Pta@CK
Pta@CV
Ptb@CK
Ptb@CV
1.3a
1.3a
2.4a
2.6a
3.2a
8.0a
5.4b
6.7b
7.4b
9.0b
6.5b
10.2b
a
From HRTEM measurements.
From PXRD measurements.
b
Catalytic GLY hydrogenolysis reactions conducted at 453 K
clearly showed for the a-type Pt catalysts, a much higher activity
compared to the b-type analogs along with a lower chemoselectiv-
ity for 1,2-PD (Table 2, entries 1/7 vs 2/8). This lower 1,2-PD
chemoselectivity found for the former catalysts is mainly due to
their high aqueous-phase reforming activity (APR) (Table 2), which
is expected to be higher for smaller Pt-NPs due to the higher num-
ber of edge and corner atoms [11–16]. Also, the EG formation is
lower in b-type compared to a-type catalysts, while the yield of
Na(LA) seems to mainly depend on the surface area of the applied
support. As a result, the carbon support with the smallest surface
area gave the highest amount of LA (i.e., 12%). More importantly,
LA is not only produced from GLY [31,32] but it is also accessible
by dehydrogenation of 1,2-PD and successive reaction with base
(Scheme 2). Accordingly, an independent experiment using 1,
2-PD as substrate (Table 2, entry 6) gave 23% of LA, even in the
presence of H2.
Fig. 6. PXRD diffractograms of as-synthesized Ptb@CV (a), recovered Ptb@CV (b),
as-synthesized Ptb@CK (c), and recovered Ptb@CK (d).
The water–gas shift (WGS) reaction [33–36] seemed slightly
more favoured by Pt@CK compared to Pt@CV, regardless of the ini-
tial size of the Pt-NPs, while on CG, WGS is almost not occurring.
Analogous catalytic GLY hydrogenolysis reactions carried out at
433 K (Table 2, entries 11–20) exhibited for Pta@CK the highest GLY
conversion (TOF = 181 hÀ1) along with the highest chemoselectiv-
ity for 1,2-PD (73%) (Table 2, entry 11). This latter chemoselectivity
dropped to 70% for a catalytic reaction lasting 8 h (entry 14).
Conversely, Ptb@CK showed under identical catalytic conditions a
much lower catalytic activity (TOF = 48 h-1) and chemoselectivity
for 1,2-PD (60%) (entry 15). Even Pta@CV (entry 16) gave scarves
results in terms of 1,2-PD chemoselectivity (53%). LA was formed
at 433 K only by Pt@CV. We recycled twice Pta@CK and Pta@CVat
433 K and observed for the former catalyst a slight decrease of
the catalytic activity as well as chemoselectivity (i.e., 68% after
the 2nd recycling experiment, entry 13), while the latter one
reached a chemoselectivity of only 40% after the 2nd recycling
experiment (entry 18).
Fig. 7. PXRD diffractograms of as-synthesized Pta@CV (a), recovered Pta@CV (4 h)
(b), recovered Pta@CV (8 h) (c), as-synthesized Pta@CK (d), recovered Pta@CK (4 h)
(e), and recovered Pta@CK (8 h) (f).
In order to study the aggregation of the Pt-NP on CK/V/G under
catalytic hydrogenolysis reaction conditions, we separated the
solid catalysts from the solution after catalysis by a simple filtra-
tion at room temperature and washed the black solids with water
(surface area reduction): 40% (mesitylene) vs 25% (pentane). A con-
trolled heating of Pta@CK (2.0 °C/min) combined with an online
mass spectroscopic analysis (i.e., [mesitylene-CH3]+ (m/z = 105))
of the mesitylene released from the carbon support showed that
at 57 °C outer pore, mesitylene is released from CK, while mesity-
lene localized inside the pores is released in an extremely broad
temperature range (i.e., from 170 to 300 °C (Fig. S6). The average
pore width distribution (i.e., bimodal pore width distribution) of
CK/V did not change upon supporting Pt.
and acetone, followed by air-drying. HRTEM-analyses of Pta@CK/V/G
,
recovered after the catalytic reactions performed at 453 K for 4 h,
were carried out and representative micrographs are shown in
Figs. 4 and 5.
In Table 3 are compared the Pd-NPs’ size of the different cata-
lysts before and after GLY hydrogenolysis reaction carried out at
453 and 433 K.
The HRTEM-micrographs evidenced a significant increase of the
Pt-NPs size on CV and CG in the course of the catalytic hydrogenoly-
sis reactions. Most importantly under the applied reaction condi-
tions, Pt-NPs on CG gave a predominant amount of particles
which form larger aggregates, precluding hence a reliable particle
size histogram (Fig. 5). In contrast, Pta@CK showed after catalysis
at 453 K, 4 h relatively small Pt-NPs of 3.2 1.2 nm. The same
trend in the Pt-NPs size was observed with b-type catalysts after
Pta,b@CK, Pta,b@CV, and Pta@CG were used to catalyze the
aqueous-phase hydrogenolysis reaction of GLY (2.4 vol%) in the
presence of NaOH (0.8 M), H2 pressure (600 psi at 303 K), and a
GLY to Pt(surface) ratio of 1274 [29]. The catalytic reactions were per-
formed at 453 K and 433 K in order the estimate the effect of the
reaction temperature on the overall catalytic activity and chemose-
lectivity toward 1,2-PD which is the target organic compound.
The liquid- and gas-phase products formed in the course of the
GLY hydrogenolysis reactions were analyzed by HPLC and mass