Full Papers
Under more severe conditions (2008C, 40 bar H ), Ru/OMC-P
promoted the further conversion of GVL to pentanoic acid
2
Table 5. LA hydrogenation over 1% Ru-based catalysts at 2008C and pH2
[a]
4
0 bar.
(
PA). This conversion did not occur with Ru/OMC. Under these
[
b]
Catalyst
t
Conversion Activity
Selectivity [%]
harsh conditions, however, the Ru/OMC-P catalyst starts to de-
activate after three consecutive cycles as a result of a combined
leaching of weakly bound phosphonic acid groups and a Ru
redox reaction with LA. Although the leaching of phosphoric
groups may favor the migration of Ru to the more accessible
surfaces of the OMC supports, the LA redox reaction promoted
the further reduction and aggregation of Ru on the carbon
surfaces. These results, however, show a major improvement in
the catalyst stability for the conversion of LA in aqueous media
compared to those on other acidic oxide supports. The ability
to further increase the conversion of GVL to PA is very stimu-
lating and of great interest for the large-scale production of bi-
ofuels from PA.
[
h] [%]
GVL PEA PA PD MTHF
Ru/OMC
2
4
2
3
6
0
79
>99
90
>99
>99
>99
3.91
2.48
4.46
3.30
96
95
85
75
68
62
0
1
4
1
2
2
0
0
10
22
29
35
0
0
0
0
0
0
1
2
2
3
1
1
Ru/OMC-P
1
À1
[
a] Reaction: [LA]=0.43 molL ; Ru/LA=1:1000 mol/mol. [b] Converted
À1 À1
molGVL
g
Ru
h .
action conditions. Such results are of great interest as PA is an
[4]
important intermediate in the production of biofuels.
The stability of Ru/OMC-P was also investigated at 2008C
and 40 bar H (Figure 6). Under these high-temperature condi-
2
tions, the catalyst exhibited a lower stability than at 708C. The Experimental Section
activity decreased slightly during the first three cycles, and
Materials
a more evident deactivation was observed after the fourth run.
The morphology of the nanoparticles changed after five con-
secutive tests (Figure 2c). Particle size analysis performed by
STEM (Table 2) showed that the Ru particles increased in size
from 1.6 to 2.5 nm with the formation of some aggregates
RuCl (99.99% purity) from Aldrich was used. Gaseous hydrogen
3
from SIAD was 99.99% pure.
Support preparation
(
Figure 2c). XPS measurements of Ru OMC-P after the catalytic
reaction revealed two Ru species with BEs of 283.4 and
Resorcinol (17.6 g) and Pluronic F127 (17.6 g) were dissolved in
ethanol (72 mL)/water (54 mL)/HCl (17.6 mL, 12.5m). Formaldehyde
(20.8 mL) was added, and phase separation was observed after
6 min. The gel was stirred for another 60 min. The top liquid phase
was separated, and the bottom gel polymer was cast on Mylar. The
film was allowed to dry overnight at RT and then at 808C for 24 h.
The obtained polymer composite was carbonized in flowing Ar
6+
0
2
80.4 eV attributed to Ru and a species that resembles Ru
(
Figure 3, inset). After the catalytic reaction, there was a clear
decrease in the PO concentration (from 0.4 to 0.2 at%) and an
4
increase in Ru concentration (0.4 to 1.2 at%). These results
point to the leaching of some of the less stable PO groups,
4
which enabled Ru to migrate to the catalyst surface. Hence,
the deactivation of the Ru-OMC-P might be because of the
leaching of phosphorylated groups. The role of LA on the de-
activation was investigated by testing the catalyst in pure
À1
À1
(
500 mLmin ) at 8508C for 120 min at a heating rate of 58Cmin .
This sample was labeled as OMC.
OMC (ꢀ2.5 g) was dispersed in concentrated H SO (25 mL) to in-
2
4
troduce sulfate groups or in H PO (25 mL) to introduce phosphate
3
4
water at 2008C and p =40 bar for 6 h. In this case the deacti-
H2
surface groups. These systems were stirred at 808C for 12 h under
vation occurred after the first run, which suggests that the de-
activation is more because of the harsh reaction conditions
than because of the chelating properties of LA.
flowing N . The solids were collected by filtration and washed with
2
water until the filtrate was neutral. The samples were dried at 808C
overnight and labeled as OMC-P and OMC-S after treatment in
H PO4 and H SO , respectively. A batch of the OMC-S material
3
2
4
(
3.0 g) was stirred in concentrated H PO (25 mL) for 6 h at 808C
3 4
Conclusions
under flowing N . The solid was recovered by filtration and washed
2
to neutral as with the other samples. This material was labeled
OMC-S/P.
We demonstrate new Ru supported on ordered mesoporous
carbon (OMC) and acid-functionalized OMC (OMC-P and OMC-
S) catalysts for the hydrogenation of levulinic acid (LA) under
aqueous conditions. The catalytic performance of the Ru nano-
particles was affected greatly by the surface chemistry of the
carbon supports. If mild reaction conditions were used (708C
Catalyst preparation by incipient wetness impregnation
À1
Solid RuCl3 was dissolved in water ([Ru]=10 mgmL ). Sufficient
and 7 bar H ), the phosphorylated OMCs accelerated the reac-
metal-containing solution was added to each of the OMC supports
2
(
OMC, OMC-P, OMC-S, and OMC-P/S) to completely fill their pores,
tion rate, and a high selectivity to g-valerolactone (GVL;
based on the total pore volume from N sorption analysis (Table 1).
The amount of support was calculated so that a final Ru loading of
1
water to remove inorganic residues (Cl , Na , etc.). The catalyst
was then collected by filtration, dried at 808C for 2 h, and reduced
in H at 2008C for 2 h. The catalysts were labeled as Ru/OMC, Ru/
OMC-P, Ru/OMC-S, and Ru/OMC-P/S.
2
>
93%) was observed. Ru/OMC-P has excellent stability and
maintains the same activity and selectivity to GVL after five
consecutive runs. However, weakly bound surface ÀS groups
leached off the surfaces of OMC-S and bound to the most
active Ru nanoparticles, which thus deactivated the Ru cata-
lyst.
wt% was obtained. The catalysts were washed several times with
À
+
2
ChemSusChem 2015, 8, 2520 – 2528
2526
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim