X. Yu et al. / Journal of Catalysis xxx (xxxx) xxx
3
2.4. Catalytic tests
range showing a potential to decrease the operating pressure in
industry.
Catalytic evaluation of vapor-phase DMO hydrogenation is con-
ducted in a continuous flow mode using a stainless-steel tubular
reactor. In a typical run, 0.8 g of calcined catalyst (40–60 mesh)
is placed in the center of the reactor (9.4 mm internal diameter),
and the top and bottom side of the catalyst bed are packed with
adequate amount of quartz powder (20–40 and 40–60 meshes).
The calcined catalyst is activated in a flow of 20% H2-80% N2 stream
(75 mL/min) controlled by mass flow controllers at 350 °C and
ambient pressure for 4 h with a ramping rate of 2 °C/min. The sys-
tem pressure is increased and precisely controlled at 2.5 MPa with
a back-pressure regulator when the reactor is cooled to the target
reaction temperature (190 °C). A 10 wt% DMO solution (in metha-
nol) is pumped into the reactor using an ISCO pump (Teledyne
ISCO). The products are collected after the reaction reaches steady
Turnover frequency is calculated according to the DMO conver-
sion data (less than 20% at 190 °C) and the surface amount of Cu0
sites determined by N2O titration method. As shown in Table 1,
the TOF value has a volcano shape with respect to the increasing
amount of In content. The value is larger on bimetallic catalysts
when the In content is not higher than 1 wt%, reaching a maximum
(70.3 5.9 hꢁ1) at 0.25 wt% In doping.
The comparison of stability is performed on 10Cu0.5In/SiO2 and
10Cu/SiO2 catalyst (Fig. 1(f)) over a 100 h run test. While both cat-
alysts are stable, the activity of 10Cu0.5In/SiO2 and its selectivity to
EG are much higher than the monometallic Cu counterpart.
3.2. Material characterizations
state and analyzed on
equipped with a flame ionization detector and RTX-Wax capillary
m). The calculation of the con-
version of DMO and selectivity to various products (methyl glyco-
late and ethylene glycol) follows formula below:
a
gas chromatograph (Agilent 5890)
Various characterization methods have been conducted to
reveal why an appropriate amount of indium species can promote
the performance of Cu/SiO2 in the hydrogenation of DMO to EG.
The chemical compositions and textural properties of all catalysts
are summarized in Table 1. ICP-OES proves the effectiveness of
the preparation method which can deposit most of the copper
and indium species onto the silica support. The N2 physisorption
measurement shows that all the calcined samples exhibit type IV
isotherms with mesopore diameter of ~20 nm, as well as some
microporosity (Fig. 2). The BET surface areas of samples with low
In loading amount (0–1 wt%) are all close to each other. The excep-
tion is the material with 2 wt% loading of indium species which
shows a 23% decrease in the BET surface area compare with
monometallic Cu catalyst, possibly arising due to pore blockage
by indium oxide particles at such high loading. Indeed, the intro-
duction of indium species decreases the pore volume by about
10–15% when compared with the pore volume of monometallic
Cu catalyst.
H2-TPR was conducted to understand the reducibility of the cat-
alysts and to elucidate surface chemical information such as the
distribution of metal species and the interaction between them.
Fig. 3(A) shows the H2-TPR profiles of 10 wt% Cu/SiO2 with differ-
ent amounts of In. For 10 wt% Cu/SiO2, the peak centered at
192 °C is assigned to the collective reduction of Cu2+ species to
Cu0/Cu+. The reduction of bulk cupric oxide usually occurs at much
higher temperatures of around 243–255 °C [11,18]. Thus, the low
reduction temperature of 10 wt% Cu/SiO2 indicates a high disper-
sion of copper species on silica, which is further supported by
TEM measurements, and a weak interaction between certain
amount of copper species (e.g. CuO) with SiO2. Interestingly, the
reduction temperature gradually increases with the amount of
indium species as listed in the inset except for the sample with
0.25 wt% In. The gradual shift of reduction temperature suggests
the existence of an interaction between copper and indium species.
Thus, the incorporation of In hinders the reduction of 10 wt%
Cu/SiO2, which is similar to modification of copper catalysts with
other promoters [11,17,18,21]. The growing peak with respect to
increasing indium content at 270 °C is ascribed to the partial
reduction of indium species, which is further confirmed by XPS
(see below). As indium species are very difficult to be well dis-
persed on silica, the size distribution is broad with a certain
amount of large grains [22,23]. The reduction behavior of indium
species is highly dependent on the particle size, with small grains
able to be reduced below 300 °C and large grains being reduced
only at very high temperatures [23,24]. The reduction of indium
species could also be initiated by the neighboring copper species
when they are in close contact. We also cannot rule out that part
of the reduction profile of indium species with small particle size
may be overlapped with the reduction profile of copper species,
given the low loading of indium species in this study.
column (30 m ꢂ 0.25 mm ꢂ 0.25
l
nDMO;in ꢁ nDMO;out
XDMO
¼
ꢂ 100
nDMO;in
nMG
SMG
¼
ꢂ 100
nDMO;in ꢁ nDMO;out
nEG
SEG
¼
ꢂ 100
nDMO;in ꢁ nDMO;out
The selectivity to other products including ethanol, 1,2-
butanediol (1,2-BDO) and 1,2-propanediol (1,2-PDO) is negligible.
Turnover frequency (TOF, molDMO/(molCu on the surfaceꢂh) or hꢁ1
)
of the reaction is measured when the DMO conversion is lower
than 20% (T = 190 °C, P(H2) = 2.5 MPa, H2/DMO molar ratio = 95,
WHSV = 4.8 hꢁ1). The TOF value is calculated based on the number
of surface Cu0 atoms determined by N2O titration method.
3. Results and discussion
3.1. Catalytic performance
The hydrogenation performance DMO over silica supported
copper and copper-indium catalysts was investigated in a fixed
bed reactor. Fig. 1(a) shows the effect of weight percentage of In
on the performance at 190 °C. Further increasing the amount of
In above 0.5 wt% decreases both the conversion and EG selectivity,
which may result from covering active sites by excess amount of
indium species.
The effect of WHSV on the catalytic performance (10Cu/SiO2
and 10Cu0.5In/SiO2) was investigated to measure the production
capacity on different catalysts. Fig. 1(b and c) shows that the per-
formance in terms of conversion and selectivity to EG decreases
on both catalysts with the increasing WHSV due to the shortening
of contact time of the reactant with catalysts [19]. However, the
performance of 10Cu0.5In/SiO2 is superior to that obtained with
10Cu/SiO2 catalyst under different WHSV, confirming the advan-
tage of bimetallic catalyst in this case. The results show that incor-
poration of an appropriate amount of In can improve the
performance of 10 wt% Cu/SiO2.
The effect of H2 pressure on the performance of catalysts was
also examined. As shown in Fig. 1(d and e), the selectivity to EG
increases with elevated H2 pressure, possibly arising from the
increased concentration of surface absorbed H2 species [20]. How-
ever, the performance of hydrogenation of DMO on 10Cu0.5In/SiO2
is much less influenced than that on 10Cu/SiO2 in the low-pressure
Please cite this article as: X. Yu, T. A. Vest, N. Gleason-Boure et al., Enhanced hydrogenation of dimethyl oxalate to ethylene glycol over indium promoted