Hydrogenation of dimethyl oxalate to ethylene glycol over Cu/KIT-6 catalysts

Article abstract of DOI:10.1039/d0cy02334e

Copper supported on KIT-6 mesoporous silica was preparedviaammonia evaporation (AE) method and applied for the catalytic hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG). The high specific surface area and interconnected mesoporous channels of the support facilitated the dispersion of copper species. The effect of AE temperature and copper loading on the structure of catalysts and induced change in hydrogenation performance were studied in detail. The results showed that both parameters influenced the overall and/or intrinsic activity. The hydrogenation of DMO to EG was proposed to proceedviathe synergy between Cu⁰and Cu⁺sites and catalysts with high surface Cu⁰/Cu⁺ratio exhibited high intrinsic activity in the investigated range.

Full text of DOI:10.1039/d0cy02334e

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Hydrogenation of dimethyl oxalate to ethylene
glycol over Cu/KIT-6 catalysts†
Cite this: Catal. Sci. Technol., 2021,
11, 2403
Xinbin Yu,a Michael Burkholder,b Stavros G. Karakalos, a Gregory L. Tate,
a
John R. Monnier,a B. Frank Guptonb and Christopher T. Williams
a
*
Copper supported on KIT-6 mesoporous silica was prepared via ammonia evaporation (AE) method and
applied for the catalytic hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG). The high specific
surface area and interconnected mesoporous channels of the support facilitated the dispersion of copper
species. The effect of AE temperature and copper loading on the structure of catalysts and induced change
in hydrogenation performance were studied in detail. The results showed that both parameters influenced
the overall and/or intrinsic activity. The hydrogenation of DMO to EG was proposed to proceed via the
synergy between Cu0 and Cu+ sites and catalysts with high surface Cu0/Cu+ ratio exhibited high intrinsic
activity in the investigated range.
Received 3rd December 2020,
Accepted 24th January 2021
DOI: 10.1039/d0cy02334e
rsc.li/catalysis
(Cu2Si2O5(OH)2) and/or Cu–O–Si unit, which are the products
of copper precursor species reacting with silica during the
1. Introduction
Ethylene glycol (EG) is a commodity widely consumed as a
solvent, building block of polyester fibers, anti-freezing agent,
alkyd resin and so forth.1 Traditionally, EG is synthesized by
hydration of ethylene oxide derived from petroleum
feedstocks.2 However, due to non-uniform global distribution
of petroleum resources, the synthesis of EG from syngas is
emerging as a rapidly growing technology, especially in
Eastern Asia.1,3 The process consists of producing syngas
from coal, coupling of CO by methyl nitride on Pd/α-Al2O3 to
form dimethyl oxalate (DMO), and subsequent hydrogenation
of DMO to EG. This process has already been implemented
on an industrial scale.4
The choice of catalysts is vital to the hydrogenation
reaction, with CuCr catalysts having been found applicable to
have excellent activity and stability.5,6 However, the Cr6+ is a
known health hazard and environmental pollutant, which
motivates research for alternative, environmentally friendly
catalysts. Along these lines, Cu/SiO2 catalysts have proved
effective for this reaction, due to the coexistence of surface
Cu0 and Cu+ species that are proposed as sites for the
activation of H2 and CO, respectively.3,4 The Cu0 species
originate primarily from the reduction of CuO, while Cu+
species come from the reduction of copper phyllosilicate
preparation of catalyst. Excellent performance of
hydrogenation of DMO to EG can be achieved over Cu/SiO2
by adjusting the balance of Cu0 and Cu+ sites.
Various promoters (B, Ni, Zn, La, Ag, Au, In, etc.) have
been studied in detail to reveal how the surface content of
Cu0 and Cu+ sites can be tuned for the optimal catalytic
performance.4,7–14 However, supports also play an important
role as they not only help disperse the active species but also
react with certain copper species to form copper
phyllosilicate and/or Cu–O–Si unit. Ordered mesoporous
silicas (e.g. MCM-41, SBA-15, and HMS) with tuneable
mesopores, easily accessible internal surface, and high
specific surface areas have been successfully applied as
supports for metal catalysts in many reactions.15,16 The
spatial confinement exerted by channels on active species
exhibits interesting phenomena, such as inhibiting particle
growth and modifying product distribution. However, most
of the studied mesoporous silicas for hydrogenation of DMO
only have one- or two-dimensional pores.11,12,17–19 These
structures may suffer from problems such as inefficient
utilization of active species due to blocked channels by
particles (during preparation and reaction), coke and
collapsed wall and inferior mass transfer kinetics.18,19 Silicas
with three dimensional, interconnected mesopores like KIT-6
would be a good choice to address these problems, and have
not yet been explored as a support for the hydrogenation of
DMO over Cu/SiO2 catalysts.
a Department of Chemical Engineering, University of South Carolina, Columbia, SC
29208, USA. E-mail: willia84@cec.sc.edu
b Department of Chemical Engineering, Virginia Commonwealth University,
Richmond, VA 23284-3068, USA
Ammonia evaporation (AE) method is typically employed
to deposit copper species onto silica supports like colloidal
silica and mesoporous silicas since its invention by Ube
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
d0cy02334e
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2403–2413 | 2403

View Article Online
Paper
Catalysis Science & Technology
Industries.3,12 Many studies utilize this synthesis protocol to
produce Cu/SiO2 catalysts for DMO hydrogenation as it has
been demonstrated to be more effective in dispersing
copper species than wet impregnation.12 As copper
phyllosilicate and/or Cu–O–Si unit are generated by the
reaction between copper–ammonia complexes and silica at
the preparation stage, their content and the attendant change
in the catalytic properties may be regulated by parameters
such as ammonia evaporation temperature, pH, types of
copper precursors, concentration of copper, calcination
temperature, etc. Chen et al. found that ammonia
evaporation temperature significantly influenced the
deposited copper amount, relative content of Cu+ and surface
area of Cu0 of reduced Cu/SiO2 catalysts and the optimal
temperature was 90 °C.3 Yin et al. found that the surface ratio
of Cu0/Cu+ could be adjusted by the copper loading, which
then influenced both DMO hydrogenation activity and
selectivity to EG.20
In the present work, ordered mesoporous silica with
interconnected channels, KIT-6, is facilely synthesized by
soft-templating method and then employed as support for
copper catalysts prepared by ammonia evaporation method.
The effect of AE temperature and copper loading on the
performance and intrinsic activity of catalysts has been
studied in detail. The catalysts are extensively characterized
to investigate how the changes in surface properties (i.e.
specific surface area of Cu0 and surface ratio of Cu0/Cu+)
induced by AE temperature and copper loading influence the
catalytic performance.
treatment). Afterward, the solid was filtrated and washed
with adequate amount of DI water to remove Cl− (tested with
0.1 M of AgNO3 solution). The obtained white powder was
dried for 12 h at 110 °C and then calcined in a furnace for 6
h at 550 °C.
Cu/KIT-6 catalysts were prepared by the ammonia
evaporation (AE) method. For example, to prepare 10 wt%
Cu/KIT-6, 3.8 g of Cu(NO3)2·3H2O was dissolved in 100 ml of
DI water. Then the pH of the solution was adjusted to 11 with
certain amount of NH3·H2O (28%) and stirred for 1 h. 9 g of
KIT-6 was added, and the suspension was stirred overnight.
The pH of mixture was adjusted to 11 and then the mixture
was heated in a silicone oil bath at 70 °C to remove NH3. The
evaporation process was terminated when the pH was less
than 7. The mixture was then filtrated, washed three times
with deionized water and dried overnight at 110 °C in an
oven. Finally, the light blue powder was calcined in air at 400
°C for 3 h. The obtained calcined sample was denoted as
10Cu-70, where 10 was the weight percent loading of Cu on
KIT-6 and 70 was the AE temperature. Samples with various
copper loading (5 wt% and 20 wt%) and different AE
temperature (80 °C, 90 °C and 100 °C) were prepared by the
same method.
2.3. Catalyst characterization
N2 physisorption at −196 °C was conducted to obtain
information about the specific surface area and pore size on
a Micromeritics ASAP 2020 instrument. Samples were heated
at 120 °C for 2 h under vacuum prior to nitrogen adsorption.
The calculation of specific surface area (SBET) followed the
2. Experimental
2.1. Materials
Brunauer–Emmett–Teller
(BET)
method.
Pore
size
distribution was evaluated by the Barrett–Joyner–Halenda
(BJH) method using the data of desorption isotherm branch.
The composition of catalysts was measured by inductively
coupled plasma-optical emission spectroscopy (ICP-OES,
Perkin Elmer Avio 200). The surface area of metallic copper
per gram of catalyst (SCu0) was analyzed by N2O titration
method at 60 °C, assuming 1.46 × 1019 copper atoms per m2.
The formation amount (mole) of N2 (equal to the
consumption amount (mole) of N2O by metallic Cu) was
measured by a thermal conductivity detector (TCD).
The Fourier-transform infrared (FT-IR) spectra of calcined
samples were obtained on Perkin Elmer Spectrum 100 FT-IR
spectrometer at room temperature (4 scans for each
spectrum). Diamond was used as an internal reflection
element (IRE).
All the reagents in this work were commercially available.
Copper(II) nitrate trihydrate (99%, Acros Organics), Pluronic®
P-123 (average Mn ∼5800, Sigma-Aldrich), HCl (37%, Sigma-
Aldrich), dimethyl oxalate (99%, Sigma-Aldrich), tetraethyl
orthosilicate (TEOS, 99.9%, Alfa Aesar), concentrated nitric
acid (68–70%, BDH), n-butanol (99.5%, ACROS), ammonia
aqueous solution (28%, BDH) were used without further
purification. H2 (UHP), He (UHP), Ar (UHP), N2 (UHP), and
air (zero grade) were purchased from Airgas. Deionized water
(18 MΩ) was supplied by using a Milli-Q water system with
Barnant B-pure filter.
2.2. Catalyst preparation
KIT-6 silica was prepared according to the reported
method.21,22 13.5 g of Pluronic® P123 was mixed with 487.5 g
of DI water and 26.1 g of HCl (37%) in a flask submerged in
a silicone oil bath at room temperature. The mixture was
vigorously stirred for 1 h and the temperature was increased
to 35 °C. 13.5 g of n-butanol was added into the flask
followed by stirring for another 1 h at 35 °C. Subsequently,
29 g of TEOS was added quickly into the solution and the
mixture was stirred for 24 h at 35 °C. Then the temperature
was elevated to 95 °C and maintained for 24 h (hydrothermal
The reducibility of catalysts was analyzed on
a
Micromeritics Autochem II 2920 instrument to obtain the H2
temperature-programmed reduction (H2-TPR) profile. In a
typical run, 100 mg of a calcined sample was reduced with
10% H2/Ar (40 mL min−1) with temperature ramping from RT
to 600 °C at 5 °C min−1. The consumption of H2 was
monitored by TCD.
The small angle X-ray scattering pattern of KIT-6 was
collected on a SAXSLab Ganesha instrument. The sample was
2404 | Catal. Sci. Technol., 2021, 11, 2403–2413
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
nDMO;in − nDMO;out
scanned from 0.18 to 4 degree at a speed of 0.2 degree per
minute at room temperature. Prior to the test, the instrument
was calibrated with 640c silicon powder with the peak
position fixed at the total scattering angle (2θ) of 28.44°
according to the criteria of National Institute of Standards
and Technology (NIST).
XDMO
¼
× 100
nDMO;in
nMG
SMG
¼
× 100
nDMO;in − nDMO;out
nEG
Wide-angle powder X-ray diffraction (XRD) patterns of
catalysts were obtained on a Rigaku SmartLab SE X-ray
diffractometer using Cu Kα radiation (λ = 0.15418 nm). For
calcined samples, the XRD patterns were collected with a
scanning angle (2θ) range from 10° to 90° at 15° min−1. Prior
to the measurement of reduced samples, calcined samples
were in situ reduced with 10% H2–90% N2 mixture (100 mL
min−1) in a XRD reactor at 350 °C for 2 h. Then sample's
XRD pattern at 350 °C and 190 °C was recorded with a
scanning angle (2θ) range from 10° to 90° at 5° min−1. The
particle size was measured in terms of Scherrer equation
using Cu(111) diffraction peak.
SEG
¼
× 100
nDMO;in − nDMO;out
Other products included 1,2-butanediol (1,2-BDO), ethanol
(EtOH) and 1,2-propanediol (1,2-PDO) and their selectivity was
negligible. The standard deviation was calculated by the
following formula:
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
N
X
u
t
1
2
̅
S ¼
×
ðxi − xÞ
N − 1
i¼1
Transmission electron microscopy (TEM) was performed
on a Hitachi HT7800 transmission electron microscope with
an acceleration voltage of 100 kV. The samples were treated
with procedures in our previous work to obtain finely
dispersed suspension.14 The suspension was then loaded
onto a copper grid coated carbon film (300 mesh) and
methanol completely evaporated at room temperature before
loading the grid into the instrument.
X-ray photoelectron spectroscopy (XPS) was measured on a
Kratos AXIS Ultra DLD XPS system with the same
attachments and operation parameters in our previous
work.14 Calcined samples were in situ reduced at 350 °C for 2
Measurements of turnover frequency (TOF, molDMO
/
(molCu
× h) or h−1) were performed when
on the surface
the conversion of DMO was below 30% at the reaction
condition (T, P(H2) and H2/DMO molar ratio) listed
above. The TOF value was calculated in terms of the
specific surface area of Cu0 measured by N2O titration.
It should be noted that the methanol produced from
the reaction cannot be quantified due to the existence
of a large excess of methanol (90 wt%) as solvent, as
is standard for this reaction on both the lab scale and
in industry.3,12
h
in a high temperature cell. Then XPS analysis was
3. Results and discussion
3.1. Catalytic performance
conducted when the samples reached room temperature. C1s
binding energy (BE) of 284.8 eV was used as a reference to
correct all the spectra.
The hydrogenation of DMO over KIT-6 supported copper
catalysts was investigated in a flow reactor. The reaction
proceeds via DMO to MG and then MG to EG, with further
dehydration of EG producing other by-products like ethanol
(Scheme 1). Fig. 1(A) shows the effect of AE temperature on
the hydrogenation performance at 190 °C at the given
conditions. All the 10 wt% Cu/KIT-6 prepared at different AE
temperatures (70, 80, 90, 100 °C) showed roughly the same
conversion of DMO (70–78%). The 10 wt% Cu/KIT-6 prepared
at 90 °C showed slightly higher selectivity to EG and lower
selectivity to MG. Thus, the AE temperature only slightly
influence the hydrogenation performance, which is different
from the significant effect of AE temperature on Cu/SiO2 with
colloidal silica as support in the hydrogenation of DMO to
EG.3 However, Fig. 1(B) and S2† show that the loading
amount of copper species has a dramatic effect on the
hydrogenation performance in terms of conversion of DMO
and selectivity to EG, which is consistent with other
reports.18–20 Increasing the loading amount of Cu from 5
wt% to 20 wt% significantly improves the conversion of DMO
and selectivity to EG. The performance of 5 wt% Cu/KIT-6 is
tested at 195 °C because it shows very low activity at 190 °C.
The positive correlation between activity and loading amount
2.4. Catalytic tests
Vapor-phase, catalytic hydrogenation of DMO was evaluated in
a stainless-steel flow reactor. Typically, 0.6 g of calcined
samples (40–60 mesh) was loaded into the reactor with internal
diameter of 9.4 mm, sandwiched with adequate amount of
quartz powder (20–40 and 40–60 meshes). Prior to the reaction,
the calcined sample was reduced in 20% H2–80% N2 stream
(75 mL min−1) at 350 °C and atmospheric pressure for 4 h. The
ramping rate was 2 °C min−1. A 10 wt% DMO solution (in
methanol) was introduced into the reactor by an ISCO pump
(Teledyne ISCO) when the reactor cooled to the 190 °C and the
system pressure reached 2.5 Mpa with the aid of a back-
pressure regulator. The H2/DMO molar ratio was kept 95 for all
the runs. The condensed products were collected when the
reaction was steady and analyzed on an Agilent 5890 Series II
gas chromatograph with a flame ionization detector (FID) and
RTX-Wax capillary column (30 m × 0.25 mm × 0.25 μm). The
conversion of DMO and analysis of products (selectivity to
methyl glycolate and ethylene glycol) were calculated in terms
of the formulas below:
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2403–2413 | 2405

View Article Online
Paper
Catalysis Science & Technology
Scheme 1 Simplified reaction pathways for the hydrogenation of DMO to MG, EG, and ethanol.
Fig. 1 (A). Effect of temperature of evaporating NH3 on hydrogenation of DMO over 10 wt% Cu/KIT-6, (B). Effect of copper weight percentage on
hydrogenation of DMO over Cu-70, reaction condition: T = 190 °C and 195 °C for 5Cu-70, P(H2) = 2.5 MPa, H2/DMO molar ratio = 95, WHSV =
0.66 h−1
.
of Cu is explained by the increase in the amount of active
sites and/or the formation of more intrinsically active sites
when elevating Cu loading amount over KIT-6, which will be
discussed in the following section.
Fig. 2 shows the stability of hydrogenation of DMO over
20 wt% Cu/KIT-6 prepared at 70 °C. The catalyst does not
show any deactivation after 221 h on stream. The fluctuations
in the selectivity are an artifact of the small uncertainties in
the GC measurements that are amplified in the calculation of
denominator in the selectivity equations above.
3.2. Material characterizations
Several characterizations have been performed to gain an
insight into the effect of AE temperature and loading amount
of Cu on the structure of Cu/KIT-6 catalysts. Fig. 3 shows the
structural characterizations of bare KIT-6. The pattern
(Fig. 3(A)) exhibits peaks at 0.92° (211), 1.01° (220) and 1.75°
(420), indicating that the support has cubic Ia3d symmetry
with interconnected channels.22 The TEM image (Fig. 3(B)),
N2 adsorption–desorption isotherm pattern (Fig. 3(C)) and
BJH pore size distribution analysis (Fig. 3(D)) confirm that
the support has mesopore structure with mesopore diameter
of 6.6 nm. Moreover, a certain amount of pores in the
micropore/small mesopore range is observed, which is
attributed to the secondary pore system in the framework
walls (i.e., intra-wall pores) providing interconnections for
adjacent mesopores (Fig. 3D).23–27
The bulk properties and compositions of representative
catalysts are summarized in Table 1. ICP-OES confirms that
the real deposited amount of copper species on KIT-6 via
ammonia evaporation method is close to the target weight
percentage in the investigated range. All the calcined samples
exhibit type IV isotherms revealed by N2 physisorption
(Fig. 4). The pore diameter distribution plot shows that all
the samples have mesopores with pore diameter of ∼6 nm
with certain amount of pores in the micropore/small
mesopore range, similar to mesopore diameter of bare
support. The bare KIT-6 support has a specific surface
area of 703 m2 g−1. The specific surface area decreases to
Fig. 2 Stability test of hydrogenation of DMO over 20 wt% Cu/KIT-6
prepared at 70 °C, reaction condition: T = 190 °C, P(H2) = 2.5 MPa, H2/
DMO molar ratio = 95, WHSV = 0.66 h−1
.
2406 | Catal. Sci. Technol., 2021, 11, 2403–2413
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
Fig. 3 Structural characterizations of KIT-6, (A). SAXS pattern (B). TEM image, (C). N2 adsorption–desorption isotherm, (D). BJH pore size
distribution.
Table 1 Physicochemical and catalytic properties of Cu/KIT-6 catalysts
d
Sample
Cu content,a /wt%
SBET,b /m2 g−1
Vp,c /cm3 g−1
SCu
,
/m2 g−1
dp,e /nm
TOF,f /h−1
KIT-6
5Cu-70
0
5.7
703
305
294
294
290
280
268
1.03
0.90
0.84
0.84
0.85
0.85
0.69
∼
∼
∼
∼
2.1
2.8
∼
∼
10Cu-70
10Cu-80
10Cu-90
10Cu-100
20Cu-70
10.4
10.3
11.3
10.5
19.7
2.81
2.93
2.84
2.47
4.55
63.3 4.5
49.9 6.9
87.8 0.5
58.5 6.5
99.4 10.4
∼
∼
3.2
a
Determined by ICP-OES analysis. b BET specific surface area. c Pore volume at P/P0 = 0.99. d Surface area of Cu0 per gram of catalyst measured
e
f
by N2O titration method. Average crystallite size measured in terms of Cu(1 1 1) peak at 190 °C. TOF value calculated in terms of specific
surface area of Cu0 from N2O titration results.
305 m2 g−1 upon the loading of 5 wt% of copper species.
However, the specific surface area does not change
significantly for samples with 10 wt% and 20 wt% of copper
species. Besides, only small change (∼10%) in pore volume
exists when comparing 5Cu-70 sample with bare support.
Based on these observations, we think that copper species
are deposited on the external surface, in intra-wall pores and
mesopores and more mesopore volume is occupied when the
deposited amount of copper is higher (e.g. 20 wt%). Similar
analysis has been reported by other researchers.20,25
Fig. 5 shows the FT-IR spectra of calcined KIT-6 and Cu-
containing samples to gain an insight into the existing
copper species on the calcined samples. It is observed that
copper could form copper phyllosilicate and/or Cu–O–Si unit
with silica at the preparation stage. The absence of the peak
at 670 cm−1 attributed to the δOH band of copper
phyllosilicate indicates that copper phyllosilicate may not
exist on the Cu-containing samples.18,20 The peaks located at
800 cm−1 and 1061 cm−1 are assigned to the symmetric and
asymmetric νSiO band of amorphous silica, respectively. The
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2403–2413 | 2407

View Article Online
Paper
Catalysis Science & Technology
Fig. 4 (A) N2 adsorption–desorption isotherms and (B) BJH pore size distribution of the calcined catalysts of different copper loadings. a. 5Cu-70,
b. 10Cu-70, c. 10Cu-80, d. 10Cu-90, e. 10Cu-100, f. 20Cu-70.
peaks at 1643 cm−1 are ascribed to the bending mode of OH
groups of adsorbed water. No peaks at 690 cm−1 belonging to
the bending absorption of the Cu–OH bond exist confirming
the absence of copper hydroxide on the calcined samples.
The peak at 970 cm−1 is assigned to the Si–O stretching of Si–
OH structure, with both the intensity and band width
influenced by the formation of Cu–O–Si, where the Si–O is
still present, but with the H is replaced by Cu.18 The ratio of
peak intensities at 970 cm−1 and 800 cm−1 (I970/I800) is
therefore calculated to estimate the amount of incorporated
copper as the peak at 800 cm−1 represents the symmetric νSiO
band of KIT-6 and the change in the intensity of peak at 970
cm−1 originates from incorporated copper.18 The gradual
increase in I970/I800 when elevating the loading amount of
copper as listed in Fig. 5 suggests that a certain amount of
copper is incorporated into the framework of KIT-6. No
copper hydroxide or copper phyllosilicates exist over the
calcined Cu/KIT-6 samples, unlike Cu/SiO2 prepared with
colloidal silica as silica source.7,9,14
Fig. 6 shows the H2-TPR profiles of different catalysts to
reveal their reducibility. All the profiles exhibit a single peak
centered at different temperature, attributed to the collective
reduction of Cu2+ and Cu+ species (see XPS section) to Cu0
and/or Cu+. The reduction temperature (185–206 °C) in our
case is significantly lower than the typical reduction
temperature of Cu/SiO2 catalysts (243–270 °C) reported in
literature, implying that copper species are well dispersed on
KIT-6.3,17–19,28 This is correlated well with TEM and XRD
measurements. Comparing the peak positions reveals that
the peak gradually shifts to higher temperature when
increasing the loading amount of Cu. It has been reported
that supported CuO with smaller crystallites is much more
facile to be reduced than bulk CuO.3,29 In situ XRD results
reveal that elevating the loading amount of Cu causes an
increase in average size of crystallites. Thus, the shift is
attributed to increasing copper particle sizes and decrease in
the dispersion.
The existing copper species on the calcined samples can
be a mixture of Cu+ species, CuO with different dispersions
Fig. 5 FT-IR spectra of different calcined samples, a. KIT-6, b. 5Cu-70,
c. 10Cu-70, d. 10Cu-80, e. 10Cu-90, f. 10Cu-100, g. 20Cu-70.
Fig. 6 H2-TPR profiles of different catalysts.
2408 | Catal. Sci. Technol., 2021, 11, 2403–2413
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
and Cu–O–Si structures, although the profile only shows one
reduction peak for every sample. Van der Grift et al. and
Marchi et al. found that the reduction of the copper
phyllosilicate and ion-exchanged Cu–O–Si species to Cu+
could overlap with the reduction of well dispersed CuO to
Cu0.30,31 FT-IR spectra show that Cu–O–Si structure rather
than copper phyllosilicate is present on the Cu-containing
samples, in accordance with others reports with ordered
mesoporous silica as support.18,20 H2-TPR might not
differentiate different copper species present on the calcined
samples due to their difference in the amount and
reducibility. XPS results show that both Cu0 and Cu+ species
exist on the surface of samples reduced at 350 °C, which
suggests that not only one type of copper species exist on the
calcined samples. Further reduction of Cu+ species
originating from Cu–O–Si structure may require a much
higher temperature to destroy their stable structures.30,31
Wide-angle powder XRD patterns of the calcined and
reduced samples were obtained to differentiate the existing
phases and calculate crystallite sizes. No noticeable peaks
assigned to metal oxides exhibit on calcined samples
(Fig. 7(A)) implying that copper species are highly dispersed
on the support. Two peaks (Fig. 7(B)) at 43.5° and 74.4°
attributed to Cu0(111) and Cu0(220) emerge after in situ
reduction at 350 °C.14,18,32,33 The small peak at around 50°
on the reduced samples may be attributed to the Cu(200)
peak, although it is partially obscured by the background
feature.12,33 The peaks become more intense and sharper as
the loading amount of copper species increase, implying the
growth in particle size. The average particle sizes are
calculated according to Scherrer equation (Table 1). It shows
the average particle sizes grow with the increase in the
loading amount of copper as it becomes more difficult to
disperse metal oxides well on the support when elevating
their loading amount.34,35 The average particle sizes are all
around 3 nm, confirming that the copper species are highly
dispersed on the catalysts. The Cu particles sizes in our case
are also smaller than Cu particles supported over other
ordered mesoporous silicas.18,20,28
Fig. 8 shows the TEM images of representative Cu/KIT-6
catalysts. All the catalysts maintain the characteristics of the
mesoporous structure, and the copper species are clearly
well-dispersed along the channels, substantiating the analysis
in H2-TPR and XRD sections. Furthermore, the uniform
distribution of copper species along the channels suggests
that the high specific surface area and interconnected
mesoporous channels of the support may facilitate the Cu
dispersion compared with similar catalysts employing
ordered mesoporous silicas as support.18–20 The copper
species distributed on KIT-6 are not in the form of typical
spherical particles which makes us unable to calculate the
average particle sizes from TEM images.
The specific surface area of Cu0 was measured by N2O
titration method as listed in Table 1. The specific surface
area of Cu0 for all the samples with 10 wt% Cu is roughly the
same and it increases significantly when elevating the
loading amount of copper species.
XPS measurements with X-ray induced Auger electron
spectroscopy (XAES) were conducted to reveal chemical states
of copper species and their relative content on calcined
samples and reduced catalysts. For Cu 2p XPS spectra of
calcined samples (Fig. 9(A)), the peak at around 934.7 eV is
assigned to Cu2+ species, confirmed by the characteristic
satellite peak at 943.0 eV.3,7,14 The peak at around 932.8 eV is
attributed to Cu+ species. The Cu+ might be generated by
thermally removing oxygen atom from CuO lattice during the
calcination
stage
as
Cu2O
species
are
more
thermodynamically stable than CuO species at high
temperature. Only reduced copper species are present on the
surface after reduction, evidenced by the existence of two
peaks at 932.8 eV and 953.0 eV (Fig. 9(B)), assigned to Cu
2p3/2 and Cu 2p1/2 of reduced copper species,
respectively.3,7,14 Consistent with these assignments, the
satellite peaks with Cu 2p binding energy of 943.0 eV assigned
Fig. 7 (A). XRD patterns of calcined samples at room temperature, (B) In situ XRD patterns of reduced catalysts at 190 °C, a. a blank XRD reactor,
b. 5Cu-70, c. 10Cu-70, d. 20Cu-70.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2403–2413 | 2409

View Article Online
Paper
Catalysis Science & Technology
Fig. 8 TEM images of (A) 5Cu-70, (B) 10Cu-70, (C) 20Cu-70, (D) 20Cu-70 after stability test.
to Cu2+ species are absent. The XPS results confirm that
surface Cu2+ species are reduced to Cu0 and/or Cu+ species.
Cu LMM XAES spectra were obtained to further
distinguish the chemical states of copper species (Cu0 and
Cu+) present on the reduced surface by the difference in their
kinetic energies. The co-existence of Cu0 and Cu+ species on
reduced samples is proved by two visible peaks with Auger
kinetic energy of 913.5 eV and 917.5 eV (Fig. 9(C)).3,7,14 Peak
deconvolution is employed to calculate the relative content of
near-surface Cu0 and Cu+ species. The results (Table 2) show
that the surface content of Cu+ species for samples prepared
at 70 °C (5 wt%, 10 wt% and 20 wt%) decreases with the
increasing loading amount of Cu. This might be explained by
the increased difficulty to disperse copper species at higher
loadings. Interestingly, 10 wt% Cu/KIT-6 prepared at 90 °C
shows lower Cu+ content than other 10 wt% Cu/KIT-6
samples prepared at 70 °C, 80 °C, 100 °C, which might be
attributed to the effect of temperature on the growth of
different copper species.
With respect to the support, it not only disperses active
species, but also can react with a certain amount of copper
species to generate the precursor of Cu+ species. In addition,
effective supports should help inhibit the growth of copper
particles, since the Hüttig temperature of copper is low and
sintering is actually the most significant problem for Cu/SiO2
catalysts. Thus, ordered mesoporous silicas with high specific
surface area and well-defined pore structure have been
utilized for the hydrogenation of DMO. Yin et al. prepared
hexagonal mesoporous silica (HMS) support copper catalysts
by an impregnation method.28 They found that HMS
supports with large pore diameters and specific surface areas
facilitated the dispersion of the active copper species and
exhibited large copper surface areas, which then improved
the performance of DMO hydrogenation.28 Ma et al.
fabricated mesoporous MCM-41 supported copper catalysts
for hydrogenation of DMO to EG, and attributed the excellent
performance to the high dispersion of copper species and
large specific surface areas of copper.18 As high metal
dispersion is one of the keys to enhance the hydrogenation
activity of Cu/SiO2 catalysts, ordered mesoporous silicas with
high specific surface area and interconnected mesoporous
channels would be an excellent choice because they may
facilitate the homogenous deposition of copper species
during the preparation stage. Characterizations (XRD, TEM
3.3. Discussion
The optimization of Cu/SiO2 catalysts for the hydrogenation
of DMO to EG has been intensively studied in terms of
promoters for copper species and morphologies of supports.
2410 | Catal. Sci. Technol., 2021, 11, 2403–2413
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
copper loading amount although the synthesized HMS and
MCM-41 have much higher specific surface area.18,28 It may
indicate that the interconnected channels can facilitate the
transfer of copper species during the preparation of catalysts.
As the AE method is widely used to prepare Cu/SiO2
catalysts, it is worthwhile to consider the effect of various
preparation parameters (e.g. loading amount of Cu and pH) on
both the absolute amount and relative content of Cu+ and Cu0
species on the surface as well as the resulting change in the
performance and intrinsic activity of the catalysts. However,
most of the related research focuses on catalysts with support
from colloidal silica. Yin et al. found that the initial
precipitation temperature significantly influenced the textural
properties of Cu/SiO2 catalysts and surface content of copper
species.36 The observed enhancement in the catalytic
performance was attributed to the high dispersion of copper
species and suitable ratio of Cu0/Cu+. Zhao et al. studied
hydrogenation of DMO over ordered mesoporous silica (OMS)
supported copper catalysts.19 They found that appropriate pH
value of the mixture of [Cu(NH3)4]2+ and OMS could preserve
the mesoporous structure and proper loading amount of
copper species could facilitate the dispersion of copper. The
sample with high surface areas of Cu0 and Cu+, and the highest
ratio of Cu+/(Cu0 + Cu+) showed the best performance. In the
present case, the loading amount of copper species exhibits a
more significant effect on the performance of Cu/KIT-6 in
terms of conversion of DMO and selectivity to EG than AE
temperature in the investigated range. It is explained by that
more active sites will be available when increasing the loading
amount of copper.
It is often proposed that hydrogenation of DMO to EG
happens via a synergism between balanced amount of Cu0
and Cu+ species.3,37–40 Poels et al. studied the hydrogenation
of methyl acetate and suggested that Cu0 species were sites
for the activation of hydrogen molecules and the Cu+ species
were active sites for adsorption of the methoxy and acyl
species.38 In this case, the two distinct peaks of the Cu LMM
XAES spectra confirm that both Cu0 and Cu+ species exist on
the surface of all the reduced Cu/KIT-6 catalysts. For Cu/SiO2
catalysts with colloidal silica as support, copper phyllosilicate
is observed over the calcined samples which accounts for the
Cu+ species over the reduced catalysts. However, our study
and the results from other groups indicate that copper
phyllosilicate is not present over the calcined Cu/SiO2
samples with ordered mesoporous silica as the support and
the Cu+ species over the reduced catalysts are proposed to
originate from the reduction of Cu–O–Si unit.18,20 The results
suggest the types of silica influence the formation of copper
phyllosilicate. As Cu+ species are Lewis acid sites, Yin et al.
proposed that the activation of CO bond might be
facilitated on Cu+ species via the electron lone pair on oxygen
atom in the hydrogenation of DMO to EG.20 Although
catalysts with balanced amount of Cu0 and Cu+ sites on the
surface could exhibit excellent performance, it is still not
clear what ratio of Cu+/(Cu+ + Cu0) is best for the catalysts
and whether the surface should be Cu0-rich or Cu+-rich to
Fig. 9 (A) Cu 2p XPS spectra of calcined samples, (B) Cu 2p XPS
spectra and (C) Cu LMM XAES spectra of reduced catalysts, a. 5Cu-70,
b. 10Cu-70, c. 10Cu-80, d. 10Cu-90, e. 10Cu-100, f. 20Cu-70.
and H2-TPR) confirm that the KIT-6 supported copper
catalysts exhibit excellent dispersions of copper. TEM
measurements shows that 20Cu-70 catalyst still maintains a
good dispersion of copper species along the channels after
221 h stability test. Compared with other mesoporous silica
(HMS and MCM-41) supported Cu catalysts, catalysts with KIT-6
as support could offer smaller metallic Cu particles at similar
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2403–2413 | 2411

View Article Online
Paper
Catalysis Science & Technology
Table 2 Deconvolution results of XPS and Cu LMM XAES of reduced catalysts
KE,a eV
Cu+
AP,b eV
Cu+
Cu 2p3/2
BE, eV
Catalyst
Cu0
Cu0
XCu+,c %
Cu0/Cu+d
5Cu-70
913.7
913.8
914.0
913.7
913.7
913.7
917.6
917.6
917.7
917.5
917.7
917.6
1846.6
1846.7
1846.8
1846.4
1846.6
1846.6
1850.5
1850.5
1850.5
1850.2
1850.6
1850.5
932.9
932.9
932.8
932.7
932.9
932.9
86
68
65
56
68
50
0.16
0.47
0.54
0.79
0.47
1.0
10Cu-70
10Cu-80
10Cu-90
10Cu-100
20Cu-70
a
b
c
Kinetic energy. Auger parameter. Ratio of peak areas between Cu+ and (Cu+ + Cu0) by deconvolution of Cu LMM XAES spectra under the
assumption that the Cu+ ions and Cu0 atoms occupy identical areas and have identical atomic sensitivity factors. d Cu0/Cu+ is calculated by (1 −
XCu+)/XCu+
.
achieve high intrinsic activity. For Cu/SiO2 catalysts with
colloidal silica and some types of ordered mesoporous silica
as support, the best performance is usually achieved on
catalysts with the highest Cu+/(Cu+ + Cu0). However, Yin et al.
found that Cu/HMS catalysts with the highest ratio of Cu0/
Cu+ were most active for DMO hydrogenation and exhibited
the highest selectivity to EG.20 Zhao et al. suggested that
surface area of Cu0 species might be the key parameter for
the DMO hydrogenation on Cu/OMS catalysts.19 In our case,
all the 10Cu/KIT-6 samples exhibit roughly the same Cu0
surface area but the highest turnover frequency is achieved
on 10Cu-90 with lowest surface Cu+ content. It is
loading amount of copper are studied to optimize the
performance of DMO hydrogenation on Cu/KIT-6 catalysts. AE
temperature does not influence the hydrogenation
performance achieved over 10 wt%/KIT-6 significantly, but it
has a noticeable effect on the intrinsic activity. The loading
amount of copper species influences both the hydrogenation
performance and intrinsic activity. The effect of AE temperature
and loading amount of copper species on intrinsic activity is
attributed to induced change in the ratio of Cu0/Cu+ on the
surface and high Cu0/Cu+ ratio on the surface favors the
hydrogenation of dimethyl oxalate in this case.
substantiated by 20Cu-70 sample with high surface Cu0/Cu+ Conflicts of interest
ratio which also shows very high intrinsic activity. Thus,
surface Cu0/Cu+ ratio is decisive to the intrinsic activity of
There are no conflicts to declare.
hydrogenation of DMO on Cu/KIT-6 catalysts. It is noted that
the conversion of DMO over 10 wt% Cu/KIT-6 is almost
Acknowledgements
independent on AE temperature. However, the turnover
The project is sponsored by National Science Foundation of
frequency over 10Cu-90 sample is much higher than the
turnover frequency obtained over other 10 wt% Cu/KIT-6
samples. The explanations are as-following. First, it is
uncertain that whether diffusion limitations are ruled out or
not at the catalytic evaluation conditions (T = 190 °C, P(H2) =
2.5 MPa, H2/DMO molar ratio = 95, WHSV = 0.66 h−1).
Second, the hydrogenation of MG to EG also happens over
Cu0 and Cu+ sites and the selectivity to EG is higher over
10Cu-90 sample when comparing with other 10 wt% Cu/KIT-
United States (#CBET-1510157) and National Science
Foundation of United States – Industry/University Cooperative
Research Center for Rational Catalyst Synthesis (#IIP-1464630).
This work made use of the South Carolina SAXS Collaborative.
The authors thank Eric Williams for the help on SAXS.
References
1 H. R. Yue, Y. J. Zhao, X. B. Ma and J. L. Gong, Chem. Soc.
Rev., 2012, 41, 4218–4244.
6
catalysts in the catalytic evaluation. Third, turnover
frequency measures the activity per site, which is not the
same as the results obtained in catalytic evaluation. Above
all, the results suggest that high surface Cu0/Cu+ ratio is
beneficial to the high intrinsic activity of Cu/KIT-6 catalysts
in the hydrogenation of DMO to EG.
2 P. P. McClellan, Ind. Eng. Chem., 1950, 42, 2402–2407.
3 L. F. Chen, P. J. Guo, M. H. Qiao, S. R. Yan, H. X. Li, W.
Shen, H. L. Xu and K. N. Fan, J. Catal., 2008, 257, 172–180.
4 Z. He, H. Q. Lin, P. He and Y. Z. Yuan, J. Catal., 2011, 277, 54–63.
5 L. R. Zehner and R. W. Lenton, US Pat., 4112245, 1978.
6 M. Haruhiko, H. Kouichi, U. Taizou, N. Yasuo, I. Seizou and
T. Takanori, JP Pat., 57123127, 1982.
7 A. Y. Yin, C. Wen, X. Y. Guo, W. L. Dai and K. N. Fan,
J. Catal., 2011, 280, 77–88.
8 W. Qi, L. Qiu, D. Ding, Y. Z. Chen, C. W. Shi, P. Cui, Y.
Wang, Q. H. Zhang, R. Liu and H. Shen, Catal. Commun.,
2018, 108, 68–72.
4. Conclusions
KIT-6 supported copper catalysts are prepared via ammonia
evaporation (AE) method. The high specific surface area and
inter-connected mesoporous channels of the support facilitate
the dispersion of copper species along the channels. The
particles still maintain a high dispersion even after nine days
during the stability test. The effect of AE temperature and
9 X. L. Zheng, H. Q. Lin, J. W. Zheng, X. P. Duan and Y. Z.
Yuan, ACS Catal., 2013, 3, 2738–2749.
2412 | Catal. Sci. Technol., 2021, 11, 2403–2413
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
10 Y. Huang, H. Ariga, X. L. Zheng, X. P. Duan, S.
Takakusagi, K. Asakura and Y. Z. Yuan, J. Catal.,
2013, 307, 74–83.
11 J. W. Zheng, H. Q. Lin, Y. N. Wang, X. L. Zheng, X. P. Duan
and Y. Z. Yuan, J. Catal., 2013, 297, 110–118.
12 A. Y. Yin, J. W. Qu, X. Y. Guo, W. L. Dai and K. N. Fan, Appl.
Catal., A, 2011, 400, 39–47.
25 C. M. A. Parlett, A. Aydin, L. J. Durndella, L. Frattini, M. A.
Isaacs, A. F. Lee, X. T. Liu, L. Olivi, R. Trofimovaite, K.
Wilson and C. F. Wu, Catal. Commun., 2017, 91, 76–79.
26 R. Guillet-Nicolas, R. Ahmad, K. A. Cychosz, F. Kleitz and M.
Thommes, New J. Chem., 2016, 40, 4351–4360.
27 R. Kishor and A. K. Ghosha, Microporous Mesoporous Mater.,
2017, 242, 127–135.
13 Y. N. Wang, X. P. Duan, J. W. Zheng, H. Q. Lin, Y. Z. Yuan,
H. Ariga, S. Takakusagi and K. Asakura, Catal. Sci. Technol.,
2012, 2, 1637–1639.
14 X. B. Yu, T. A. Vest, N. Gleason-Boure, S. G. Karakalos, G. L.
Tate, M. Burkholder, J. R. Monnier and C. T. Williams,
J. Catal., 2019, 380, 289–296.
28 A. Y. Yin, X. Y. Guo, W. L. Dai, H. X. Li and K. N. Fan, Appl.
Catal., A, 2008, 349, 91–99.
29 Z. Liu, M. D. Amiridis and Y. Chen, J. Phys. Chem. B,
2005, 109, 1251–1255.
30 C. J. G. Van Der Grift, A. Mulder and J. W. Geus, Appl. Catal.,
1990, 60, 181–192.
15 H. J. Chae, T. W. Kim, Y. K. Moon, H. K. Kim, K. E. Jeong,
C. U. Kim and S. Y. Jeong, Appl. Catal., B, 2014, 150,
596–604.
16 J. Liang, Z. Liang, R. Zou and Y. Zhao, Adv. Mater., 2017, 29,
1701139.
31 A. J. Marchi, J. L. G. Fierro, J. Santamaría and A. Monzon,
Appl. Catal., A, 1996, 142, 375–386.
32 S. Zhao, H. R. Yue, Y. J. Zhao, B. Wang, Y. C. Geng, J. Lv, S. P.
Wang, J. L. Gong and X. B. Ma, J. Catal., 2013, 297, 142–150.
33 J. Ding, T. Popa, J. K. Tang, K. A. M. Gasem, M. H. Fan and
Q. Zhong, Appl. Catal., B, 2017, 209, 530–542.
34 J. X. Chen, Y. Yang, H. Shi, M. F. Li, Y. Chu, Z. Y. Pan and
X. B. Yu, Fuel, 2014, 129, 1–10.
35 X. B. Yu, J. X. Chen and T. Y. Ren, RSC Adv., 2014, 87,
46427–46436.
36 A. Y. Yin, X. Y. Guo, W. L. Dai and K. N. Fan, Catal.
Commun., 2011, 12, 412–416.
37 J. W. Zheng, J. F. Zhou, H. Q. Lin, X. P. Duan, C. T. Williams
and Y. Z. Yuan, J. Phys. Chem. C, 2015, 119, 13758–13766.
38 E. K. Poels and D. S. Brands, Appl. Catal., A, 2000, 191,
83–96.
39 Y. Wang, Y. L. Shen, Y. J. Zhao, J. Lv, S. P. Wang and X. B.
Ma, ACS Catal., 2015, 3, 6200–6208.
40 X. P. Wang, M. Chen, X. K. Chen, R. H. Lin, H. J. Zhu, C. Q.
Huang, W. S. Yang, Y. Tan, S. S. Wang, Z. N. Du and Y. J.
Ding, J. Catal., 2020, 383, 254–263.
17 X. Y. Guo, A. Y. Yin, W. L. Dai and K. N. Fan, Catal. Lett.,
2009, 132, 22–27.
18 X. B. Ma, H. W. Chi, H. R. Yue, Y. J. Zhao, Y. Xu, J. Lv, S. P.
Wang and J. L. Gong, AIChE J., 2013, 59, 2530–2539.
19 Y. J. Zhao, Y. Q. Zhang, Y. Wang, J. Zhang, Y. Xu, S. P. Wang
and X. B. Ma, Appl. Catal., A, 2017, 539, 59–69.
20 A. Y. Yin, X. Y. Guo, W. L. Dai and K. N. Fan, J. Phys. Chem.
C, 2009, 113, 11003–11013.
21 S. Schünemann, G. Dodekatos and H. Tüysüz, Chem. Mater.,
2015, 27, 7743–7750.
22 F. Kleitz, S. H. Choi and R. Ryoo, Chem. Commun.,
2003, 2136–2137.
23 K. Soni, B. S. Rana, A. K. Sinha, A. Bhaumik, M. Nandi, M.
Kumar and G. M. Dhar, Appl. Catal., B, 2009, 90, 55–63.
24 B. Zhou, C. Y. Li, N. Qi, M. Jiang, B. Wang and Z. Q. Chen,
Appl. Surf. Sci., 2018, 450, 31–37.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2403–2413 | 2413