An environmentally benign and low-cost approach to synthesis of thermally stable industrial catalyst Cu/SiO2 for the hydrogenation of dimethyl oxalate to ethylene glycol

Article abstract of DOI:10.1016/j.apcata.2015.07.026

This investigation focused on developing an environmentally benign approach to synthesis of nanostructured Cu/SiO₂ catalysts for hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG). It was accomplished by using (NH₄)₂CO₃, instead of the conventional ammonia evaporation method, for making nanostructured Cu/SiO₂ exhibiting good performance for DMO conversion and EG selectivity in the desired temperature range, with increased temperature stability. The resulting catalysts were characterized for catalytic activity, optimal reaction temperature, specific surface area, crystalline structure, surface composition and the influence of sodium in support raw material. The new method generated active catalysts with improved thermal stability linked to better dimension control and more Cu₂O content, together with improved morphology of the supported copper particles.

Full text of DOI:10.1016/j.apcata.2015.07.026

Applied Catalysis A: General 505 (2015) 52–61
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
An environmentally benign and low-cost approach to
synthesis of thermally stable industrial catalyst Cu/SiO2 for the
hydrogenation of dimethyl oxalate to ethylene glycol
Tiberiu Popaa, Yulong Zhangb, Erlei Jina, Maohong Fana,c,d,∗
a Department of Chemical and Petroleum Engineering, University of Wyoming, 1000 East University Avenue, Laramie 82071, WY, USA
b Western Research Institute, 3474 North 3rd Street, Laramie 82072, WY, USA
c School of Energy Resources, University of Wyoming, 1000 East University Avenue, Laramie 82071, WW, USA
d School of Civil and Environmental Engineering, Georgia Institute of Technology, Mason Building, 790 Atlantic Drive, Atlanta 30332, GA, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 1 May 2015
Received in revised form 17 July 2015
Accepted 20 July 2015
Available online 26 July 2015
This investigation focused on developing an environmentally benign approach to synthesis of nanostruc-
tured Cu/SiO2 catalysts for hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG). It was
accomplished by using (NH4)2CO3, instead of the conventional ammonia evaporation method, for mak-
ing nanostructured Cu/SiO2 exhibiting good performance for DMO conversion and EG selectivity in the
desired temperature range, with increased temperature stability. The resulting catalysts were character-
ized for catalytic activity, optimal reaction temperature, specific surface area, crystalline structure, surface
composition and the influence of sodium in support raw material. The new method generated active cata-
lysts with improved thermal stability linked to better dimension control and more Cu2O content, together
with improved morphology of the supported copper particles.
Keywords:
Dimethyl oxalate hydrogenation catalyst
Ethylene glycol
Cu/SiO2
© 2015 Elsevier B.V. All rights reserved.
Ammonium carbonate
The resulting EG can be further hydrogenated/dehydrated to
ethanol:
1. Introduction
Large variations in oil prices have recently led to many eco-
nomic issues for petrochemical industries. Upcoming possible CO2
emissions regulations [1–3] may aggravate these issues. There-
fore, seeking alternative raw materials for production of chemicals
has become imperative, among these being coal, natural gas and
biomass [4]. Syngas produced from these resources can be used
for producing ethylene glycol (EG) in two steps [5]: CO oxida-
tive coupling to dimethyl oxalate (DMO) [4] and hydrogenation of
DMO to EG.
HO-CH2-CH2-OH + H2 → CH3-CH2-OH + H2O
ꢁH0 = −87.20 kJ/mol
(R3)
In addition, other products might be generated, such as 1,2-
butanediol, ethane [6,8], etc.
The key to the second step of EG production is to find a high-
performance and environmentally friendly catalyst, and thus avoid
the use of the established copper/chromia (Cu/Cr2O3) catalyst.
Although there are Ru-ligands homogenous catalysts with good
catalytic activity, the separation of the catalysts from the products
of reaction can be difficult [8–11], thus heterogeneous catalysts are
preferred.
The reaction generating ethylene glycol (EG) occurs in two major
steps [6,7], having as an intermediary methyl glycolate (MG):
CH3OOCCOOCH3 + 2H2 → CH3OOCCH2OH + CH3OH
As catalysts used for DMO hydrogenation to EG, supported Au,
Ag and Au-Ag catalysts have been developed, also with remarkable
activity. However, because they can have 10 wt.% Ag (Ag/MCM41
[12]) and for Au-Ag combination 8 wt.% Au and 4.5 wt.% Ag (Au-
Ag/SBA15 [13]), these might not be the best choice as industrial
catalysts, particularly if a less expensive option is present, such as
Cu/SiO2 catalysts.
As catalysts used for DMO hydrogenation directly to ethanol,
successful were Cu/Al2O3 prepared using sol–gel methods [14], Cu
phyllosilicates prepared by ammonia evaporation hydrothermal
ꢀH0 = −30.03 kJ/mol
(R1)
(R2)
CH3OOCCH2OH + 2H2 → HOCH2CH2OH + CH3OH
ꢁH0 = −28.70 kJ/mol
∗
Corresponding author at: Department of Chemical and Petroleum Engineering,
University of Wyoming, 1000 East University Avenue, Laramie, WY 82071, USA.
Tel.: +1 307 766 5633/404 385 4577.
E-mail addresses: mfan@uwyo.edu, mfan3@mail.gatech.edu (M. Fan).
http://dx.doi.org/10.1016/j.apcata.2015.07.026
0926-860X/© 2015 Elsevier B.V. All rights reserved.

T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
53
(AEH) method [15] and copper-phyllosilicate core-sheath nanore-
actor using hydrothermal methods [16].
1–2. The precipitation pH 6–7 (6.8 [27]) is obtained via urea
decomposition by heating at ∼80–90 ◦C, followed by hot filtra-
tion, precipitate washing with large quantities of water [28], or
re-slurrying of filter cake and re-filtration [27]. The second tech-
nique uses urea decomposition [29,30], starting from basic pH,
of a solution of copper nitrate with added ammonia, together
with silica sol and urea, heating at 80 ◦C for four hours until the
formation of the precipitate, which is again followed by hot filtra-
tion. However, the details of the mechanism of the decomposition
process are complex and not well understood [31,32]. Further,
the urea decomposition reaction is relatively slow and contin-
ues as well as during the filtration step, thus the process requires
immediate filtration while in the preferred end-pH range, fol-
lowed by very good washing process to remove urea [27,28]. For
larger sizes filter cakes, this can generate uneven copper con-
centrations, and the structures generated also demonstrate large
deactivation (with DMO conversions as low as 60% [29]) after
90 h [19] to 100 h [29] of reactions at 190 ◦C.
This research is aimed at designing a method that can be
scaled up for synthesis of Cu/SiO2 as an industrial catalyst for
dimethyl oxalate (DMO) hydrogenation to ethylene glycol (EG).
Numerous synthesis methods were examined to select the most
suitable catalysts, using several selection criteria; among these
criteria are current and announced environmental regulations, reg-
ulations governing the transport and storage of the raw materials
(e.g. ammonia), the number and content effluents requiring treat-
ment, the practicality of the method and ease of synthesis, the
ability for automation, and the activity and selectivity of the catalyst
samples.
The conventional Cu/SiO2 synthesis procedure includes forma-
tion of tetraamino cupric complex by mixing copper nitrate with
ammonia solutions to pH ∼11, the addition of silica sol to the result-
ing solution, and ammonia evaporation (AE), achieved by boiling
the suspension until the pH reaches 6–7 for the precipitation of
copper compounds [9,10]. This method was developed in order
to generate smaller crystalline structures and a wider coverage of
Cu compounds on the support, compared with incipient wetness
impregnation methods [7,17], and became the common method of
Cu compounds deposition on support [7,10,11,18–21]. However,
this hydrothermal method has a number of shortcomings:
– Other methods attempted for the synthesis of ester hydro-
genation catalysts involved precipitation using Na2CO3, NaOH
[33–35], NaHCO3, and KOH [35];
– (NH4)2CO3 was tried as coprecipitation agent for Cu-Zn-Si
oxides/hydroxides system at pH = 8, but the product had low Cu
coverage due to the formation of copper complex (Cu(NH3)42+).
In addition, it was stated that “(NH4)2CO3 is not a suitable pre-
cipitant for copper ions and Cu2Zn1Si1-(NH4)2CO3 catalyst is not
worth studying” [35], which is surprising because urea hydroly-
sis at the end of the complicated mechanism generates precisely
NH3 and CO2 [27,29–32].
– The whole conventional catalyst preparation process takes four
hours of mixing before the ammonia evaporation step, and
then, although evaporation in our lab conditions is faster due
to the lower local atmospheric pressure caused by altitude
(∼780 mbar), another three hours at 90 ◦C was required, followed
by a filtration and washing step, which required another hour.
– The quality of the catalyst is very sensitive to boiling temperature
of the ammonia evaporation operation. For example, Chen et al.
obtained three different catalyst morphologies in the 60–100 ◦C
range [10] and Yin et al. described the influence of temperature
on texture and Cu species distribution [18].
– The process requires constant supervision and cannot be accel-
erated because of the risk of hot spots associated with water
evaporation and overflow of the viscous suspensions.
– The method generates fumes with high concentration of ammo-
nia during the ammonia evaporation step [9,22], which definitely
poses problems for large batches, and the waste gas stream needs
to be treated in order to limit its ammonia concentrations below
25 or 50 ppm [23].
The methods in use tend to be fairly complicated and sensitive
to various influences. Consequently, a less complicated approach
to the synthesis of the Cu/SiO2 catalyst is desirable for scale-up and
industrial utilization. The catalyst synthesis process should:
(1) Have the ability for automation.
(2) Generate effluents needing treatment that are low in number
and toxicity.
(3) Generate catalysts having good performance on DMO conver-
sion, good EG, MG and ethanol selectivity over the desired
temperature range, with increased temperature stability.
2. Experimental
– The hydrogenation of DMO, pure Cu/SiO2 (i.e., without stabi-
lizer additives) prepared with ammonia evaporation methods,
showed marked deactivation after ∼16 h at 200 ◦C [7,9,11], thus
a method producing more stable structures is desirable.
2.1. Catalyst preparation
The catalysts were prepared following multiple routes. Their
designation contains the percent of copper in the fresh (unreduced
and unused) catalyst, the raw material used as source of silica and
the process used for precipitation. For example, a catalyst con-
taining 20 wt.% copper using ammonia stabilized silica sol AS30
and using ammonium carbonate as synthesis route was designated
20Cu-AS30-AC. The deposition precipitation using ammonium
carbonate and ammonia evaporation are described below; the
methods using tetraethoxysilane (TEOS) through sol–gel approach
are placed in the electronic supplementary material section.
Deposition–precipitation was employed using (NH3)2CO3 as
precipitation agent (deposition using ammonium carbonate – AC):
7.6 g Cu(NO3)2·xH2O (Puratronic®, 99.999% – metals basis, Alfa
Aesar) were dissolved in 26.25 g water and the pH was cor-
rected to pH = 1 using nitric acid 69.0–70.0% (J.T. Baker “Nitric
Acid, 69.0–70.0%, BAKER INSTRA-ANALYZED® Reagent”). Next, the
appropriate quantities of support generating substances: silica sol
(24.90 g LUDOX® AS-30 colloidal silica 30% ammonia stabilized
or 25.15 g LUDOX®AM-30 colloidal silica 30%, sodium stabilized,
There are several possible alternatives to the ammonia evapo-
ration method:
– Incipient wetness impregnation of silica, using various soluble
copper compounds. However, this method generates less active
catalysts, due to large distributions of the copper particles [17]
and Cu particles agglomeration at higher loading [18].
– Sol–gel methods give active catalysts, but they tend to be slow,
sensitive to other influences [24,25], and preparation methods
are complex, reducing reproducibility [18].
– Cationic exchange methods [17] and chemisorption hydrolysis
[26] leading to only 11–12 wt.% Cu loading.
– An environmentally friendly method appears to be precipitation
using urea decomposition. This method employs two main tech-
niques. The first one starts at low pH [27,28], using fine silica
or silica sol as sources of SiO2 and a solution of copper nitrate,
where the starting pH can be adjusted with nitric acid [27] to

54
T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
silica in water, Sigma-Aldrich) were added and stirred for 30 min.
Then (NH3)2CO3 (ammonium carbonate, ACS reagent, NH3>30.0+%,
Sigma-Aldrich) saturated solution was added slowly to end-pH 6–7.
The resulting suspension containing the precipitate continued to be
stirred for 4 h for crystals maturation, then it was filtered, washed
and dried at 89 ◦C for 10 h, and the resulting materials were calcined
in air at 450 ◦C for four hours. Finally, after cooling in desiccator,
the catalyst was crushed and sieved, and the fractions 125–250 ␮m
were used for catalytic testing; the resulting catalysts were denom-
inated 20Cu-AS30-AC and 20Cu-AM30-AC, respectively.
Deposition–precipitation using NH3 (ammonia evaporation-
AE): 7.6 g of Cu(NO3)2·xH2O (Puratronic®, 99.999% – metals basis,
Alfa Aesar) were dissolved in 26.25 g water, followed by adding
ammonia solution (Ammonia, 30%, Baker) to pH = 10–11 and stir-
ring another 30 min resulting in a blue solution. Then 24.90 g silica
sol (Silica sol LUDOX® AS-30 colloidal silica 30% silica in water,
ammonia stabilized, Sigma-Aldrich) were added and stirred for four
more hours. After this period the temperature was raised to 90 ◦C
for the evaporation of ammonia until pH 6–7, when precipitation
occurred. The precipitates were filtered, washed, dried at 89 ◦C for
10 h and the resulting materials were calcined in air at 450 ◦C for
four hours. Finally, after cooling in a desiccator, they were crushed
and sieved, and the fractions 125–250 ␮m were used for catalytic
testing. The resulting catalyst was denominated 20Cu-AS30-AE.
analysis for particles in one layer in samples led to the dimensions
of the particles; the shape factor used was circularity fcirc = 4ꢂA/P2
where A is the area of the shape and P is the perimeter of the shape
(for a perfect circle this has value 1); the equivalent diameter de is
ꢀ
Feret’s diameter de
=
P/ꢂ where P is the perimeter of the shape.
2.3. Catalytic activity tests
The hydrogenation of dimethyl oxalate was performed on the
experimental set-up shown in the schematic drawing Fig. S1. The
system has four main gas lines (H2, N2, Ar and optional gas,
shown at the top left corner of the drawing) with flow rates
controlled by corresponding mass flow controllers, and a liquid
injection line for DMO 10 wt.% in methanol (Dimethyl oxalate,
99%, VWR and Methanol, Certified ACS >99.8%, Fisher Chemical)
with flow rate regulated by a high pressure pump (LabAlliance
Series 1). H2 (5.0 UHP, US Welding) was the reaction gas, N2
(5.0 UHP, US Welding) was the carrier gas, while He (5.0 UHP,
US Welding) or Ar (5.0 UHP, US Welding) served as internal gas
standard. The reaction gases were supplied from pressurized gas
cylinders by high pressure reducers, regulated by high pressure
mass flow controllers (Parker-Porter), and then mixed with the
vaporized liquid. The catalyst was synthesized in oxide form and
required reduction four hours at 350 ◦C in H2 before being used
as catalyst for DMO hydrogenation. The DMO hydrogenation reac-
tion occurred in a ½ inch fixed bed reactor with 11 mm inner
diameter, and its temperature was set by a controller regulated
furnace. The catalyst was introduced into the reactor without
dilution with quartz sand and it was supported by inert mate-
rial (ceramic wool). Gaseous products were analyzed by online
gas chromatograph SRI 8610C, SRI Instruments, equipped with
two Restek packed columns 1/8 in. × 6 in. (3.175 mm × 182.88 cm)
Hayesep D and 1/8 in. × 6 in. (3.175 mm × 182.88 cm) MolSieve 13×
and a capillary column 0.53 mm ID × 60 m × 5 ␮ Restek MXT-1, with
a standard error ∼5%, while liquid products were collected and
analyzed by offline GC–MS Agilent 7890A-5975C VL MSD (mass
spectrometer) using an Agilent HP-5MS (30 m × 0.250 mm) capil-
lary column, equipped with Automatic Liquid Sampler (ALS) and
standard error ∼2%. The reaction system was controlled by a com-
puter through a National Instruments DAQ system.
Test procedure: 1.51 g catalyst corresponding to 0.3 g Cu was
loaded into the reactor with the thermocouple in the middle of the
catalyst bed. The top and bottom of the catalyst bed was packed
with inert material (ceramic wool). After leak check the system
was pressurized with N2 to the reaction pressure for additional
leak check and the catalyst reduced in 92.6 vol.% H2 and 7.4 vol.%
N2 at 350 ◦C for 4 h. After reduction step, the pressure and temper-
ature were set to test parameters and 10 wt.% DMO/methanol was
introduced. During testing, the pressure was maintained at speci-
fied pressure and the lowest test temperature of 170 ◦C (above the
vaporization temperature of DMO), and the maximum tempera-
ture was established as selectivity to EG and MG decreased to less
than ∼30%, typically 240 ◦C, while 20Cu-AM30-AC was tested up
to 260 ◦C. The test conditions were: catalyst mass corresponding to
0.3 g Cu, pressure 2 or 3 MPa, temperature at least 170 ◦C, weight
hourly space velocity (WHSV) 0.8 (g DMO/(h*g cat.)), 92.6 vol.% H2,
7.4 vol.% N2, H2/DMO mol ratio 80/1, 10 wt.% DMO/methanol. Gas
products were analyzed by inline GC and liquid products were col-
lected from sample cylinder at specified intervals (typically every
hour) and analyzed by offline GC–MS.
2.2. Characterization of catalysts
The total metal loading of the catalysts was analyzed by
inductively coupled plasma spectrometry (ICP) using PerkinElmer
Optima 8300 ICP-OES Spectrometer.
To obtain specific surface area of the selected catalyst, N2
adsorption isotherms were performed on a volumetric system
AutosorbIQ ASIQC0100-4 Quantachrome Instruments, using N2
(5.0 UHP, US Welding) and He (5.0 UHP, US Welding) after a
degassing procedure at 150 ◦C for 4 h.
X-ray diffraction (XRD) tests were performed on a Rigaku
Smartlab X-ray diffractometer system in thin layer powder con-
figuration, 2Theta/Theta mode, 20–100◦ range, 0.02◦ step, 1◦/min
scanning speed. The software used for refinements was Crystal-
Sleuth/American Mineralogist Crystal Structure Database [36].
X-ray Photoelectron Spectroscopy (XPS) and X-Ray-Excited
Auger Electron Spectroscopy (XAES) data was collected using
a Kratos Axis Ultra DLD X-ray Photoelectron Spectrometer. It
employs a monochromated Al K-alpha source running at 150 W.
Survey scans were acquired at 80 eV pass energy at 1 eV resolu-
tion. High resolution elemental scans were acquired at 40 eV pass
energy at a resolution of 0.05 eV. The reduced and used samples
were stored in vials under Ar.
Temperature programmed reduction tests were performed on
a AutosorbIQ ASIQC0100-4, Quantachrome Instruments, equipped
with thermal conductivity detector (TCD), using a mixture of 5 vol.%
H2 (5.0 UHP, US Welding)–95 vol.% N2 (5.0 UHP, US Welding) and
He (5.0 UHP, US Welding). After one hour at 110 ◦C in He, the sample
was cooled to 50 ◦C, then the gas was changed to 5 vol.% H2–95 vol.%
N2 and, after stabilization, the temperature was raised at 5 ◦C/min
to above 900 ◦C (typically to 950 ◦C). The mass of the sample was
calculated to contain ∼10 mg Cu, thus the mass of 20Cu-AS30-AC,
20Cu-AM30-AC and 20Cu-AS30-AE was ∼0.05 g, equivalent sam-
ples containing 10 wt.%Cu/SiO2 introduced had a mass of ∼0.1 g.
The morphology and the particle size as well as the dispersion
of the catalysts were studied by transmission electron microscopy
(TEM) using FEI- Tecnai G2 F20 S-Twin 200 kV equipment. Samples
for TEM observations were prepared by dispersing the catalysts in
ethanol and drying one drop of the solution on copper grids. The
particle size distribution was calculated on monolayer particles by
using ImageJ 1.48v [37] and Icy 1.5.3.0 [38] software. The image
After analysis of the liquid fraction using GC–MS, data was used
to generate values for DMO conversion, selectivity and yield values
to EG, MG, ethanol, and also molar flow rates of the analyzed prod-
ucts. From reactions (R1) to (R3) it can be observed that 2 mol H2 are
consumed to hydrogenate 1 mol DMO to produce 1 mol MG, 4 mol

T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
55
20Cu-AS30-AC-fresh
20Cu-AM30-AC-fresh
20Cu-AS30-AE-fresh
SiO2-Cristobalite
CuO-Tenorite
H2 are consumed for 1 mol DMO generating 1 mol EG and 5 mol H2
for 1 mol DMO is required to generate 1 mol ethanol. Knowing the
flow rate of DMO it means catalytic activity of the hydrogenation
catalyst can be expressed in terms of (mol H2/(h*g cat)) used to gen-
erate EG, MG and ethanol thus eliminating the influence of other
side-reactions (thermal decompositions, intra and inter-molecular
dehydrations, etc.) in the analysis of the hydrogenation catalysts.
Cu silicate
Cu2O-Cuprite
3. Results and discussion
3.1. Characteristics of the catalysts
20
40
60
80
100
One important issue of earlier attempts to use ammonium car-
bonate was the large loss in copper during precipitation step,
making the concentration of copper compounds very low in com-
parison with the other methods [35]. Table 1 shows the results of
ICP-OES for metals loading and it demonstrates that, when using
ammonium carbonate, the precipitate copper content is similar to
the sample prepared by the ammonia evaporation method. The
concentrations of Cu in prepared catalysts are similar with those
reported in literature for other methods [10,18,28].
BET analyses (Table 2) show that the catalyst obtained using
ammonia evaporation has the highest surface area for the fresh
catalyst (334.817 m2/g), but after ∼16 h at high temperature (4 h,
350 ◦C reduction in H2 then above 170 ◦C in DMO-H2 stream) the
catalyst loses 35.45% from the initial specific surface area, while in
the case of the AC precipitation the catalyst 20Cu-AS30-AC loses
28.01% under the same conditions, while 20Cu-AM30-AC starts
with the lowest specific surface area (150.0 m2/g) and also loses
22.4% of the initial specific surface area.
2θ
Fig. 1. XRD spectra for fresh 20Cu-AS30-AC, 20Cu-AM30-AC and 20Cu-AS30-AE.
signifying less crystallinity compared with the other two samples.
The large peak centered near 23◦ can be assigned to poorly crys-
talized SiO2 structures [21,24,39]. All three samples display this
peak, while the rest of the larger peaks displayed by the sam-
ples can be assigned to CuO (tenorite) [21,24,39] and some small
peaks can be assigned to copper silicates [10]. As can be seen
in Fig. 1, the crystallinity of the samples is in order from more
crystalline to less crystalline: 20Cu-AS30-AC-fresh > 20Cu-AM30-
AC-fresh > 20Cu-AS30-AE-fresh. In the case of the catalyst prepared
using the Na+ stabilized silica sol, the XRD spectra suggests it
depressed the proper crystallization of tenorite (CuO) during pre-
cipitation.
Fig. S4 and S5 illustrate the spectra of the XRD test for reduced
and for used catalysts. The large peaks at ∼43.4◦, 50.6◦ and 74◦
can be assigned to metallic Cu (Cu0), while the peak at 36.5◦ and
the wide peak centered around 61.5◦ can be assigned to cuprite
(Cu2O as mineral-cuprite has largest peak at 36.46◦ and other large
peaks at 42.35◦, 61.44◦ and 73.6◦). Compared to the peaks gener-
ated by metallic Cu, the cuprite peaks are much wider, signifying
a poorly crystalline and small crystal domains phase. Comparing
spectra of the catalysts, it can be seen that in the case of the catalyst
prepared using ammonia evaporation, the peaks for Cu0 are much
smaller, compared with the catalyst prepared using ammonium
carbonate precipitations, signifying smaller crystalline size, while
the peaks assigned to cuprite tend to be much larger. The catalyst
prepared using Na+ stabilized silica sol (AM30) displays initially a
larger cuprite wide peak at ∼36.5◦ after reduction, but after use it
shows peaks similar to the peaks generated by the other catalyst
prepared using ammonium carbonate precipitation.
The values generated using Scherrer’s equation (Table 3) for all
catalysts tested illustrate larger Cu0 particles for the reduced and
used catalyst compared with the CuO particles of the fresh cata-
lyst suggesting a sintering process. Cu2O peaks were very broad
and could not be analyzed, which underlines the difficulty of XRD
analysis for systems having mixed crystalline with almost amor-
phous/vitreous structures, where the amorphous fraction does not
have enough regularly repeated atomic layers to generate by inter-
ference significant XRD signal, and reveals that for Cu2O component
XRD might not be the most suitable technique.
XRD patterns obtained for the unreduced (fresh), reduced and
used catalyst show the growth of metal crystalline phase for the
used catalysts, and confirms sintering as one of the deactivation
mechanisms. Fig. 1 shows the XRD patterns obtained for the unre-
duced (fresh) samples as they resulted after calcination in air for
4 h at 450 ◦C in the form intensity vs. 2ꢃ. The sample obtained
by ammonia evaporation technique using the ammonia stabi-
lized silica sol (20Cu-AS30-AE-fresh) displays less evident peaks,
Table 1
Total loading as determined by ICP-OES.
Sample
Cu (wt.%)
Na (wt.%)
Si (wt.%)
Cu/Si
(wt/wt)
20Cu-AS30-AC fresh
18.37
18.16
18.67
19.08
18.14
18.38
0.03
0.05
0.14
0.14
0.02
0.02
35.97
36.08
35.73
35.55
36.12
35.98
0.5
0.5
0.5
0.5
0.5
0.5
20Cu-AS30-AC used 200 h
20Cu-AM30-AC fresh
20Cu-AM30-AC used 200 h
20Cu-AS30-AE fresh
20Cu-AS30-AE used 16 h
Table 2
Specific surface areas for the fresh and used catalysts.
Sample
BET surface
area (m2/g)
BJH total pore
BJH average
pore diameter
(nm)
volume (cm3/g)
20Cu-AS30-AC
fresh
20Cu-AM30-AC
fresh
20Cu-AS30-AE
fresh
20Cu-AS30-AC
used
20Cu-AM30-AC
used
20Cu-AS30-AE
used
215.6
150.0
334.8
155.2
116.4
216.1
0.509
0.873
1.187
0.420
0.720
0.902
9.52
12.25
3.04
Temperature programmed reduction (TPR-H2) test results for
the fresh (unreduced and unused) samples are shown in Fig. 2.
For the sample 20Cu-AS30-AE it shows only one peak at 240 ◦C,
corresponding to reduction of the finely disseminated CuO [10,26],
while the sample containing half the metal coverage, but having the
same quantity of metal (∼10 mg Cu) in the sample introduced in the
test equipment, shows also one relatively sharp peak but at lower
temperature (226 ◦C) In the case of 20Cu-AS30-AC there are two
features, one peak representing a maximum of H2 consumption at
12.17
12.25
3.40

56
T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
Table 3
Crystal dimensions determined using Scherrer equation.
Catalyst
20Cu-AS30-AC
Fresh
20Cu-AM30-AC
20Cu-AS30-AE
Fresh
Reduced
Used
Fresh
Reduced
Used
Reduced
Used
Particle
Size (nm)
CuO
13.98
Cu0
31.90
Cu0
31.92
CuO
8.80
Cu0
23.52
Cu0
24.82
CuO
4.71
Cu0
3.45
Cu0
3.52
20Cu-AS30-AC
20Cu-AS30-AE, AC based catalysts have sharper peaks, meaning
more compact and better crystallized CuO in the case of fresh cat-
alysts.
10Cu-AS30-AC
20Cu-AM30-AC
10Cu-AM30-AC
20Cu-AS30-AE
10Cu-AS30-AE
This translates into better crystallized Cu0 phases after reduc-
tion, a process that is slightly depressed by Na+ in 20Cu-AM30-AC.
However, XPS results show that in reduced and used catalysts the
majority of copper is Cu+, with Cu2+ also having a significant frac-
tion, likely as copper silicates. Additionally, XPS data show that,
for AC prepared catalysts, the effect of the partial decomposi-
tion of copper silicate and the following copper migration during
reduction offsets sintering. However, after being used, the copper
coverage decreases again, but only at about 72.9% of the original
coverage for 20Cu-AS30-AC and just at 96.5% of the original cover-
age for 20Cu-AM30-AC.
0
200
400
600
800
Temperature (oC)
As can be seen from TEM images (Fig. 4), the catalysts obtained
using ammonia evaporation (AE) differ in morphology from both
silica support particles (light gray) and also in the deposited cop-
per containing material (dark gray), from the catalysts generated
using ammonium carbonate precipitation. The supported material
morphology in 20Cu-AS30-AE fresh sample is widely distributed on
support at nanometer size and largely featureless, which prevented
obtaining proper images required for the particle identification.
Although XRD is unreliable for poorly crystalized materials as
is 20Cu-AS30-AE fresh sample, a value of 4.71 nm diameter was
obtained for CuO using the peak at 2ꢃ = 58.7◦. The material cover-
ing the silica starts aggregating during reduction after four hours
at 350 ◦C in ∼3.3 nm structures, and it will accumulate in masses
of uneven size and an equivalent mean value diameter de of 3.0 nm
for the used catalysts. In the case of the AC samples, the supported
material is concentrated in smaller regions of almost round shape
with de = 2.8 nm, and the structures become 2.9 nm after reduction,
as sintering occurs. The catalyst used in-stream shows aggregates
of around 3.9 nm. The morphology of the support differs; while AC
samples have rounded supporting silica structures, the AE sam-
ples were elongated, some slab-like, but, after use, the morphology
changes into more round structures. It is possible that this behavior
is due to the partial decomposition of phyllosilicate during reduc-
tion and reconstruction of the structures of copper oxides, silica
and again phyllosilicate during use or because of the substance
loss due to the reaction with methanol generating the observed
tetramethoxysilane [10].
Fig. 2. TPR of the 20Cu-AS30-AC, 10Cu-AS30-AC, 20Cu-AM30-AC, 10Cu-AM30-AC,
20Cu-AS30-AE and 10Cu-AS30-AE (∼10 mg Cu, 5 vol.% H2 balance N2, 5 ◦C/min).
232 ◦C, corresponding to the reduction of finely disseminated CuO
on support [10,26], and a shoulder at 248 ◦C, corresponding to the
reduction of bulk CuO in the structure [10,26] while 10Cu-AS30-
AC has one peak at 242 ◦C. The sample 20Cu-AM30-AC shows one
peak centered at 249 ◦C with a possible small shoulder at 236 ◦C
while 10Cu-AM30-AC has the main peak at 226 ◦C and a shoulder
at 244 ◦C.
XPS/XAES tests results illustrated in Fig. 3 give more informa-
tion about the composition to an analysis depth of a maximum
20 nm [40]. The peaks centered at 933 eV binding energy (Cu2p),
representing Cu in terms of Cu wt.% on the surface (Fig. 3a), can
be deconvoluted into peaks representing Cu2+, but the peaks posi-
tion (binding energy) for the Cu0 and Cu1+ are almost identical [30].
To obtain the concentration values for Cu0 and Cu1+, the XAES Cu
LMN spectra were used (Fig. 3b), and the results of the combined
XPS/XAES tests are shown in Table 4. The surface of calcined 20Cu-
AS30-AE has 27.76 Cu wt.% as 23.76 wt.% Cu2+, but also 4 wt.% Cu+,
suggesting that some tetraamino cupric material was encapsulated
in silica gel [24] and during calcination its decomposition generated
a reductive environment, but the value obtained by XPS is larger
than the value obtained from ICP suggesting that most of copper
compounds are on the surface of the support. Secondly, the disper-
sion of the total copper on the silica support drops by almost 50%
from initial value (13.89 wt.%) after the reduction step at 350 ◦C in
hydrogen, although the decomposition of copper silicate and cop-
per migration during reduction should make available more copper
materials [10]. Thirdly, after an additional ∼8 h in-stream the cov-
erage becomes only 46.1% of the original coverage (12.79 wt.%)
because of sintering. In contrast, for the AC prepared catalysts, in
the case of fresh 20Cu-AS30-AC, Cu coverage (11.51 wt.%) is lower
than the value obtained using ICP suggesting that a large fraction
is below the surface, also copper is more reduced after calcina-
tion step suggesting encapsulation, a similar case being observed
for 20Cu-AM30-AC-fresh where part of total Cu (10.77 wt.%) could
be attributed also to more reduced copper. Copper coverage for
20Cu-AS30-AC catalyst can be estimated at only 41.5% of the Cu
wt.% coverage of the AE prepared catalyst, meaning that Cu con-
centrates in three-dimensional structures instead of layers, which
is in agreement with TEM. Further, XRD shows that compared to
3.2. Effect of different factors on dimethyl oxalate hydrogenation
3.2.1. Pressure and temperature
For this research, published and new methods were used to gen-
erate 20 wt.% Cu/SiO2 catalysts, which were tested using the same
method using catalyst mass corresponding to 0.3 g Cu and reactor
pressure: 2 MPa (Fig. S2 and S3). The activity of the catalysts was
analyzed as function of the DMO conversion and catalytic activity
toward the economically desirable EG, MG and ethanol.
In terms of DMO conversion as a function of temperature, the
catalyst prepared using AS-30 silica sol (ammonia stabilized) as
support and precipitation using ammonium carbonate at end-pH
6–7 was very active, and maintained good activity after 14–16 h
on-stream, while most of the unstabilized AE prepared Cu/SiO2
catalysts in the literature reported loss of activity after 10–16 h
on-stream [7,9,11]. A catalyst with the same raw material as

T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
57
Cu2+
A)
Cu1+/0
a)
B)
Cu0
Cu1+
a)
b)
b)
c)
d)
e)
c)
f)
d)
e)
g)
h)
i)
f)
960
940
920
906
911
916
921
Kinetic Energy(eV)
Binding Energy(eV)
Fig. 3. (A) XPS Cu 2p spectra of the catalysts; (B) XAES Cu LMN spectra of the reduced and used catalysts; (a) 20Cu-AS30-AC used, (b) 20Cu-AM30-AC used, (c) 20Cu-AS30-AE
used, (d) 20Cu-AS30-AC reduced, (e) 20Cu-AM30-AC reduced, (f) 20Cu-AS30-AE reduced, (g) 20Cu-AS30-AC fresh, (h) 20Cu-AM30-AC fresh, (i) 20Cu-AS30-AE fresh.
Table 4
Atom coverage (wt.%) on the catalyst surfaces resulting from XPS (Cu 2p peak) deconvolution.
wt.%
Catalyst
20Cu-AS30-AC
Fresh
20Cu-AM30-AC
Fresh
20Cu-AS30-AE
Fresh
Reduced
Used
Reduced
Used
Reduced
Used
Cu2+
5.79
4.97
0.63
2.56
7.87
2.62
1.80
3.55
3.07
8.39
5.46
4.43
0.88
4.26
4.95
3.74
2.37
6.58
1.45
23.76
4.00
0.00
3.21
6.09
4.59
3.03
6.78
2.98
Cu1+
Cu0
Total Cu(2p)
11.51
13.05
10.77
12.95
10.40
27.76
13.89
12.79
20Cu-AS30-AC, but having end-pH 5.5 showed lower conversion,
signifying the optimal end-pH is 6–7; above pH = 7 there were Cu
loses in filtrate.
Plotting the result in terms of catalytic activity to desirable
products (mol H2/(h*g cat.)), the tests at 2 MPa shows that 20Cu-
AS30-AC at 210 ◦C is 9.56% more active than 20Cu-AS30-AE at its
highest activity (200 ◦C) while 20Cu-AM30-AC at 210 ◦C is 5.18%
less active than the AE prepared catalyst at 200 ◦C.
purposes the catalyst prepared using ammonia evaporation
(20CuAS30-AE), were the object of further investigations.
The evaluation tests of the catalysts prepared with both AC and
AE methods were performed with a set-up similar with the installa-
tion shown in Fig. S1, but equipped for 3 MPa. The conditions were
the same as used previously, but a pressure of 3 MPa was employed.
Fig. 5 shows that the performance of the catalysts prepared with
conventional AE method and the new AC method are significantly
different. The DMO conversions achieved with 20Cu-AS30-AE
catalyst reached 90% or higher percentages when reaction temper-
atures were above 200 ◦C, Further, at 190 ◦C, the DMO conversion
of 20Cu-AS30-AC is already as high as 96.4%, higher than that of
20Cu-AS30-AE, and with a yield of EG of 76.36%, compared with
the value of 28.86% in the case of 20Cu-AS30-AE. 20Cu-AS30-AC is
substantially improved over that prepared using AE, both in terms
of DMO conversion and EG and MG selectivity. Furthermore, 20Cu-
AS30-AC also exhibits an increase in selectivity to EG over a wider
temperature range. The catalyst 20Cu-AM30-AC has lower conver-
sion at the same temperature, but it also shows reduced ethanol
production. It reaches a maximum yield of EG at 230 ◦C which may
be attributed to the 0.1% sodium content in the silica sol used to gen-
erate the support, pointing to the role of the support in modulating
the catalytic reaction [19].
Above 220 ◦C byproducts were identified as traces of methyl
acetate, methylal, 1-propanol, 2-butanol, 2-methyl-1-butanol,
acetic acid, methoxy-methyl ester, 1,2-butanediol etc. As temper-
ature increased above 200 ◦C, tetramethoxysilane (TMOS) was also
detected in liquid phase, but as the test advanced, TMOS quantity
decreased, although still within the detection limit of the GC–MS.
As the temperature increased above 200 ◦C, the gaseous products
increasingly showed CO and CO2 from 210 ◦C, a sign of the onset
of decomposition, excessive hydrogenation/dehydration (ethane,
methane), molecular and/or intermolecular dehydration (ethylene,
dimethyl ether).
The results of these preliminary tests showed that catalysts
prepared using ammonia carbonate precipitation and silica sol AS-
30 performed very well, both in terms of EG and MG selectivity
and DMO conversion. The catalysts precipitated using ammonium
carbonate at end-pH 6–7 using ammonia stabilized silica sol (20Cu-
AS30-AC) was selected, to test the effect of increased quantities of
sodium in the structure the catalyst prepared using sodium sta-
bilized AM-30 colloidal silica (20Cu-AM-AC) and for comparison
3.2.2. Thermal stability
According to literature [11,19,29], pure Cu/SiO2 (i.e., without
stabilizer additives) for the hydrogenation of DMO, prepared with

58
T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
Fig. 4. TEM images of Cu/SiO2 catalysts (a) 20Cu-AS30-AC; (b) 20Cu-AM30-AC; (c) 20Cu-AS30-AE: (1) fresh, (2) reduced, (3) used.
ammonia evaporation methods, showed markedly deactivation
after ∼16 h at 200 ◦C [7,9,11]. In addition, Cu/SiO2 prepared though
urea decomposition had considerable deactivation (with DMO con-
versions as low as 60% [29]) after 90 [19] to 100 [29] hours of
reaction at 190 ◦C. A thermal stability test for 20Cu-AS30-AC was
performed at 2 MPa and 210 ◦C for 200 h, and is illustrated in
Figs. 6 and 7. Scattering of the results is attributed to the pres-
sure variation associated with collecting samples, but the behavior
of the catalyst remained stable. The result of the test (Fig. 6) indi-
cates that the conversions of DMO with the catalyst 20Cu-AS30-AC
prepared in this research only decrease from 99.5% to 97.5%, even
after 200 h, and the selectivity to EG decreases from 85.5% only
to 79%. The better performance of 20Cu-AS30-AC is attributed to
the compact nanosized crystalline copper structures containing a
large fraction of Cu+ and low Cu coverage between Cu regions,
which can reduce Cu migration and thus the deactivation of the
catalyst due to sintering [33]. As can be observed, the catalytic
activity, expressed as mol H2/(h*g cat.) for the production of EG,
MG and EtOH (Fig. 7), remains almost constant, mainly because
as the test advanced the quantity of ethanol decreased in favor
of the production of MG, while EG selectivity remained almost
constant.
In the case of the 20Cu-AM30-AC, the test for thermal deactiva-
tion was also performed, at the temperature of maximum activity
230 ◦C and 2 MPa and is illustrated in Figs. 8 and 9. Although the
catalyst had very good DMO conversion during the entire test, as
presented in Fig. 8, the catalyst started showing signs of deacti-
vation after only ∼85 h in-stream, when it also showed increased
sensitivity to the pressure variation in the test installation associ-
ated with collecting samples, resulting in an increased scattering
of the results.
The deactivation is more pronounced as it is analyzed in terms
of catalytic activity toward EG, MG and ethanol expressed as mol
H2/(h*g cat) and illustrated in Fig. 9, which shows not only a delay
in reaching maximum activity, but also a pronounced loss of more
than 17% from the maximum activity after 200 h.
Both catalysts produced using AC method performed better than
the catalyst produced using AE method, which was also tested at
210 ◦C and 2 MPa (Figs. 10 and 11), which, as expected [7,9,11],
displays the onset of deactivation marked by the decrease in DMO
conversion after 15 h at 210 ◦C.
Weight liquid hourly space velocity (WLHSV) tests in the range
0.2–4.2 h−1 were performed at 3 MPa and 210 ◦C for 20Cu-
AS30-AC (Figs. 12 and 13), and 3 MPa and 230 ◦C in the case of

T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
59
100
80
60
40
20
0
100
80
60
EG- 20Cu-AS30-AC
3 MPa 20Cu-AS30-AC
3MPa 20Cu-AM30-AC
3MPa 20Cu-AS30-AE
MG- 20Cu-AS30-AC
40
EG- 20Cu-AM30-AC
MG- 20Cu-AM30-AC
20
EG- 20Cu-AS30-AE
MG- 20Cu-AS30-AE
0
170 190 210 230 250 270 290
Temperature ( oC)
170 190 210 230 250 270 290
Temperature ( oC)
▪10-2
3
2
1
0
100
80
60
40
20
0
EG- 20Cu-AS30-AC
20Cu-AS30-AC
20Cu-AM30-AC
20Cu-AS30-AE
MG- 20Cu-AS30-AC
EG- 20Cu-AM30-AC
MG- 20Cu-AM30-AC
EG- 20Cu-AS30-AE
MG- 20Cu-AS30-AE
170
190
210
230
250
270
290
170 190 210 230 250 270 290
Temperature ( oC)
Temperature ( oC)
Fig. 5. The performances of the 20 wt.% Cu/SiO2 catalysts prepared with conventional and new methods (a) selectivity to EG and MG; (b) DMO conversion; (c) yield of EG and
MG; (d) catalytic activity to EG + MG + Ethanol [catalyst mass: 0.3 g Cu, gas condition (pressure: 3 MPa; WHSV: 0.8 g DMO/(h*g cat.); H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO
mol ratio: 80/1].
·10-2
3
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
DMO conversion
Selectivity to EG
Selectivity to MG
Selectivity to Ethanol
2
1
0
DMO conversion
mol H2 /(h*g cat)
0
50
100
Time (hours)
150
200
0
50
100
150
200
Time (hours)
Fig. 6. Stability test of 20Cu-AS30-AC [catalyst mass: 0.3 g Cu, gas condition (pres-
sure: 2 MPa; WHSV: 0.8 g DMO/(h*g cat.); H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol
ratio: 80/1, 210 ◦C].
Fig. 7. Stability test of 20Cu-AS30-AC as DMO conversion and catalytic activity to
EG + MG + ethanol [catalyst mass: 0.3 g Cu, gas condition (pressure: 2 MPa; WHSV:
0.8 g DMO/(h*g cat.); H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol ratio: 80/1, 210 ◦C].
20Cu-AM30-AC (Figs. 14 and 15). As can be seen from figures, the
catalysts performed better at higher values of WLHSV due to the
reduction in over-hydrogenation of the EG produced connected to
the production of ethanol. At WLHSV 4.2 h−1 the DMO conversion
was still 100% in the case of both catalysts prepared using ammo-
nium carbonate methods and selectivity to EG reached 94.22% at
3.9 h−1 in the case of 20Cu-AS30-AC and 97.2% at 4.2 h−1 using
20Cu-AM30-AC. Analysing the hydrogenation catalysts in term of
catalytic activity toward EG, MG and ethanol (mol H2/(h*g cat))
both 20Cu-AS30-AC (Fig. 13) and 20Cu-AM30-AC (Fig. 15) show
a surprisingly straight line, signifying additional unused catalytic
capabilities, but possibly at lower EG selectivity.
The method using ammonium carbonate, besides being much
faster and with results only tied to end-pH, gives characteristics
more desirable for industrial use, for which this class of catalysts
is aimed. The method also produced comparably more active cat-
alysts. The analysis of the material is somewhat surprising. TPR
suggests the presence of a CuO bulk phase for the more active
20Cu-AS30-AC, while only one peak for the similar catalyst pre-
pared using AE. The XRD also shows the presence of well crystalized
CuO in the case of fresh (unused and unreduced) 20Cu-AS30-AC,
with tall and narrow Cu peaks for both the reduced and used cat-
alyst. Different morphology, compared with the catalyst prepared
using ammonia evaporation, is exposed by TEM. There is more free

60
T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
100
80
60
40
20
0
100
·10-2
100
80
3
2
1
0
80
60
40
20
0
60
40
20
0
DMO conversion
Selectivity to EG
Selectivity to MG
Selectivity to
Ethanol
DMO conversion
mol H2 /(h*g cat)
0
5
10
Time (hours)
15
20
0
50
100
150
200
Time (hours)
Fig. 11. Stability test of 20Cu-AS30-AE as DMO conversion and catalytic activity to
EG + MG + ethanol [catalyst mass: 0.3 g Cu, gas condition (pressure: 2 MPa; WHSV:
0.8 g DMO/(h*g cat.); H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol ratio: 80/1, 210 ◦C].
Fig. 8. Stability test of 20Cu-AM30-AC [catalyst mass: 0.3 g Cu, gas condition (pres-
sure: 2 MPa; WHSV: 0.8 g DMO/(h*g cat.); H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol
ratio: 80/1, 230 ◦C].
100
80
60
40
20
0
100
80
60
40
20
0
·10-2
100
80
60
40
20
0
3
2
1
0
DMO conversion
Selectivity to EG
Selectivity to MG
Selectivity to Ethanol
DMO conversion
mol H2 /(h*g cat)
0
50
100
150
200
0
1
2
3
4
5
Time (hours)
WLHSV (DMO g/(h*g cat.))
Fig. 9. Stability test of 20Cu-AM30-AC as DMO conversion and catalytic activ-
ity to EG + MG + ethanol [catalyst mass: 0.3 g Cu, gas condition (pressure: 2 MPa;
WHSV: 0.8 g DMO/(h*g cat.); H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol ratio: 80/1,
230 ◦C].
Fig. 12. WLHSV test of 20Cu-AS30-AC [catalyst mass: 0.3 g Cu, gas condition (pres-
sure: 3 MPa; H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol ratio: 80/1, 210 ◦C].
·10-2
100
80
60
40
20
0
15
10
5
100
80
60
40
20
0
100
80
60
40
20
0
DMO conversion
Selectivity to EG
Selectivity to MG
Selectivity to Ethanol
DMO conversion
mol H2 /(h*g cat)
0
5
10
Time (hours)
15
20
0
0
1
2
3
4
5
WLHSV (DMO g/(h*g cat))
Fig. 10. Stability test of 20Cu-AS30-AE [catalyst mass: 0.3 g Cu, gas condition (pres-
sure: 2 MPa; WHSV: 0.8 g DMO/(h*g cat.); H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol
ratio: 80/1, 210 ◦C].
Fig. 13. WLHSV test of 20Cu-AS30-AC as DMO conversion and catalytic activity to
EG, MG and ethanol [catalyst mass: 0.3 g Cu, gas condition (pressure: 3 MPa; H2:
92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol ratio: 80/1, 210 ◦C].
support surface available for the AC prepared catalysts, while the
AE prepared catalysts have a surface largely occupied by layers of
copper and copper compounds (oxides and copper silicates). The
more compact structure of AC prepared material seems to ensure
a smaller loss of the initial surface area, 28.01% of the original sur-
face compared with 35.45% in case of the AE prepared. In the case
of the catalyst 20Cu-AM30-AC, although it is also prepared using
AC method, the presence of sodium seems to generate a structure
which has a less compact crystallization for copper compounds.
This will result in catalyst having a lower initial specific surface, but
also having a 22.4% loss of the initial value of specific surface area. In
the case of AE prepared catalyst, for which the intention is to create
a widespread layer of Cu compounds, thus large surface area avail-
able for reaction comparable with the surface of the support, the
TEM images show that the main cause of deactivation occurs by sin-
tering of the layer into more compact masses containing Cu [9]. The
cause of deactivation for the catalysts prepared using AC method
it is more complex. Although they are more thermally stable, as
illustrated by the long term tests and TEM images, the detection of

T. Popa et al. / Applied Catalysis A: General 505 (2015) 52–61
61
100
80
60
40
20
0
100
80
60
40
20
0
Acknowledgements
The authors gratefully acknowledge the support of the U.S.
Department of Energy and the State of Wyoming. They also appre-
ciate the great help the members of Dr. Fan’s research group that
was provided during this research.
DMO conversion
Selectivity to EG
Selectivity to MG
Selectivity to Ethanol
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.apcata.2015.07.
026
0
1
2
3
4
5
References
WLHSV (DMO g/(h*g cat.))
[1] Y. Kong, G. Jiang, M. Fan, X. Shen, S. Cui, A.G. Russell, Chem. Commun. 50 (2014)
12158–12161.
[2] Y. Kong, Xi. Shen, S. Cui, M. Fan, Green Chem. (2015), http://dx.doi.org/10.1039/
C5GC00354G
Fig. 14. WLHSV test of 20Cu-AM30-AC [catalyst mass: 0.3 g Cu, gas condition (pres-
sure: 3 MPa; H2: 92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol ratio: 80/1, 230 ◦C].
[3] S. Cui, Y. Liu, M. Fan, A.T. Cooper, B. Lin, X. Liu, G. Han, X. Shen, Mater. Lett. 65
(2011) 606–609.
[4] E. Jin, Y. Zhang, L. He, B. Teng, M. Fan, Appl. Catal. A 476 (2014) 158–174.
[5] E. Jin, L. He, Y. Zhang, A.R. Richard, M. Fan, RSC Adv. 4 (2014) 48901–48904.
[6] H. Yue, Y. Zhao, L. Zhao, J. Lv, S. Wang, J. Gong, X. Ma, AIChE J. 58 (9) (2012)
2798–2809.
[7] C. Wen, Y. Cui, A. Yin, K. Fan, W.-L. Dai, ChemCatChem 5 (2013) 138–141.
[8] A. Yin, X. Guo, W.-L. Dai, H. Li, K. Fan, Appl. Catal. A 349 (2008) 91–99.
[9] X. Zhang, B. Wang, Y. Guo, G. Xu, J. Fuel Chem. Technol. 39 (9) (2011) 702–705,
ISSN: 0253-2409 CN:14-1140/TQ.
[10] L.-F. Chen, P.-J. Guo, M.-H. Qiao, S.-R. Yan, H.-X. Li, W. Shen, H.-L. Xu, K.-N. Fan,
J. Catal. 257 (2008) 172–180.
[11] J. Lin, X. Zhao, Y. Cui, H. Zhang, D. Liao, Chem. Commun. 48 (2012) 1177–1179.
[12] A. Yin, C. Wen, W.-L. Dai, K. Fan, Appl. Catal. B 108–109 (2011) 90–99.
[13] J. Zheng, H. Lin, Y. Wang, X. Zheng, X. Duan, Y. Yuan, J. Catal. 297 (2013) 110–118.
[14] Y. Zhu, X. Kong, X. Li, G. Ding, Y. Zhu, Y.-W. Li, ACS Catal. 4 (2014) 3612–3620.
[15] J. Gong, H. Yue, Y. Zhao, S. Zhao, L. Zhao, J. Lv, S. Wang, X. Ma, J. Am. Chem. Soc.
134 (2012) 13922–13925.
[16] H. Yue, Y. Zhao, S. Zhao, B. Wang, X. Ma, J. Gong, Nat. Commun. 4 (2013)
2339–2346.
[17] T. Toupance, M. Kermarec, C. Louis, J. Phys. Chem. B 104 (2000) 965–972.
[18] A. Yin, X. Guo, W.-L. Dai, K. Fan, Catal. Commun. 12 (2011) 412–416.
[19] H. Lin, X. Zheng, Z. He, J. Zheng, X. Duan, Y. Yuan, Appl. Catal. A 445–446 (2012)
287–296.
[20] F. Li, C.-S. Lu, X.-N. Li, Chin. Chem. Lett. 25 (2014) 1461–1465.
[21] C. Wen, Y. Cui, W.-L. Dai, S. Xie, K. Fan, Chem. Commun. 49 (2013) 5195–5197.
[22] Y. Zhu, S. Wang, L. Zhu, X. Ge, X. Li, Z. Luo, Catal. Lett. 135 (2010) 275–281.
[23] NIOSH Pocket Guide to Chemical Hazards, DHHS (NIOSH) Publication No. 2005-
149, 3’rd Printing, NIOSH Publications, Cincinnati, OH, 2007, pp. 15.
[24] L. Lin, P. Pan, Z. Zhou, Z. Li, J. Yang, M. Sun, Y. Yao, Chin. J. Catal. 32 (2011)
957–969.
[25] L. He, X. Chen, J. Ma, H. He, W. Wang, J. Sol-Gel Sci. Technol. 55 (2010) 285–292.
[26] F. Zaccheria, N. Scotti, M. Marelli, R. Psaro, N. Ravasio, Dalton Trans. 42 (2013)
1319–1328.
[27] C.J.G. Van Der Grift, P.A. Elberse, A. Mulder, J.W. Geus, Appl. Catal. 59 (1990)
275–289.
[28] S. Wang, X. Li, Q. Yin, L. Zhu, Z. Luo, Catal. Commun. 12 (2011) 1246–1250.
[29] Y. Huang, H. Ariga, X. Zheng, X. Duan, S. Takakusagi, K. Asakura, Y. Yuan, J. Catal.
307 (2013) 74–83.
·10-2
100
80
60
40
20
0
15
10
5
DMO Conversion
mol H2 /(h*g cat)
0
0
1
2
3
4
5
WLHSV (DMO g/h g cat)
Fig. 15. WLHSV test of 20Cu-AM30-AC as DMO conversion and catalytic activity to
EG, MG and ethanol [catalyst mass: 0.3 g Cu, gas condition (pressure: 3 MPa; H2:
92.6 vol.%; N2: 7.4 vol.%); H2/DMO mol ratio: 80/1, 230 ◦C].
tetramethoxysilane (TMOS) in the product liquid phase indicates
dissolution of the support, while the larger presence of Cu+ in the
structure as revealed by XPS/XAES seem to lead to an increase in the
thermal stability of the copper structures [33]. According to Wen
et al. [21], this type of deactivation by the transformation of the sil-
ica support in TMOS is specific to the DMO hydrogenation systems
where methanol is used as solvent for DMO, but is improbable for
the systems where ethanol is used.
[30] Z. He, H. Lin, P. He, Y. Yuan, J. Catal. 277 (2011) 54–63.
[31] A. Lundström, B. Andersson, L. Olsson, Chem. Eng. J. 150 (2009) 544–550.
[32] A.N. Alexandrova, W.L. Jorgensen, J. Phys. Chem. B 111 (4) (2007) 720–730.
[33] Z. Huang, F. Cui, H. Kang, J. Chen, X. Zhang, C. Xia, Chem. Mater. 20 (2008)
5090–5099.
[34] H.-M. Chen, Y.-L. Zhu, G.-Q. Ding, H-.Y. Zheng, Y.-W. Li, J. Fuel. Chem. Technol.
39 (2011) 519–526.
[35] K. Zhong, X. Wang, Int. J. Hydrogen Energy 39 (2014) 10951–10958.
[36] T. Laetsch, R.T. Downs, Abstracts from the 19th General Meeting of the Inter-
national Mineralogical Association, Kobe, 2006.
[37] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, Nat. Methods 9 (2012) 671.
[38] F. De Chaumont, S. Dallongeville, N. Chenouard, N. Hervé, S. Pop, T. Provoost,
V. Meas-Yedid, P. Pankajakshan, T. Lecomte, Y. Le Montagner, T. Lagache, A.
Dufour, J.C. Olivo-Marin, Nat. Methods 9 (2012) 690–696.
[39] S. Wang, W. Guo, H. Wang, L. Zhu, S. Yin, K. Qiu, New J. Chem. 38 (2014)
2792–2800.
4. Conclusion
The ammonium carbonate precipitation method for synthe-
sis of nanostructured Cu/SiO2 catalyst for the hydrogenation of
dimethyl oxalate (DMO) to ethylene glycol (EG) was successfully
developed. It is quicker, safer, and less expensive for preparing
nano Cu/SiO2 catalyst for DMO hydrogenation to EG, an impor-
tant commercial chemical. The AC based nano Cu/SiO2 catalysts
have similar Cu content, compared to the AE prepared catalyst, but
they are characterized by more compacted nanosized crystalline
copper structures with more Cu2O content, which lead to the bet-
ter DMO conversion and EG selectivity and more thermally stable
structures.
[40] A.G. Shard, Surf. Interface Anal. 46 (2014) 175–185.