G Model
CATTOD-9790; No. of Pages7
ARTICLE IN PRESS
2
R. van den Berg et al. / Catalysis Today xxx (2015) xxx–xxx
[30,31]. In impregnation, a support is contacted with a metal solu-
reduction [32]. The interface area between the metal particles and
the support is dependent on the preparation route [33]. It might
be larger in case of precipitation due to incomplete segregation
[34–36]. During reaction the interface area might change due to a
change in particle size, particle shape or support structure [28,37].
To study the effect of the synthesis method on the
as a model system. In such catalysts, the activity for CO/CO2 hydro-
genation scales with the copper surface area, and the loss of activity
during reaction is often mainly due to copper particle growth
[11,38,39]. Moreover, recent developments in the synthesis of well-
defined copper on silica catalysts allow careful tuning of the size
and distribution of the copper particles [40,41]. In this work Cu/SiO2
catalysts with similar copper particle sizes and local copper weight
loadings were synthesized via impregnation and via precipitation.
Electron tomography was used to characterize the SiO2 support
structure and the location of the Cu particles within the support.
The performance in the methanol synthesis reaction of both cata-
lysts was investigated at 40 bar at 260 ◦C for a period of 10 days.
The catalysts were retrieved after reaction to determine the extent
of copper particle growth and change in support structure.
that underwent reduction in a flow of 100 ml/min of 20% H2 in Ar
at 250 ◦C (2 ◦C/min) for 2½ h. After the reduction treatment, the
resulting Cu/SiO2 was passivated for 15 min by slowly exposing
the sample to diluted air/N2 at room temperature. The sample was
stored in a glove box under argon atmosphere.
2.3. Characterization
N2-physisorption measurements were performed at −196 ◦C
using a Micromeritics Tristar 3000 apparatus. The BET method was
used to calculate the specific surface areas. The pore volumes were
determined at p/p0 = 0.983. Pore size distributions were determined
from the adsorption branch by the BJH method [45]. X-ray diffrac-
tion (XRD) was performed with a Bruker-Nonius D8 Advance X-ray
˚
diffractiometer using Co-K␣12 (ꢀ = 1.79026 A) radiation. Diffrac-
tograms were collected at room temperature from 20◦ to 70◦ (2ꢁ).
either after reduction and passivation or catalysis and passivation
and was subsequently sealed. Copper crystallite sizes were esti-
mated by applying the Debye–Scherrer equation to the (1 1 1)
diffraction of Cu (2ꢁ = 50.5◦, k = 0.9) [46]. H2-TPD was performed
on the catalysts before and after methanol synthesis [47,48]. Mea-
surements were carried out in a fixed-bed flow setup and online
gas analysis was performed by a quadrupole mass spectrometer
(Balzers GAM 445). The copper phyllosilicate was reduced in situ
at 250 ◦C (2 ◦C/min) in 100 ml/min of 1% H2/He. At 250 ◦C the gas
flow was changed to 100 ml/min of 100% H2 for 2½ h. The calcined
copper on silicagel and the samples after catalysis and passivation
were reduced in situ at 220 ◦C (2 ◦C/min) in a flow 100 ml/min of 1%
H2/He for 7 h. Then the samples were flushed for ½ h in 100 ml/min
of He to remove any adsorbed hydrogen and subsequently cooled
down to 0 ◦C. At this temperature the reactor was pressurized to
1.6 MPa of H2, before it was cooled down further to −33 ◦C. After
½ h, the catalyst was rapidly cooled to −196 ◦C and depressurized.
The gas flow was changed to 100 ml/min of He to flush out the
excessive hydrogen. After 1 h the temperature was increased up
to 210 ◦C (6 ◦C/min) and the H2 desorption profile was recorded.
Copper surface areas (SACu, H2 ) were calculated from the amount
of desorbed hydrogen assuming a H/Cu ratio of 0.5 and 1.47 × 1019
copper surface atoms per m2. Temperature programmed reduc-
tion (TPR) was performed using an Autochem II ASAP 2910 from
micromeritics. The H2 concentration during the experiment was
measured with a thermal conductivity detector. About 0.05 g of
sample was put on top of a quartz wool bed in a glass reactor tube
and subsequently heated to 500 ◦C (5 ◦C/min) under a (50 ml/min)
flow of 5% H2/Ar. The copper loading was estimated from the H2
consumption by assuming the following reduction stoichiome-
try: CuO + H2 → Cu + H2O. Transmission electron microscopy (TEM)
images were acquired with a Tecnai 12 (FEI) microscope operated
at 120 kV with a pixel size of 0.45 nm. Precipitated Cu/SiO2 samples
before and after catalytic tests were prepared by grinding followed
by sonication in ethanol. A droplet of the ethanol suspension was
deposited on a carbon coated copper TEM grid (Agar S162 200 Mesh
Cu). The resolution and contrast were sufficient to detect parti-
cles larger than 2 nm (4 × 4 pixels). The impregnated Cu/SiO2 was
ground, embedded in a two component epoxy resin (Epofix, EMS)
and cured at 60 ◦C overnight, and cut into thin sections (50–100 nm)
using a Diatome Ultra 35◦ diamond knife mounted on an Ultra-
cut E microtome (Reichert-Jung). Sections were deposited on a
TEM grid. The resolution and contrast were sufficient to detect and
measure particles larger than 3 nm (6 × 6 pixels). Copper surface
2. Material and methods
2.1. Synthesis of Cu/SiO2 via precipitation
Precipitated Cu/SiO2 was synthesized via a modified method
of van der Grift et al. [42]. 20.1 g LUDOX-AS 30 (Sigma–Aldrich,
30 wt% SiO2), 16.1 g Cu(NO3)2·3H2O (Acros Organics, 99% for analy-
sis) and 12.1 g Urea (Acros Organics, 99.5% for analysis) were added
to 1.7 l of demineralized water in a 2 l reaction vessel. The pH was
adjusted to 2–3 with a few drops of HNO3 (Merck, 65% for analysis)
to prevent premature hydrolysis of copper nitrate. The suspen-
sion was then heated to 90 ◦C over 1 h under stirring. At 90 ◦C the
hydrolysis of urea led to an increase in pH resulting in precipita-
tion of Cu2(NO3)(OH)3. The well-stirred reaction vessel was kept
at 90 ◦C for 7 days to allow recrystallization of precipitated copper
and silica, which resulted in the formation of copper phyllosilicate
[43,44]. The precipitate was obtained by hot filtration of the sus-
pension and washed three times with demineralized water, filtered
and dried overnight at 60 ◦C. The yield was 10 g. Part of the as-
prepared copper phyllosilicate was used for catalytic reaction (with
in situ reduction to obtain metallic Cu on SiO2) as described below.
Another part (2 g) intended for characterization was reduced at
250 ◦C (2 ◦C/min) in a flow of 100 ml/min of 20% H2 in Ar for 2½ h.
After the reduction treatment the resulting Cu/SiO2 was passivated
for 15 min by slowly exposing the sample to diluted air/N2 at room
temperature. The sample was stored in a glove box under argon
atmosphere.
2.2. Synthesis of Cu/SiO2 via impregnation
Copper was deposited on a commercial silica gel (Davicat 1454,
Grace-Davison, pore volume p/p0 < 0.95 = 0.81 cm3/g, pore diam-
eter 9 nm) via incipient wetness impregnation with an aqueous
solution of 2 M Cu(NO3)2·3H2O (Acros Organics, 99% for analy-
sis) and 0.1 M HNO3 (Merck, 65% for analysis), followed by drying
overnight under vacuum at room temperature and calcination in
375 ml/min (GHSV = 15,000 h−1) of 2% NO/N2 at 350 ◦C (2 ◦C/min)
for 1 h. Part of the as-prepared sample was used for catalytic reac-
tion (with in situ reduction to obtain metallic Cu on SiO2) as
described below and a portion (2 g) intended for characterization
areas (SACu, TEM) were calculated from the particle size distribu-
tions obtained via TEM, and assuming spherical particles. Energy
dispersive X-ray (EDX) spectroscopy was performed on a Technai
20FEG (FEI) electron microscope equipped with a field emission
gun and an EDAX Super Ultra Thin Window EDX detector and
Please cite this article in press as: R. van den Berg, et al., Impact of the synthesis route of supported copper catalysts on the performance