Synthesis and characterization of supported copper phyllosilicate catalysts for acetic ester hydrogenation to ethanol

Article abstract of DOI:10.1016/j.apcata.2015.10.026

Supported copper phyllosilicate (Cu/SiO₂) catalysts were prepared by a simple co-precipitation method (CP), and compared with the urea hydrolysis deposition-precipitation method (DP) in terms of phase evolution during synthesis, physicochemical properties, chemical states of copper species after reduction, and catalytic performance in ester hydrogenation to ethanol. Catalysts were characterized by N₂-physisorption, X-ray powder diffraction (XRD), IR spectroscopy, H₂ temperature programmed reduction (TPR), transmission electron microscopy (TEM), in situ X-ray photoelectron spectroscopy (XPS), and FTIR/CO spectroscopy. It was confirmed that copper phyllosilicates species existed in all catalysts but the exact form differed in the Cu-CP catalysts and the Cu-DP catalysts. While the Cu-DP catalyst favored the form of 1:1 copper phyllosilicate, the Cu-CP catalysts tended to have more species of 2:1 copper phyllosilicate which led to only small amount of active copper species on surface after reduction and thus low ester conversion. Addition of ammonium ion and zinc ion during the CP method synthesis promoted the copper silicate phase evolution from 2:1 copper phyllosilicate to 1:1 copper phyllosilicate, and the reactivity of ester hydrogenation was greatly enhanced. Conversion of 99.67%, selectivity of 99.07%, and space-time yield of 1.23 g-EtOH/(g-cat·h) were achieved at 250 °C using a Cu-Zn-NCP catalyst.

Full text of DOI:10.1016/j.apcata.2015.10.026

Accepted Manuscript
Title: Synthesis and characterization of Supported copper
phyllosilicate catalysts for Acetic Ester Hydrogenation to
Ethanol
Author: Wei Di Jihong Cheng Shuxun Tian Jing Li Jingyun
Chen Qi Sun
PII:
S0926-860X(15)30203-9
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.10.026
APCATA 15602
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
23-7-2015
14-10-2015
18-10-2015
Please cite this article as: Wei Di, Jihong Cheng, Shuxun Tian, Jing Li, Jingyun
Chen, Qi Sun, Synthesis and characterization of Supported copper phyllosilicate
catalysts for Acetic Ester Hydrogenation to Ethanol, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2015.10.026
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Synthesis and characterization of Supported
copper phyllosilicate catalysts for Acetic Ester
Hydrogenation to Ethanol
Wei Di, Jihong Cheng*,Shuxun Tian, Jing Li, Jingyun Chen, Qi Sun.
National Institute of Clean-and-Low carbon Energy, Beijing, P. R.China
1
*
02211
Corresponding author.
E-mail address: chengjihong@nicenergy.com. Phone: +86-10-57339629

Graphical abstract
Highlights:




Supported copper phyllosilicate catalysts were successfully synthetized by a simple co-
precipitation method.
The formation mechanism of copper phyllosilicates during co-precipitation was
proposed.
The co-precipitation method favored the form of the 2:1 copper phyllosilicate rather than
the 1:1 copper phyllosilicate
Addition of ammonium ion and zinc ion drastically promoted copper silicate phase
evolution from the 2:1 copper phyllosilicate to the 1:1 copper phyllosilicate during
precipitation.


Supported 1:1 copper phyllosilicate catalysts showed superior activity for acetic ester
hydrogenation to ethanol.
Abstract
Supported copper phyllosilicate (Cu/SiO ) catalysts were prepared by a simple co-
2
precipitation method (CP), and compared with the urea hydrolysis deposition-precipitation
method (DP) in terms of phase evolution during synthesis, physicochemical properties,
chemical states of copper species after reduction, and catalytic performance in ester
hydrogenation to ethanol. Catalysts were characterized by N
2
-physisorption, X-ray powder
diffraction (XRD), IR spectroscopy, H temperature programmed reduction (TPR),
2
transmission electron microscopy (TEM), in situ X-ray photoelectron spectroscopy (XPS),
and FTIR/CO spectroscopy. It was confirmed that copper phyllosilicates species existed in all
catalysts but the exact form differed in the Cu-CP catalysts and the Cu-DP catalysts. While
the Cu-DP catalyst favored the form of 1:1 copperphyllosilicate, the Cu-CP catalysts tended
to have more species of 2:1 copperphyllosilicate which led to only small amount of active
copper species on surface after reduction and thus low ester conversion. Addition of
ammonium ion and zinc ion during the CP method synthesis promoted the copper silicate
phase evolution from 2:1 copperphyllosilicate to 1:1 copper phyllosilicate, and the reactivity
of ester hydrogenation was greatly enhanced. Conversion of 99.67%, selectivity of 99.07%,
and space-time yield of 1.23 g_EtOH/(g_cat﹒h) were achieved at 250 °C using a Cu-Zn-
NCP catalyst.
Keywords:Cu/SiO
2
Catalyst, Copper phyllosilicate, Co-precipitation, Phase evolution,
0
+
Cu /Cu , Acetic Ester Hydrogenation, EthanolSynthesis.

1
. Introduction
Ethanol was an important industrial commodity and a potential alternative synthetic fuel
to be used in automobiles. Currently, ethanol was produced by two major processes: (1)
fermentation of sugars derived from corn or sugar cane, and (2) hydration of petroleum-based
ethylene. Due to the diminishing crude oil resource and potential food crisis, gasification of
coal or biomass to syngas (CO + H
became more promising for large ethanol production. However, due to the slow kinetics of
the initial C–C bond formation and fast chain growth of C intermediates, the direct
conversion process suffered from low yield and poor selectivity . Alternatively, several
indirect, multi-step processes, such as via hydrogenation of C intermediates such (acetic
esters and oxalic esters) to ethanol, had attracted extensive attentions
2
), followed by catalytic conversion of syngas to ethanol,
2
[1]
2
[2-3]
. Among these key
steps, ester hydrogenation to ethanol was the only step that was still under research and
development, and thus was the technical bottleneck we needed to overcome for a viable coal-
to-ethanol process. This work presented the recent development of hydrogenation catalysts
for acetic ester to ethanol conversion.
Catalysts for ester hydrogenation had generally been prepared by mixing hydrogenating
metals (e.g., Fe, Co, Ni, Cu, and Zn) with various oxides (e.g., SiO
2
, ZrO
) . Typically, Copper-based catalysts were considered to be suitable for selective ester
hydrogenation, because they allowed for selective hydrogenation of C=O bonds but were
2
, V
2
O
5
, Cr
2
O
3
, and
[4]
WO
3
[5]
relatively inactive in C-C bond hydrogenolysis . However, due to the low melting point (ca.
083 °C) and thus low Thȕttig (ca. 178 °C) and TTammann (ca. 405 °C) of copper, single copper
nano particles were not sintering-resistant or mechanically stable for industrial applications
1
[6]
.
Adkins et al. firstly described the application of a copper-chromium catalyst for ethyl ester

[7]
hydrogenation . Later on, Cr-free catalysts had been developed due to strong environmental
restrictions, among which Cu/SiO catalysts stood out in hydrogenation of dimethyl oxalate
2
[2-3, 8-9]
to ethylene glycol and ethanol
. Research on silica supported copper catalysts indicated
that both high dispersion of copper species and strong metal-support interaction were vital to
high activity and stability in vapor-phase hydrogenation of ester to alcohols or glycol, which
[3,10-11]
was similar to the Cu-Cr catalyst system
. In addition, because of the weak acidic and
basic properties of SiO , Cu/SiO catalysts had demonstrated highest yield of desired
2
2
[
12]
products in hydrogenation of esters at mild conditions
In order to obtain desired performance of Cu/SiO
.
2
catalysts, it was necessary to disperse
fine particles onto the support. Conventional methods such as wet impregnation often led to
broad size distribution of metal particles and large particle size when the loading was high.
Ion-exchange approach could well disperse and stabilize copper species on silica support, but
the metal loading was low due to limited amount of terminal –OH groups on the surface of
[13]
silica . As early as 1990, Van der Grift confirmed that Cu/SiO catalysts prepared by urea
2
hydrolysis deposition-precipitation method (DP) could offer good metal dispersion, even at
high metal loading, partially due to the formation of “chrysocolla (copper phyllosilicates) -
[10]
like” precursors . Almost at the same time, copper phyllosilicates were reported in Cu/SiO
2
[3, 12, 14]
catalysts by ammonia evaporation (AE) methods
and hydrothermal precipitation
[5, 10, 15-18]
techniques
.
The chemical states of copper in both calcined and reduced copper phyllosilicates were
also of great importance to the catalytic hydrogenolysis of esters. N. Takezaw et al. claimed
that isolated cupric precursors were reduced to monovalent copper while small clusters of
cupric oxide (CuO) were reduced to metallic copper in copper/silica catalysts prepared by
[19]
ion-exchange approaches . C. J. G. Van Der Grift suggested that calcination at 550 °C, as

well as direct reduction at 277 °C, the “chrysocolla-like” catalyst precursors prepared by DP
method led to the decomposition of copper hydrosilicate and reorganization of silica support,
furthermore, calcination of the same sample at 427 °C caused the subsequent reduction to
occur at a lower temperature, which indicated the presence of small copper oxide particles
[17]
over silica support . J. Gong obtained Cu/SiO
2
catalysts by ammonia evaporation
hydrothermal (AEH) method, and found the coexistence of copper phyllosilicates and cupric
0
+
oxide even after calcination at 400 °C. After reduction, Cu and Cu species co-existed and
were primarily originated from well–dispersed copper oxide nano-particles and copper
[3]
0
+
phyllosilicates, respectively . The synergetic effect between Cu and Cu was believed to
play an excellent role for Cu/SiO catalysts during the hydrogenolysis of dimethyl oxalate
DMO) to ethylene glycol or ethanol
good dispersion of metal particles and provide active sites upon reduction, they were often
considered a key component in Cu/SiO catalysts.
2
[3, 12, 15]
(
. Since copper phyllosilicates contributed to the
2
Copper phyllosilicates were copper silicates with a lamellar structure that consists of
layers of SiO tetrahedra and layers of CuO octahedra. The 1:1 copper phyllosilicate, which
was often known as chrysocolla [Cu Si (OH) ], was a poorly ordered mineral containing
Si/Cu ≈ 1 and one of the typical copper phyllosilicates available in nature
of copper phyllosilicate was the 2:1 copper phyllosilicate, also called Cu-talc or Cu-smectie,
4
6
2
2
O
5
2
[10, 20]
. Another type
[20]
which was less common and only forms when there is large excess amount of Si . Several
[21-23]
[5, 10, 15-18]
[24]
researchers revealed that silica may react with certain metal (Ni
, Cu
, Co
)
precursors to form phyllosilicates during preparation, depending on the reactivity of silicic
[21]
species with the metal complexes . Formation of silica-supported copper phyllosilicates
and nickel phyllosilicates was a well-known case during catalyst preparation by DP method
[21-23]
[25, 26]
and selective adsorption of metal tetraammine complex
. Characterization studies
[3, 10, 15, 21, 26]
[20]
[25, 26]
had confirmed the formation of phyllosilicates by IR
, EXAFS , and ESR
.

Siffert et al. found that the prerequisites for the formation of phyllosilicates include: (1)
silicic species in solution must be monomeric; (2) metal hydroxo complexes must be present
in solution; and (3) The pH must be lower than that of the hydroxide precipitation. The
mechanism proposed was based on the formation of two types of monomeric species in
solution, i.e., monomer 1 [(HO)
3
Si]-O-[M(H
2
O)
4
(OH)] and monomer 2 [(HO)
3
Si]-O-
[
M(H O) ]-O-[Si(OH) ], whose polymerization would lead to the formation of 1:1 and 2:1
2
4
3
[27]
phyllosilicates, respectively . In addition, in the case of Ni/SiO
2
and Cu/SiO systems, it
2
+
was shown that addition of NH
4
could inhibit the formation of phyllosilicates, which could
be attributed to the formation of “competitor ion” or the decrease of the concentration of
0
[26]
[
Cu(OH)
2
(H
2
O)
4
] complex
.
In the present work, silica supported copper catalysts were prepared by adding acidic
copper nitrate solution to a basic sodium silicate solution to co-precipitate copper hydroxo
complexes and monomeric silica. As a comparison, another kind of Cu/SiO catalyst with the
2
same composition was prepared by urea hydrolysis deposition-precipitation method (DP).
These two types of catalysts were calcined at the same temperature, and the catalytic
performance was evaluated by using gas phase hydrogenation of ethyl acetate as the probe
reaction. Characterization of physicochemical properties of catalysts by N
2
-physisorption, X-
ray powder diffraction (XRD), IR spectra, H temperature programmed reduction (TPR),
2
transmission electron microscopy (TEM), in situ X-ray photoelectron spectroscopy (XPS),
and FTIR/CO spectroscopy were performed to correlate with activity performance. Special
attention was paid to the nature of copper precursors and the valence states of active sites on
the surface of reduced catalysts.
2
. Experimental Section
2
.1. Catalyst preparation

The Cu/SiO
precipitation (CP) and urea hydrolysis deposition precipitation (DP). Copper nitrate
Cu(NO •3H O (A.R., Sinopharm Chemical Reagent, China) was used as copper source.
Silica sol (JN-30, 30.62 wt% SiO , Qingdao Haiyang Chemical) and sodium silicate
Na SiO •9H O (A.R., Sinopharm Chemical Reagent, China) were used as silicon source.
2
Catalysts studied were prepared by two different precipitation methods co-
3
)
2
2
2
2
3
2
After drying, the catalyst precursors were calcined in static air at 450 °C for 4 h. All catalysts
were denoted as Cu-x, where x represents the preparation technique (CP, NCP, DP)
respectively.
a) Co-precipitation of copper hydroxo complexes and monomeric silica. The Cu-CP
sample was prepared as following description: 31.2 g of Na
mL deionized water to form Solution A. According to stoichiometric ratio, 13.0 g of Cu
NO •3H O and nitric acid [nHNO3=2(nSi-n ), M represents bivalent metal cation] were
2
SiO
3
•9H O was dissolved in 200
2
(
3
)
2
2
M
dissolved in 150 mL deionized water to form Solution B. Solution A was transferred to an oil
bath preheated at 90 °C, then Solution B was added drop-wise into Solution A and stirred to
form a suspension, with the pH of the solution changing from 14 to 7. After the precipitation,
the suspension was aged at 90 °C with stirring for 5 h. The mixture was filtrated and washed
with 500 mL of deionized water five times, and then was dried at 110 °C overnight. When
2
+
+
ammonium nitrate was added to the solution B with a Cu :NH
4
molar ratio of 1:5, the
precipitated sample was denoted as Cu-NCP; The Cu-Zn-NCP sample was prepared by
2
+
2+
+
adding zinc nitrate and ammonium nitrate to Solution B with a Cu :Zn :NH
4
molar ratio of
1
:0.1:5.
b) Urea hydrolysis deposition precipitation (DP) method. 13 g of Cu (NO
3
)
2
•3H O and
2
urea (per equivalent of copper three equivalents of urea) were dissolved in 150 mL deionized
water. According to the desired formula, 22.5g of acidulated silica sol was added into the

mixed solution and the pH was adjusted to 2 - 3 by adding nitric acid at room temperature.
Then the suspension was stirred and kept at 90 °C in an oil bath for 24 h, allowing for urea
hydrolysis. The pH of the suspension solution increased gradually, and when it reached 6 -7,
the heating process was terminated. The mixture was filtrated and washed with 500 mL of
deionized water five times, and then dried at 110 °C overnight. The sample was denoted as
Cu-DP.
For comparison, a Cu/SiO
2
sample prepared by the co–precipitation method described
was coded as Cu-JTS. In this preparation, sodium silicate solution was
added into a copper nitrate solution and the pH of the solution started from 1 and ended at
.3. Amorphous silica (SiO -450) was prepared by adding ammonium nitrate to a sodium
silicate solution, which was followed by drying and then calcination 450 °C.
[16]
by Van der Grift
4
2
2
.2. Catalyst characterization
The chemical composition of the catalysts were determined by X-ray fluorescence
spectroscopy (XRF), using a Rigaku ZSX PrimusⅡX-Ray spectrometer which is equipped
with an Rh target X-ray tube at the power of 3 kW. The Cu weight loading is expressed in
wt% of Cu per gram of the sample：
wt%Cu=mCu⁄ (mCu+mSiO2)×100%
Specific surface area and pore volume of the samples were measured from Nitrogen
adsorption isotherms at 196 °C, using MicromeriticsTristar 3000 equipment. Sample
degassing was carried out at 300 °C for 3h prior to adsorption. BET surface area was
calculated using the BET equation, the pore size distribution was calculated following BJH
method according to the desorption isotherm branch, and the total pore volume (Vp) was
derived from the adsorbed N volume at a relative pressure of approximately 0.99.
2

IR spectra of powder samples with KBr wafer were recorded at room temperature using
a Bruker Vertex 70 FTIR Spectrometer that is equipped with DTGS detector. Spectra were
-1
-1
recorded in the 4000 - 400 cm range at a resolution of 4 cm .
X-ray powder diffraction (XRD) patterns of samples were scanned using a Bruker D8
Advance X-ray diffractometer with Ni-filtered Cu K
α
radiation source（λ = 0.15406 nm）at
4
0 kV and 40 mA, The scanning 2θ range was 10 - 80°, and the scanning speed 2 °/min. For
measurements of reduced catalyst samples, Ar was used to protect the cooling sample from
oxidation during transfer and sample preparation before XRD analysis.
Transmission electron microscopy (TEM) images were taken with JEOL 2011 Electron
Microscope operating at 200 kV. This microscope also has an attached X-ray energy-
dispersive spectrometer (EDS). Before being transferred into the TEM chamber, reduced
catalysts samples were quickly dispersed with ethanol and then dropped onto the nickel grid
for TEM/EDS analysis.
The reducibility of the calcined catalysts was determined by H temperature
2
programmed reduction (TPR) on a MicromeriticsAutochem II 2920 auto adsorption
apparatus. 40 mg of sample was degassed at 120 °C under Ar for 1 h, and after cooling to
room temperature the gas was switched to 10 vol% H
temperature was ramped linearly to 800 °C with a heating rate of 10 °C/min. The thermal
conductivity detector (TCD) was used to determine the H consumption during the run.
2
/Ar with a flow rate of 50 mL/min. The
2
The dispersion and distribution of the active surface metallic copper sites were
determined by dissociative N O adsorption at 60 °C and 90 °C, using the procedures
described by Van Der Grift et.al. . Firstly, The calcined sample was reduced at 300 °C
under 10 vol% H /Ar (50mL/min) for 1h.and cooled down to the designated oxidation
2
[18]
2

temperature, the gas was switched to pure Ar, and purged for 20 min. Then, the sample was
exposed in the 10% N O/Ar (30 mL/min) atmosphere and oxidized for 1h, ensuring that
surface Cu atoms were completely oxidized according to the reaction: 2Cu + N O→Cu
, Finally, sample was purged with Ar and cooled to room temperature. Following the same
procedures as the first one, a second TPR run was performed, the consumed amount of H
consumption was calibrated using known amount of surface Cu O, for the sample subjected
to oxidation at 60 °C, it was denoted as Y , while it was denoted as Y for oxidation at 90 °C.
2
2
2
O +
N
2
2
2
1
2
The quantity of loading copper in all reduced catalysts was measured by XRF and denoted as
X, and copper dispersion (D) was calculated by dividing the amount of surface copper atoms
into total supported copper atoms as follows:
D=(2×Y⁄X)×100%
To obtain the copper chemical states on surface for catalysts upon calcination, after
reduction and after re-oxidation by N O, in situ X-ray photoelectron spectroscopy (XPS) and
2
Auger electron spectroscopy (XAES) were carried out on Thermo Scientific K-alpha
spectrometer system, which is equipped with monochromatic Al Kα (1486.6 kV) X-ray
source, a hemispherical energy analyzer and multi-stage samples cabin. All binding energies
were calibrated by using the C1s peak at 284.6 eV as the reference. Each calcined catalyst
was pressed into pellet and equally placed into the sample chamber. The samples were
-7
transferred to the vacuum chamber (2×10 Pa) and the first XPS measurement was
performed. Then, the samples were drawn back into the sample chamber and reduced at 300
°
C in 10 vol% H
measurement was performed following the same procedures. Finally, the reduced samples
were equally exposed to the 10% N O/Ar (30 mL/min) atmosphere at 90 °C for oxidation for
h, and the third XPS/XAES measurement was performed.
2
/Ar (50mL/min) atmosphere for 1 h and cooled down. Second XPS/XAES
2
1

The infrared studies of adsorptive CO on the surface of catalysts were conducted with
Bruker Vertex 70 FTIR Spectrometer system which is equipped with an in situ transmission
cell (Specac GS05850). The cell is connected to a vacuum-desorption apparatus with a
-4
residual pressure below 10 Pa, can be heated up at different atmospheres. To study reduced
samples, each catalyst was pressed into disk and pretreated in situ in the cell at 300 °C under
1
0 vol% H /Ar (50 mL/min) for 1 h. After cooling down to 25 °C, the spectrum was taken as
2
the reference for subsequent spectra of the sample. The catalyst was then exposed to a
flowing mixture of 1% CO/Ar for 30 min at 25 °C to reach the adsorption equilibrium. While
staying at the constant desorption temperature of 25 °C, spectra were collected at different
evacuation pressures. To investigate the CO adsorption behavior of catalyst upon surface re-
oxidation by N
evacuated, and surface oxidation was accomplished by flowing 10% N
0 °C for 1 h. After cooling down to 25 °C, spectra could be collected following the above
mentioned procedures.
2
O, the sample after reduction was first cooled in H
2
/Ar and subsequently
2
O/Ar (30mL/min) in
9
2
.3. Catalytic reaction
The vapor-phase hydrogenation of ethyl acetate (EA) was tested in a continuous-flow
fixed-bed Micro-reactor. Typically, 1g catalyst (40 - 60 meshes) was placed between plugs of
quartz sand in a steel tubular reactor. Before the experiment, all catalysts were activated in
situ at 300 °C with a ramping rate of 1 °C/min under pure H
then cooled under H to the reaction temperature. Pure ethyl acetate (A.R, purity > 99%) and
were mixed and preheated, and fed into reduced catalyst bed with a H /EA molar ratio of
9 at a total pressure of 3.0 MPa and in a temperature range of 210 °C to 280 °C. Steady-
2
(50 mL/min) for 4 hours, and
2
H
2
2
2

state was usually obtained after 4 h on stream reaction. Products were condensed to allow for
liquid and gas separation. Liquid and gas products were analyzed separately by an off-line
gas chromatograph (Shimadzu GC17-A) equipped with a HP-FFAP capillary column (30 m ×
0
.25 mm × 0.3 μm) and a flame-ionization detector (FID). The products were quantified via
internal calibration method.
3
. Results and Interpretation
3
.1. Physicochemical and Morphological properties of catalysts
The chemical compositions and the textural properties of the two types of catalysts were
listed in the Table 1. The actual copper loading determined by XRF was close to the nominal
value. For calcined co-precipitation (CP) catalysts, the BET surface area of the Cu-CP sample
2
was 515.7 m /g, which was noticeably higher than that of the Cu-DP sample prepared by urea
+
2+
hydrolysis deposition precipitation (DP) method. By adding NH
precipitation process, the structure of the Cu/SiO catalysts changed rapidly, and their BET
surface area decreased to 423.4 m /g. However, they were all much higher than that of
amorphous silica (SiO -450).
4
and Zn to the co-
2
2
2

After calcinations, similar to the DP sample, the color of all co-precipitation catalysts
Cu-CP, Cu-NCP, and Cu-Zn-NCP) changed from bright blue to olive green. The high
(
specific surface area and olive green or bright blue color were often attributed to the presence
of copper phyllosilicates and/or copper oxide with a high specific surface area and a two-
dimensional dispersion of copper species over the corrugated silica sheets [28]. According to
Van der Grift, the bright blue color of the precipitates indicated the presence of hydrated,
octahedrally surrounded copper(II) ions in the catalyst precursors, and the olive green was
characteristic of decomposed chrysocolla (copper phyllosilicates) with a two dimensional
[10, 16]
tenorite on silica
. In the Cu-JST sample, the color changed from bright blue to dark
[28]
gray, suggesting the presence of CuO with some degree of aggregation in silica matrix
.
Hence it was reasonable to assume that the formation of copper phyllosilicates in the co-
precipitated catalyst precursors which was dehydrated and partly decomposed after
calcination.
It was noticeable that the BET surface area of Cu-DP sample was the higher than Cu-DP
+
sample even with the same copper loading, and could be adjusted by adding NH
4
and zinc to
the co-pricipitation process. Because copper phyllosilicates possess high specific surface area
[10]
,
this larger surface area may be attributed to more copper phyllosilicates, or special type
of copper phyllosilicate species preserved in the co-precipitation catalysts.
The N
calcined Cu/SiO
dominatedas implied in Fig. 1. The Cu-DP sample was found to possess H
loop and show a bimodal pore-size distribution with maxima at radii 3.0 and 7.0 nm, whereas
for the co-precipitation catalysts, all samples exhibited H -type hysteresis loop and single
2
adsorption-desorption isotherms and pore-size distributions curves of all
catalysts samples exhibited a type IV Langmuir isotherm and mesopores
-type hysteresis
2
1
2

pore-size distribution centered at 3.0 nm, implying that “parallel plate narrow slit channels”
dominated in these catalysts.
During the co-precipitation of sodium silicate and copper nitrate solution, copper nitrate
hydroxide, the mixture of copper nitrate hydroxide with silica gel, or copper phyllosilicates
were probably formed. As mentioned in the introduction, the FTIR spectroscopy was well
[6.26]
adopted to distinguish between these three copper species in the sample
. As shown in
Figure 2a, the copper nitrate hydroxide was characterized by OH bands of structural hydroxyl
-1
-1
-1
groups at 677, 776 and 887 cm , and two sharp bands at 1423 cm and 1347 cm
corresponded to the molecular vibration of nitrate. The discrimination between copper
hydroxide and copper phyllosilicates can be made from the OH bands and the presence or
[26]
-1
absence of SiO band . The band at around 690 cm was often attributed to the structural
[6,29]
hydroxyl group of Cu(OH)
2
, while the presence of chrysocolla (1:1 copper phyllosilicate)
-1
was often characterized by a SiOband of SiO
4
tetrahedra at ～1040 cm and by OH bands of
-1
-1[3,20,26]
structural hydroxyl groups at 670 cm and 3620 cm
. Typically, a sharp band at 3620
-1
cm was attributed to O-H stretching of the structural OH, and a medium intensity band at
-1
[20]
6
70-675 cm was assigned to the deformation mode of OH linked to 3 Cu atoms
.
So for silica supported copper catalyst precursors, in addition to the bands of amorphous
-1
silica at about 1100, 800 and 470 cm (Figure 2a), the Cu-JTS sample only exhibited bands
-1
at 700, 1088, 1347 and 1423 cm , indicating only mixture of copper nitrate hydroxide was
formed in the silica gel. While for the dried precursors of Cu-CP and Cu-DP sample, the
-1
typical bands at 670, 1024, and 3623 (3645) cm were observed, suggesting the formation of
copper phyllosilicates in the catalyst precursors.
However, it was hard to further identify which type of copper phyllosilicates preserved
in the Cu-CP and Cu-DP precursors , since the ill-crystalllized 1:1 and 2:1 phyllosilicates

-1
cannot be distinguished from their OH vibration bands which were very close (~ 670 cm ),
[21]
and their SiO stretching vibration bands which were very broad . Although the syntheses
started with Si/Cu ratios of 2 - 4 which would otherwise lead to the formation of 2:1 copper
[10,12,20]
phyllosilicate, most of products obtained from previous reports
were chrysocolla (1:1
copper phyllosilicate) ,or sometimes copper oxides, unless very high temperature were used,
in which case the reaction mixtures contained a large excess of active Si , and the 2:1 copper
phyllosilicate (Cu-talc) was obtained, when the reaction mixtures did not contain a large
excess of active Si, the products were mixture of copper oxides and amorphous silicates, with
[20]
perhaps some phyllosilicates .
After calcination, copper phyllosilicates were also preserved in all precipitated catalysts,
which were characterized by the IR spectra shown in Figure 3. Comparing to the dried
-1
samples, the intensity of absorption bands at 670 ,1024, 3623 cm decreased, and the
-1
shoulder peak at 1110 cm became obvious. This indicated that part of the copper
phyllosilicates decomposed and rearranged after calcination at 450 °C. The relative amount
of supported phyllosilicates before and after calcination can be roughly estimated by
-1
considering the integrated intensity of OH band at 670 cm normalized to the integrated
-1
intensity of the symmetric SiO band of amorphous silica at 800 cm , which was defined as
[26]
I
670/I800 . The results were summarized in the Table 2. For co-precipitated catalysts, the Cu-
CP sample seemed to contain larger amount of copper phyllosilicates and to be more
thermally stable than others in terms of phyllosilicates decomposition. Introduction of NH
+
4
2
+
and Zn had inhibited the formation of stable type of copper phyllosilicates during the co-
pricipitation process.
To further investigate the cupreous species in co-precipitated Cu/SiO
2
catalysts, XRD
patterns of bulk CuO and two types of calcined Cu/SiO catalysts were collected and shown
2

in Fig. 4. The Cu/SiO
patterns, but all were different from the characteristic diffraction peaks of CuO. The broad
peak at 22° was generally considered to be amorphous silica as the support in Cu/SiO
catalysts, while the strong and broad peaks at ca. 30.5° , 35.8°, 56.9°, 63.3°, and 72°
2
catalysts by two types of precipitation method exhibited similar XRD
2
[3,26,30]
suggested the predominant presence of copper phyllosilicates with poor crystallinity
.
No diffraction peaks for copper oxide were observed, indicating that copper oxide species
was highly dispersed in co-precipitated catalysts.
The TEM images of calcined Cu-DP and Cu-CP catalysts were shown in Figure 5. Upon
calcination, the dark spherical particles irregularly dispersed on the silica support. The
average particle size was about 8.4 nm for the Cu-DP catalyst and 4.2 nm for the Cu-CP
catalyst. According to the EDS results, the main component of this particles was copper
oxide, which should be attributed to the decomposition products of copper silicates or
Cu(OH) on silica. It may be worth noting that some core-shell particles were distributed in
2
the Cu-CP catalyst with a Si/Cu ratio about 11, which suggested that some copper species
were buried into silica species in the Cu-CP catalyst after calcination.
The temperature programmed reduction (TPR) profiles for CuO and the synthesized
Cu/SiO
2
catalysts were presented in Fig. 6. The reduction peak for bulk CuO was around 225
°C while the peak for the Cu-CP sample 243 °C, implying there was a strong interaction
between copper species and silica support. Literatures had suggested that the main peak could
0
be assigned to the reduction of well-dispersed CuO particles to copper metal (Cu ) and
[
3, 12, 15]
phyllosilicates to cuprous oxide (Cu
may need much higher temperature and the peak may not be visible on the profile. The
reduction temperature for the Cu-CP catalyst was higher than other Cu/SiO catalysts,
probably because it had more stable copper phyllosilicates. This was consistent with the
2
O)
. Reduction of cuprous oxide to copper metal
2

+
2+
semi-quantitative results of IR characterization. Similarly, adding NH
Cu-NCP and Cu-Zn-NCP catalysts) inhibited the formation of stable phyllosilicates and thus
led to lower reduction peak.
4
and Zn to samples
(
3
3
.2. Characterization of reduced catalysts
.2.1 Crystalline size
The two types of Cu/SiO
shown in Fig. 7. While copper oxide and phyllosilicates disappeared, there were broad peaks
for Cu metal and cuprous oxide (Cu O). Peaks at 43.3°, 50.3°, and 73.9° could be assigned to
Cu metal (111), (200), and (220) plane diffraction, respectively (JCPDS 04-0836). Peaks at
6.4° and 61.4° were characteristic of Cu O crystalline. The broad peak at 21° was from
amorphous silica. Crystalline size of Cu metal and Cu O was calculated using Scherrer
Formula and listed in Table 3. It can be seen that both Cu size and Cu O size in the DP
catalyst were similar to the CP catalysts, with Cu metal in the range of 7.8 – 8.4 nm and Cu
.5 – 7.9 nm.
2
catalysts were reduced at 300 °C, and their XRD patterns were
2
3
2
2
2
2
O
7
The TEM images of reduced catalysts were shown in Fig. 8. The light gray silica
supports were identified along with highly dispersed dark spherical particles assignable to Cu
species. It was apparent that metal particles were uniformly dispersed onto the silica support
in all samples, for the co-precipitation catalysts. The statistical results based on TEM images
showed that the copper particles sizes were about 4.3 nm to 5 nm, which were smaller and
+
2+
more uniform than the DP catalyst. It was also noticeable that adding NH
4
and Zn during
precipitation (Cu-NCP and Cu-Zn-NCP catalysts) did not change the metal particle size.
3
.2.2 Dispersion and distribution of metallic copper on surface
To investigate the metal species and their distribution on surface for the reduced
Cu/SiO
2
catalysts, oxidation by N O was performed for the reduced catalysts at 60 °C and 90
2

°C, respectively, which was followed by 2nd reduction using the same procures as the 1st
TPR. Copper oxide (CuO) was used as reference. Literature had suggested that copper metal
[18]
species could be oxidized to Cu
2
O by N O under 90 °C . If the quantity of loading copper
2
in all reduced catalysts was denoted as X, and the total hydrogen consumption of the 2nd
TPR as Y (in which Y60 referred to the 2nd reduction at 60 °C and Y90 at 90 °C), dispersion
of copper metal can thus be calculated as 2Y/X. The dispersion results of all catalysts were
summarized in Table 3.
The 2nd TPR profiles of the samples were shown in Fig. 9. For pure CuO, copper metal
species from the 1st TPR could not be oxidized by N O even at 90 °C, resulting in no further
2
hydrogen consumption during the 2nd TPR and therefore virtually zero dispersion. For the
Cu-DP sample prepared by urea hydrolysis deposition precipitation, the 2nd TPR profiles,
including the reduction peak position and area, looked similar when the oxidation treatment
was performed at 60 °C and 90 °C. While for the Cu-CP sample prepared by co-precipitation,
the 2nd TPR profiles differed in both peak position and area. For instance, in the case of
oxidation treatment at 60 °C, the reduction peak was broad spanning from 100°C to 175 °C
and the dispersion was 7.47%; while in the case at 90 °C, the peak shifted to 145 °C and the
dispersion increased to 19.72%. This suggested that there was less copper metal species on
the surface of the Cu-CP catalyst comparing to the Cu-DP one. It should be understandable
that the co-precipitation method gave out a more copper enrichment on the surface of silica
support than the deposition precipitation method .When the oxidation temperature rose from
6
0 °C to 90 °C, N
the Cu-CP catalyst, which soundly explained why more hydrogen was consumed during the
nd TPR run. Besides, the reduction peak shift to higher temperature was an indication of
enhanced interaction between cuprous species and the support.
2
O penetrated deeper and more copper beneath the surface was oxidized in
2

+
2+
Adding NH and Zn to the co-precipitated catalysts seemed to help shift more copper
4
species to the surface. For the Cu-NCP catalyst, the dispersion was 16.14% for the case of
oxidation at 60 °C and increased to 19.75% when at 90 °C. For the Cu-Zn-NCP catalyst, the
dispersion was high in both cases but the difference already diminished. In summary, with
same copper loading, using additives during the co-precipitation synthesis could change the
microstructure evolution and enrich copper species onto the surface, and thus greatly
improved the copper metal dispersion upon reduction.
3
.2.3 In situ XPS analysis of surface chemical states
XPS and XAES were employed to study the chemical status of copper species on
surface of four catalysts, i.e., Cu-DP, Cu-CP, Cu-NCP and Cu-Zn-NCP, which were subject
to in situ reduction at 300 °C and in situ oxidation by N O at 90 °C when needed. The Cu2p
2
XPS spectra of the four as-calcined catalysts were illustrated in Fig. 10. The binding energy
for Cu2p3/2 and Cu2p1/2 were centered at 935.1 eV and 954.9 eV, respectively, and the
satellite peaks of 2p-3d orbitals were in the binding energy range of 941 – 944 EV. These
2
+
features were all due to the characteristics of Cu species. Moreover, the Cu2p3/2 binding
energy of CuO is typically around 933.5 EV, so the main peak at 935.1 eV observed in this
2
+
study was likely due to the strong interaction of Cu and the silica support, i.e., the formation
[10]
of copper phyllosilicates . This is consistent with the XRD and IR results.
When the catalysts were reduced in hydrogen at 300 °C, the Cu2p XPS spectra were
shown in Fig. 11A. The Cu2p3/2 binding energy shifted to 932.4 – 932.8 eV and the Cu2p1/2
binding energy to 952.1 – 952.5 eV. Meanwhile, the satellite peaks disappeared, suggesting
2
+
0
+
Cu species were reduced to Cu and/or Cu . When the reduced catalysts were kept in the
chamber and further oxidized by N O at 90 °C, the Cu2p XPS spectra were shown in Fig.
2

1
2A. It could be found the spectra were similar to Fig. 11A and lack of the satellite peaks
2
+
characteristic to Cu . This confirms that oxidation of copper metal by N O at 90 °C could
2
[3, 12]
only go to Cu
2
O
.
0
+
The amount of Cu and Cu on surface of the catalysts after reduction and after further
oxidation by N O could be determined by Cu LMM XAES, as in Fig. 11B and Fig. 12B,
2
respectively. In both figures, the Auger kinetic energy peaks were broad and asymmetric, but
could be de-convoluted into two symmetrical peaks centered at 916 and 918 eV
+
0
corresponding to Cu and Cu species, respectively. Quantitative results were summarized in
+
Table 4. The reduced Cu-DP catalyst had dominant Cu species on surface (96.0%) which
further increased to nearly 100% coverage after oxidation by N O ; For co-precipitation
2
+
catalysts, the reduced Cu-CP catalysts only had Cu surface coverage of 60.4%. As the
+
promoters were added during the precipitation , however, the Cu surface coverage was
increased, the reduced Cu-NCP and Cu -Zn-NCP catalysts exhibited 71.4% and 94.9 % of
+
Cu surface coverage respectively, which, upon oxidation, also dramatically increased to
+
2+
9
2.9% and 100%, respectively. It is thus found that using NH and Zn during the co-
4
precipitation synthesis could not only enrich copper onto the support surface, but also tune
the percentage of different copper species on surface.
3
.2.4 IR study of CO adsorption on catalysts
IR of CO adsorption on catalysts was also used to probe the chemical status and
distribution of copper species on surface. The IR spectra of CO adsorption on the reduced
Cu-Zn-NCP catalyst at different equilibrated pressure were shown in Fig. 13A. When the CO
pressure was at 1 atm, physisorption and chemisorption co-existed on the catalyst surface,
-1
-1
where the bands at 2120 cm and 2170 cm were due to stretching of gaseous CO, and band
-1
at 2054 cm was due to the CO adsorbed on the silica support via hydroxyl group bonding.

With vacuuming, the gaseous CO and physical adsorbed CO were removed, and chemisorbed
-1
CO peak started to appear at 2113 cm . When the pressure was further reduced, this weak
peak quickly disappeared and a strong and stable chemisorbed CO peak emerged at 2124 cm
-
1
-3
-1
-1
.
When the pressure dropped to 4.5x10 Pa，2124 cm blue shifted to 2130 cm . CO
0
+
chemisorption is usually considered to occur on the copper species to form Cu -CO , Cu -CO
2
+
,
Cu -CO species via σ- and π-back bonds, as a result of the synergistic effect between two
+
0
bonds, the Cu -CO species are characterized by the higher thermal stability than Cu -CO and
2
+
[31-35]
Cu -CO species
; besides, it is often thought that the stretching frequency of CO species
2
+
+
0
[36]
absorbed onto cupreous surfaces follows the sequence Cu -CO > Cu -CO > Cu -CO
.
-
1
-1
According to these assignments , band at 2130 cm and 2113 cm can be assigned to linear
+
0
+
stretching of Cu -CO and Cu -CO species respectively, and the only Cu -CO carbonyls could
be preserved at the end of high vacuum.
When the Cu-Zn-NCP catalyst was reduced and further oxidized by N O at 90 °C, the
2
IR spectra of CO adsorption were shown in Fig. 13B. The evolution of sorption bands as a
function of CO pressure was the same as for the reduced catalyst. The only difference was
+
-1
that, under high vacuum, the Cu -CO band shifted to 2132 cm and the sorption peak area
+
was bigger than the reduced one, which was a good indication of more Cu species that was
0
due to some Cu species oxidation by N
2
O.
IR of CO adsorption on the Cu-DP, Cu-CP and Cu-NCP catalysts were also performed
following the same procedures. The spectra for these reduced catalysts and further oxidized
ones were presented in Fig. 14A and B, respectively. A qualitative analysis on the amount of
+
Cu on surface was given in Table 5, which was found to be in good agreement with the
+
quantitative analysis by H
2
-TPR and XPS. For the Cu-DP catalyst, there was a big Cu -CO
-1
sorption peak at 2132 cm after it was reduced, and the peak area increased after it was

+
0
further oxidized by N
2
O, suggesting large amount of Cu species, along with some Cu
0
species, existed on surface. For the Cu-CP catalyst, only a trace of Cu -CO sorption peak at
-1
-1
2
107 cm could be seen when it was reduced, and a weak shoulder peak at 2126 cm
+
appeared after oxidation which could be presumably assigned to Cu -CO sorption. This
0
+
indicated that the amount of both Cu and Cu on the Cu-CP catalyst surface was actually
0
small though the relative percentage of Cu was high. For the Cu-NCP catalyst, it can be
+
qualitatively seen that the amount of Cu on surface was more than the Cu-CP catalyst,
+
implying the copper species was more concentrated on surface via the assistance of NH
4
during the synthesis.
3
.3. Catalytic performances
The reactivity of the catalysts was probed using vapor-phase hydrogenation of ethyl
acetate (EA), and was correlated with the surface structure and active metal species of the
catalysts. Under the reaction conditions, the main product was ethanol and ethane was the
byproduct that could be detected. The conversion, selectivity, and space time yield over
different catalysts were summarized in Table 6.
For the Cu-DP catalyst, the reactivity at 250 °C was as high as 94.0% which was close
the thermodynamic equilibrium conversion. As a comparison, for the Cu-CP catalyst, the
conversion of EA at 250 °C was only 11.4% but could reach 95.3% when the temperature
+
2+
was raised to 280 °C. Using additives such as NH and Zn remarkably improved the
4
reactivity. For example, the Cu-Zn-NCP catalyst behaved much better than the Cu-DP one,
and when at 250 °C, the conversion of EA reached 99.7% and STY 1.23 g_EtOH/(g_cat﹒h)
4
. Discussions

4
.1. Formation of copper phyllosillicates in Cu/SiO
2
catalysts prepared by co-
precipitation
The formation of copper phyllosilicates was well known during the preparation of
Cu/SiO catalysts by urea hydrolysis deposition-precipitation (DP) method, ammonia
evaporation (AE), sol-gel, and selective adsorption of copper tetraammine complex
2
2
+
[
Cu(NH
3
)
4
(H
2
O) ] on silica. In the present work, copper phyllosilicates were also obtained
2
by co-precipitation method, despite the pH range for precipitation and the precursor for silica
was different. Our study showed these preparation methods followed the same type of
elementary reaction mechanism. Firstly, monomeric silicic acid reacts with hydrolyzed
copper complexes containing H
2
O and OH ligands through a hetero-condensation reaction to
form Si-O-Cu monomers or Si-O-Cu-O-Si monomers, and secondly, monomers polymerize
[26]
to form copper phyllosilicates . It has been reported that the prerequisites for the formation
of copper phyllosilicates include: (1) formation of monomeric silicic species in solution; and
(
2) formation of neutral copper complex with hydroxyl ligands. In addition, the formation of
copper phyllosilicates must be faster than the precipitation of copper hydroxide species and
[26-27]
gelation of silicic species
. Thus, the pH of the solution is critically important, especially
[27]
for DP method in which the pH must be lower than that of the hydroxide precipitation
.
0
During the preparation by CP method, the copper hydroxo complexes ([Cu(OH)
2
] ) and
monomeric silicic acid were formed and co-precipitated by adding acidified copper nitrate
solution drop-wise into the alkaline sodium silicate solution. With the pH of the solution
gradually decreasing from 14 to 8, the two precipitates reacted to form copper phyllosilicates.
Their characteristic IR peaks were strong in all three samples including Cu-CP，Cu-NCP and
Cu-Zn-NCP after filtering and drying. Meanwhile, IR peaks for Cu(OH)
2
were not obvious,
suggesting most copper species exist as in the compound of copper phyllosilicates.

When the alkaline sodium silicate solution was added drop-wise into the acidified
copper nitrate solution, the pH of the solution gradually increased from 3 to 7, which was
(
2x-y)+
accompanied by the gradual change of copper species from [Cu
x
(OH)
] monomers. Meanwhile, silicate hydrolyzed quickly to silicic acid
monomers which further polymerized and condensed along with pH change. The rate
difference in forming the two types of monomers led to the precipitation of Cu(OH) and
y
] cations to
0
neutral [Cu(OH)
2
2
amorphous silica separately, as in the case of the reference sample Cu-JTS where these two
components were confirmed by IR.
After calcination at 450 °C, copper phyllosilicates tended to decompose but the catalysts
prepared by co-precipitation retained more than other catalysts, suggesting the Cu-CP
catalysts had higher thermal stability. As shown in the IR, BET, XRD, XPS and TPR
characterization, Cu-CP catalysts had higher I670/I800 ratio, larger BET surface areas, more
complete layer structure and higher reduction temperature.
4
.2. Structure of the Cu/SiO catalysts prepared by co-precipitation
2
It was known that during the urea hydrolysis deposition precipitation, with the increase
of copper loading, the phase of chrysocollas (1:1 copper phyllosilicates) became more
predominant in catalysts as evidenced by the increase of the characteristic I670/I800 ratio in IR
[10,15]
and the BET surface area
. On the other hand, Van der Grift et al. identified that the TPR
profiles of the Cu/SiO catalysts prepared via homogeneous deposition precipitation were
2
independent of metal loading within the range of 1 - 40 wt%, and the strong reduction peak
+
could be ascribed to the reduction of copper phyllosilicates to Cu and well-dispersed CuO to
0
[3, 12, 15]
Cu
.Themonovalent copper species were formed during the reduction of ion-
catalysts, due to the strong
catalysts prepared by the
exchanged Cu-O-Si species and copper phyllosilicates in Cu/SiO
2
[12,19]
interaction between copper species and SiO
2
. For Cu/SiO
2

same procedure, more copper phyllosilicates often led to more monovalent copper under
[12, 15, 30]
identical reduction conditions
.
As described above, the physicochemical properties and surface structure of the Cu-CP
and Cu-DP catalysts were totally different, even with the same copper loading. There were
more copper phyllosilicates in the Cu-CP catalyst than in the Cu-DP one, based on which one
+
may think there would be more Cu species in the Cu-CP catalyst after reduction. However,
our results of N O surface oxidation, In-situ XPS, and FT-IR CO-adsorption showed that the
2
0
+
amount of both Cu and Cu on surface of the Cu-CP catalyst was lower than that of the Cu-
DP one. Obviously, the difference in the content of copper phyllosilicates could not account
for the different properties between the two types of catalysts.
One possible explanation was that these two types of catalysts contained different types
of copper phyllosilicates. Paolo Burattin et al. showed that two types of monomeric species of
nickel phyllosilicates could be formed under aqueous condition depending on the Metal: Si
ratio in the solution, and polymerization of each would lead to the formation of 1:1 and 2:1
[21]
phyllosilicates, respectively . It was also suggested that the formation of phyllosilicates
was a kinetically limited process, and was especially governed by the kinetics of silica
dissolution. The 1:1 phyllosilicate was formed dominantly during deposition-precipitation of
[21-22]
silica and copper species under normal temperature and atmospheric pressure
. Jun Xiao
and Gilles Villemure found that 2:1 trioctahedral copper phyllosilicate (Cu-talc or Cu-
smectie) was obtained with the chemical composition of Si/Cu ratios of approximately 1.33,
under the hydrothermal treatment of aqueous mixtures of sodium hydroxide, copper chloride
and excess sodium silicate at 150 °C. While the 1:1 copper phyllosilicate (chrysocolla) was
predominate when increasing the hydrothermal treatment to 250 °C, and it had a lower Si/Cu
[20]
ratio of approximately 1 . In the present work, all the samples had similar copper loading at

about 35 wt% (Si/Cu ratio = 1.8), but the initial concentration of monomeric silicic acid in
the two types of precipitation methods was very different, which thus likely resulted in the
formation of two types of copper phyllosilicates. In the co-precipitation (CP) method, the
acidified copper nitrate solution was added drop-wise into the alkaline sodium silicate
solution. It was easier to form silicic acid with high concentration in this case, as compared
with the deposition-precipitation (DP) method. Therefore, the 2:1 copper phyllosilicate rather
than 1:1 copper phyllosilicate might be expected to form dominantly in co-precipitation (CP)
samples.
Due to poor crystallinity, our IR and XRD results did not distinguish between 1:1 copper
[21]
phyllosilicate and 2:1 copper phyllosilicate in the Cu-CP samples . The Cu-CP catalyst had
a reduction peak at 243 °C and was more difficult to reduce than the Cu-DP catalyst. It has
been suggested in the literature that the 2:1 phyllosilicate was more difficult to reduce than
the 1:1 phyllosilicate, which implied that more stable 2:1 phyllosilicate exist in the Cu-CP
[21]
catalyst . In addition, it was well-known that the 1:1 phyllosilicate consisted of a brucite-
type sheet containing metal (M) in octahedral coordination and a sheet containing linked
tetrahedral SiO
sandwiched one brucite-type sheet
copper phyllosilicates often decomposed by dehydration and partly reorganization of the
4
units, while in the 2:1 phyllosilicate, two sheets of linked SiO
4
units
[21-23]
. After high temperature calcination or reduction,
[
16, 17, 23 ]
silica support
morphology of core-shell structure in Cu-CP catalysts, which would be easily interpreted as
the copper species buried into silica support. In the 2:1 copper phyllosilicate, all the CuO
octahedrons were trapped into the two layers of SiO rather than linked in one layer, this also
. Characterization of calcined catalysts by TEM and EDS had shown a
4
4
explained why smaller amount of active copper was detected on the surface of the Cu-CP
catalysts and why enhanced interaction between cuprous species and the support was
observed.

+
2+
4
.3. Effect of NH
4
and Zn on the evolution of the structure of Cu/SiO catalysts
2
prepared by co-precipitation
+
2+
Addition of NH
4
and Zn during co-precipitation changed the structure of Cu/SiO
2
catalysts dramatically, such as lower reduction peak and more amounts of active copper
species after reduction on surface. It should be noted that the Cu-Zn-NCP catalyst actually
exhibited almost the same physicochemical properties as the Cu-DP catalyst in which the 1:1
phyllosilicate was predominant. Comparing to the Cu-CP catalyst, the Cu-NCP and Cu-Zn-
NCP catalysts seemed to contain more 1:1 phyllosilicate and less 2:1 phyllosilicate.
Thierry Toupance et al. found that addition of ammonium nitrate to the precursor
solution prevented the formation of copper phyllosilicates during the preparation of Cu/SiO
2
by selective adsorption of copper tetraammine complex on silica, and in the case of Ni/SiO
2
+
[26]
catalysts, addition of NH could inhibit the formation of nickel phyllosilicates . Previous
4
reports had suggested that the ammonium ion acted as “a competitor ion”, but there were two
explanations:
+
One possible rationalization was that the NH
4
ions prevented the formation of
phyllosilicates by inhibiting silica dissolution. According to Stumm et al., nucleophiles and
anionic ligand adsorption increased the rate of silicate dissolution, whereas cation adsorption
[37]
had an opposite effect . Dove showed that alkaline ions increased the rate of silica
2
+
2+
3+
3+
dissolution, while multivalent ions such as Zn , Cu , Al , and Fe inhibited dissolution.
+
+
Alkaline ions (such as Na and Li ) were small cations with a localized positive charge which
could easily approach negatively charged silica surface to adsorb and weaken the Si-O-Si
[38]
+
bonding and thus to favor silica dissolution . In contrast, NH was a much larger cation
4
with less localized positive charge. In consequence, when it approached the silica surface, it
was less strongly adsorbed, and therefore, silica dissolution is less favored.

Another explanation was that the formation of copper phyllosilicates required the
[26]
presence of OH bridging ligands in the copper complex . It is known that the tetraammine
copper complex, which only had non-bridging NH ligands, cannot be involved in the
formation of phyllosilicates, only the [Cu(OH) (H O) ] complex is zero charged, and it
contains two OH bridging ligands. Therefore, it can be considered as a good candidate for
3
2
2
4
[25]
+
participation in the nucleation stage with monomeric silicic species . Adding NH
4
to the
solution caused the concentration of [Cu(OH) (H O) ] to decrease dramatically and the pH to
2
2
4
shift to a lower value, which slowed down the rate of silica dissolution and in turn speeded up
the polymerization and condensation of monomeric silicic acid. The formation of copper
[26]
phyllosilicates was thus inhibited
.
Unlike the soluble silica sol in alkaline solution, the monomeric silicic acid came from
+
2+
the hydrolysis of silicate during the co-precipitation. Cations such as NH , Zn , and even
4
2
+
Cu could also act as a “promoter” to accelerate the polymerization reaction to form poly
silicic acid which further coagulated to inertial silica gel. So introduction of ammonium and
zinc ions might have reduced the number of free monomeric silicic acid and changed the ratio
of Si and M in the solution. The concentration of free monomeric silicic acid was lower; the
probability of forming 1:1 phyllosilicates was higher.
It was reported that zinc species can also react with silica to form the zinc phyllosilicates
[21]
during impregnation or cation exchange processes . But in this study there was no
evidence to identify this compound, probably due to the low composition. However, due to
the difference in ionic radius between copper and zinc, it still could be inferred that
introduction of zinc cation would interfere with the ordered lamellar structures of copper
phyllosilicates, and the 1:1 phyllosilicate structure would be more likely to form.