Formic acid directly assisted solid-state synthesis of metallic catalysts without further reduction: As-prepared Cu/ZnO catalysts for low-temperature methanol synthesis

Article abstract of DOI:10.1016/j.jcat.2013.02.025

Metallic catalysts (Cu/ZnO) and pure metals (Co, Ni, and Ag) without any impurities are directly prepared by a novel formic acid-assisted solid-state method without further reduction. During the decomposition of metal-formic acid precursors at 523 K under argon, H₂ and CO are liberated and act in situ as reducing agents to obtain pure metals and metallic catalysts (C _argon). X-ray diffraction, X-ray photoelectron spectroscopy, energy-dispersive spectroscopy, and temperature-programmed reduction analysis reveal that the as-prepared catalyst C_argon without further reduction is converted into metallic Cu⁰ and ZnO species. TPR analysis results, Fourier transform infrared analysis, and the thermal decomposition behavior in air illustrate that no amorphous carbon or carbonic residues are left in C_argon when formic acid is used as the chelating agent and reductant, because formic acid is the simplest organic acid. The as-prepared Cu/ZnO catalyst is tested for low-temperature methanol synthesis at 443 K from syngas containing CO₂ and using ethanol as a solvent and catalyst; it exhibits much higher activity and methanol selectivity than catalysts prepared by conventional solid-state methods.

Full text of DOI:10.1016/j.jcat.2013.02.025

Journal of Catalysis 302 (2013) 83–90
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Formic acid directly assisted solid-state synthesis of metallic catalysts without
further reduction: As-prepared Cu/ZnO catalysts for low-temperature
methanol synthesis
Lei Shi a,b, Wenzhong Shen , Guohui Yang , Xiaojing Fan , Yuzhou Jin , Chunyang Zeng ,
c
a
d
a
a
c
a,e,
⇑
Kenji Matsuda , Noritatsu Tsubaki
a
Department of Applied Chemistry, School of Engineering, University of Toyama, Gofuku 3190, Toyama 930-8555, Japan
Department of Applied Chemistry, Shenyang University of Chemical Technology, Shenyang 110142, People’s Republic of China
Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, People’s Republic of China
Department of Material Engineering, School of Engineering, University of Toyama, Gofuku 3190, Toyama 930-8555, Japan
JST, ACT-C, Tokyo, 102-8666, Japan
b
c
d
e
a r t i c l e i n f o
a b s t r a c t
Article history:
Metallic catalysts (Cu/ZnO) and pure metals (Co, Ni, and Ag) without any impurities are directly prepared
by a novel formic acid-assisted solid-state method without further reduction. During the decomposition
of metal–formic acid precursors at 523 K under argon, H and CO are liberated and act in situ as reducing
2
agents to obtain pure metals and metallic catalysts (Cargon). X-ray diffraction, X-ray photoelectron spec-
troscopy, energy-dispersive spectroscopy, and temperature-programmed reduction analysis reveal that
the as-prepared catalyst Cargon without further reduction is converted into metallic Cu0 and ZnO species.
TPR analysis results, Fourier transform infrared analysis, and the thermal decomposition behavior in air
illustrate that no amorphous carbon or carbonic residues are left in Cargon when formic acid is used as the
chelating agent and reductant, because formic acid is the simplest organic acid. The as-prepared Cu/ZnO
Received 20 September 2012
Revised 26 February 2013
Accepted 27 February 2013
Keywords:
Metallic catalysts
Pure metals
Formic acid
Solid-state method
Waste free
Reduction-free
Cu/ZnO
Low-temperature methanol synthesis
2
catalyst is tested for low-temperature methanol synthesis at 443 K from syngas containing CO and using
ethanol as a solvent and catalyst; it exhibits much higher activity and methanol selectivity than catalysts
prepared by conventional solid-state methods.
Ó 2013 Elsevier Inc. All rights reserved.
1
. Introduction
and selectivity of the reaction. A simpler and wastewater-free so-
lid-state method [8] using citric acid and metal nitrates was pro-
Metallic catalysts are widely used in industrial catalytic pro-
cesses, such as hydrogenation, dehydrogenation, isomerization
1], reforming, and selective hydrocarbon oxidation [2]. General
methods [3] for the synthesis of metallic catalysts are to reduce
metal oxides with H or CO at high temperatures. However, the ex-
posed to prepare metallic Cu/ZnO catalysts that displayed much
higher activity and methanol selectivity than the sol–gel autocom-
bustion method [5], but the influence of amorphous carbon and
carbonic residues cannot be eliminated. For the industrial prepara-
[
2
2 3
tion of many metallic catalysts [9], such as Cu/ZnO/Al O catalysts,
tra cost for reduction energy and engineering makes the whole cat-
alyst preparation process inefficient.
large amounts of concentrated nitrate wastewater, which is envi-
ronmentally harmful, are produced through the application of ni-
Jiang et al. reported a citric acid–nitrate sol–gel autocombustion
method for preparing pure metals and alloys [4]. We developed
this technology (sol–gel autocombustion method) to directly pre-
trate salts as soluble metal sources.
A
nitrate-free co-
precipitation synthesis route [10] for Cu/ZnO catalysts using Cu
and Zn formate as precursors has been developed. However, the
formate precursors are much more expensive than nitrate salts.
Furthermore, the resulting CuO/ZnO precursors require an addi-
tional reduction step to yield the active Cu/ZnO phase.
2
pare metallic Cu/ZnO [5,6] and Co/SiO catalysts [7] without fur-
ther reduction. However, small amounts of the amorphous
carbon and carbonic residues derived from decomposition of citric
acid, as identified by Raman and FT-IR spectroscopy [5–7], are left
in the as-prepared catalysts, which seriously lowers the activity
Here, a novel formic acid-assisted solid-state method is pro-
posed. In contrast to the solid-state oxalate-precursor method
[
11] and the polyol preparation process [12] for preparing CuO-
based catalysts, this novel proposed method can prepare metallic
catalysts (Cu/ZnO) and pure metals (Co, Ag, and Ni) directly
⇑
Corresponding author. Fax: +81 76 445 6846.
E-mail address: tsubaki@eng.u-toyama.ac.jp (N. Tsubaki).
0
021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.02.025

8
4
L. Shi et al. / Journal of Catalysis 302 (2013) 83–90
without any impurity and further reduction. During the solid-state
process, brownish-yellow NO gas is released and absorbed by
sors (nitrate-free) with uniform color are generated. Subsequently,
the muddy state metal–formic acid precursors are dried at 393 K
for 10 h. Then, they are transformed to foamy bluish metal–formic
acid precursors, as shown in Fig. 2a. Next, the precursors are cal-
cined under argon at 673 K for 3 h and then passivated by 1% oxy-
gen diluted by nitrogen for 4 h. The as-prepared catalysts (denoted
as Cargon, Fig. 2b) with homogeneous dark red color are directly
used for low-temperature methanol synthesis without hydrogen
x
deionized water. As compared with the industrial preparation of
Cu/ZnO catalysts by a co-precipitation method that creates large
amounts of concentrated nitrate wastewater (sodium nitrate), this
novel formic acid-assisted solid-state method produces only a
small amount of wastewater, which can be recycled and used as ni-
tric acid. Homogeneous metal–formic acid precursors (nitrate-free)
with uniform color are generated. Pure Ni, Co, Cu, and Ag metals
without any impurity can be obtained directly when the metal–
formic acid precursors are calcined in argon at temperatures lower
than 673 K. Formic acid decomposition can occur in two different
reduction. The Cu/formic acid molar ratio must be controlled to
0
get metallic Cu . As reference 1 (Cnone), only Cu(NO
3
)
2
2
Á3H O and
3
Zn(NO )
2
Á6H
2
O are physically mixed without formic acid. The Cnone
catalysts are reduced by a flow of 5% hydrogen under nitrogen at
553 K for 6 h and successively passivated by 1% oxygen diluted
by nitrogen (noted as Cnone-reduction). As reference 2, the as-pre-
ways, either by decarboxylation affording H
onylation producing CO and H O [13]. During decomposition of the
metal–formic acid precursors, CO and H act as reductants to ob-
2 2
and CO or by decarb-
2
2
2
pared Cargon is also reduced by H under the same conditions as ref-
tain metallic catalysts. This novel formic acid-assisted solid-state
method has the following features: (1) Highly homogeneous solid
metallic catalysts without any impurities and without further
reduction can be synthesized directly; (2) The catalyst synthesis
process is easily operated, and inexpensive metal nitrates and for-
mic acid are used as raw materials; (3) During the whole catalyst
preparation process, only a small quantity of water with absorbed
NO gases is produced, which can also be recycled; (4) No carbon or
x
carbonic residues are formed on the final catalyst.
In this study, metallic Cu/ZnO catalysts are synthesized by this
novel formic acid-assisted solid-state method and used directly in
low-temperature methanol synthesis without further reduction.
This new route of low-temperature methanol synthesis is proposed
erence 1 (denoted as Cargon-reduction). As reference 3, the former
reported catalyst [8] prepared by a citric acid-assisted solid-state
method is denoted as Ccitricacid
.2. Characterization of catalysts
The characterization conditions and methods of X-ray diffraction,
.
2
temperature-programmed reduction, Fourier transform infrared
spectroscopy, scanning electron microscopy, thermogravimetry dif-
ferential thermal analysis, the Brunner–Emmett–Teller method,
2
and N O pulse chemisorption techniques are the same as those re-
ported in our previous work [5–7]. X-ray photoelectron spectroscopy
is used to study the chemical composition and oxidation state of cat-
alyst surfaces. Spectra are recorded on a Thermo ESCALAB 250 Xi
from CO/CO
lyst and solvent, by which methanol is produced at 443 K and
.0 MPa [14–18]. CO in the syngas not only participates in the
2 2
/H on Cu/ZnO-based catalyst using alcohol as a cata-
using monochromatized AlKa radiation (1486.6 eV) with a spot size
5
2
of 320 m. Before being transferred to the analysis chamber, the
l
reaction, but also promotes the activity of the reaction. As the alco-
hol solvent contains a small amount of water, it is considered that
this new low-temperature process can directly use low-grade syn-
samples are first placed in the pretreatment chamber and outgassed
under vacuum. The C1s, Cu2p, Zn2p, and O1s are recorded with pass
energy 20 eV, dwell time 50 ms, and step size 0.1 eV, and the corre-
sponding binding energies are referenced to the C1s line at 284.88 eV.
2 2
gas containing CO and H O, without purification.
2
. Experimental
2.3. Catalytic activity tests with as-prepared Cu/ZnO catalysts
2
.1. Preparation of catalysts
A batch reactor with inner volume 85 ml and a stirrer is used to
study the catalytic activity. A sample of 1 g catalyst and 40 ml eth-
anol is first poured into the autoclave. The air inside is purged by
A schematic flow chart of the catalyst prepared by a novel for-
mic acid-assisted solid-state method is displayed in Fig. 1. In this
method, 0.025 mol analytical grade Cu(NO (P99%),
Á3H
.025 mol Zn(NO O (P99%), and 0.6 mol formic acid (88%)
Á6H
are physically mixed in the mortar and ground for 10 min in air.
During the solid-state process, brownish-yellow NO gas is re-
2 2
the reactant gas, Ar/CO/CO /H = 3.13/33/5.23/58.64; then, the
3
)
2
2
O
pressure in the reactor is raised to 5.0 MPa at 298 K, and the reac-
tion takes place at 443 K for 4 h. The stirring speed is fixed at
2000 rpm to prevent the diffusion-controlled regime from devel-
oping. Continuous low-temperature methanol synthesis is also
conducted for 40 h at 443 K and 5.0 MPa using a flow-type semi-
batch reactor with 3 g catalyst and 40 ml ethanol. The flow rate
of the syngas is fixed at 20 ml/min. All products are determined
on a GC–MS (Shimadzu GCMS 1600). A TCD (Shimadzu GC-320)
is used to analyze gaseous products, and a FID (Shimadzu GC-8A)
is used to analyze liquid products.
0
3
)
2
2
x
leased, and the gases are absorbed by deionized water, which
can be recycled. Finally, homogeneous metal–formic acid precur-
3
. Results and discussion
3.1. XRD analysis of the metallic catalysts
As compared in Fig. 3a, the XRD patterns of Cnone without formic
acid assistance are indexed to CuO and ZnO phases. No metallic Cu
phase is found. However, in Fig. 3c, all diffraction peaks of Cargon are
attributed to Cu and ZnO phases. These findings evidently prove
that copper–formic acid precursors are reduced to metallic Cu dur-
ing decomposition of the copper–formic acid precursors in argon.
Fig. 3b and d compare the XRD patterns of Cnone-reduction and Car-
gon-reduction after 6 h reduction in 5% hydrogen at 553 K. It is clear
Fig. 1. Schematic flow chart of the catalysts prepared by a novel formic acid-
assisted solid-state method.

L. Shi et al. / Journal of Catalysis 302 (2013) 83–90
85
Cnone-reduction (80 nm), Cargon (18 nm), and Cargon-reduction (20 nm) is
distinctive. Formic acid with its carboxylate group is a complexing
2+
ligand. It forms very stable bidentate complexes with Zn ions.
2
+
Without the assistance of formic acid, Zn is the main dissolved
species, in equilibrium with ZnO as the precipitated phase. During
the formic acid-assisted solid-state process, the surface adsorption
of aqueous formic acid via ligand exchange process is followed by
the subsequent dissolution of the surface as a result of polarized
2Àx
and weakened metal–oxygen bonds. Zn(HCO
2
)x
become the
main dissolved species in equilibrium with ZnO. As reported, the
morphology and crystallite sizes of ZnO are significantly affected
by the pH value [19,20], soft templates (organic species) [19,20],
and calcination temperature. In this study, the presence of formic
acid remarkably influences the dissolution and solid-state pro-
cesses, and the amount of formic acid obviously affects the pH va-
lue, resulting in much smaller ZnO crystallite sizes.
Fig. 2. Photos of (a) the metal–formic acid precursors of Cargon and (b) as-prepared
C
argon
.
3.2. XPS analysis of the as-prepared samples
The chemical composition and oxidation state of as-prepared
catalyst surfaces are studied by X-ray photoelectron spectroscopy.
As compared in Fig. 4a for Cnone, the peak positions (933.6 and
9
43.5 eV) and the spectral shapes are in good agreement with
those of standard CuO, as shown in Fig. S3 (Supporting informa-
tion), based on standard spectroscopy and the literature [21,22].
In Fig. 4b for Cargon, a shift of the binding energy of Cu2p3/2 from
9
33.6 to 932.7 eV can be observed. The latter indicates that only
0
metallic Cu exists on the surface of Cargon [21,22]. This result con-
forms that when this novel formic acid-assisted solid-state method
is used to prepare Cu/ZnO catalyst, the surface state (Cu species) of
0
as-prepared catalysts is metallic Cu without CuO or Cu
2
O. In fact,
Fig. 3. XRD patterns of (a) Cnone, (b) Cnone-reduction, (c) Cargon, and (d) Cargon-reduction
s, CuO; I, Cu; j, ZnO).
(
the surface state is also influenced by the molar ratio of Cu/HCOOH,
as displayed in Table 1 in the Supporting information.
that no obvious difference is observed between Fig. 3c and d, which
further confirms that Cu2+ ions is already reduced to metallic Cu
3.3. TPR analysis of Cnone, Cargon, and Cargon-reduction
when this novel formic acid-assisted solid-state method is used.
After 6 h reduction, the CuO phase in Cnone-reduction is also reduced
to metallic Cu. As shown in Fig. S1 (Supporting information), not
only metallic Cu/ZnO, but also pure Co with fcc and hcp, Ag with
fcc, and Ni with fcc phases are identified without any impurity.
The preparation methods are also presented in the Supporting
information.
The average Cu and ZnO crystallite sizes of the as-prepared cat-
alysts calculated using the Scherrer formula are presented in Ta-
ble 1. The Cu crystallite size of Cnone-reduction is about 56 nm,
much larger than that of Cargon (36 nm) or Cargon-reduction (40 nm).
It is interesting that the difference in ZnO crystallite sizes in
2
The TCD signals from H TPR of Cnone, Cargon and Cargon-reduction
are compared in Fig. 5 in terms of hydrogen consumption. As
shown in Fig. 5a, the TPR profile of Cnone consists of one main peak
at 556 K and a small shoulder peak at about 504 K, which are re-
lated to the reduction in different CuO species. The main TPR pro-
file at about 556 K corresponds to the reduction in bulk CuO
[
23,24]. The small shoulder peak at about 504 K indicates the pres-
ence of Cu–ZnO interaction in Cnone, which facilitates the reduction
of CuO species. As shown in Fig. 5b for Cargon, a small broad peak
can be observed at lower temperature (428 K), which is attributed
to the reduction of small quantities of the surface CuO phase
formed during passivation by 1% oxygen. The TPR profile of
Table 1
Characterization results for the physical properties and the batch reaction results of low-temperature methanol synthesis.
a
2
b
2
Cuc
ZnOc
À3 À1
)e
Catalysts
S
BET (m /g)
S
Cu (m /g)
Conversion (%)
TOF (10
s
Selectivity (%)
Yield (%)
d
Dia. (nm)
Dia. (nm)
CO
CO
–
56.5
71.9
70.9
2
C
–
29.5
42.2
41.6
Total
CH
3
OH
HCOOC
2
H
5
CH
3
OH
2 5
HCOOC H
C
C
C
C
none
0.95
4.16
15.29
14.99
–
–
78
80
18
20
–
–
–
–
–
none-reduction
argon
2.65
4.42
4.27
56
36
40
25.2
37.4
36.9
12.76
10.94
11.17
59.4
78.2
76.9
40.6
21.8
23.1
17.5
33.0
32.0
12.0
9.2
9.6
argon-reduction
Note: Reaction conditions: T = 443 K, P = 5.0 MPa, catalyst weight: 1 g, ethanol solvent: 40 ml, stirring speed: 2000 rpm, reaction time: 4 h, syngas: Ar/CO/CO
.23/58.64.
2 2
/H = 3.13/33/
5
a
Determined by N
Determined from N
Calculated by Scherrer formula.
2
physical adsorption–desorption at 77 K.
O pulse chemisorption.
b
c
2
d
e
C
Total = CO conv. Â a/(a + b) + CO
2
conv. Â b/(a + b) (a, b were the contents of CO, CO
2
in the feed gas).
TOF is the total carbon turnover frequency.

8
6
L. Shi et al. / Journal of Catalysis 302 (2013) 83–90
the chelating/reducing agent. Cu and Zn weight ratio are about
3.93% and 44.41%, which are nearly the same as the initial catalyst
Cu 2p3/2
9
33.6
4
943.5
design with a Cu/Zn mole ratio of 1/1. O is about 11.07 wt.%, which
equals the amount (44.41 wt.% Â 16/65.39 = 10.87 wt.%) of that in
the form of ZnO, indicating that surface Cu species are in the state
a
9
32.7
0
of metallic Cu , which is consistent with XRD, XPS, and TPR analy-
sis. According to the EDS analysis, it can be seen that almost no C is
left in the as-prepared catalyst Cargon
.
b
9
45
940
935
930
3.5. TG-DTA analysis of the precursors for Cnone and Cargon in argon to
trace the reduction process
Binding Energy (eV)
Fig. 4. XPS spectra of Cu2p3/2 for the as-prepared samples: (a) Cnone and (b) Cargon
.
Thermal behavior of the precursors was investigated by TG-DTA
measurement from 298 to 873 K under argon. As shown in Fig. 7a
for the precursor of Cnone, the DTA plot shows one main endother-
mic peak at about 523 K. The TG plot shows a sustained weight loss
from about 490 to 550 K, which is attributed to decomposition of
the metal nitrates. However, for the metal–formic acid precursors
of Cargon, one exothermic peak is observed at about 503 K in
5
56 K
5
04 K
2+
Fig. 7b, which corresponds to the redox process between Cu
and reductant. As compared in Fig. S2 (Supporting information),
Cu and ZnO phases without any impurities can be obtained when
the metal–formic acid precursors of Cargon are calcined at a lower
a
b
c
4
28 K
temperature (523 K). During this process, gases such as H
2 2
, H O,
CO, and CO are liberated. H and CO are the reducing agents that
can be used in this process to obtain metallic Cu from Cu in the
metal–formic acid precursors. The whole reduction process can be
2
2
0
2+
3
50 400 450 500 550 600 650 700 750 800
Temperature (K)
separated into two steps: (1) generation of H
2
and CO by decompo-
Fig. 5. TPR profiles of (a) Cnone, (b) Cargon, and (c) Cargon-reduction
.
sition of the metal–formic acid precursors at about 490 K; (2)
2
+
reduction of Cu in the precursors by H
2
and CO, resulting in the
0
formation of metallic Cu .
Cargon-reduction after 6 h reduction by a flow of 5% hydrogen is dis-
played in Fig. 5c. It is clear that the shape and reduction tempera-
ture peak maxima in Fig. 5b and c are similar. This result confirms
that Cu2+ is reduced to metallic Cu0 during decomposition of the
copper–formic acid precursors, similar to that observed with
3
.6. FT-IR analysis of the precursors, Ccitricacid, and Cargon
The FT-IR spectra of the precursors, Ccitricacid, and Cargon are com-
pared in Fig. 8. Assignment of the bands for adsorbed species is made
by analogy with the spectra of known compounds and by compari-
son with published the literature. In Fig. 8a for the metal–formic acid
precursors of Ccitricacid, the bands at 820 and 1380 cm are assigned
to NO . Other complicated bands are clearly identified in previous
Cnone-reduction, as confirmed by XRD and XPS analysis.
À1
3.4. SEM-EDS analysis of the as-prepared catalyst Cargon
À
3
À
An SEM micrograph, magnified 500 times, of the catalyst Cargon
studies [5,8], which are attributed to carboxyl ions (COO ) and free
is shown in Fig. 6. The crushed section image displays bound
agglomerate with a compact aspect. The EDS analysis of the as-dis-
played SEM micrograph is also displayed in Fig. 6. Only 0.59 wt.% C
carboxylic groups. However, for the metal–formic acid precursors of
À1
C
argon in Fig. 8b, no bands are found at 820 or 1380 cm , suggesting
À
that no NO3 is left in the precursor of Cargon. This phenomenon also
indicates that during the solid-state process, NO3 in the metal ni-
À
stays on the surface of the
1
C
C
argon catalyst, compared with
4.14 wt.% C that is left on the surface of the formerly reported
citricacid catalyst [8]. This result shows that the amount of residual
trate is released in the form of NO
1351 and 1542 cm are assigned to the asymmetric and symmetric
stretching vibrations of the bidentate formate species (OCO) on Cu,
x
(gases). The bands at about
À1
surface carbon is significantly reduced when formic acid is used as
1
000
9
8
7
6
00
00
00
00
Weight (%)
Cu
O
C
O
0.59%
11.07%
Zn
5
4
3
2
1
00
00
00
00
00
0
Cu 43.93%
Zn 44.41%
Zn
Cu
C
0
.00
2.00
4.00
6.00
8.00
10.00
keV
Fig. 6. SEM-EDS analysis of the as-prepared catalyst Cargon
.

L. Shi et al. / Journal of Catalysis 302 (2013) 83–90
87
1
1
0
0
0
0
0
0
.1
.0
.9
.8
.7
.6
.5
.4
3.7. TG-DTA analysis of the as-prepared catalysts Ccitricacid and Cargon in
air to study the carbon content left in the catalysts
(a)
80
40
0
TG
The content of amorphous carbon left in the as-burnt catalysts
DTA
is investigated by TG-DTA measurements in air with a flow of
À1
5
0 ml/min and at a heating rate of 10 K min
from 293 to
-40
1
073 K. As seen in Fig. 9a1 for Ccitricacid, about a 2.7% weight in-
crease is observed from 430 to 530 K, and simultaneously, the
DTA curve in Fig. 9a2 displays a broad exothermic peak that is
attributed to the oxidation of Cu. With continually increasing tem-
perature, the TG plot exhibits an obvious weight loss of about
-
-
80
120
3
00
400
500
600
700
800
1
2.5%, with two exothermic peaks at about 540 and 628 K ob-
Temperature (K)
served in the corresponding DTA plot. The former exothermic peak
at 540 K arises from the oxidation of amorphous carbon, which is
supported by the experimental phenomena reported by Liu et al.
1
1
0
0
0
0
0
0
.1
.0
.9
.8
.7
.6
.5
.4
(b)
80
[
6
28] and Pranda et al. [29]. The latter exothermic peak at about
28 K is derived from the pyrolysis of organic residues, which is
4
0
0
DTA
consistent with previously reported data [5] showing that if the
precursor is burnt directly in air, the exothermic peak of the pyro-
lysis of organic residues is just about at 630 K. This phenomenon
demonstrates that when citric acid is used as chelating agent and
reductant, small amounts of the amorphous carbon and carbonic
residues are left in the as-burnt catalyst Ccitricacid, which is also sup-
ported by FT-IR and EDS analysis.
-40
TG
-
-
80
120
3
00
400
500
600
700
800
However, for Cargon in Fig. 9b1, a gradual weight increase from
30 to 706 K is observed. The DTA curve in Fig. 9b2 displays two
Temperature (K)
4
Fig. 7. TG-DTA curves of (a) the precursors for Cnone and (b) the metal–formic acid
precursors for Cargon calcined in argon.
main exothermic peaks. The first exothermic peak is the same as
in Fig. 9b1. The second broad exothermic peak from about 530 to
6
20 K is attributed to the oxidation of Cu to CuO. The direct proof
is that the TG plots display a continuous weight increase from
about 530 to 620 K. The weight increase ratio is about 9.2% [de-
rived from (0.491–0.45)/0.45 = 9.2%], which equals that of O from
Cu to CuO in the as-burnt catalyst Cu/ZnO, calculated as 16/
(63.55 + 65.39 + 16.00) = 11%, as only O can be contributed to the
weight increase in the formation from Cu to CuO. Metallic copper
2
2
1
1
0
0
.5
.0
.5
.0
.5
.0
1
380
8
20
a
1
542
1570
595
1351
370
1
b
c
1
undergoes a two-step oxidation [7,30] from Cu to Cu
2
O (480–
1565
1400
8
40
520 K) and from Cu O to CuO (530–620 K). The second broad exo-
2
thermic peak from about 530 to 620 K is attributed to the oxidation
of metallic Cu to CuO, because this oxidation process happens
d
4
000 3500 3000 2500 2000 1500 1000 500
-
1
Wavenumber (cm )
1
1
1
1
1
0
0.90
0
0
.20
.15
.10
.05
.00
.95
TG
Fig. 8. FT-IR analysis of (a) the metal–citric acid precursors of Ccitricacid and (b) the
a1 C
citric acid
06 K b1 Cargon
metal–formic acid precursors of Cargon, (c) Ccitricacid, and (d) Cargon
.
7
À1
+ 9.2%
(
H–COO–Cu) [16]. The band at about 1370 cm comes from the
asymmetric stretching vibrations of the bidentate formate species
+
2.7%
- 12.5%
on Zn, (H–COO–Zn), while two bands occur near 1570 and
À1
1
595 cm , representing the asymmetric stretching vibrations of
formate on Zn. It is clear that the FT-IR spectra of the metal–formic
acid precursors for Cargon in Fig. 8b are all attributed to metal–for-
.85
.80
mate species, without any NO
3
.
300 400 500 600 700 800 900 1000
For the burnt catalyst Ccitricacid (in argon), two broadbands at about
DTA
2
2
150
100
50
00
exo
a2 Ccitric acid
628 K b2 Cargon
À1
1
400 and 1565 cm are presented in Fig. 8c, which are assigned to the
540 K
asymmetric and symmetric stretching vibrations for carboxyl ions
À
(
COO ) [25,26] from metal carboxylates. These results clarify that
4
71 K
small amounts of carbonic residues (carboxyl ions) are left in Ccitricacid
.
À1
A broadband at about 840 cm is a further proof of the formation of
5
0
0
carbonates [27], which are generated during the combustion process
and left in Ccitricacid. However, for Cargon in Fig. 8d, almost no bands
555 K
À1
-50
are observed at about 840, 1351, 1400, and 1560 cm , which indicates
that when this formic acid-assisted solid-state method is applied to
prepare metallic catalysts directly, almost no carbonic residues are left
in the as-burnt products. This is thought to be due to the fact that no C–
C covalent bonds exist in formic acid.
-100
3
00 400 500 600 700 800 900 1000
Fig. 9. TG curves of (a1) Ccitricacid and (b1) Cargon as well as the DTA curves of (a2)
citricacid and (b2) Cargon
C
.

8
8
L. Shi et al. / Journal of Catalysis 302 (2013) 83–90
above 523 K [28] and displays a much broader and smaller exo-
thermic peak than that for the oxidation of amorphous carbon in
Fig. 9a2. Further proof of copper oxidation is given by the TPR anal-
ysis of Cargon at higher temperatures, as displayed in Fig. S4 (Sup-
porting information). According to a previous work [7] using
citric acid as a chelating/reducing agent and burnt in argon, a broad
peak at about 790 K is assigned to the hydrogenation of amorphous
carbon. This is confirmed by the formation of methane, as observed
by TPR–MS analysis of the off gas. In Fig. S4, no peaks are observed
at higher than 800 K, suggesting that no carbon is left in Cargon. As
displayed in Fig. 9b2, no peaks are seen at about 628 K, which is
attributed to the oxidation of organic residues. These phenomena
indicate that when formic acid is used as chelating agent and
reductant, no carbon or carbonic residues are left in Cargon. As pre-
viously stated, the lack of carbon/carbonic residues is thought to be
due to the fact that no covalent C–C bonds exist in formic acid. Car-
bon residues can also potentially be formed via the Boudouard
reaction, but this reaction is unlikely under the present conditions
because of the following: (1) During the decomposition of the me-
tal–formic acid precursors, CO is generated from the decarbonyla-
tion of formic acid. According to the TG-DTA and MS analysis in
argon, CO is produced in a split second at about 523 K and is rap-
idly taken away by argon with a flow rate of 50 ml/min. (2) CO is
Characterization results and the batch reaction records are dis-
2
À1
À1
played in Table 1. The BET surface area of Cnone is only 0.95 m g
.
.
2
After reduction, the BET surface area increases to 4.16 m g
2
À1
However, for Cargon, the BET surface area is 15.29 m g , which is
much larger than that of Cnone. No significant BET surface area dif-
ferences are observed between Cargon and Cargon-reduction
.
As shown in Table 1, for this low-temperature methanol synthe-
sis method, only methanol and ethyl formate are the liquid-phase
products for all the as-prepared catalysts. The total carbon conver-
sion and the methanol selectivity of Cargon and Cargon-reduction after
2
+
reduction are consistent which confirms that Cu is reduced to
metallic Cu0 during decomposition of the metal–formic acid pre-
cursors at 523 K in an argon atmosphere. The total carbon conver-
sion of Cargon (42.2%) is much higher than that of Cnone-reduction
0
(29.5%) because of its larger BET and copper (Cu ) surface area. This
phenomenon is in good agreement with the literature data [31,32],
because higher Cu surface area provides more active sites. Cu0 is
the active site for the ester hydrogenation, a rate-determining step
in this low-temperature methanol synthesis.
On the other hand, it is notable that the methanol selectivity
(78.2%) of Cargon is much higher than that of Cnone-reduction (59.4%).
The methanol selectivity of formerly reported [8] Ccitricacid is lim-
ited to about 37.6%, because small amounts of carbonic residues
left in the as-burnt catalyst significantly influence the valence state
of Cu, which will affect the formation and hydrogenation of ethyl
formate, resulting in much lower methanol selectivity. The pro-
posed mechanism [15,16] for this low-temperature methanol syn-
accompanied by a large amount CO
influence the Boudouard reaction in the formation of C and CO
3) According to our knowledge, no reports of the Boudouard reac-
tion proceeding on Cu-based catalysts at temperatures of about
23 K have been reported.
2 2
and H O gases, which will
2
.
(
5
2 2
thesis from syngas (CO/CO /H ) using ethanol as a catalyst and
+
solvent is that Cu and ZnO are the active sites for the formation
0
of formate and ethyl formate, but only Cu is active for the hydro-
3.8. TEM and HRTEM analysis of the as-prepared catalyst Cargon
genation of ethyl formate. Higher methanol selectivity may be de-
rived from smaller Cu crystallite sizes. It is considered that the
methanol selectivity is associated with the crystallite sizes of Cu.
Smaller copper crystallite sizes favor methanol formation, as they
exhibit a higher hydrogenation capability. This phenomenon is
accordance with our precious reports [5,6,8].
TEM analysis of the as-prepared Cargon is shown in Fig. S5a. The
as-prepared catalyst Cargon exhibits a relatively homogeneous Cu
and ZnO particle distribution. HRTEM of Cargon is compared in
Fig. S5b. This microstructure, which displays a characteristic Cu/
ZnO structure, is similar to that observed in Cu/ZnO catalysts pre-
pared by conventional methods. The crystallite size of Cu (HRTEM)
is about 30 nm, which is similar to that calculated by the Scherrer
formula.
1
00
100
90
80
70
60
50
40
30
20
10
0
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
3.9. Activity of the as-prepared catalysts
The activity of Cnone-reduction, Cargon, and Cargon-reduction is investi-
gated for low-temperature methanol synthesis from syngas con-
taining CO using ethanol as a solvent and catalyst at 443 K and
.0 MPa for 4 h. This synthetic route [14–18] with alcohols as sol-
vent and catalyst allows using syngas containing CO and H O di-
2
5
2
2
2
4
6
8
10 12 14 16 18 20
rectly from an industrial natural gas reformer or biomass/coal
gasifier without deep purification. This new route can operate at
significantly low temperature to remarkably increase one-step
CO conversion, as the efficiency of methanol synthesis at high tem-
perature is severely limited by thermodynamics because it is an
extremely exothermic reaction.
Reaction time (h)
Fig. 10. The batch results for different reaction times for low-temperature
methanol synthesis. Reaction conditions: T = 443 K, P = 5.0 MPa, catalyst weight:
1 g, ethanol solvent: 40 ml, stirring speed: 2000 rpm, syngas: Ar/CO/CO /H2 = 3.13/
2
33/5.23/58.64.
Table 2
Continuous reaction results for low-temperature methanol synthesis with Cargon and Cnone-reduction
.
TOFa (10
À3 À1
s
)
Selectivity%
CH OH
Yield%
CH OH
Catalyst
Conversion%
CO
CO
2
C
total
3
HCOOR
3
HCOOR
C
C
argon
28.9
13.5
À2.0
À0.8
24.6
11.5
4.4
3.4
98.6
75.6
1.4
24.4
24.3
8.7
0.3
2.8
none-reduction
Note: Reaction conditions: T = 443 K, P = 5.0 MPa, catalyst weight: 3 g, solvent: ethanol 40 ml, stirring speed: 2000 rpm, reaction time: 40 h, syngas: Ar/CO/CO
.23/58.64, 20 ml/min.
2
/H
2
= 3.13/33/
5
a
2
Based on N O pulse method. TOF is total carbon turnover frequency.

L. Shi et al. / Journal of Catalysis 302 (2013) 83–90
89
5
23 K in an argon atmosphere, H
2
and CO are liberated and act as
3
2
1
0
0
0
0
reducing agents to obtain pure metal or metallic catalysts. The X-
ray diffraction, X-ray photoelectron spectroscopy, EDS, and temper-
CO %
Ctotal%
ature-programmed reduction analysis reveal that Cargon without
further reduction is converted into metallic Cu0 and ZnO species.
CO2%
TPR analysis results, Fourier transform infrared analysis, and ther-
mal decomposition behavior in air illustrate that no amorphous
carbon or carbonic residues are left in Cargon when formic acid is
used as chelating agent and reductant because it is the simplest or-
ganic acid (HCOOH) with no C–C covalent bonds. The Boudouard
reaction is not feasible under the present conditions.
1
0
20
30
40
-
10
Time on stream (40 h)
The activity of Cargon with and without reduction is tested for
2
Fig. 11. Variations of CO, CO , and total carbon conversions with reaction time for
continuous synthesis of methanol. (Reaction conditions: T = 443 K, P = 5.0 MPa,
catalyst weight: 3 g, solvent: ethanol 40 ml, stirring speed: 2000 rpm, reaction
time: 40 h, syngas: Ar/CO/CO /H = 3.13/33/5.23/58.64, 20 ml/min.
2 2
low-temperature methanol synthesis from syngas containing CO
using ethanol as a solvent and catalyst at 443 K and 5.0 MPa for
h. The total carbon conversion and methanol selectivity are con-
2
4
sistent. Compared with Cnone-reduction and Ccitricacid, the activity and
methanol selectivity here are significantly enhanced.
Batch results for different reaction times are displayed in
Fig. 10. The Ctotal conversion increases rapidly from 2 h (about
The advantages of this formic acid-assisted solid-state method
are as follows: (1) Highly homogeneous metallic solid catalysts
without any impurity and without further reduction can be syn-
thesized directly; (2) The catalyst synthesis process is easily oper-
ated, and inexpensive metal nitrates and formic acid are used as
raw materials; (3) During the whole catalyst preparation process,
2
2%) to 8 h (about 56%). After 8 h, the Ctotal conversion is steady,
because the pressure in the batch reactor continually decreases
as the reaction proceeds. Methanol synthesis is a molecule-reduc-
ing reaction. Consequently, esterification of CO or CO does not
2
proceed at lower pressure. However, the methanol selectivity in-
creases gradually with time-on-stream, because hydrogenation of
ethyl formate proceeds continually at 443 K on Cu-based catalysts.
After 20 h reaction, the methanol selectivity is about 94%, indicat-
ing that the hydrogenation ability of this novel Cu-based catalyst is
excellent. TOF (total carbon) of Cu atom, which is based on the to-
x
only a small quantity of water with absorbed NO gases is pro-
duced, which can also be recycled and used as nitric acid; (4) No
carbon or carbonic residues are formed on the final catalyst.
Appendix A. Supplementary material
0
tal conversion and metallic Cu surface area in Fig. 10, changes
À3 À1
À3 À1
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.02.025.
from 11.4 Â 10
s
at 2 h to 2.9 Â 10
s
at 20 h.
In order to give insight into the stability of the as-synthesized
catalysts, the continuous reaction results for low-temperature
methanol synthesis at 443 K and 5.0 MPa for 40 h using Cargon or
References
C
none-reduction are compared in Table 2 and Fig. 11. Fig. 11 expresses
[
[
[
[
1] O. Deutschmann, H. Knözinger, K. Kochloefl, T. Turek, Heterogeneous Catalysis
and Solid Catalysts, Wiley-VCH, Weinheim, 2009, pp. 38–39.
2] O. Deutschmann, H. Knözinger, K. Kochloefl, T. Turek, Heterogeneous Catalysis
and Solid Catalysts, Wiley-VCH, Weinheim, 2009, pp. 82–93.
3] G. Ertl, H. Knözinger, F. Schüth, J. Weitkamp, Handbook of Heterogeneous
Catalysis, Wiley-VCH, Weinheim, 2008, pp. 57–720.
the time-on-stream activity change of the reaction during 40 h.
Analysis of the effluent gas after the trap shows that only CO,
CO , and H exist. No alcohols or esters are observed in the tail
2 2
gases. At the initial stage of the reaction, 45 ml dead volume of
the reactor and 18 ml cold trap must be filled by the pressurized
feed gas. Thus, the apparent conversions are initially low but in-
crease gradually. After 10 h, the conversions are stable. CO conver-
sion is about 28.9%, whereas CO
total carbon conversion is about 24.6%. At the initial 10 h, the
CO
conversion drops to the minimum À11% then returns to about
À2%. The negative CO conversion proves that CO is first converted
to CO through the water-shift reaction, and then, CO is hydroge-
nated to methanol. The balance between CO formation and for-
mate formation is reached after 10 h. After 40 h reaction, the
methanol selectivity calculated is as high as 98.6%. The total carbon
TOF calculated by N
pared with Cnone-reduction, the stable Ctotal conversion and methanol
selectivity are much higher.
4] Y.W. Jiang, S.G. Yang, Z.H. Hua, H.B. Huang, Angew. Chem. Int. Ed. 48 (2009)
8681–8683.
[5] L. Shi, K. Tao, R.Q. Yang, F.Z. Meng, C. Xing, N. Tsubaki, Appl. Catal. A 401 (2011)
6–55.
6] L. Shi, R.Q. Yang, K. Tao, Y. Yoneyama, Y.S. Tan, N. Tsubaki, Catal. Today 185
2012) 54–60.
[7] L. Shi, K. Tao, T. Kawabata, T. Shimamura, X.J. Zhang, N. Tsubaki, ACS Catal. 1
2011) 1225–1233.
4
2
conversion is about À2%. The
[
(
2
(
2
[
[
8] L. Shi, Y.S. Tan, N. Tsubaki, ChemCatChem 4 (2012) 861–871.
9] M. Behrens, F. Studt, I. Kasatkin, S. Kühl, M. Hävecker, F.A. Pedersen, S. Zander,
F. Girgsdies, P. Kurr, B.L. Kniep, M. Tovar, R.W. Fischer, J.K. Nørskov, R. Schlögl,
Science 336 (2012) 893–3897.
2
2
2
[
10] M. Behrens, S. Kißner, F. Girsgdies, I. Kasatkin, F. Hermerschmidt, K. Mette, H.
Ruland, M. Muhler, R. Schlögl, Chem. Commun. 47 (2011) 1701–1703.
[11] L.G. Wang, Y.M. Liu, M. Chen, Y. Cao, H.Y. He, G.S. Wu, W.L. Dai, K.N. Fan, J.
Catal. 246 (2007) 193–204.
12] C.Y. Lu, H.H. Tseng, M.Y. Wey, T.W. Hsueh, Chem. Eng. J. 145 (2009) 461–467.
13] C. Fellay, N. Yan, P.J. Dyson, G. Laurenczy, Chem. Eur. J. 15 (2009) 3752–3760.
À3 À1
2
O pulse method is about 4.4 Â 10
s
. Com-
[
[
For this novel formic acid-assisted synthesis method, the Cu/
formic acid ratio is critical preparation condition that must be con-
trolled in order to get metallic Cu/ZnO and not the oxide versions
CuO/ZnO. The effect of the Cu/formic acid ratio and the reaction re-
sults are discussed in the Supporting information, as shown in
Table S1.
[14] L. Fan, Y. Sakaiya, K. Fujimoto, Appl. Catal. A 180 (1999) L11–L13.
[
[
[
15] N. Tsubaki, M. Ito, K. Fujimoto, J. Catal. 197 (2001) 224–227.
16] R.Q. Yang, Y.L. Fu, Y. Zhang, N. Tsubaki, J. Catal. 228 (2004) 23–35.
17] N. Tsubaki, J.Q. Zeng, Y. Yoneyama, K. Fujimoto, Catal. Commun. 2 (2001) 213–
217.
[18] J.Q. Zeng, N. Tsubaki, K. Fujimoto, Fuel 81 (2002) 125–127.
[
19] A. Aimable, M.T. Buscaglia, V. Buscaglia, P. Bowen, J. Eur. Ceram. Soc. 30 (2010)
91–598.
20] I.A. Mudunkotuwa, T. i Rupasinghe, C.M. Wu, V.H. Grassian, Langmuir 28
2012) 396–403.
5
[
(
4
. Conclusions
[
[
21] W.L. Dai, Q. Sun, J.F. Deng, D. Wu, Y.H. Sun, Appl. Surf. Sci. 177 (2001) 172–179.
22] J. Agrell, H. Birgersson, M. Boutonnet, I. Melián-Cabrera, R.M. Navarro, J.L.G.
Fierro, J. Catal. 219 (2003) 389–403.
Metallic catalysts (Cu/ZnO) and pure metals (Co, Ni, and Ag) are
[
[
23] R. Kama, C. Selomulya, R. Amal, J. Scott, J. Catal. 273 (2010) 73–81.
24] M. Turco, G. Bagnasco, U. Costantino, F. Marmottini, T. Montanari, G. Ramis, G.
Busca, J. Catal. 228 (2004) 43–55.
directly prepared by a new formic acid-assisted solid-state method
without further reduction using metal nitrates and formic acid.
During the decomposition of metal–formic acid precursors at about
[25] W.D. Li, J.Z. Li, J.K. Guo, J. Eur. Ceram. Soc. 23 (2003) 2289–2295.

9
0
L. Shi et al. / Journal of Catalysis 302 (2013) 83–90
[
26] S. Vivekanandhan, M. Venkateswarlu, N. Satyanarayana, J. Alloys Compd. 462
2008) 328–334.
27] M.C. Chang, J. Tanaka, Biomaterials 23 (2002) 4811–4818.
28] J.H. Zhao, Z.Y. Liu, D.K. Sun, J. Catal. 227 (2004) 297–303.
[29] P. Pranda, K. Prando v´ , V. Hlavacek, Fuel Process. Technol. 61 (1999) 211–221.
[30] X.J. Duan, R.J. Gao, Y.D. Zhang, Z. Jian, Mater. Lett. 65 (2011) 3625–3628.
[31] G.C. Chinchen, K.C. Waugh, D.A. Whan, Appl. Catal. 25 (1986) 101–107.
[32] W.X. Pan, R. Cao, D.L. Roberts, G.L. Griffin, J. Catal. 114 (1988) 440–446.
(
[
[