Rh-catalyzed syngas conversion to ethanol: Studies on the promoting effect of FeOx

Article abstract of DOI:10.1016/j.cattod.2011.03.023

Rhodium catalysts loaded on silica modified by various transition metal oxides were investigated for the conversion of syngas to ethanol. Iron oxide was found to be an efficient promoter for ethanol formation. The increase in iron content up to 5 wt% significantly increased CO conversion and ethanol selectivity. The preparation method used for introducing FeO_x was found to affect both the conversion and the selectivity significantly. The catalyst prepared by the impregnation of a FeO_x-SiO₂ composite, which was synthesized by a sol-gel technique preliminarily, with Rh(NO₃)₃ aqueous solution provided better ethanol formation activity than those prepared by co-impregnation and co-sol-gel methods. An ethanol selectivity of 42% was achieved at CO conversion of 12% over a 5 wt% Rh/(5 wt% FeO_x-SiO₂) catalyst prepared by this method. Larger interfaces between Rh and FeO_x species were proposed to be a crucial factor for obtaining higher ethanol selectivity. The co-existence of Rh³⁺ with Rh⁰ and the size of Rh particles also played key roles in ethanol formation.

Full text of DOI:10.1016/j.cattod.2011.03.023

Catalysis Today 171 (2011) 257–265
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Rh-catalyzed syngas conversion to ethanol: Studies on the promoting effect of
FeOx
Jingjuan Wang, Qinghong Zhang, Ye Wang∗
State Key Laboratory of Physical Chemistry of Solid Surfaces, National Engineering Laboratory for Green Chemical Productions of Alcohols,
Ethers and Esters, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Rhodium catalysts loaded on silica modified by various transition metal oxides were investigated for the
conversion of syngas to ethanol. Iron oxide was found to be an efficient promoter for ethanol formation.
The increase in iron content up to 5 wt% significantly increased CO conversion and ethanol selectivity. The
preparation method used for introducing FeOx was found to affect both the conversion and the selectivity
significantly. The catalyst prepared by the impregnation of a FeOx–SiO2 composite, which was synthesized
by a sol–gel technique preliminarily, with Rh(NO3)3 aqueous solution provided better ethanol formation
activity than those prepared by co-impregnation and co-sol–gel methods. An ethanol selectivity of 42%
was achieved at CO conversion of 12% over a 5 wt% Rh/(5 wt% FeOx–SiO2) catalyst prepared by this method.
Larger interfaces between Rh and FeOx species were proposed to be a crucial factor for obtaining higher
ethanol selectivity. The co-existence of Rh3+ with Rh0 and the size of Rh particles also played key roles in
ethanol formation.
Received 30 October 2010
Received in revised form 19 February 2011
Accepted 9 March 2011
Available online 8 April 2011
Keywords:
Syngas conversion
Ethanol
Rhodium
Iron oxide
Promoting effect
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
formation activity [3,4,7–14]. However, the main products over the
supported single Rh catalysts are typically hydrocarbons (particu-
Catalytic transformations of syngas (H2 + CO), which can be pro-
duced from non-petroleum carbon resources including natural gas,
coal, and biomass, into clean fuels and valuable chemicals have
attracted much attention in recent years because of the global
demand for the decrease in the dependence on petroleum [1,2].
Ethanol is one of the attractive target products in syngas trans-
formations, because it can be used as a fuel or fuel additive and a
potential source of hydrogen for fuel cells, and it can also serve as a
feedstock for the production of a variety of chemicals and polymers
[3,4]. Many catalysts, particularly Rh-, Co-, Cu- and Mo-based cata-
lysts, have been reported to be capable of catalyzing the conversion
of syngas to ethanol [3,4]. However, these catalysts still suffer from
lower productivity. Relatively high ethanol selectivity (40–50%) can
only be achieved at low CO conversions (typically <10%). It is wor-
thy mentioning that Tsubaki and co-workers [5,6] have developed
an intriguing method for ethanol synthesis from dimethyl ether
and syngas using the combined zeolite (e.g., H-ZSM-5) and metal
(e.g., Cu/ZnO) catalysts.
larly methane) [3,4]. The presence of a promoter (e.g., transition
metal oxide such as MnOx, VOx, or FeOx) or the combination of
several promoters is required for obtaining higher ethanol selec-
tivity [7–14]. The understanding of the functioning mechanism of
the promoter will certainly be helpful for the rational design of
efficient catalysts for ethanol synthesis. A few studies have demon-
strated that the location of promoter (or the contact between
promoter and Rh) is quite important for obtaining better catalytic
performances [8,15,16]. Thus far, many of the reported stud-
ies have employed co-impregnation or sequential impregnation
for introducing the transition metal promoters [7–14]. However,
these methods cannot ensure the high dispersion of promoters on
support and the contact between Rh and transition metal oxide
promoters.
The sol–gel technique is known to be capable of produc-
ing catalysts with homogeneously distributed supported species
[17]. Recently, we prepared transition metal oxide-containing
SiO2 composites using the sol–gel technique and investigated
the catalytic performances of Rh catalysts supported on these
composite oxides with highly dispersed transition metal oxide
promoters for syngas conversions. Herein, we report the cat-
alytic behaviors and the structural features of these catalysts
with finely dispersed promoters. The effect of catalyst prepara-
tion methods is also discussed to gain insights into the active
sites.
Among the catalysts reported to date for the direct conversion of
syngas to ethanol, supported Rh shows the most promising ethanol
∗
Corresponding author. Tel.: +86 592 2186156; fax: +86 592 2183047.
E-mail address: wangye@xmu.edu.cn (Y. Wang).
0920-5861/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2011.03.023

258
J. Wang et al. / Catalysis Today 171 (2011) 257–265
2. Experimental
2.1. Catalyst preparation
a Quantum 2000 Scanning ESCA Microprob instrument (Physi-
cal Electronics) using Al–K␣ radiation. The binding energy was
calibrated using C1s photoelectron peak at 284.6 eV as a ref-
erence. Transmission electron microscopy (TEM) measurements
were performed on a JEM-2100 electron microscope operated at
an acceleration voltage of 200 kV. The mean sizes of Rh particles
were estimated from TEM micrographs by counting ca. 150–200
particles.
H2 temperature-programmed reduction (H2-TPR) was per-
formed using a Micromeritics AutoChem 2920 II instrument.
Typically, after the pretreatment of sample loaded in a quartz
reactor and cooling to 303 K, a H2–Ar gas mixture was intro-
duced into the reactor, and the temperature was raised to 1073 K
at a rate of 10 K min−1. H2 consumption was monitored by a
thermal conductivity detector. CO chemisorption was carried
out with a Micromeritics ASAP 2010 C. After the sample was
pretreated by H2 reduction at 573 K and evacuation, CO chemisorp-
tion was performed at 308 K. After the first isotherm (total CO
uptake), the sample was evacuated for 10 min, and the second
isotherm (reversible CO uptake) was measured. The amount of the
chemisorbed CO (irreversible CO uptake) was calculated using the
difference between the total and reversible CO uptakes.
Transition metal oxide (MOx)–SiO2 composites (denoted as
MOx–SiO2 hereafter) were prepared by the sol–gel technique [17].
Typically, tetraethyl orthosilicate (TEOS) and the precursor of tran-
sition metal oxide (metal nitrates except for NH4VO3 used for
VOx-modified SiO2) were dissolved in the mixture of water and
ethylene glycol, and a homogeneous sol was obtained. The sol was
then heated at 343 K for 16 h to form a homogenous gel. After being
dried at 383 K in air for 12 h, the gel was calcined at 623 K for 6 h
in air. The supported Rh catalysts were prepared by incipient wet-
ness impregnation method using Rh(NO3)3 as the precursor of Rh.
After being dried at 383 K for 12 h in air, the supported catalyst was
calcined at 623 K in air for 6 h and finally reduced by H2 at 573 K
for 2 h. The catalyst prepared by this procedure was denoted as
Rh/(MOx–SiO2).
Co-impregnation and co-sol–gel methods were also employed
for the preparation of FeOx-promoted Rh catalysts supported on
SiO2. For the co-impregnation method, the powdery SiO2, which
was prepared preliminarily by the sol–gel technique described
above, was added into the mixed aqueous solution containing
Rh(NO3)3 and Fe(NO3)3, followed by drying at 383 K for 12 h, cal-
cination at 623 K for 6 h, and H2 reduction at 573 K for 2 h. For the
co-sol–gel method, TEOS, Rh(NO3)3, and Fe(NO3)3 were dissolved
in the mixture of water and ethylene glycol, and the mixture under-
went heating at 343 K to form a homogenous gel. After being dried
at 383 K in air for 12 h, the sample was calcined at 623 K for 6 h in
air, followed by H2 reduction at 573 K for 2 h. The catalysts prepared
by the co-impregnation and co-sol–gel methods were denoted as
(Rh–FeOx)/SiO2 and Rh–FeOx–SiO2, respectively.
3. Results and discussion
3.1. Catalytic behaviors of promoted Rh catalysts supported on
SiO2
3.1.1. Catalytic performances of Rh catalysts loaded on various
MOx–SiO2 composites
Table 1 shows the catalytic performances of Rh catalysts loaded
on various MOx–SiO2 (M = transition metal) composites prepared
by the sol–gel method. Under our reaction conditions, the 1.0 wt%
Rh/SiO2 exhibited a CO conversion of 0.7% and a CH4 selectivity
of 59%. The selectivities of C2H5OH and CH3OH over this cata-
lyst were 9.5% and 4.9%, respectively. The presence of a transition
metal oxide (MOx) modifier listed in Table 1 increased the conver-
sion of CO and decreased the selectivity of CH4, which is a highly
undesirable by-product. A CO conversion of 93% was obtained
over the Rh/(CoOx–SiO2) catalyst, and this catalyst showed higher
2.2. Catalytic reaction
Catalytic reactions were performed on a fixed-bed reactor oper-
ated at 2 MPa. The catalyst loaded in the reactor was pretreated
by H2 at 573 K for 2 h. After the catalyst was cooled down to the
reaction temperature (typically 523 K), the syngas with a H2/CO
ratio of 2.0 was introduced into the reactor. Typical reaction condi-
tions were as follows: pressure (P) = 2 MPa, H2/CO = 2, temperature
(T) = 523 K, WHSV = 8000 mL g(cat)−1 h−1. The products were ana-
lyzed by on-line gas chromatography.
+
selectivities to CO2 and C2 hydrocarbons. We speculate that the
outstandingly high activity of this catalyst is due to the catalytic
functions of cobalt, since cobalt is a well-known active catalyst in
Fischer–Tropsch (FT) synthesis for the production of linear long-
chain hydrocarbons [1,2]. However, this catalyst only provided a
quite low selectivity to C2H5OH (1.8%). On the other hand, the
catalyst containing MnOx, CrOx, FeOx, or VOx promoter exhib-
ited relatively higher C2H5OH selectivity (≥20%). The selectivity
of CH4 became significantly lower over the Rh/(MnOx–SiO2) and
Rh/(FeOx–SiO2) catalysts. Among the catalysts showing C2H5OH
2.3. Catalyst characterization
XRD patterns were collected on a Philips X’Pert Pro Super X-
ray diffractometer equipped with X’Celerator and Xe detection
systems. Cu K␣ radiation (40 kV and 30 mA) was used as the X-
ray source. X-ray photoelectron spectra (XPS) were recorded with
Table 1
Catalytic performances of Rh catalysts loaded on various transition metal oxide-modified SiO2.a
Catalystsb
CO conv. (%)
Selectivity (%)
CO2
EtOH yield (%)
CH4
C2+ HCc
MeOH
EtOH
Rh/SiO2
0.7
5.0
3.0
1.5
6.3
93
0
0
0
4.1
3.8
21
1.3
4.3
59
45
38
15
21
46
44
49
7.1
20
5.1
6.0
12
27
17
22
4.9
3.7
28
17
33
1.9
15
14
9.5
20
22
39
21
1.8
19
5.6
0.1
1.0
0.7
0.6
1.3
1.6
0.9
0.4
Rh/(VOx–SiO2)
Rh/(CrOx–SiO2)
Rh/(MnOx–SiO2)
Rh/(FeOx–SiO2)
Rh/(CoOx–SiO2)
Rh/(ZrOx–SiO2)
Rh/(MoOx–SiO2)
4.9
7.1
a
Reaction conditions: H2/CO = 2; WHSV = 8000 mL g(cat)−1 h−1; P = 2.0 MPa; T = 523 K.
b
c
The loading of Rh was 1.0 wt%; the content of MOx was 10 wt%.
+
C2 hydrocarbons.

J. Wang et al. / Catalysis Today 171 (2011) 257–265
259
Table 2
Catalytic performances of Rh catalysts loaded on FeOx–SiO2 with different Fe contents.a
Fe content (wt%)
CO conv. (%)
Selectivity (%)
CO2
EtOH yield (%)
+
CH4
C2 HCb
MeOH
EtOH
0
0.7
2.7
3.5
5.3
6.3
8.2
0
59
30
29
25
21
18
7.1
4.5
3.0
5.8
12
4.9
23
22
25
33
9.5
31
32
37
21
0.1
0.8
1.1
2.0
1.3
1.6
1.0
2.0
5.0
10
3.7
2.3
2.3
3.8
4.2
20
13
35
19
a
Reaction conditions: H2/CO = 2; WHSV = 8000 mL g(cat)−1 h−1; P = 2.0 MPa; T = 523 K. Rh loading, 1.0 wt%. The Fe content was calculated on the basis of Fe2O3.
C2 hydrocarbons.
b
+
Table 3
Catalytic performances of Rh catalysts loaded on FeOx–SiO2 with different Rh loadings.a
Rh loading (wt%)
CO conv. (%)
Selectivity (%)
CO2
EtOH yield (%)
+
CH4
C2 HCb
MeOH
EtOH
0
Trace
5.3
7.6
9.5
12.4
–
–
25
24
23
26
–
–
25
20
21
18
–
37
37
44
42
–
1.0
2.0
4.0
5.0
2.3
2.3
3.0
3.5
5.8
7.5
3.9
5.2
2.0
2.8
4.2
5.2
a
Reaction conditions: H2/CO = 2; WHSV = 8000 mL g(cat)−1 h−1; P = 2.0 MPa; T = 523 K. Fe content, 5.0 wt%.
C2 hydrocarbons.
b
+
selectivities ≥20%, the Rh/(FeOx–SiO2) catalyst provided the high-
a C2H5OH selectivity was better than those reported over VOx-
, LaOx-, FeOx-, and MnOx-promoted Rh/SiO2 catalysts prepared
by co-impregnation or sequential impregnation methods [12,13],
and was similar to that reported over a 2% Rh–5% Fe/TiO2 cata-
lyst [10,16]. The increase in Rh loadings to 4 wt% or 5 wt% in our
case further raised C2H5OH selectivity, and a 42% C2H5OH selectiv-
ity was achieved at 12.4% CO conversion over our 5 wt% Rh/(5 wt%
FeOx–SiO2) catalyst. To the best of our knowledge, this is one of
the best combinations of CO conversion and C2H5OH selectivity
reported to date.
We have also investigated the effect of temperature of H2
reduction for the 1.0 wt% Rh/(5.0 wt% FeOx–SiO2) catalyst on cat-
alytic performances, and the results are shown in Table 4. This
series of catalysts were calcined at 573 K before reduction. With-
out H2 pre-reduction, CO conversion and C2H5OH selectivity were
5.0% and 26%, respectively. In this case, the catalyst could still be
reduced in situ in the reactant gas (H2 + CO) flow. The pre-reduction
of catalyst by H2 at 373–573 K enhanced both CO conversion
and C2H5OH selectivity. However, higher reduction temperatures
(≥673 K) decreased CO conversion and C2H5OH selectivity. Thus,
a proper reduction temperature is also crucial for obtaining better
C2H5OH selectivities and higher CO conversions.
est CO conversion. Therefore, considering the balance of CO
conversion and C2H5OH selectivity, we selected the Rh/(FeOx–
SiO2) catalyst for further studies.
3.1.2. Catalytic performances of Rh catalysts loaded on FeOx–SiO2
We have examined the effect of Fe content in the Rh/(FeOx–SiO2)
catalysts on their catalytic performances, and the results are shown
in Table 2. The loading of Rh was fixed at 1.0 wt%. The increase in
Fe content significantly increased the conversion of CO. The pres-
ence of FeOx led to the formation of CO2, but the selectivity of CO2
kept at <5% with Fe contents ranging from 0 to 20 wt%. The selectiv-
ity of CH4 decreased monotonically with increasing Fe content. On
the other hand, the selectivities to CH3OH and C2H5OH increased
significantly after the addition of FeOx into SiO2. The selectivity of
CH3OH increased from 4.9% to 35% as the Fe content rose from 0
to 20 wt%, while that of C2H5OH reached a maximum (37%) at a
Fe content of 5.0 wt%. In other words, the catalyst with an appro-
priate content of Fe is required for obtaining the highest C2H5OH
formation activity.
We further investigated the effect of Rh loadings on cat-
alytic behaviors of the Rh/(FeOx–SiO2) catalysts with a Fe content
of 5.0 wt%. As shown in Table 3, with increasing Rh loadings,
CO conversion increased remarkably, but the product selectivity
did not undergo significant changes at the same time. Over the
2 wt% Rh/(5 wt% FeOx–SiO2) catalyst, the conversion of CO and
the selectivity of C2H5OH were 7.6% and 37%, respectively. Such
3.1.3. Catalytic performances of FeOx-promoted Rh catalysts
prepared by different methods
Table 5 shows the catalytic behaviors of FeOx-promoted Rh cat-
alysts prepared by different methods. The catalytic performance
Table 4
Catalytic performances of the 1.0 wt% Rh/(5.0 wt% FeOx–SiO2) catalysts pre-reduced at different temperatures.a
Reduction temp.b (K)
CO conv. (%)
Selectivity (%)
CO2
EtOH yield (%)
+
CH4
C2 HCc
MeOH
EtOH
Without pre-reduction
5.0
6.1
6.5
6.0
3.2
3.0
3.8
2.7
2.0
2.1
2.7
2.4
22
21
22
16
25
24
8.9
4.8
4.8
3.3
7.6
6.5
31
21
23
29
23
31
26
45
41
42
38
31
1.3
2.7
2.7
2.5
1.2
0.9
373
473
573
673
773
a
Reaction conditions: H2/CO = 2/1; WHSV = 8000 mL g(cat)−1 h−1; P = 2.0 MPa; T = 523 K.
b
c
The catalyst was calcined at 573 K before reduction.
+
C2 hydrocarbons.

260
J. Wang et al. / Catalysis Today 171 (2011) 257–265
Table 5
Catalytic performances of FeOx-promoted Rh/SiO2 catalysts prepared by different methods.a
Catalystsb
CO conv. (%)
Selectivity (%)
CO2
EtOH yield (%)
+
CH4
C2 HCc
MeOH
EtOH
Rh/SiO2
0.7
5.3
3.3
2.2
0
59
25
28
28
7.1
5.8
5.0
5.4
4.9
25
31
9.5
37
26
0.1
2.0
0.8
0.6
Rh/(FeOx–SiO2)
(Rh–FeOx)/SiO2
Rh–FeOx–SiO2
2.3
0.2
0
31
27
a
Reaction conditions: H2/CO = 2/1; WHSV = 8000 mL g(cat)−1 h−1; P = 2.0 MPa; T = 523 K.
b
c
The loading of Rh was 1.0 wt%; the content of Fe was 5.0 wt%.
+
C2 hydrocarbons.
of Rh/SiO2 without FeOx modification is also listed in Table 5
for comparison. In each case, the presence of FeOx could signifi-
cantly increase CO conversion and C2H5OH selectivity. However,
the catalytic performance also depended strongly on the prepa-
ration method. The catalyst prepared by the impregnation of the
FeOx–SiO2 composite, which was prepared by the sol–gel technique
preliminarily, with Rh(NO3)3 aqueous solution followed by drying,
calcination, and H2 reduction, afforded the highest CO conversion
and the best C2H5OH selectivity. The catalysts prepared by the
co-impregnation [(Rh–FeOx)/SiO2] and co-sol–gel (Rh–FeOx–SiO2)
methods exhibited significantly lower CO conversions and lower
C2H5OH selectivities.
3.2. Structural features of FeOx-promoted Rh catalysts supported
on SiO2
3.2.1. XRD
Fig. 2. XRD patterns for (0–5 wt%)Rh/(5 wt% FeOx–SiO2) catalysts with different Rh
loadings.
In the XRD patterns for the 1.0 wt% Rh/(FeOx–SiO2) catalysts
with different Fe contents (0–20 wt%) (Fig. 1), a broad peak at
2Â of ∼23◦ attributed to amorphous SiO2 was clearly observed. A
very weak peak at 2Â of ∼42◦, which was assignable to the (1 1 1)
reflection of metallic Rh, could also be discerned for some of these
samples. No diffraction lines ascribed to any FeOx-related phases
can be observed even when the content of FeOx was as high as
20 wt%, suggesting that the high dispersion of FeOx in SiO2. This is
in agreement with the general consensus that the sol–gel technique
leads to highly dispersed supported species [17]. With increasing
Rh loadings over the 5 wt% FeOx–SiO2 composite, the intensity of
the weak diffraction peak at 2Â of ∼42◦ ascribed to metallic Rh
was not significantly enhanced and the peak was still quite broad
(Fig. 2). This indicates that the size of Rh particles has not increased
significantly even in the catalysts with higher Rh loadings.
3.2.2. H2-TPR
H2-TPR studies were employed to characterize the reducibility
of Rh and Fe species and their interactions in our catalyst without
H2 reduction. The 1.0 wt% Rh/SiO2 catalyst without FeOx showed
a single peak for the reduction of Rh3+ to Rh0 at 426 K (Fig. 3).
The incorporation of FeOx by each method shifted the reduction
peak to higher temperatures, indicating the inhibited reduction
of Rh3+. The inhibition of Rh3+ reduction in the presence of FeOx
was also reported elsewhere [9,12], and such an inhibition was
believed to reflect the interaction between Rh and FeOx species
in the catalyst. Burch and Hayes [9] once argued that the inhibited
reduction of Rh3+ by the presence of FeOx could be attributed to
Fig. 3. H2-TPR profiles for 1 wt% Rh/SiO2, 5 wt% FeOx–SiO2, (1 wt% Rh–5 wt%
FeOx)/SiO2, 1 wt% Rh/(5 wt% FeOx–SiO2), and 1 wt% Rh–5 wt% FeOx–SiO2.
Fig. 1. XRD patterns for 1 wt% Rh/(0–20 wt% FeOx–SiO2) catalysts with different Fe
contents.

J. Wang et al. / Catalysis Today 171 (2011) 257–265
261
covering of Rh species by a layer of FeOx. The (Rh–FeOx)/SiO2 cata-
lyst, which was prepared by the co-impregnation method, showed
three overlapped reduction peaks in the lower-temperature region
(400–800 K) and a reduction peak in the higher-temperature region
(>800 K). Our quantitative calculations uncovered that the first peak
(439 K) could be ascribed to the reduction of Rh3+ to Rh0, whereas
the other two lower-temperature peaks at 515 K and 593 K might
be assignable to the reductions of Fe3+ into Fe2+ and Fe2+ partially
into Fe0. The higher-temperature reduction peak at 1123 K for this
sample might be assigned to the reduction of the remaining Fe2+
to Fe0 or the reduction of Fe3+ possessing strong interactions with
the SiO2 matrix [18–20].
On the other hand, the Rh/(FeOx–SiO2) catalyst exhibited a sin-
gle stronger reduction peak at 449 K in the lower-temperature
region and a reduction peak at 1113 K in the higher-temperature
region. Our quantitative analysis suggests that the peak at 449 K
should comprise not only the reduction of Rh3+ into Rh0 but also
the partial reduction of Fe3+ into Fe2+, while the reduction peak at
1113 K may correspond to the partial reduction of Fe2+ to Fe0 or
the reduction of Fe3+ possessing strong interactions with the SiO2
matrix. Fig. 3 also shows the H2-TPR profile for the FeOx–SiO2 sam-
ple prepared by the sol–gel method. The peak for the reduction of
Fe3+ into Fe2+ in this sample without Rh occurred at 690 K. Thus,
Rh species significantly accelerated the reduction of Fe species in
this sample. For the Rh–FeOx–SiO2 catalyst prepared by the co-
sol–gel method, two peaks at 478 K and 637 K were observed in
lower-temperature region. The quantitative analysis clarified that
the first peak comprised the reductions of Rh3+ into Rh0 and Fe3+
partially into Fe2+, while the second peak corresponded to the
reductions of the remaining Fe3+ into Fe2+ and also a small part
of Fe2+ into Fe0. The reduction of Rh3+ into Rh0 further shifted to
higher temperatures in this sample. The observation that the reduc-
tions of Rh3+ and Fe3+ into Rh0 and Fe2+ occurred simultaneous over
the Rh/(FeOx–SiO2) catalyst indicates the largest contact boundary
between Rh and FeOx species in this catalyst.
in this sample. The higher-temperature peak was observed for
the samples with Fe contents ≥5.0 wt%, and this peak mainly
corresponded to the partial reduction of Fe2+ into Fe0 or the
reduction of Fe3+ possessing strong interaction with SiO2 matrix.
The reductions of Rh3+ into Rh0 and Fe3+ into Fe2+ occurred mainly
at <1000 K. A single reduction peak was observed for this series
of catalysts with Fe contents ≤10 wt%. The position of this peak
shifted to higher temperatures on increasing Fe content, indicating
that the degree of inhibition of Rh3+ reduction by FeOx became
larger at higher Fe contents. For the catalyst with a Fe content of
20 wt%, the lower-temperature peak became significantly broader,
and a shoulder peak at 728 K appeared. This may imply that some
FeOx species in the catalyst are not in intimate contact with Rh
particles.
3.2.3. XPS
XPS studies have been performed to gain information about the
chemical state of Rh on catalyst surfaces after reduction at 573 K.
Typically, after reduction by H2 at 573 K, followed by cooling to
room temperature in H2 gas flow, the catalysts were transferred
to XPS chamber as soon as possible to avoid the exposure in air
for a long time. As shown in Fig. 5, the binding energy of Rh 3d5/2
over the 1.0 wt% Rh/SiO2 was 307.2 eV, which could be assigned
to that of Rh0 [21]. Thus, Rh0 was the dominant Rh species over
this catalyst without FeOx promoter. After the introduction of FeOx,
besides the peak at 307.2–307.4 eV for Rh 3d5/2, a shoulder peak at
309.3 eV, which could be attributed to Rh3+ [21], also appeared.
This suggests that some oxidized Rh species also co-exist with Rh0
on the surfaces of the reduced catalysts containing FeOx. More-
over, the relative intensity of the shoulder peak attributable to
oxidized Rh species depended on the preparation method. We
have roughly estimated the molar ratio of Rh3+/Rh0 from XPS,
and the results are summarized in Table 6. The ratio of Rh3+/Rh0
over the Rh/(FeOx–SiO2) catalyst, which exhibited the best cat-
alytic performance for C2H5OH formation, was higher than that
over the (Rh–FeOx)/SiO2 catalyst prepared by the co-impregnation
but lower than that over the Rh–FeOx–SiO2 catalyst prepared by
the co-sol–gel method.
We further performed H2-TPR studies for 1.0 wt%
Rh/(FeOx–SiO2) catalysts with different Fe contents. As shown
in Fig. 4, with increasing Fe content, the intensity of the lower-
temperature peak (400–500 K) increased gradually, further
confirming that this peak also comprised the reduction of Fe3+
species. No higher-temperature peak was observed for the sample
with a Fe content of 2.0 wt%. Our quantitative analysis clarified
that the lower-temperature peak for this sample comprised the
reductions of Rh3+ into Rh0 and Fe3+ into Fe0. We speculate that
most of the FeOx species are in intimate contact with Rh particles
We also performed XPS studies for the 1.0 wt% Rh/(FeOx–SiO2)
series of catalysts with different Fe contents. As shown in Fig. 6,
the ratio of Rh3+/Rh0 depended on Fe contents in these sam-
ples. The ratio of Rh3+/Rh0 increased with increasing Fe content
up to 10 wt%, and a further increase in Fe content did not sig-
nificantly alter the ratio of Rh3+/Rh0. XPS studies for the 1.0 wt%
Fig. 5. XPS spectra. (a) 1 wt% Rh/SiO2, (b) (1 wt% Rh–5 wt% FeOx)/SiO2, (c) 1 wt%
Fig. 4. H2-TPR profiles for 1 wt% Rh/(FeOx–SiO2) samples with different Fe contents.
Rh/(5 wt% FeOx–SiO2), (d) 1 wt% Rh–5 wt% FeOx–SiO2.

262
J. Wang et al. / Catalysis Today 171 (2011) 257–265
Table 6
The reducibility, metal dispersion and particles size of Rh/SiO2 and Fe-modified Rh-based samples prepared by different methods.
Catalysta
Rh3+/Rh0b
Rh sizec
Rh dispersiond
Chemisorbed CO/Rh
Rh/SiO2
∼0
2.8
3.4
6.5
5.0
2.6
4.5
4.8
2.2
–
0.27
0.22
0.12
0.15
0.29
0.17
0.16
0.34
–
0.28
0.14
0.070
0.048
0.26
0.07
0.02
0.40
0.29
0.060
0.03
Rh/(5 wt% FeOx–SiO2)
(Rh–5 wt% FeOx)/SiO2
Rh–5 wt% FeOx–SiO2
Rh/(2 wt% FeOx–SiO2)
Rh/(10 wt% FeOx–SiO2)
Rh/(20 wt% FeOx–SiO2)
Rh/(5 wt% FeOx–SiO2)e
Rh/(5 wt% FeOx–SiO2)f
Rh/(5 wt% FeOx–SiO2)g
Rh/(5 wt% FeOx–SiO2)h
0.75
0.25
1.1
0.57
1.1
1.1
1.1
0.72
0.23
0.11
11
–
0.068
–
a
Rh loading was 1.0 wt%.
Estimated from XPS spectra
Evaluated from TEM measurements.
b
c
d
Estimated by the following relationship, Rh dispersion = 0.75/[Rh size/nm] [22].
Reduced at 373 K (the standard reduction temperature was 573 K).
Reduced at 474 K (the standard reduction temperature was 573 K).
Reduced at 673 K (the standard reduction temperature was 573 K).
Reduced at 773 K (the standard reduction temperature was 573 K).
e
f
g
h
Rh/(5.0 wt% FeOx–SiO2) catalysts reduced at different temperatures
(373–773 K) further uncovered that a higher reduction temperature
resulted in a lower Rh3+/Rh0 ratio (Table 6).
cles. Particularly, when the reduction temperature was raised from
573 K to 673 K, the mean size of Rh particles increased remarkably
from 3.4 nm to 10.8 nm.
By assuming the spherical metal particles, Rh dispersion could
be estimated by the following relationship, dispersion = 0.75/d (nm)
[22]. The value of Rh dispersion thus estimated for each catalyst
is listed in Table 6. We have also measured the amount of CO
chemisorption, and the molar ratios of chemisorbed CO/Ru for
different catalysts are summarized in Table 6. The chemisorbed
CO/Rh ratio was almost the same with the Rh dispersion evaluated
from the mean size of Rh particles for the Rh/SiO2 catalyst. How-
ever, for most of the FeOx-promoted catalysts listed in Table 6, the
chemisorbed CO/Rh ratios were lower than the Rh dispersions esti-
mated from the mean Rh particle sizes. We speculate that this may
be because a fraction of Rh species are in oxidized state (Rh3+), and
CO may not be chemisorbed on these Rh3+ species. By taking into
account the Rh3+/Rh0 value evaluated from XPS studies (Table 6),
we calculated the corrected Rh dispersion, which only considers
the fraction of surface Rh0 in the total Rh atoms. The corrected
Rh dispersion was similar to the chemisorbed CO/Rh ratio for the
Rh/(5 wt% FeOx–SiO2) catalyst. However, the chemisorbed CO/Rh
ratios were still lower than the corrected Rh dispersions for the
(Rh–5 wt% FeOx)/SiO2 and the Rh–5 wt% FeOx–SiO2 catalysts. It is
likely that such lower chemisorbed CO/Rh ratios may arise from the
partial covering of the Rh species by FeOx or SiO2 in these samples.
3.2.4. TEM and CO chemisorption
The size and dispersion of Rh particles are also expected to influ-
ence the catalytic performances. Thus, we have carried out TEM
and CO chemisorption measurements for some of our catalysts.
Fig. 7 shows the TEM micrographs and the corresponding Rh par-
ticle size distributions for the 5.0 wt% FeOx-promoted 1.0 wt% Rh
catalysts prepared by different methods. Without FeOx, the mean
size of Rh particles (d) over the 1.0 wt% Rh/SiO2 sample was 2.8 nm.
The mean size of Rh particles in the 1.0 wt% Rh/(FeOx–SiO2) cata-
lyst became slightly larger (3.4 nm), whereas the co-impregnation
method resulted in the biggest Rh particles (mean size, 6.5 nm).
TEM studies for the 1.0 wt% Rh/(FeOx–SiO2) catalysts with different
Fe contents showed that the mean size of Rh particles increased
with Fe content (Table 6). This observation further confirms that
the presence of FeOx slightly increases the size of Rh particles.
Fig. 8 shows the TEM images and the corresponding Rh particle
size distributions for the 1 wt% Rh/(5 wt% FeOx–SiO2) catalysts after
H2 reduction at three different temperatures. The increase in H2
reduction temperature caused the rise in the mean size of Rh parti-
3.3. Discussion on roles of FeOx promoters
Our catalytic studies have clarified that, among various tran-
sition metal oxides examined (including VOx, CrOx, MnOx, FeOx,
CoOx, ZrOx, and MoOx), FeOx is the most efficient promoter for
SiO2-supported Rh catalysts for ethanol formation (Table 1). The
presence of FeOx significantly increased CO conversion and ethanol
selectivity, whereas CH4 selectivity was markedly decreased at the
same time. There exists an appropriate Fe content for obtaining the
maximum ethanol selectivity. A further increase in Fe contents to
≥10 wt% increased methanol selectivity at the sacrifice of ethanol
selectivity (Table 2). Furthermore, we have clearly demonstrated
that the catalytic performance is strongly dependent on the method
of preparing the FeOx-promoted catalysts. The Rh/(FeOx–SiO2) cat-
alyst prepared by the impregnation of the FeOx–SiO2 composite,
which was synthesized preliminarily by a sol–gel technique, with
Rh(NO3)3 aqueous solution, followed by drying, calcination, and
H2 reduction afforded significantly better catalytic performances
for ethanol formation than the (Rh–FeOx)/SiO2 and Rh–FeOx–SiO2
Fig. 6. XPS spectra for 1 wt% Rh/(FeOx–SiO2) samples with different Fe contents. Fe
contents: (a) 0, (b) 2 wt%, (c) 5 wt%, (d) 10 wt%, (e) 20 wt%.

J. Wang et al. / Catalysis Today 171 (2011) 257–265
263
Fig. 7. TEM micrographs and Rh particle size distributions. (a) 1 wt% Rh/SiO2, (b) 1 wt% Rh/(5 wt% FeOx–SiO2), (c) (1 wt% Rh–5 wt% FeOx)/SiO2, (d)1 wt% Rh–5 wt% FeOx–SiO2.

264
J. Wang et al. / Catalysis Today 171 (2011) 257–265
Fig. 8. TEM micrographs and Rh particle size distributions for 1 wt% Rh/(5 wt% FeOx–SiO2) samples reduced at different temperatures. Reduction temperatures: (a) 373 K,
(b) 573 K, (c) 673 K.
catalysts prepared by co-impregnation and co-sol–gel methods,
respectively. The temperature for catalyst reduction has also been
found to exert a significant effect on catalytic behaviors.
the decrease in ethanol selectivity. For example, we observed that
the increase in Fe content in the 1.0 wt% Rh/(FeOx–SiO2) cata-
lysts to ≥10 wt% resulted in the increase in the ratio of Rh3+/Rh0
and the decrease in ethanol selectivity. The hydrogenation of non-
dissociative CO may give high methanol selectivity over these
catalysts (Table 2).
Our H2-TPR studies also indicated larger interfaces between
Rh and FeOx species in the Rh/(FeOx–SiO2) catalyst than in the
(Rh–FeOx)/SiO2 and Rh–FeOx–SiO2 catalysts because the former
catalyst showed a single H2-TPR peak for the simultaneous reduc-
tions of Rh3+ and Fe3+. Such larger interfaces in the Rh/(FeOx–SiO2)
catalyst may account for its better catalytic performances (Table 5).
Moreover, the catalysts prepared by the co-impregnation and
the co-sol–gel methods showed quite lower CO chemisorption
amounts (Table 6). Our analysis suggests that this is not only
because of the relatively bigger Rh sizes in these samples (Fig. 7)
H2-TPR studies have clarified that the presence of FeOx inhibits
the reduction of Rh3+ into Rh0 due to the interaction between Rh
and Fe species, and the remaining of a part of Rh3+ on catalyst sur-
faces after H2 reduction has been confirmed by XPS. However, only
Rh0 species was observed over the 1.0 wt% Rh/SiO2 catalyst with-
out FeOx promoter. It is established that the selective formation
of ethanol requires the balance of CO dissociation, CO insertion,
and hydrogenation abilities [3,4,23]. The metallic and oxidized Rh
species have been proposed to favor CO dissociation and CO inser-
tion, respectively [23]. We speculate that the co-existence of Rh3+
with Rh0 may increase the CO insertion ability, and thus increase
ethanol formation activity. However, a too high ratio of Rh3+/Rh0
may decrease the CO dissociation ability. This may also lead to

J. Wang et al. / Catalysis Today 171 (2011) 257–265
265
but also probably due to the covering of Rh species by FeOx or SiO2.
These might result in the lower CO conversion activity for these
two catalysts (Table 5).
Higher reduction temperatures (≥673 K) caused the aggregation of
Rh particles, resulting in lower CO conversions and ethanol selec-
tivities. The interfaces between Rh and FeOx species, the ratio of
Rh3+/Rh0, and the Rh particle size are proposed to be important
factors for the selective conversion of syngas to ethanol.
Higher reduction temperatures (≥673 K) for our 1.0 wt%
Rh/(5.0 wt% FeOx–SiO2) catalyst caused the formation of larger Rh
particles with lower fractions of Rh3+. These catalysts provided
lower CO conversions and lower ethanol selectivities (Table 4). The
bigger Rh particles may not only decrease the hydrogenation abil-
ity but also decrease the interface between Rh and FeOx species,
leading to a lower CO insertion ability. Table 4 shows that these
catalysts exhibit relatively higher selectivities to C2+ hydrocarbons.
This observation is in agreement with those reported for supported
Co and Ru catalysts, where larger metal nanoparticles also favor
the chain-growth probability [1,2]. The formation of ethanol is a
complex interplay among CO dissociation, CO insertion, and hydro-
genation. Our present studies have demonstrated that the interface
between Rh and FeOx species, the Rh3+/Rh0 ratio, and the size of Rh
particles are all important factors determining the catalytic behav-
iors in the conversion of syngas to ethanol.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (nos. 20873110, 20923004, and 21033006),
the National Basic Program of China (no. 2010CB732303), the
Research Fund for the Doctoral Program of Higher Education (no.
20090121110007), and the Key Scientific Project of Fujian Province
(no. 2009HZ0002-1).
References
[1] A.Y. Khodakov, W. Chu, P. Fongarland, Chem. Rev. 107 (2007) 1692–1744.
[2] Q. Zhang, J. Kang, Y. Wang, ChemCatChem 2 (2010) 1030–1058.
[3] J.J. Spivey, A. Egbebi, Chem. Soc. Rev. 36 (2007) 1514–1528.
[4] V. Subramani, S.K. Gangwal, Energy Fuels 22 (2008) 814–839.
[5] X. San, Y. Zhang, W. Shen, N. Tsubaki, Energy Fuels 23 (2009) 2843–2844.
[6] X. Li, X. San, Y. Zhang, T. Ichii, M. Meng, Y. Tan, N. Tsubaki, ChemSusChem 3
(2010) 1192–1199.
4. Conclusions
[7] M.M. Bhasin, W.J. Bartley, P.C. Ellgen, T.P. Wilson, J. Catal. 54 (1978) 120–128.
[8] R. Burch, M.I. Petch, Appl. Catal. A Gen. 88 (1992) 39–60.
[9] R. Burch, M.J. Hayes, J. Catal. 165 (1997) 249–261.
[10] M.A. Haider, M.R. Gogate, R.J. Davis, J. Catal. 261 (2009) 9–16.
[11] J. Gao, X. Mo, A.C. Chien, W. Torres, J.G. Goodwin Jr., J. Catal. 262 (2009)
119–126.
[12] X. Mo, J. Gao, N. Umnajkaseam, J.G. Goodwin Jr., J. Catal. 267 (2009) 167–176.
[13] D. Mei, R. Rousseau, S.M. Kathmann, V.A. Glezakou, M.H. Engelhard, W. Jiang,
C. Wang, M.A. Gerber, J.F. White, D.J. Stevens, J. Catal. 271 (2010) 325–342.
[14] N.D. Subramanian, J. Gao, X. Mo, J.G. Goodwin Jr., W. Torres, J.J. Spivey, J. Catal.
272 (2010) 204–209.
[15] M. Ichikawa, T. Fukushima, J. Phys. Chem. 90 (1986) 1222–1224.
[16] M.R. Gogate, R.J. Davis, ChemCatChem 1 (2009) 295–303.
[17] M.V. Landau, in: G. Ertl, H. Knözinger, F. Schüth, J. Weitkamp (Eds.), Handbook
of Heterogeneous Catalysis, vol. 1, second ed., Wiley-VCH, Weinheim, 2008, pp.
119–160.
[18] C. Alberto, G. Fajardo, D. Niznansky, Y.N. Guyen, C. Courson, A.C. Roger, Catal.
Commun. 9 (2008) 864–869.
[19] T. Kobayashi, N. Guilhaume, J. Miki, N. Kitamura, M. Haruta, Catal. Today 32
(1996) 171–175.
[20] K. Tabata, Y. Takemoto, T. Suzuki, E. Banares, M.A. Pena, J.L.G. Fierro, Catal. Rev.
44 (2002) 1–58.
[21] M. Ojeda, M.L. Granados, S. Rojas, P. Terreros, F.J. García-García, J.L.G. Fierro,
Appl. Catal. A Gen. 261 (2004) 47–55.
Rh catalysts supported on transition metal oxide-modified silica
were studied for ethanol synthesis from syngas. FeOx was found to
be an efficient promoter for ethanol formation. The presence of FeOx
increased CO conversion and ethanol selectivity, and decreased
methane selectivity. An appropriate Fe content was required for
obtaining the best catalytic performance. Our characterizations
revealed that the presence of FeOx inhibited the reduction of Rh3+
,
and a part of Rh3+ remained on catalyst surfaces after reduction.
The co-existence of metallic and oxidized Rh species may enhance
the CO insertion step and increase ethanol formation activity. Our
present work clearly demonstrated that the preparation method
for introducing FeOx played a significant role in determining the
catalytic performance. The catalyst prepared by the impregnation
of the FeOx–SiO2 composite, which was synthesized by a sol–gel
technique, with Rh(NO3)3 aqueous solution followed by drying,
calcination, and H2 reduction afforded better ethanol formation
activity than those prepared by co-impregnation and co-sol–gel
methods. Our characterizations indicated that this catalyst con-
tained larger interfaces between Rh and FeOx species. The reduction
temperature was also found to affect the catalytic performance.
[22] C.K. Lambert, R.D. Gonzalez, Micropor. Mater. 12 (1997) 179–188.
[23] S.S.C. Chuang, R.W. Stevens Jr., R. Khatri, Top. Catal. 32 (2005) 225–232.