High activity of au/γ-Fe2o3 for co oxidation: Effect of support crystal phase in catalyst design

Article abstract of DOI:10.1021/cs5020496

Au/γ-Fe₂O₃ and Au/α-Fe₂O₃ catalysts with identical size of Au nanoparticles, chemical state of Au species, and amount of surface OH^- group were prepared. The Au/γ-Fe₂O₃ catalyst exhibited exceptionally high activity, regardless of the heat treatments. The CO-TPR, sequential pulse reaction, and in situ Raman spectra demonstrate that the much higher activity of Au/γ-Fe₂O₃ originated from its higher redox property at low temperature. Systematic study shows that this higher-redox-property-based higher activity could be extended to γ-Fe₂O₃-supported Pt-group metals and to other reactions that follow Mars-Van Krevelen mechanism. This finding may provide a new avenue for catalyst improvement or development by choosing the suitable crystal phase of the oxide support.

Full text of DOI:10.1021/cs5020496

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
High Activity of Au/#-Fe2O3 for CO Oxidation:
Effect of Support Crystal Phase in Catalyst Design
Kunfeng Zhao, Hailian Tang, Botao Qiao, Lin Li, and Junhu Wang
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/cs5020496 • Publication Date (Web): 30 Apr 2015
Downloaded from http://pubs.acs.org on May 2, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 12
ACS Catalysis
1
2
3
4
5
6
7
8
9
High ActivityofAu/ γꢀFe2O3 forCO Oxidation: Effect of Support Crysꢀ
tal Phase in Catalyst Design
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
§, ,
§
§
, †,
#
, †
Kunfeng Zhao,†, ψ Hailian Tang,†, , ‡, # Botao Qiao,* Lin Li† and Junhu Wang*
† State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy Sciences, Dalian 116023, China
§ Mössbauer Effect Data Center & Laboratory of Catalysts and New Materials for Aerospace, Dalian Institute of Chemical
Physics, Chinese Academy of Sciences, Dalian 116023, China
‡ University of Chinese Academy of Sciences, Beijing 100049, China
other noble metal catalysts. For example, by noticing the
ABSTRACT: Au/ γꢀFe2O3 and Au/ αꢀFe2O3 catalysts with identiꢀ
important roles of ferrihydrite and ferrihydriteꢀderivated FeOx
cal size of Au nanoparticles, chemical state of Au species and
supports in the high activity of supported Au catalysts for CO
amount of surface OHꢀ group were prepared. Au/ γꢀFe2O3 catalyst
oxidation or CO preferential oxidation in H2 (PROX),13ꢀ17 a series
exhibited exceptionally high activity, regardless of the heat treatꢀ
of ferrihydrite and FeOx supported metal catalysts (Pt, Pd, Ir, Cu)
ments. The COꢀTPR, sequential pulse reaction and inꢀsitu Raman
with similar or even better performance have already been
spectra demonstrate that the much higher activity of Au/ γꢀFe2O3
developed,18ꢀ24 although the exact catalytic mechanism was not
originated from its higher redox property at low temperature.
known then. A recent program entitled “After the Goldrush”
Systematic study shows that this higher redox property resultant
which aims to design and develop catalytic materials by exploiting
higher activity could be extended to γꢀFe2O3 supported Ptꢀgroup
the learning gained from gold catalysis suggests clearly that the
reference role of study on gold catalysts has been realized.25
metals and to other reactions that follow MarsꢀVan Krevelen
mechanism. This finding may provide a new avenue for catalyst
improvement or development by choosing the suitable crystal
phase of the oxide support.
The CO oxidation on supported gold NPs/ clusters therefore has
been a paradigm in the last two decades.26ꢀ27 Besides its great
importance in practical applications19,
and the fact that
28ꢀ30
Key words: gold, CO oxidation, redox, γꢀFe2O3, catalyst design
supported Au catalysts are most active for this reaction, a more
important reason is that CO oxidation is an ideal model reaction
to fundamentally investigate the heterogeneous catalytic
mechanism and study the nature of the catalyst.27 The reaction
and product molecules are simple and no side reaction occurs.
However, the most critical issue for oxidation reaction on
supported Au catalysts, how to activate the O2, was involved.
Therefore, to well address the CO oxidation mechanism on
supported Au catalyst may provide guidance/ reference to other
oxidation reaction on supported Au, and even other metal
catalysts.
INTRODUCTION
Heterogeneous catalysis is critical for many important industrial
processes and environmental issues that closely related to our
daily life.1ꢀ2 The demand to develop new catalytic materials may be
endless due to our continuous pursuit of better catalytic
performance (e.g. higher activity/ selectivity and better stability,
cheaper and abundance elements etc.) or of new and green
catalytic processes.3 Finding the right catalytic materials, however,
has many elements of trial and error, thus is usually a tedious,
timeꢀconsuming process.4ꢀ5 Therefore, the catalyst development
by rational design has been a longꢀterm dream and been regarded
as one of the greatest challenges.3, 5ꢀ6
In the past decades, oxides supported gold nanoparticles (NPs) or
clusters as catalysts have attracted a great deal of and yet
increasing interest in catalysis field. On the one hand, supported
Au catalysts have demonstrated unique catalytic performance for
numerous important chemical reactions.7ꢀ12 On the other hand,
the performances of supported Au catalysts were dramatically
influenced by many factors such as the size of gold NPs, the
valence state of the gold and the nature of the support oxides as
well as the interaction between gold and them. The latter is not
only fundamentally interesting but also practically important
because an inꢀdepth understanding of why and how these factors
influence the catalyst could be helpful to modify the catalyst
formula thus to promote the catalyst performance. Of more
importance, the learning gained from the studies of supported Au
catalysts may provide reference to the design and development of
Recently, increasing evidences show that CO oxidation followed,
at least partially, the MarsꢀVan Krevelen (redox) mechanism on
reducible oxides supported Au catalysts.31ꢀ36 Especially, Li et al
demonstrated that the CO oxidation on the Au/ ferrihydrite
catalyst was mainly according to a redox mechanism, even at
temperatures as low as –60 C.37 It suggested that the very high
o
activity of Au/ ferrihydrite catalyst actually originated from the
high redox property of the ferrihydrite support. Furthermore, this
mechanism works for ferrihydrite supported other noble metals
such as Pd and Pt. This finding may provide a new avenue to
fabricate highly active supported Au catalysts by using iron
oxides/ hydroxides with higher redox property. Maghemite (γꢀ
Fe2O3) is one of the most common iron oxide polymorphs and has
the reverse spinel structure similar to magnetite (Fe3O4).38
Therefore, the reduction of γꢀFe2O3 is just related to the losing of
O but does not involve any lattice rearrangement. An existed fact
is that despite the similar reducibility of γꢀFe2O3 and hematite (αꢀ
Fe2O3) support, the Au/ γꢀFe2O3 catalyst was much more
reducible than the Au/ αꢀFe2O3 catalyst,39 which implied a
potential higher activity of Au/ γꢀFe2O3 for CO oxidation
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 12
Table 1. Physicochemical properties and catalytic activities of Fe2O3 and Au/ Fe2O3 catalysts
1
2
3
4
5
6
7
8
9
a
Surface area
(m2 gꢀ1)
Au loading DAu
(wt%)
B.E. of Au4f7/ 2
(eV)
Aux+/ Auo
Specific rateb
(molCO hꢀ1 gAu
)
TOF
(sꢀ1)
Sample
ꢀ1
(nm)
αꢀFe2O3
49
46
44
92
84
78
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
This work
Au/ αꢀFe2O3
Au/ αꢀFe2O3ꢀC
γꢀFe2O3
2.9
2.9
ꢀꢀ
U.D.c
2.8
85.0, 86.3
50/ 50
0/ 100
ꢀꢀ
2.12
0.40
ꢀꢀ
0.13d
0.068e
ꢀꢀ
This work
This work
This work
This work
This work
Ref 40
84.4
ꢀꢀ
ꢀꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Au/ γꢀFe2O3
Au/ γꢀFe2O3ꢀC
2.9
2.9
4.4
3.7
1.0
2.9g
2.9g
1.0
U.D.
1.9
84.3, 85.6
53/ 47
0/ 100
ꢀꢀ
33.3
9.85
0.185
3.78
3.6ꢀ7.3f
2.02d
1.14e
0.045
0.23
0.5ꢀ1
3.5h
83.8
ꢀꢀ
Au/ Fe2O3ꢀWGC ꢀꢀ
3ꢀ4
U.D.c
2ꢀ3
~0.5
ꢀꢀ
Au/ FeOx
Au/ FeOx
Au/ FeOx
Au/ FeOx
Au/ Fe2O3
ꢀꢀ
Ref 40
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
Ref 47
ꢀꢀ
Ref 56
ꢀꢀ
0.027i
Ref 56
109
7.4
0.94
ꢀꢀ
Ref 60
a Measured by TEM; b measured at 24 oC; cundetected; d using a dispersion of 90%; e the gold dispersion was calculated according to D =
0.9/ dAu, where d means diameter; f calculated from the TOF data by assuming a dispersion of 40%; g atomic percent; h TOF of bilayer Au
cluster; i TOF value of the whole catalyst, i.e. the average of the gold atoms, clusters and NPs
according to the redoxmechanism.37
sample.
In this work, we used commercial γꢀFe2O3 as support to develop
an exceptionally active Au/ γꢀFe2O3 catalyst by a facile depositionꢀ
precipitation method. By using commercial αꢀFe2O3 for
comparison and through a systemic study, we demonstrated that
γꢀFe2O3 is a highly active support for a number of metals (Au, Pt,
Rh etc.) for CO oxidation and other reaction that follows redox
mechanism. More importantly, it may provide a feasible avenue to
design and develop supported metal catalysts by choosing suitable
crystal phase of the oxide supports.
The surface composition and the binding energy (B.E.) of the
catalysts were determined by Xꢀray photoelectron spectra (XPS)
on an ESCALAB250 Xꢀray photoelectron spectrometer with
contaminated C as the internal standard (C1s = 284.5 eV).
Xꢀray diffraction (XRD) patterns were recorded on a PW3040/ 60
X’ Pert PRO (PANalytical) diffractometer equipped with a Cu Kα
radiation source (λ = 0.15432 nm), operating at 40 kV and 40 mA.
A continuous mode was used for collecting data in the 2θ range
from 20 ◦ to 80 ◦ at a scanning speed of 5o min−1.
Room temperature 57Fe Mössbauer spectra were recorded by using
a Topologic 500A spectrometer and a proportional counter.
57Co(Rh) moving with a constant acceleration mode was used as
the γꢀray radioactive source. The velocity was calibrated by a
standard αꢀiron foil. The spectra were fitted with the appropriate
superpositions of Lorentzian lines using the MossWinn 3.0i
computer program. In this way, the 57Fe Mössbauer spectral
parameters could be determined, including the isomer shift (IS),
the electric quadrupole splitting (QS), effective magnetic field
(H), the full width at half maximum (FWHM), and the relative
resonance areas of the different components of the absorption
patterns.
Transmission electron microscope (TEM) analyses were
performed with a JEOL JEMꢀ2000EM microscope operated at 120
kV. The samples were suspended in ethanol with an ultrasonic
dispersion for 5 ꢀ 10 min. And then a few droplets of the
suspension were put on a microgrid carbon polymer supported on
a copper grid and allowed to dry at room temperature for TEM
observations.
Temperatureꢀprogrammed reduction with CO (COꢀTPR) was
performed in a quartz microreactor. The samples were purged
under He flow at ꢀ75 oC or 25 oC for 15 min. And then the gas flow
was switched to a flow of 5% CO/ He (30 mL minꢀ1), and the
sample was heated to 900 oC at a rate of 10 oC minꢀ1.
EXPERIMENTALSECTION
Catalyst preparation. The commercial nanoꢀsphere iron oxides (αꢀ
Fe2O3 and γꢀFe2O3) were purchased from Aladdin Reagent
Company and employed directly as supports without any
pretreatment.
Gold NPs were deposited onto αꢀFe2O3 and γꢀFe2O3 oxides via a
depositionꢀprecipitation process targeting at a gold loading of 5.0
wt%, which was denoted as Au/ αꢀFe2O3 and Au/ γꢀFe2O3,
respectively. Detailed procedure see our previous reports.40ꢀ41 For
investigating the effect of heatꢀtreatment, the two catalysts were
o
further calcinated at 300 C under N2 atmosphere for 3 hours,
denoted as Au/ αꢀFe2O3ꢀC and Au/ γꢀFe2O3ꢀC, respectively.
αꢀFe2O3 and γꢀFe2O3 supported other noble metal (Pt and Rh)
catalysts were also prepared with the same procedure and denoted
as M/ αꢀFe2O3ꢀC and M/ γꢀFe2O3ꢀC (M = Pt and Rh), respectively.
Sample characterization. The actual metal loadings (Au, Pt, Rh)
of various catalysts were determined by inductively coupled
plasma atomic emission spectrometer (ICPꢀAES) on an IRIS
Intrepid II XSP instrument (Thermo Electron Corporation).
BrunauerꢀEmmettꢀTeller (BET) surface areas of samples were
measured by nitrogen adsorption at ꢀ196 oC using a Micromeritics
ASAP 2010 apparatus. Prior to the measurements, the samples
were outgassed at 110 ◦C to 0.03 Torr for the calcined samples or
outgassed at ambient temperature for the uncalcined samples for
12 hours to ensure no or little change of the structure of the
In situ diffuse reflectance infrared Fourier transform (DRIFT)
spectra were acquired with a BRUKER Equinox 55 spectrometer
ACS Paragon Plus Environment

Page 3 of 12
ACS Catalysis
equipped with a MCT detector and operated at a resolution of 4
the corresponding fitting analysis (Table S1), we can confirm the
cmꢀ1. All spectra were obtained at room temperature. Before the
experiment, the sample (40 mg) in a powder form was purged for
asꢀreceived γꢀFe2O3 phase and claim that there is no phase
transformation for the Au/ γꢀFe2O3ꢀC catalyst after being calcined
1
2
o
o
30 min with He at 60 C. After cooled to room temperature, the
at 300 C under N2 atmosphere. In addition, no diffraction peaks
3
4
5
6
7
8
9
background spectrum was recorded and then the 1 vol% CO + 1
vol%O2 balanced with He were introduced into the reaction cell at
a total flow rate of 66.7 mL minꢀ1 (space velocity, SV = 100,000
mL hꢀ1 gcatꢀ1).
In situ Raman spectra were collected using an in situ Raman
reactor (Linkam CCR1000) on a LabRam HR800 confocal
microprobe Raman instrument (HORIBA Jobin Yvon, France)
with laser excitation at 633 nm (HeꢀNe laser) at a laser power of ca.
0.1 mW. The spectra were obtained by exposure to air then
switching to 10.0%CO/ He at room temperature.
of Au or Au oxides were observed on all αꢀFe2O3 and γꢀFe2O3
supported Au catalysts, suggesting that the Au was highly
dispersed on the iron oxides, even after being calcined at 300 oC.
(a)
γ
ꢀFe
2
O
3
•
•
Au(111)
•
•
•
•
Au/
γ
ꢀFe
ꢀFe
ꢀFe
2
O
3
ꢀC
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Au/
γ
2O
3
3
γ
2O
The pulse reactions of CO oxidation on various catalysts were
measured with a HTꢀ1000 calorimeter (Setaram, France) which is
connected to a pulse reactor system, details of this apparatus see
previous literatures.37, 42 Typically, prior to measurements, 20 mg
of sample was purged in flowing He gas at room temperature for
(b)
αꢀFe2O3
♦
♦
♦
♦
♦
♦
♦
♦♦
♦
Au
/
α
ꢀ
Fe
Fe
Fe
2
O
3
ꢀC
o
Au
/
α
ꢀ
2
O
3
30 min, cooled to ꢀ20 C, and then pulses of the CO and O2 (0.5
mL) were sequentially admitted to the sample at an interval of 30
min at ꢀ20 oC and other temperatures. The effluents were onꢀline
analyzed by a gas chromatograph (GC, Agilent 6890N). Coꢀpulse
of CO + O2 was performed with the same procedure except in each
pulse the CO or O2 was replaced with 0.5 mL CO + 0.5 mL O2.
Catalytic performance test. The catalytic activity for CO
oxidation was measured using a continuousꢀflow fixedꢀbed reactor
system. For each test, 50 mg of the catalyst sample powder diluted
with about 350 mg SiO2 was loaded into a Uꢀshape quartz reactor.
After purged with He for 30 min at room temperature, a feed
stream containing 1.0 vol% CO and 1.0 vol% O2 balanced with He
was allowed to pass through the catalyst sample at a flow rate of
α
ꢀ
2
O
3
20
30
40
50
60
70
80
2ꢀtheta (o)
Figure 1. XRD patterns of (a) γꢀFe2O3, Au/ γꢀFe2O3 and Au/ γꢀ
Fe2O3ꢀC and (b) αꢀFe2O3, Au/ αꢀFe2O3 and Au/ αꢀFe2O3ꢀC.
γ
ꢀFe
2O
3
ꢀ
66.7 mL minꢀ1, resulting in a space velocity (SV) of 80,000 mL gcat
1
hꢀ1. The inlet and outlet gas compositions were analyzed by an
onꢀline gas chromatograph with TCD detector using helium as
carrier gas (HP 6890, TDXꢀ01 column).
The activity for water gas shift (WGS) reaction was measured in
the same procedure except that the reaction gas composition was 2
vol%CO + 10 vol%H2O balanced with He.
The specific rates were measured under a differential model where
the CO conversions were controlled below 25%. The conversions
of CO during the first 60 min were averaged and used for calculaꢀ
tions of the specific reaction rates.
Au/
γ
ꢀFe
2
O ꢀC
3
RESULTS
Catalyst structure and surface property. Some physicochemical
properties of the iron oxides and Au supported catalysts are
presented in Table 1. It shows that the surface areas of both iron
oxides are small (< 100 m2 gꢀ1), although the surface area of γꢀ
Fe2O3 is higher than that of αꢀFe2O3. After loading of Au, the
surface areas of both oxides decreased slightly, suggesting that
some of the pores or the adsorption sites were occupied by the
gold species. The actual Au loadings on both iron oxides are 2.9
wt%, much lower than the nominal value. This might be due to the
loss of Au during the deposited process.40, 43ꢀ44
Figure 1 shows the XRD patterns of the iron oxides and the iron
oxides supported Au catalysts. Both oxides presented their typical
diffraction patterns, suggesting the pure phase of the iron oxides.
After loading of Au, neither the crystal phase nor the nanoparticle
size of αꢀFe2O3 and γꢀFe2O3 oxides changes. Since γꢀFe2O3 has the
same XRD patterns with that of Fe3O4, and γꢀFe2O3 tends to
transform to αꢀFe2O3 under calcination at elevated temperature,
we further performed 57Fe Mössbauer measurements to confirm
the γꢀFe2O3 phase before and after calcination. From Figure 2 and
Au/
αꢀFe
2O
3ꢀC
-12 -10 -8 -6 -4 -2
0
2
4
6
8
10 12
Velocity (mm sꢀ1)
Figure 2. Room temperature 57Fe Mössbauer spectra of asꢀreceived
γꢀFe2O3, Au/ γꢀFe2O3ꢀC and Au/ αꢀFe2O3ꢀC.
The size of the Au NPs has been regarded as one of the most
important factors to determine their catalytic activity.7, 45ꢀ46 To
measure the size of the gold NPs, we examined the catalysts with
TEM. The typical images and the size distribution of Au NPs are
presented in Figure 3. It shows that no Au NPs could be observed
on the Au/ αꢀFe2O3 and Au/ γꢀFe2O3 catalysts, Figures 3a, b,
suggesting that the Au species were highly dispersed on both iron
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 12
oxides as clusters smaller than 1 nm that cannot be detected by the
indicated that with or without calcination the chemical states of
TEM.14, 17, 40 After calcination at 300 oC, visible Au NPs with size of
1 ꢀ 4 nm are clearly observed on both catalysts, Figures 3c, d,
suggesting that the Au clusters aggregated during calcination. The
size distributions, Figures 3e, f, show that the Au NPs on Au/ γꢀ
Fe2O3ꢀC were slightly smaller than that on Au/ αꢀFe2O3ꢀC (mean
diameter of 1.9 nm vs 2.8 nm), indicating that the degree of
aggregation was slightly different. This suggests that the
interaction between Au and γꢀFe2O3 might be stronger than that
between Au and αꢀFe2O3. However, despite the slight difference in
Au NP size, these values are both in the optimize size of < 5 nm for
a highly active Au catalyst.7, 47 Therefore, the size effect on the two
type oxides supported catalysts, regardless of calciation or not,
could be to a large extent excluded.
Au are similar on both iron oxides. The chemical state effect of Au
species thus could be ruled out as well. It should be noted that the
B.E. of Au4f for αꢀFe2O3 supported catalysts is slightly higher than
that of γꢀFe2O3 supported ones. This should be due to the support
charging effect which arouse the shifting of the peak position and
the broadening of the peak width.49 It was reported that the B.E.
shift of Au is size and supportꢀdependent: An increased (positive)
B.E. shift correlated with the decreasing of the Au cluster size and
the support conductivity.50 Although αꢀFe2O3 and γꢀFe2O3 are
both semiconductors, the band gap of αꢀFe2O3 is higher than that
of γꢀFe2O3, suggesting its conductivity is lower.51 Therefore the
B.E. of Au supported on αꢀFe2O3 should be higher than that on γꢀ
Fe2O3. Furthermore, the Au4f peak for Au/ αꢀFe2O3ꢀC is obviously
wider than that for Au/ γꢀFe2O3ꢀC, also indicating a higher chargꢀ
ing effect on Au/ αꢀFe2O3ꢀC than on Au/ γꢀFe2O3ꢀC sample. It also
shows that the uncalcined sample have higher Au4f B.E. than their
calcined counterparts. This should be due to the size effect50, 52
since the size of the Au clusters on uncalcined samples are much
smaller than that on the calcined ones.
It has been reported that surface OHꢀ group had great influence on
the activity of CO oxidation on supported Au catalyst.53ꢀ55 The
XPS spectra of O1s for the four samples were also measured and
analyzed in detail. As shown in Figure S2, the O1s XPS spectra
exhibited a broadened peak including a main peak centered at
~530 eV and a shoulder peak centered at ~532 eV. After fitting
analysis, as shown in Table 2, the O species were composed of
lattice O (~530 eV), OHꢀ group (~531.5 eV)56ꢀ58 and adsorbed
H2O (~533.0 eV).59 The ratios of these three species in all
catalysts are 72ꢀ75% for lattice O, 20ꢀ24% for OHꢀꢀgroup O and
~5% for adsorption H2O. Clearly, the amount of OHꢀ group
species in all the catalysts were similar. Therefore, the effect of
OHꢀꢀgroup could also be ruled out.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Catalyst activity. The samples were tested for CO oxidation with a
gas composition of 1 vol% CO + 1 vol% O2 + He, Figure 4. It
shows that both αꢀFe2O3 and γꢀFe2O3 are inactive for CO
o
oxidation at temperature below 150 C. At temperature range of
70
150 ꢀ 300 oC αꢀFe2O3 exhibits higher activity than that of γꢀFe2O3.
After Au was loaded, the catalytic activities increase dramatically,
suggesting that the Au or the entities of Au/ iron oxides are much
more active than iron oxides themselves. However, the degrees of
activity increase are significantly different. In contrary to the
activity order of the bare iron oxides, γꢀFe2O3 supported Au
catalysts exhibit much higher activity than that of αꢀFe2O3
supported Au catalysts: Au/ γꢀFe2O3 and Au/ γꢀFe2O3ꢀC can totally
oxidize CO at temperature lower than or around ꢀ20 oC; the Au/ αꢀ
Fe2O3 and Au/ αꢀFe2O3ꢀC catalysts, however, can only completely
oxidize CO at 10oC and 60 oC, respectively.
(e)
(f)
60
50
40
30
20
10
0
1
2
3
4
5
1
2
3
4
5
Particle size (nm)
Figure 3. Typical TEM images of Au/ γꢀFe2O3 (a), Au/ αꢀFe2O3
(b), Au/ γꢀFe2O3ꢀC (c) and Au/ αꢀFe2O3ꢀC (d) and size
distribution of Au NPs on Au/ γꢀFe2O3ꢀC (e) and Au/ αꢀFe2O3ꢀC
(f).
To obtain the intrinsic activity of these catalysts for a more
accurate comparison and to compare with the catalysts reported in
literatures,60 we measured their specific reaction rates for CO
o
oxidation at 24 C and calculated the corresponding TOF values,
The chemical valence of the Au species is another important factor
that affect their CO oxidation activity dramatically.48 The
measurements of Au4f XPS spectra were therefore performed. As
shown in Figure S1 and Table 1, the Au species on Au/ γꢀFe2O3
and Au/ αꢀFe2O3 catalysts are mixture of Au0 and Aux+ without
calcination treatment, in line with the previous iron oxides
supported Au catalysts.17, 40 A quantitative analysis shows that the
ratios of Aux+/ Au0 were about 1 for both samples, suggesting that
without calcination treatment the chemical states of Au species on
both samples are similar. After being calcined at 300 oC, as
expected, the Au oxides/ hydroxides decomposed completely on
both iron oxides, also in line with the previous reports.17, 40 This
Table 1. The activity comparison can be based on either the initial
activities or the steadyꢀstate activities. Here we’d like to measure
their initial specific rate for comparison although the reaction
stabilities of all the catalysts are good (stability tests are shown
below). However, to avoid the probably occasional abnormal or
inaccurate data obtained in one time, we used the average activity
at 20, 40 and 60 min. We also compared the activity at 20 min and
the average activity at 20, 40 and 60 min, Figure S3, Table S2.
Although the activity at 20 min is slightly higher than that of the
average of 60 min, they are indeed very similar, suggesting that the
activity didn’t change much during the first 1 hour. As shown in
Table 1, the Au/ αꢀFe2O3 and Au/ αꢀFe2O3ꢀC catalysts exhibit
ACS Paragon Plus Environment

Page 5 of 12
ACS Catalysis
o
o
similar or higher TOFs compared with the standard Au/ Fe2O3ꢀ
peak at 250ꢀ300 C and a broad band at temperature > 350 C,
which correspond to the reduction of αꢀFe2O3/ γꢀFe2O3 to Fe3O4
and further to FeO/ Fe species, respectively.64ꢀ65 The temperature
of the sharp peak for αꢀFe2O3 is lower than that for γꢀFe2O3,
suggesting that αꢀFe2O3 is easier to be reduced with CO to form
Fe3O4. This is in line with the activity test result that αꢀFe2O3 is
more active than γꢀFe2O3. However, after introduction of Au
species, the reduction behavior of the catalysts changed. For Au/ αꢀ
Fe2O3 and Au/ γꢀFe2O3 catalysts, there are two sharp reduction
peaks and a broad reduction band. The first peak (denoted as peak
I) centered at about ꢀ5 oC can be ascribed to the reduction of Aux+
species. The second peak (denoted as peak II) corresponds to the
reduction of αꢀFe2O3/ γꢀFe2O3 to Fe3O4. Compared with the bare
iron oxides the reduction temperatures of peak II for both catalysts
decrease, suggesting that the presence of Au promoted the
reduction of iron oxides. However, the decrease extent varies
1
2
3
4
5
6
7
8
WGC catalyst (entry 7, Table 1) and higher specific rates.
ꢀ1
Especially, the Au/ αꢀFe2O3 yields a specific rate of 2.12 molco gAu
hꢀ1 (corresponding to ~0.06 molco gcatꢀ1 hꢀ1), suggesting it possessed
high activity according to Schuth’s definition (0.02 molco gcat hꢀ
ꢀ1
1).61 Compared with αꢀFe2O3 supported Au catalysts, γꢀFe2O3
supported Au catalyst exhibits much higher activity: it yields TOFs
of 2.02 and 1.14 sꢀ1 for Au/ γꢀFe2O3 and Au/ γꢀFe2O3ꢀC,
respectively, which was about 15 ꢀ 20 times higher than that of
Au/ αꢀFe2O3 and Au/ αꢀFe2O3ꢀC catalysts. They are even more
active than 2 ꢀ 3 nm Au colloidal on FeOx which is one of the most
active Au catalysts (entry 9, Table 1,).47 The highly active Au
species in the optimized size range of 0.5 ꢀ 5 nm usually generate
the TOF of around 1 sꢀ1.61ꢀ62 Therefore, our γꢀFe2O3 supported Au
catalysts are among the most active supported Au catalysts. Since
the TOF of Au/ γꢀFe2O3 catalyst was obtained by using a 90%
dispersion, their specific rate is particularly high, reaching to 33.3
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
significantly: for Au/ αꢀFe2O3 it decreased slightly from 280 C to
250 C while for Au/ γꢀFe2O3 it decreased dramatically from 300
oC to only 95 C. This result indicates clearly that although the
ꢀ1
ꢀ1
o
molco gAu hꢀ1 (corresponding to 1 molco gcat hꢀ1, about 50 times
higher than the highly active Au catalyst61). In fact, to our best
knowledge, it is the most active catalyst at room temperature
reported so far. Chen et al have reported a bilayer structure of Au
species possessed a TOF of 4 sꢀ1.63 However, their data were
obtained from model catalyst. Herzing et al subsequently reported
a similar data of ~3.5 sꢀ1 of the bilayer structure in a real catalyst of
Au/ FeOx (entry 10, table 1),56 while the fraction of this bilayer
structure in all Au species was only 0.6 atom% which resulted in a
much lower activity of the whole catalyst (0.027 sꢀ1, entry 11,
Table 1).
o
reducibility of γꢀFe2O3 is slightly lower than that of αꢀFe2O3, it
becomes much higher after it was collaborated with Au species.
The different promoting effect is probably due to the different
interaction between Au and the iron oxides, which will be
discussed later. As to the Au/ αꢀFe2O3ꢀC and Au/ γꢀFe2O3ꢀC
catalysts, their reduction behaviors are similar to that of the Au/ αꢀ
Fe2O3 and Au/ γꢀFe2O3, except that the sharp peak I disappeared.
This is as expected since the XPS data proved that the Aux+ species
was reduced after calcination at 300 o, Table 1. However, there is
o
o
still a visible broad band from ꢀ20 C to 60 C on Au/ γꢀFe2O3ꢀC
catalysts (Figure S4) which proved that some lattice oxygen, most
probable the surface O that closely contacted with Au species, can
Table 2. Oxygen species in the two iron oxides supported Au
catalysts obtained from the O1s XPS analysis
o
be reduced by CO at temperature as low as ca. ꢀ20 C. The area
Sample
B.E.
(eV)
O species
Percent
(%)
72
24
4
73
21
6
76
20
4
78
17
5
74
21
5
72
22
6
ratio of this band to peak II was calculated to be about 1 : 7,
showing that the percentage of these O species is about 14%. In
the case of Au/ αꢀFe2O3ꢀC, however, this band is ambiguous and
the area ratio to peak II is about 1 : 25, suggesting a much lower
percentage of reactive lattice O species.
Au/ γꢀFe2O3
529.9
531.5
533.0
530.1
531.5
533.1
530.0
531.6
532.9
529.9
531.2
533.1
529.8
531.3
533.1
529.7
531.5
532.8
Lattice O
OHꢀ
H2O
Lattice O
OHꢀ
H2O
Lattice O
OHꢀ
H2O
Lattice O
OHꢀ
H2O
Lattice O
OHꢀ
H2O
Lattice O
OHꢀ
H2O
Au/ γꢀFe2O3ꢀ
used
100
80
Au/ γꢀFe2O3ꢀC
Au/γꢀFe2O3ꢀUC
60
Au/γꢀFe2O3ꢀC3
Au/ γꢀFe2O3ꢀCꢀ
used
Au/αꢀFe2O3ꢀUC
40
Au/αꢀFe2O3ꢀC3
γꢀFe2O3
Au/ αꢀFe2O3
20
αꢀFe2O3
0
Au/ αꢀFe2O3ꢀC
-60 -40 -20
0
20 40 60 150 200 250 300
Temperature (oC)
Figure 4. CO conversion as a function of reaction temperature on
different Au/ Fe2O3 catalysts for CO oxidation. 1 vol% CO + 1 vol%
O2 + 98 vol%He, SV = 80,000 mL gcatꢀ1 hꢀ1.
Correlation of activity and redox property of the catalysts. It was
reported that the iron oxides supported Au catalysts were mainly
according to a redox mechanism at low temperature and the high
activity of Au/ ferrihydrite catalysts originated from its high redox
property.37 To demonstrate that this redox mechanism is general
for iron oxides supported Au catalysts and to correlate the redox
property with activity, we measured the reducibility of these αꢀ
Fe2O3 and γꢀFe2O3 supported catalysts with COꢀTPR, Figure 5.
Prior to the catalyst measurements, the bare iron oxides of αꢀFe2O3
and γꢀFe2O3 were firstly measured as reference for comparison. As
shown in Figure 5a, both αꢀFe2O3 and γꢀFe2O3 exhibit one sharp
The COꢀTPR results show clearly that the reducibility of iron
oxides and the catalysts are well consistent with their activity for
CO oxidation. To further corroborate that the high activity of γꢀ
Fe2O3 supported catalysts actually originated from their high redox
properties, the pulse reactions of CO oxidation were further
performed at ꢀ20 oC where the activity of CO oxidation on αꢀFe2O3
and γꢀFe2O3 supported Au catalysts had distinct difference
according to Figure 4. This method is effective to verify the redox
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 12
mechanism.37 Figure 6 illustrates the consumption of CO and O2
280 oC
303 oC
γ
α
ꢀ
ꢀ
Fe
Fe
2
2
O
O
3
3
1
2
3
4
5
6
7
8
as well as the production of CO2 on various catalysts. As shown in
Figures 6a and 6b, it consumes about 200 and 160 ꢁmol gcatꢀ1 CO
and meanwhile produced about 120 and 100 ꢁmol gcatꢀ1 CO2 with
each pulse of CO on Au/ γꢀFe2O3 and Au/ γꢀFe2O3ꢀC catalysts,
respectively. In each O2 pulse it consumed about half mole amount
of that of CO consumption or less under steady state while
without CO2 producing. This clearly suggested that the CO
oxidation accomplished in the CO pulse process by reacting with
the surface/ subꢀsurface lattice O of the γꢀFe2O3 support while the
O2 replenished the consumed lattice O in the O2 pulse process,
demonstrating the redox mechanism of CO oxidation. The
amount of CO2 production in each pulse was lower than that of the
CO consumption because part of the produced CO2 was absorbed
on the oxide support to form carbonates.37 The nonstoichiometric
consumption of O2 may be due to that the adsorption and accumuꢀ
lation of CO2 occupies some of the adsorption sites. Similarly, on
Au/ αꢀFe2O3 catalyst (Figure 6c) the same phenomenon is
observed with the sequential pulses of CO and O2 except that the
amounts of CO consumption and CO2 production are much lower
than that on Au/ γꢀFe2O3 and Au/ γꢀFe2O3ꢀC catalysts, in line with
the activity results. Furthermore, the amounts of CO consumption
and CO2 production increase with the reaction temperature,
indicating that the redox process is more favorable at elevated
temperatures. As to the Au/ αꢀFe2O3ꢀC catalyst (Figure 6d), there
is no CO2 production in both CO and O2 pulses, which is well
consistent with the fact that this catalyst is inactive for CO
(a)
250 oC
9
Au/γ
Au/α
ꢀ
ꢀ
Fe
Fe
2
2
O
O
3
3
(b)
-5 oC
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
95 oC
265 oC
Au/γ
Au/α
ꢀFe
2
O
O
3
ꢀC
ꢀC
(c)
ꢀ
Fe
2
3
105 oC
20oC
o
oxidation at ꢀ20 C (conversion < 5%). The redox reaction only
0
100 200 300 400 500 600 700 800 900
occures when reaction temperature increased to ca. 20 oC and the
amounts of CO consumption and CO2 production also increased
with the increase of the reaction temperature.
Temperature (oC)
Figure 5. COꢀTPR profiles of (a) αꢀFe2O3 and γꢀFe2O3, (b) Au/ αꢀ
Fe2O3 and Au/ γꢀFe2O3 and (c) Au/ αꢀFe2O3ꢀC and Au/ γꢀFe2O3ꢀC.
The above pulse reactions unambiguously demonstrated that CO
oxidation could follow a redox mechanism. However, this method
could not exclusively prove that CO reacted with the lattice O of
the iron oxides (FeꢀO species) because a possibility of CO
reaction with other O2ꢀcontaining species (for example a possible
stably adsorbed O species) could not be thoroughly ruled out.
However, the redox reaction with FeꢀO species must result to the
change of the iron oxides. We therefore employed the Raman
spectroscopy which is a powerful technique to detect the surface
change of oxides to confirm the redox reaction between Fe3+ and
Fe2+. Figure 7 presents the in situ Raman spectra of Au/ γꢀFe2O3ꢀC
and Au/ αꢀFe2O3ꢀC exposed to different atmosphere. For Au/ γꢀ
Fe2O3ꢀC in air, it exhibits four bands (370, 490, 667, 720 cmꢀ1)
which are typical γꢀFe2O3 feature.66 It should be particularly noted
that the 720 cmꢀ1 band were attributed to the vibrational modes of
local FeꢀO structures in the vicinity of cation vacancies.67 When
the air flow is replaced with 10 vol% CO/ He, the 720 cmꢀ1 band
disappears firstly, suggesting that the O species located at the
cation vacancies were most activated and had the highest activity.
Then other bands disappear gradually with time evaluation and
only the band at 667 cmꢀ1 remains. The 667 cmꢀ1 band is the typical
feature of Fe3O4, suggesting that the surface Fe3+ has been partially
reduced into Fe2+ under CO atmosphere to form Fe3O4.66, 68 When
the gas flow is switched from 10 vol% CO/ He to air, the feature
bands of γꢀFe2O3 appear again, indicating that surface Fe2+ is reꢀ
oxidized to Fe3+ under air flow. The in situ Raman spectra
apparently demonstrate the redoxprocess of Fe3+↔Fe2+ existed on
the surface of Au/ γꢀFe2O3ꢀC catalyst. On the contrary, for Au/ αꢀ
Fe2O3ꢀC catalyst it shows the typical αꢀFe2O3 feature bands in both
air and 10 vol% CO/ He gas flow, suggesting that no redox
reaction occured or the reaction was too weak to be detected at
room temperature.
Comparison of sequential pulse and coꢀpulse of CO and O2
reaction. To roughly estimate what percentage of the redox
pathway accounts for in the overall reaction, we further compared
the CO consumption and CO2 generation in sequential pulse of
CO and O2 and coꢀpulse of CO + O2 reaction.37 As shown in
Figure S5a, for Au/ γꢀFe2O3 catalyst at steady state, it consumes
about 210 ꢁmol gcatꢀ1 CO in each CO pulse in sequential pulse
ꢀ1
reaction while consumes about 380 ꢁmol gcat in each CO + O2
pulse. Meanwhile, it separately generates about 130 and 360 ꢁmol
gcatꢀ1 CO2. The much lower CO2 yields (denoted as percentage of
CO2 generation to CO consumption) in sequential pulse process
(~62% vs ~95%) suggests that CO2 adsorption and accumulation
is much more serious than that during coꢀpulse process which is
general for all the samples (Figure S5). From the CO consumption
ratio of sequential pulse to coꢀpulse process we can roughly
estimate that the redox process accounts for at least 55% in the
overall reaction.37 Considering that the CO2 accumulation in
sequential pulse reaction may hinder the subsequent O adsorption
and reaction and that in coꢀpulse process the O2 species can be
replenished timely thus each active site may participate in the
reaction more than once during a single pulse, it’s reasonable to
conjecture that the redox process might take an even higher
percentage. Similarly, for Au/ γꢀFe2O3ꢀC the redox process
accounts for about 50% percent. For Au/ αꢀFe2O3 and Au/ αꢀ
Fe2O3ꢀC catalyst, the redox process also account for about 50% or
o
even higher at ꢀ20 C. However, at elevated temperatures, the
percentages decrease significantly, suggesting that the dominating
reaction mechanim changes at higher temperature, espeically for
Au/ αꢀFe2O3ꢀC catalyst. This also explains the above result that no
redox process was detected in in situ Raman measurement (Figure
ACS Paragon Plus Environment

Page 7 of 12
ACS Catalysis
7) while there was high CO conversion (~50%, Figure 4) for
Au/ αꢀFe2O3ꢀC catalyst at 20 oC.
still valid at their steadyꢀstate and wonder whether the
1
2
3
4
5
6
7
8
characterizations are representative. In addition, the stability of a
catalyst is a crucial factor for its catalytic application. This is
particularly true for the supported Au catalysts since their reaction
stabilities are usually low and the stability issue has been a major
barrier to their practical application.69 Therefore, we further tested
the stabilities of these Au catalysts by significantly increasing the
SV to avoid the activity saturation. As shown in Figure 8, all
samples show good reaction stability: the CO conversions drop
only 10 ꢀ 20% during 100 hourꢀtest. Of more importance, the γꢀ
Fe2O3 supported catalysts show similar or even better stability
compared to the αꢀFe2O3 supported ones, demonstrating that they
are certainly more active, either for intial activity or in their
steadyꢀstate. The good stabilities make them more applicable. The
samples after stability test were examined by TEM and XPS to
characterize the possible changes in size and oxidation state of Au.
Figure S6 presents the TEM images of various fresh and used
samples and their corresponding Au particle size distribution. For
Au/ αꢀFe2O3 and Au/ γꢀFe2O3 catalysts, some ambiguous Au NPs
could be observed after 100 hourꢀtest, suggesting that sintering of
Au species occurred during reaction which may mainly accounts
for the rapid initial deactivation. Comparatively, there are no
significant size changes on Au/ αꢀFe2O3ꢀC and Au/ γꢀFe2O3ꢀC since
they were preꢀheated at 300 oC. The Au4f XPS (Figure S7) show
that after 100 hourꢀtest, Au species on all samples are Au0. This
suggests that Aux+ species on Au/ αꢀFe2O3 and Au/ γꢀFe2O3
catalysts were reduced during reaction which is consistent with
previous studies.37, 70 The O1s XPS were also measured (Figure S8)
and fitted (Table 2). Clearly for Au/ γꢀFe2O3 and Au/ γꢀFe2O3ꢀC
catalyst there is no detectable O specie change. For Au/ αꢀFe2O3
and Au/ αꢀFe2O3ꢀC catalysts, unfortunately, the diluent SiO2
contributed to the O spectra (γꢀFe2O3 supported catalysts can be
purified by magnetic separation) so the fitting analysis is not
available. However, based on the results of γꢀFe2O3 supported
catalysts, it is reasonable to conjecture that there’s no significant
change of O species on αꢀFe2O3 supported Au catalysts either.
CO CO CO CO CO CO CO CO CO CO
240
200 (a)
160
120
80
O2 O2 O2 O2 O2 O2 O2 O2 O2 O2
40
0
9
160
120
80
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
40
(b)
(c)
0
240
200
160
120
80
CO/O2
CO2
40
0
22-22, 60-60,100 --100oC
100
80
60
40
20
0
(d)
18------18, 65-----65, 100--100oC
8 10 12 14 16 18 20
-20
0
2
4
6
Pulse number
Figure 6. Sequential pulses of CO and O2 (square symbol) and
generation of CO2 (circle symbol) on (a) Au/ γꢀFe2O3, (b) Au/ γꢀ
Fe2O3ꢀC at ꢀ20 oC, (c) Au/ αꢀFe2O3 and (d) Au/ αꢀFe2O3ꢀC at ꢀ20,
ca. 20, ca. 60 and 100 oC.
100
80
Au/
γ
γ
α
ꢀFe
ꢀFe
ꢀFe
Au/αꢀFe
2
O
3
, SV = 32,300 L gAuꢀ1 hꢀ1
ꢀC, SV = 32,300 L gAuꢀ1 hꢀ1
, SV = 9,800 L gAuꢀ1 hꢀ1
ꢀC, SV = 2,500 L gAuꢀ1 hꢀ1
(a)
-1
(b)
490
667
Au/
2
O
3
720 cm
370
60
40
20
0
Au/
2
O
O
3
Air
2
3
CO
Air
CO
CO
CO
CO
0
10
20
30
40
50
60
70
80
90 100
Time (h)
Figure 8 CO oxidation stabilities of various Au/ Fe2O3 catalysts at
30 oC.
Air
200 400 600 800 1000
250 500 750 100012501500
To detect the change of oxidation state of Au species, we further
performed in situ DRIFT spectroscopy measurements of different
Au catalysts under CO oxidation reaction. As shown in Figure S9a
and b, on Au/ γꢀFe2O3 and Au/ αꢀFe2O3 catalysts three CO
adsorption bands centered at 2180, 2160 and 2110 cmꢀ1 are
observed, which could be attributed to gas CO, CO adsorption on
Au+ and Au0 species,71 respectively. It shows clearly that the Au+
species are rapidly reduced to Au0 within 2 min. Comparatively, on
Au/ γꢀFe2O3ꢀC and Au/ αꢀFe2O3ꢀC catalysts, there are only two
bands at 2180 and 2110 cmꢀ1 (Figure S9c and d), suggesting that
Raman shift (cmꢀ1)
Raman shift (cmꢀ1)
Figure 7. In situ Raman spectra of (a) Au/ γꢀFe2O3ꢀC and (b)
Au/ αꢀFe2O3ꢀC obtained in different atmosphere at room
temperature.
Catalyst stabilities and characterizations of used samples. It
should be mentioned that all the reaction data were tested for
initial activity and the characterization were performed for the
fresh sample. One may argue whether the activity comparison is
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 12
the Au species existed as Au0 at the beginning. This is well
To prove that on these catalysts CO oxidation also followed redox
1
2
3
4
5
6
7
8
consistent with the XPS results. In addition, on all samples
obvious adsorption bands at carbonate region (1200 ꢀ 1600 cmꢀ1)
are observed, suggesting the adsorption and accumulation of CO2
occurred during reaction, which is in line with the CO pulse
reaction results.
mechanism and positively correlated with their reducibility, we
further performed COꢀTPR and CO pulse reaction measurements.
From Figure S11 we can see that both Pt and Rh promoted the
o
reduction of γꢀFe2O3 to Fe3O4 significantly (lower than 200 C)
while less or didn’t promote the reduction of αꢀFe2O3. This results
are similar to that of the supported Au catalysts. Similarly, the CO
pulse reactions (Figure S12) also show that the redox cycles
occured on γꢀFe2O3 supported catalysts more than it did on αꢀ
Fe2O3 supported catalysts, although the difference is different
from that of supported Au catalysts.
100
Pt/γꢀFe
Pt/αꢀFe
2
O
3
ꢀC
ꢀC
80
60
40
20
0
2
O
3
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Specific rates and TOFs of iron oxides supported Pt and
Rh catalysts for CO oxidation at 150 oC
Sample
Metal
Specific rate
TOF × 10 3
loading
(wt%)
(molCO gmetalꢀ1 hꢀ1) (sꢀ1)
Pt/ αꢀFe2O3ꢀC
Pt/ γꢀFe2O3ꢀC
Rh/ αꢀFe2O3ꢀC
Rh/ γꢀFe2O3ꢀC
0.6
1.3
4.0
4.4
2.9
2.9
0.24
1.91
0.09
0.27
0.08
0.42
14.2
115.2
2.6
7.8
5
0
40
80
120
160
200
240
280
Temperature (oC)
Figure 9. CO conversion as function of reaction temperature for
CO oxidation on 1.3 wt% Pt/ γꢀFe2O3ꢀC and 0.6 wt% Pt/ αꢀFe2O3ꢀ
C catalysts. 1 vol% CO + 1 vol% O2 + 98 vol% He, SV = 80,000
mL gcatꢀ1 hꢀ1.
a
Au/ αꢀFe2O3
a
Au/ γꢀFe2O3
25
a for WGS reaction at 120 oC.
Iron oxides supported other metal catalysts for CO oxidation.
Above activity and characterization data unambiguously
demonstrated that i) the redox mechanism was universal for iron
oxides such as γꢀFe2O3 supported Au catalysts; ii) γꢀFe2O3
supported Au catalysts had better reducibility and redox property
than αꢀFe2O3 supported ones; iii) γꢀFe2O3 supported Au catalysts,
therefore, had higher activity for CO oxidation than αꢀFe2O3
supported ones. It was reported that the redox mechanism of
Au/ ferrihydrite for CO oxidation can be extended to other noble
metals such as supported Pd, Pt catalysts.37 Accordingly we can
expect that γꢀFe2O3 supported Pt, Pd catalysts would be more
active for CO oxidation than αꢀFe2O3 supported ones. To verify
this, we prepared αꢀFe2O3 and γꢀFe2O3 supported Pt and Pd
catalysts with the same method and tested their activities for CO
oxidations. As shown in Figure 9, as expected, the Pt/ γꢀFe2O3ꢀC
certainly exhibits much higher activity than that of the Pt/ αꢀ
Fe2O3ꢀC at the measurement temperature range. However, it
should be noted that the Pt loading on Pt/ γꢀFe2O3ꢀC is double of
that on Pt/ αꢀFe2O3ꢀC (Table 3) although the nominal Pt loadings
(5.0 wt%) were same. This indicated that Pt losing also occurred
during the prepared process and the losing extent were different,
probably due to the different interaction between Pt and the iron
oxides. Therefore, to give a more reasonable comparison, we
measured their specific rates for CO oxidation, Table 3. Clearly,
the activity of Pt/ γꢀFe2O3ꢀC is about 8 times higher than that of
Pt/ αꢀFe2O3ꢀC in terms of per weight of Pt. The Pd/ αꢀFe2O3ꢀC and
Pd/ γꢀFe2O3ꢀC catalysts were also prepared but unfortunately the
Pd metal could not be deposited on both iron oxides. The reason
is not clear yet and need to be further studied. We thus chose Rh,
another Ptꢀgroup metal which has been regarded as a good
candidate for CO oxidation,22, 72ꢀ74 to prepare Rh/ αꢀFe2O3ꢀC and
Rh/ γꢀFe2O3ꢀC to further verify this trend. As shown in Table 3,
although the difference is not as large as that of Pt and Au catalysts,
the activity of Rh/ γꢀFe2O3ꢀC is still 3 times higher than that of the
Rh/ αꢀFe2O3ꢀC catalyst. The TOFs were calculated based on a 90%
dispersion since the TEM images showed that there were no
visible NPs on all samples, Figure S10.
Activity of other reaction followed redox mechanism. All
experimental data above show very good positive correlation
between the redox properties and the activity for CO oxidation
due to that CO oxidation is mainly according to the redox
mechanism on the iron oxides supported catalysts in the
measurement temperature range. Therefore, we can further
conjecture that for other reactions that follow redox mechanism
the γꢀFe2O3 supported Au catalysts would also have higher activity
than that of αꢀFe2O3 supported ones. WGS is an old but important
industrial processes employed in energy conversion and H2 gas
purification.75 This reaction has been well accepted to follow redox
mechanism on iron oxides at high temperatures. Recently, it has
also proposed a dominant redox (regeneration) mechanism for
iron oxide supported Au catalysts.75ꢀ78 We therefore chose WGS as
a probe reaction to verify this prediction. The results presented in
Table 3 clearly show that, although their activities are much lower
than that of CO oxidation, the Au/ γꢀFe2O3 is about 5 times more
active than Au/ αꢀFe2O3, suggesting the critical role of the high
redox property of γꢀFe2O3. However, it should be noted that
although the redox mechanism has been well accepted for
Au/ Fe2O3 catalysts,75ꢀ78 a possibility that the redox properties of
the support material are crucial in the WGS reaction with an assoꢀ
ciative mechanism that invokes the formation of stable surface
intermediates located at oxygen vacancies cannot be exclusively
ruled out. Therefore, a detailed mechanism study is needed.
DISCUSSION
It was reported that CO oxidation on ferrihydrite supported Au
catalysts obeyed the redox mechanism.37 Therefore, the catalyst
with higher redox properties would have better activity. However,
in that report,37 all samples were prepared with a coꢀprecipitation
method. To compare the ferrihydrite and αꢀFe2O3 supported
catalysts, the samples were treated at different temperatures (e.g.
o
o
dried at 60 C vs. calcined at 400 C). Therefore, some other
factors such as the size of Au NPs, the chemical state of Au species
and surface OHꢀ groups which were thought to dramatically affect
ACS Paragon Plus Environment

Page 9 of 12
ACS Catalysis
the activity of supported Au catalysts varied. In this work, to
was much more reducible than that of αꢀFe2O3, Figure 5, COꢀTPR.
1
2
3
4
5
6
7
8
exclude the effect of these factors, we use commercial iron oxides
as supports and deposit the Au NPs by a depositionꢀprecipitation
method which ensured that all Au species were dispersed on the
surface of the supports. Furthermore, we made the same heat
treatments on both samples, i.e., dried at 60 oC only and calcined
A similar trend of H2ꢀTPR was also observed for Au/ αꢀFe2O3 and
Au/ γꢀFe2O3.39 Since the size and chemical state of Au were almost
same, this different metal promoting effect must originate from the
different metalꢀsupport interaction or the special crystal structure
of γꢀFe2O3. γꢀFe2O3 has an inverse spinel structure similar to that
of Fe3O4 except that it has cation vacancies that compensate for
the oxidation of FeII.38, 67 The same crystal structure allows γꢀFe2O3
to be reduced to Fe3O4 by losing of O while without lattice
rearrangement (see Figure S13). The losing of O could occur at
low temperature under the assistant of Au species either by
H2/ CO spillover or by weakening the strength of the FeꢀO band.19,
37 Furthermore, the stronger interaction between Au and γꢀFe2O3,
as suggested by the less aggregation degree of Au NPs (Figures 3e,
f), might activate the FeꢀO band in γꢀFe2O3 more than that in αꢀ
Fe2O3. Especially, γꢀFe2O3 has cation vacancies which may adsorb
metal to form some special/ stronger interaction different from
that between Au and αꢀFe2O3, as evidenced by the in situ Raman
spectra. As shown in Figure 7, the O species located in the vicinity
o
at 300 C under N2 atmosphere. By taking these procedures, αꢀ
Fe2O3 and γꢀFe2O3 supported Au catalysts, with or without
calcination, have similar size of Au NPs (Figure 3), chemical state
of Au species (Table 1, Figure S1) and amount of surface OHꢀ
group (Table 2, Figure S2). In this case, we can claim that the
effect of these factors has been, at least majorly, ruled out. The
activity of CO oxidation can be therefore directly correlated to the
redox properties of the catalysts. The subsequent activity test
results had very good correlation with the COꢀTPR
characterization results, as shown in Figure 4 and Figure 5. The
surface pulse reaction of CO and O2 and in situ Raman spectra
further confirmed that the high activity of γꢀFe2O3 supported Au
catalysts originated from their high redox properties, Figure 6 and
Figure 7.
It was suggested that the redox mechanism was universal for
ferrihydrite supported other metal (such as Pt and Pd) catalysts.37
Since we have demonstrated in this work that redox mechanism
also works for γꢀFe2O3 supported Au catalysts, we can therefore
expect that γꢀFe2O3 supported other metal catalysts would have
higher activity for CO oxidation than that of αꢀFe2O3 supported
ones. The certain examples (Pt and Rh in this work) have further
verified the prediction, although their activities and differences in
activity varied, Figure 9 and Table 3. Moreover, based on these
success predictions, we can further expect that other reactions that
follow redoxmechanism have the same activity order. As expected,
the γꢀFe2O3 supported Au catalyst certainly showed higher activity
for the selected probe reaction (WGS reaction in this work) than
that of αꢀFe2O3 supported Au catalyst, Table 3.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
of cation vacancies was most activated, indicating
a
stronger/ special interaction between gold and these lattice O
species. The high activity of Au/ γꢀFe2O3 clearly suggests that an
oxide with high reducibility is not indispensable, as long as the
metal/ oxide entity has higher reducibility/ redoxproperty.
Besides the high activity and the wide application of γꢀFe2O3 as a
potential support material, this work demonstrated the dramatic
effect of crystal phase of iron oxides on CO oxidation activity. It
has been noticed recently that CO oxidation on other reducible
oxides (for example CeO2,31 MnOx,32ꢀ34 and TiO235ꢀ36) supported
Au catalysts also followed the redox mechanism, at different
reaction temperature region and to different extent. We can
therefore in some degree predict the catalytic performance of
these oxides supported metal catalysts based on their
reducibility/ redox properties. We believe that metal/ oxide entity
with higher redox property would have higher activity for CO
oxidation. Indeed, there is a good example that various singleꢀ
phase manganese oxides supported Au catalysts exhibited a good
correlation between their activity for CO oxidation and their
reducibility (measured by H2ꢀTPR), although the highest activity
was mainly attributed to the smaller size of Au NPs by the
authors.33 Perhaps the higher redox property is the main
origination of the high activity. Apart from CO oxidation, many
important chemical reaction were proposed to follow the redox
mechanism such as WGS reaction75 and selective oxidations which
usually use oxides as catalysts.87 Obviously, an inꢀdepth
understanding of the importance of the redox property of oxides
supported noble metal catalysts would be helpful in development
of new catalysts or improvement of the catalyst formula. Moreover,
recently a correlation between the reducibility of iron oxide crystal
phase and the catalytic activity of iron oxides supported gold cataꢀ
lysts was also observed for hydrogenation reaction,88 although
their reaction mechanism is different from ours. This suggested
that in other types of reaction such as hydrogenation, the catalyst’s
reducibility also took important role in determining their catalytic
performance, which will make it more valuable to understand the
reducibility of the catalysts.
The study in this work revealed that γꢀFe2O3 could be a good
support for many metals to prepare highly active catalysts which
79ꢀ81
was rarely noticed before.39,
Particularly, the Au/ γꢀFe2O3
showed exceptionally high activity for CO oxidation and presented
the highest activity so far except the bilayer structure of Au in a
model catalyst.63 γꢀFe2O3 is one of the most common polymorphs
in many iron oxides.82 Although it is less stable than αꢀFe2O3, the
most stable polymorph among iron oxides, it’s much more stable
than ferrihydrite. Ferrihydrite is highly reducible and the
metal/ ferrihydrite catalyst had very good redox property.37
Therefore, metal/ ferrihydrite entities are highly active for CO
24, 83
oxidation and PROX reaction.13ꢀ22,
And theoretically, they
should also have very high activity for other reactions that follow
the redox mechanism. However, ferrihydrite is not heating
o
resistant. Even middleꢀtemperature heat treatment (e.g. 200 C)
may arouse the phase transformation and result in the activity
37
decrease.15ꢀ17, This heating nonresistance severely limited its
practical application, especially at elevated temperatures.
Relatively, γꢀFe2O3 is much more thermodynamically stable and
can stand at least 300 oC calcination without any phase
transformation, as confirmed in this work. With special prepare
method to form special morphology, they can even stand
calcination at 550 ꢀ 600 oC without phase change.84 In addition, γꢀ
Fe2O3 is ferromagnetic at a wide temperature range which renders
it very suitable as catalyst support for easy separation in liquid
phase reactions.85ꢀ86 All these features make γꢀFe2O3 more
applicable as a catalyst support than ferrihydrite, especially for
those highꢀtemperature or liquidꢀphase reactions.
It is worth pointing out that Acerbi et al recently proposed that the
reducibility enhancement of oxides (e.g. CeO2) by metals
depended on the work function89 and the dꢀband center90 of the
supported metals. This means that we can even predict the
reducibility/ redox property of metal/ oxide and subsequently the
activity of metal/ oxides catalysts simply by theoretical calculation
based on their work function and dꢀband center.
Different from ferrihydrite, γꢀFe2O3 itself is less reducible
compared with αꢀFe2O3. However, after incorporation of Au, it
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 12
5. van Santen, R. A.; Future Perspectives in Catalysis. Dutch National
CONCLUSIONS
Research School Combination, Catalysis Controlled by Chemical
Design: Eindhoven, Netherlands, 2009.
6. Derouane, E. G., CATTECH 2001, 5, 214ꢀ225.
7. Haruta, M., Chem. Rec. 2003, 3, 75ꢀ87.
1
2
3
4
5
6
7
8
Commercial αꢀFe2O3 and γꢀFe2O3 oxides supported Au catalysts
with identical size of Au NPs, chemical state of Au species and
amount of surface OHꢀ group were prepared. It was found that the
Au/ γꢀFe2O3 catalysts were much more active than Au/ αꢀFe2O3
catalysts for CO oxidation, presenting one of the most active
catalysts so far. The high activity originated from the higher redox
property of Au/ γꢀFe2O3, demonstrating that the CO oxidation on
Au/ γꢀFe2O3 mainly followed the redox mechanism. More
importantly, this phenomenon could be extended to γꢀFe2O3
supported other noble metals and to other reaction that followed
redox mechanism. This finding may provide a new avenue for
catalyst development by choosing the correct oxide crystal phase.
8. Hughes, M. D.; Xu, Y.ꢀJ.; Jenkins, P.; McMorn, P.; Landon, P.; Enache,
D. I.; Carley, A. F.; Attard, G. A.; Hutchings, G. J.; King, F.; Stitt, E.
H.; Johnston, P.; Griffin, K.; Kiely, C. J., Nature 2005, 437, 1132ꢀ1135.
9. Corma, A.; Serna, P., Science 2006, 313, 332ꢀ334.
10. Turner, M.; Golovko, V. B.; Vaughan, O. P. H.; Abdulkin, P.;
BerenguerꢀMurcia, A.; Tikhov, M. S.; Johnson, B. F. G.; Lambert, R.
M., Nature 2008, 454, 981ꢀ983.
11.Zhong, D.; Franke, J.ꢀH.; Podiyanachari, S. K.; Blömker, T.; Zhang, H.;
Kehr, G.; Erker, G.; Fuchs, H.; Chi, L., Science 2011, 334, 213ꢀ216.
12. Ball, L. T.; LloydꢀJones, G. C.; Russell, C. A., Science 2012, 337, 1644ꢀ
1648.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13. Finch, R. M.; Hodge, N. A.; Hutchings, G. J.; Meagher, A.; Pankhurst,
Q. A.; Siddiqui, M. R. H.;Wagner, F. E.; Whyman, R., Phys. Chem.
Chem. Phys. 1999, 1, 485ꢀ489.
14. Hodge, N. A.; Kiely, C. J.; Whyman, R.; Siddiqui, M. R. H.; Hutchings,
G. J.; Pankhurst, Q. A.; Wagner, F. E.; Rajaram, R. R.; Golunski, S. E.,
Catal. Today 2002, 72, 133ꢀ144.
15. Qiao, B.; Deng, Y., Chem. Commun. 2003, 2192ꢀ2193.
16. Qiao, B.; Deng, Y., Appl. Catal., B 2006, 66, 241ꢀ248.
17. Qiao, B.; Zhang, J.; Liu, L.; Deng, Y., Appl. Catal., A 2008, 340, 220ꢀ
228.
18. Qiao, B.; Liu, L.; Zhang, J.; Deng, Y., J. Catal. 2009, 261, 241ꢀ244.
19. Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu, J.; Li,
J.; Zhang, T., Nat. Chem. 2011, 3, 634ꢀ641.
20. Qiao, B.; Wang, A.; Takahashi, M.; Zhang, Y.; Wang, J.; Deng, Y.;
Zhang, T., J. Catal. 2011, 279, 361ꢀ365.
21. Qiao, B.; Wang, A.; Lin, J.; Li, L.; Su, D.; Zhang, T., Appl. Catal., B
2011, 105, 103ꢀ110.
22. Lin, J.; Qiao, B.; Liu, J.; Huang, Y.; Wang, A.; Li, L.; Zhang, W.; Allard,
L. F.; Wang, X.; Zhang, T., Angew. Chem. Int. Ed. 2012, 51, 2920ꢀ2924.
23. Liu, L.; Qiao, B.; He, Y.; Zhou, F.; Yang, B.; Deng, Y., J. Catal. 2012,
294, 29ꢀ36.
24. Qiao, B.; Lin, J.; Li, L.; Wang, A.; Liu, J.; Zhang, T., ChemCatChem
2014, 6, 547ꢀ554.
ASSOCIATED CONTENT
Supporting Information
Mössbauer parameters of bare γꢀFe2O3 and Au/ Fe2O3 after 300 oC
calcination under N2 atmosphere, XPS spectra of Au4f and O1s of
various supported Au catalysts before and after reaction, CO conꢀ
version curves for specific rate test and activity comparison of
various supported Au catalysts, MS signals of CO and CO2 for COꢀ
TPR profiles of Au/ αꢀFe2O3ꢀC and Au/ γꢀFe2O3ꢀC, curves of CO
consumption and CO2 generation during coꢀpulse and sequential
pulse of CO and O2 reaction over various supported Au catalysts,
TEM images of various supported Pt, Rh and Au catalysts before
or after reaction, in situ DRIFT spectra of various supported Au
catalysts under reaction, COꢀTPR profiles and CO pulse reactions
of supported Pt and Rh catalysts, and crystal structures of various
iron oxides. This material is available free of charge via the Interꢀ
net at http:/ / pubs.acs.org.
AUTHORINFORMATION
Corresponding Authors
25. Golunski, S., Platinum Met. Rev. 2013, 57, 82ꢀ84.
26. Liu, X.; Liu, M.ꢀH.; Luo, Y.ꢀC.; Mou, C.ꢀY.; Lin, S. D.; Cheng, H.;
Chen, J.ꢀM.; Lee, J.ꢀF.; Lin, T.ꢀS., J. Am. Chem. Soc. 2012, 134, 10251ꢀ
10258.
27. Freund, H. J.; Meijer, G.; Scheffler, M.; Schlogl, R.; Wolf, M., Angew.
Chem. Int. Ed. 2011, 50, 10064ꢀ94.
wangjh@dicp.ac.cn
bqiao@dicp.ac.cn
Notes
The authors declare no competing financial interests.
28. Shelef, M.; McCabe, R. W., Catal. Today 2000, 62, 35ꢀ50.
29. Twigg, M. V., Appl. Catal., B 2007, 70, 2ꢀ15.
# These authors contribute equally to this work.
ψ Present address: National Engineering Research Center for Nanꢀ
otechnology, Shanghai, 200241, China.
30. Liu, K.; Wang, A. Q.; Zhang, T., ACS Catal. 2012, 2, 1165ꢀ1178.
31. Widmann, D.; Leppelt, R.; Behm, R. J., J. Catal. 2007, 251, 437ꢀ442.
32.Ma, Z.; Liang, C.; Overbury, S. H.; Dai, S., J. Catal. 2007, 252, 119ꢀ126.
33. Wang, L.ꢀC.; Liu, Q.; Huang, X.ꢀS.; Liu, Y.ꢀM.; Cao, Y.; Fan, K.ꢀN.,
Appl. Catal., B 2009, 88, 204ꢀ212.
34. Morgan, K.; Cole, K. J.; Goguet, A.; Hardacre, C.; Hutchings, G. J.;
Maguire, N.; Shekhtman, S. O.; Taylor, S. H., J. Catal. 2010, 276, 38ꢀ
48.
35. Widmann, D.; Behm, R. J., Angew. Chem. Int. Ed. 2011, 50, 10241ꢀ
10245.
36. Widmann, D.; Liu, Y.; Schüth, F.; Behm, R. J., J. Catal. 2010, 276, 292ꢀ
305.
ACKNOWLEDGMENT
This work was supported by the Chinese Academy of Sciences for
“100 Talents” Project, National Natural Science Foundation of
China (21173218, 21203181, 11079036, 21476232).
REFERENCES
1. Deutschmann, O.; Knözinger, H.; Kochloefl, K.; Turek, T.,
Heterogeneous Catalysis and Solid Catalysts. WileyꢀVCH Verlag GmbH
& Co. KGaA: Weinheim, German, 2000.
2. Ertl, G.; Knözinger, H.; Schüth, F.; Weitkamp, J., Handbook of
heterogeneous catalysis. 2 ed.; WileyꢀVCH Verlag GmbH& Co. KGaA:
Weinheim, German, 2008; Vol. 1.
3. Norskov, J. K.; Bligaard, T.; Rossmeisl, J.; Christensen, C. H., Nat.
Chem 2009, 1, 37ꢀ46.
4. Aris, R.; Bell, A. T.; Boudart, M.; Chen, N. Y.; Gates, B. C.; HAAG, W.
O.; Somorjar, G. A.; Wei, J., Catalyst Design: Progress and Perspective
WileyꢀInterscience: 1987.
37. Li, L.; Wang, A.; Qiao, B.; Lin, J.; Huang, Y.; Wang, X.; Zhang, T., J.
Catal. 2013, 299, 90ꢀ100.
38. Cornell, R. M.; Schwertmann, U., In The Iron Oxides, WileyꢀVCH
Verlag GmbH & Co. KGaA: Weinheim, German, 2004; pp 9ꢀ38.
39. Huang, J.; Dai, W.ꢀL.; Fan, K., J. Catal. 2009, 266, 228ꢀ235.
40. Zhao, K.; Qiao, B.; Wang, J.; Zhang, Y.; Zhang, T., Chem. Commun.
2011, 47, 1779ꢀ1781.
41. Zhao, K.; Qiao, B.; Zhang Y.; Wang, J., Chin. J. Catal. 2013, 34, 1386ꢀ
1394.
42. Lin, J.; Li, L.; Huang, Y.; Zhang, W.; Wang, X.; Wang, A.; Zhang, T., J.
Phys. Chem. C 2011, 115, 16509ꢀ16517.
43. Wolf, A.; Schuth, F., Appl. Catal., A 2002, 226, 1ꢀ13.
44. Moreau, F.; Bond, G. C.; Taylor, A. O., J. Catal. 2005, 231, 105ꢀ114.
ACS Paragon Plus Environment

Page 11 of 12
ACS Catalysis
45. Haruta, M., Catal. Today 1997, 36, 153ꢀ166.
70. FierroꢀGonzalez, J. C.; Gates, B. C., J. Phys. Chem. B 2004, 108, 16999ꢀ
17002.
71. Mihaylov, M.; Knözinger, H.; Hadjiivanov, K.; Gates, B. C., Chem. Ing.
Tech. 2007, 79, 795ꢀ806.
46. Valden, M.; Lai, X.; Goodman, D. W., Science 1998, 281, 1647ꢀ1650.
47. Liu, Y.; Jia, C.ꢀJ.; Yamasaki, J.; Terasaki, O.; Schüth, F., Angew. Chem.
Int. Ed. 2010, 49, 5771ꢀ5775.
1
2
3
4
5
6
7
8
9
48. Guzman, J.; Gates, B. C., J. Am. Chem. Soc. 2004, 126, 2672ꢀ2673.
49. Yu, X.; Hantsche, H., Fresenius J. Anal. Chem. 1993, 346, 233ꢀ236.
50. Cuenya, B. R.; Baeck, S.ꢀH.; Jaramillo, T. F.; McFarland, E. W., J. Am.
Chem. Soc. 2003, 125, 12928ꢀ12934.
51. Cornell, R. M.; Schwertmann, U., In The Iron Oxides, WileyꢀVCH
Verlag GmbH & Co. KGaA: Weinheim, German, 2004; pp 111ꢀ137.
52. Peters, S.; Peredkov, S.; Neeb, M.; Eberhardt, W.; AlꢀHada, M., Surf.
Sci. 2013, 608, 129ꢀ134.
53. Costello, C. K.; Kung, M. C.; Oh, H. S.; Wang, Y.; Kung, H. H., Appl.
Catal., A 2002, 232, 159ꢀ168.
54. Smit, G.; Strukan, N.; Craje, M. W. J.; Lazard, K., J. Mol. Catal. A:
Chem. 2006, 252, 163ꢀ170.
55. Veith, Gabriel M.; Lupini, Andrew R.; Pennycook, Stephen J.; Dudney,
Nancy J., ChemCatChem 2010, 2, 281ꢀ286.
56. Herzing, A. A.; Kiely, C. J.; Carley, A. F.; Landon, P.; Hutchings, G. J.,
Science 2008, 321, 1331ꢀ1335.
57. Jing, L. Q.; Wang, D. J.; Wang, B. Q.; Li, S. D.; Xin, B. F.; Fu, H. G.;
Sun, J. Z., J. Mol. Catal. A: Chem. 2006, 244, 193ꢀ200.
58. Haber, J.; Stoch, J.; Ungier, L., J. Electron. Spectrosc. Relat. Phenom.
1976, 9, 459ꢀ467.
59. Martensson, N.; Malmquist, P. A.; Svensson, S.; Basilier, E.; Pireaux, J.
J.; Gelius, U.; Siegbahn, K., Nouveau, J. Chim. 1977, 1, 191ꢀ195.
60. Carabineiro, S. A. C.; Bogdanchikova, N.; Tavares, P. B.; Figueiredo, J.
L., RSC Adv. 2012, 2, 2957ꢀ2965.
72. Ligthart, D. A. J. M.; van Santen, R. A.; Hensen, E. J. M., Angew. Chem.
Int. Ed. 2011, 50, 5306ꢀ5310.
73. Grass, M. E.; Zhang, Y. W.; Butcher, D. R.; Park, J. Y.; Li, Y. M.;
Bluhm, H.; Bratlie, K. M.; Zhang, T. F.; Somorjai, G. A., Angew. Chem.
Int. Ed. 2008, 47, 8893ꢀ8896.
74. Ghosh, T. K.; Nair, N. N., ChemCatChem 2013, 5, 1811ꢀ1821.
75. Ratnasamy, C.; Wagner, J. P., Catal. Rev. Sci. Eng. 2009, 51, 325 ꢀ 440.
76. Andreeva, D.; Idakiev, V.; Tabakova, T.; Andreev, A.; Giovanoli, R.,
Appl. Catal., A 1996, 134, 275ꢀ283.
77. Boccuzzi, F.; Chiorino, A.; Manzoli, M.; Andreeva, D.; Tabakova, T., J.
Catal. 1999, 188, 176ꢀ185.
78. Aeijelts Averink Silberova, B.; Mul, G.; Makkee, M.; Moulijn, J. A., J.
Catal. 2006, 243, 171ꢀ182.
79. Kozlova, A. P.; Sugiyama, S.; Kozlov, A. I.; Asakura, K.; Iwasawa, Y., J.
Catal. 1998, 176, 426ꢀ438.
80. Smit, G., Croat. Chem. Acta 2003, 76, 269ꢀ271.
81. Gawande, M. B.; Rathi, A. K.; Tucek, J.; Safarova, K.; Bundaleski, N.;
Teodoro, O. M. N. D.; Kvitek, L.; Varma, R. S.; Zboril, R., Green Chem.
2014, 16, 4137ꢀ4143.
82. Cornell, R. M.; Schwertmann, U., In The Iron Oxides, WileyꢀVCH
Verlag GmbH & Co. KGaA: Weinheim, German, 2004; pp 1ꢀ7.
83. Golunski, S.; Rajaram, R.; Hodge, N.; Hutchings, G. J.; Kiely, C. J.,
Catal. Today 2002, 72, 107ꢀ113.
84.Mou, X.; Zhang, B.; Li, Y.; Yao, L.; Wei, X.; Su, D. S.; Shen, W., Angew.
Chem. Int. Ed. Engl. 2012, 51, 2989ꢀ2993.
85. Mori, K.; Kanai, S.; Hara, T.; Mizugaki, T.; Ebitani, K.; Jitsukawa, K.;
Kaneda, K., Chem. Mater. 2007, 19, 1249ꢀ1256.
86. Mou, X.; Wei, X.; Li, Y.; Shen, W., CrystEngComm 2012, 14, 5107ꢀ
5120.
87. Mars, P.; van Krevelen, D. W., Chem. Eng. Sci. 1954, 3, Supplement 1,
41ꢀ59.
88. Milone, C.; Crisafulli, C.; Ingoglia, R.; Schipilliti, L.; Galvagno, S.,
Catal. Today 2007, 122, 341ꢀ351.
89. Acerbi, N.; Tsang, S. C.; Golunski, S.; Collier, P., Chem. Commun.
2008, 0, 1578ꢀ1580.
90. Acerbi, N.; Tsang, S. C. E.; Jones, G.; Golunski, S.; Collier, P., Angew.
Chem. Int. Ed. 2013, 52, 7737ꢀ7741.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61. Schüth, F., Phys. Status Solidi B 2013, 250, 1142ꢀ1151.
62. Haruta, M., Faraday Discuss. 2011, 152, 11ꢀ32.
63. Chen, M.; Goodman, D. W., Science 2004, 306, 252ꢀ255.
64. Li, P.; Miser, D. E.; Rabiei, S.; Yadav, R. T.; Hajaligol, M. R., Appl.
Catal., B 2003, 43, 151ꢀ162.
65. Zielinski, J.; Zglinicka, I.; Znak, L.; Kaszkur, Z., Appl. Catal., A 2010,
381, 191ꢀ196.
66. Legodi, M. A.; de Waal, D., Dyes Pigm. 2007, 74, 161ꢀ168.
67. Long, J. W.; Logan, M. S.; Rhodes, C. P.; Carpenter, E. E.; Stroud, R.
M.; Rolison, D. R., J. Am. Chem. Soc. 2004, 126, 16879ꢀ16889.
68. de Faria, D. L. A.; Venâncio Silva, S.; de Oliveira, M. T., J. Raman
Spectrosc. 1997, 28, 873ꢀ878.
69. Corti, C. W.; Holliday, R. J.; Thompson, D. T., Top. Catal. 2007, 44,
331ꢀ343.
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12
ACS Paragon Plus Environment