Discriminating Catalytically Active FeNx Species of Atomically Dispersed Fe-N-C Catalyst for Selective Oxidation of the C-H Bond

Article abstract of DOI:10.1021/jacs.7b05130

Nanostructured Fe-N-C materials represent a new type of "platinum-like" non-noble-metal catalyst for various electrochemical reactions and organic transformations. However, no consensus has been reached on the active sites of the Fe-N-C catalysts because of their heterogeneity in particle size and composition. In this contribution, we have successfully prepared atomically dispersed Fe-N-C catalyst, which exhibited high activity and excellent reusability for the selective oxidation of the C-H bond. A wide scope of substrates, including aromatic, heterocyclic, and aliphatic alkanes, were smoothly oxidized at room temperature, and the selectivity of corresponding products reached as high as 99%. By using sub-?ngstr?m-resolution HAADF-STEM in combination with XPS, XAS, ESR, and M?ssbauer spectroscopy, we have provided solid evidence that Fe is exclusively dispersed as single atoms via forming FeN_x (x = 4-6) and that the relative concentration of each FeN_x species is critically dependent on the pyrolysis temperature. Among them, the medium-spin Fe^IIIN₅ affords the highest turnover frequency (6455 h^-1), which is at least 1 order of magnitude more active than the high-spin and low-spin Fe^IIIN₆ structures and 3 times more active than the Fe^IIN₄ structure, although its relative concentration in the catalysts is much lower than that of the Fe^IIIN₆ structures.

Full text of DOI:10.1021/jacs.7b05130

Article
pubs.acs.org/JACS
Discriminating Catalytically Active FeN Species of Atomically
x
Dispersed Fe−N−C Catalyst for Selective Oxidation of the C−H Bond
§
,‡,†
§,†
§
§
§
§
§
Wengang Liu,
Leilei Zhang, Xin Liu, Xiaoyan Liu, Xiaofeng Yang, Shu Miao, Wentao Wang,
,
§
,§
Aiqin Wang,* and Tao Zhang*
§State Key Laboratory of Catalysis, iChEM (Collaborative Innovation Center of Chemistry for Energy Materials), Dalian Institute of
Chemical Physics, Chinese Academy of Sciences, Dalian, 116023, China
‡University of Chinese Academy of Sciences, Beijing, 100049, China
*
S Supporting Information
ABSTRACT: Nanostructured Fe−N−C materials represent a
new type of “platinum-like” non-noble-metal catalyst for
various electrochemical reactions and organic transformations.
However, no consensus has been reached on the active sites of
the Fe−N−C catalysts because of their heterogeneity in
particle size and composition. In this contribution, we have
successfully prepared atomically dispersed Fe−N−C catalyst,
which exhibited high activity and excellent reusability for the
selective oxidation of the C−H bond. A wide scope of
substrates, including aromatic, heterocyclic, and aliphatic
alkanes, were smoothly oxidized at room temperature, and
the selectivity of corresponding products reached as high as 99%. By using sub-an
combination with XPS, XAS, ESR, and Mossbauer spectroscopy, we have provided solid evidence that Fe is exclusively dispersed
as single atoms via forming FeN (x = 4−6) and that the relative concentration of each FeN species is critically dependent on
̊
̈
gstrom-resolution HAADF-STEM in
̈
x
x
III
−1
the pyrolysis temperature. Among them, the medium-spin Fe N affords the highest turnover frequency (6455 h ), which is at
5
III
II
least 1 order of magnitude more active than the high-spin and low-spin Fe N structures and 3 times more active than the Fe N
structure, although its relative concentration in the catalysts is much lower than that of the Fe N structures.
6
4
III
6
2
3
INTRODUCTION
adsorption configurations after a thorough XANES analysis.
■
Su et al. argued that the real active site for the ORR in the Fe−
Noble metals are widely used as catalysts in energy conversion,
chemicals production, and emission control.
III
1
−3
N−C catalysts was FeN , a [Fe (porphyrin)(pyridine) ]
6
2
The scarcity
24
̈
complex, on the basis of Mossbauer spectroscopy. One key
and high cost of noble metals, however, constitute major
barriers to large-scale practical applications. Searching for earth-
abundant yet efficient catalysts to replace noble metals has
therefore become one of the central tasks in the catalysis
reason for the discrepancy about the active sites of M−N−C
materials is that the catalysts are heterogeneous in both
composition and particle size as a result of an uncontrollable
pyrolysis process. In fact, in most cases multispecies of M, such
4
−7
community.
Among others, M−N−C (M typically refers to
as metallic M nanoparticles, MO nanoparticles, and mono-
x
Fe and Co) materials, which were originally obtained by
pyrolysis of transition-metal-containing macrocycles and later
could be prepared by simply heating mixtures of more available
and cheaper metal and nitrogen precursors on carbon supports,
have attracted a great deal of attention thanks to their
promising catalytic performances in a variety of typical
platinum-catalyzed reactions, such as the oxygen reduction
dispersed and aggregated M−N , are simultaneously present in
x
24−27
one catalyst,
which poses a significant challenge for the
identification of the active site.
On the other hand, single-atom catalyst (SAC), in which
catalytically active metal is exclusively dispersed as single
2
,28−37
atoms,
provides a good entry to identify the structure of
8
−11
12
the active site without interference from other metal species.
M−N−C catalysts with M dispersed as exclusively single atoms
can be viewed as a special SAC if the surrounding N and C
atoms with which the metal atoms bond are considered as a
In spite of the distinctive advantages of uniform
dispersion of active sites, the successful synthesis of M−N−C
SAC remains challenging and critically depends on the rational
reaction (ORR),
CO2 electroreduction reaction, and
1
3−16
hydrogen evolution reaction (HER),
as well as fine
However, despite extensive inves-
17−21
chemical synthesis.
tigations of M−N−C catalysts, no consensus has been reached
on the active sites. For example, Mukerjee and co-workers
stated that the active site of Fe−N−C materials was FeN C in
16,38
ligand.
4
10
crystallographic atomic defects or FeN C in armchair edges of
4
8
22
graphitic surfaces. Zitolo et al. proposed that the Fe−N4
moieties in Fe−N−C catalysts should be assigned to in-plane
Fe−N4 porphyrinic architectures with two different O2
Received: May 24, 2017
Published: July 26, 2017
©
2017 American Chemical Society
10790
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
Article
design of the metal−support interaction as well as controllable
diffraction peaks corresponding to Fe or FeO were detected,
x
pyrolysis. Previously, we employed Mg(OH) instead of carbon
excluding the presence of any large crystalline particles of Fe-
containing species. In accordance with the XRD result, the
Raman spectra showed only two bands characteristic of carbon
2
as the sacrificial support for the preparation of self-supporting
Co−N−C SAC, which exhibited excellent performances in the
39
−1
−1
hydrogenative coupling of nitroarenes. By using aberration-
corrected HAADF-STEM, X-ray absorption, and DFT
calculations, we were able to identify, without much ambiguity,
the structure of the active site as CoN C -1−2O . Chen et al.
materials (G band at 1600 cm and D band at 1350 cm ,
43
Figure S2, SI) and no formation of any FeO species. The N2
x
adsorption−desorption isotherms and pore size distributions
revealed that the three samples all had large surface areas
4
8
2
2
reported atomically dispersed Fe−N on N and S codecorated
(620−835 m /g depending on the pyrolysis temperature) and
x
hierarchical carbon layers that showed good performances in
contained both micro- and mesopores (Figure S3, SI). By
scanning electron microscopy (SEM) under different detector
modes we could confirm that there were no visible particles
40
both OER and ORR reactions. Also aiming for the ORR, Zhu
et al. prepared Fe−N SAC by using a one-step hydrothermal
x
4
1
44
method. Nevertheless, in the above two examples and other
larger than 10 nm in the three samples (Figure S4, SI).
2
0,38
previous literature,
the exact structure of the active sites is
Nevertheless, from HAADF-STEM images one can clearly see
that small particles with sizes of 3−5 nm are sparsely dispersed
in the Fe−N−C-800 sample (Figure 1c inset and Figure S5a,b,
yet to be determined. Moreover, in comparison with the
extensively studied electrocatalysis, the atomically dispersed
M−N−C catalyst has rarely been explored for organic
transformations, especially for the challenging reaction of C−
H bond activation.
In this contribution, we report the preparation, character-
ization, and catalytic performances of atomically dispersed Fe−
N−C catalyst. With nano-MgO as a sacrificial template,
atomically dispersed self-supporting Fe−N−C catalyst was
successfully prepared, and its structure was identified by a
combinatorial use of aberration-corrected HAADF-STEM,
electron spin resonance (ESR), X-ray absorption spectroscopy
(
̈
XAS), and Mossbauer spectroscopy. To our surprise, even if
all the Fe atoms are atomically dispersed, there are four types of
different FeNx (x = 4−6) structures, and the relative
concentration is critically dependent on the pyrolysis temper-
ature. The most active site for the selective oxidation of the C−
III
H bond is the medium-spin Fe N structure, which is at least 1
5
order of magnitude more active than the high-spin and low-spin
III
II
Fe N structures and 3 times more active than the Fe N
6
4
structure. To the best of our knowledge, this is the first report
that different FeNx structures are well-discriminated and
correlated with the catalytic activity in the atomically dispersed
Fe−N−C material.
Figure 1. HAADF-STEM images of Fe−N−C-600 (a), Fe−N−C-700
RESULTS AND DISCUSSION
■
(
(
b), and Fe−N−C-800 (c) and elemental mapping of Fe−N−C-700
Preparation of Atomically Dispersed Fe−N−C Cata-
lyst. The Fe−N−C catalyst with atomically dispersed Fe was
prepared by a template-sacrificial approach [Scheme S1,
d). The inset in part c shows the Fe-containing particles highlighted
with red circles.
39
Supporting Information (SI)]. Briefly, the Fe(phen) (phen
x
=
1,10-phenanthroline) complex supported on nano-MgO
SI) while they are completely absent in the other two samples
(Figure S6, SI). From the image contrast we could assign the
very small particles in Fe−N−C-800 to the Fe-containing
particles. HRTEM images of Fe−N−C-800 (Figure S5c,d, SI)
further reveal that the Fe-containing particles are encapsulated
with carbon layers of 3−5 nm thickness, which protect them
from being leached by acid. The results demonstrate an
aggregation tendency of Fe atoms upon elevating the pyrolysis
temperature. Since no visible clusters/particles are observed in
either Fe−N−C-600 or Fe−N−C-700 with both SEM and
template was pyrolyzed at 600−800 °C under N atmosphere,
2
followed by acid leaching to remove the MgO template (for
more details, see the Supporting Information). The samples
pyrolyzed at different temperatures are denoted as Fe−N−C-X
(X = pyrolysis temperature). ICP-AES (inductively coupled
plasma atomic emission spectroscopy) analysis revealed Fe
loadings of 1.8, 1.6, and 1.4 wt % for the Fe−N−C-600, Fe−
N−C-700, and Fe−N−C-800 samples, respectively. A slight
decrease of the Fe loading with pyrolysis temperature suggested
that more metallic Fe particles are formed at higher pyrolysis
temperatures and they are subsequently removed by acid
leaching, leading to a decrease of the Fe loading.
Identification of Atomic Dispersion of Fe. The XRD
patterns of the samples showed only two broad peaks at 25°
and 43° (Figure S1, SI), which were assigned to reflections of
̊
̈
normal HAADF-STEM, we then resort to sub-angstrom-
resolution aberration-corrected HAADF-STEM to identify the
dispersion of Fe species. As shown in Figure 1, a high density of
Fe single atoms are uniformly dispersed in both Fe−N−C-600
and Fe−N−C-700, while both Fe single atoms and nano-
particles are observed in Fe−N−C-800. On the basis of the Fe
loading and surface area of each sample, we could roughly
estimate the density of Fe single atoms0.31, 0.21, and 0.20
42
the (002) and (004) planes of carbon. The carbon was
formed as a result of pyrolysis of 1,10-phenanthroline that was
not fully ligated to iron cations. It is noted that not any
2
Fe atoms per nm for Fe−N−C-600, Fe−N−C-700, and Fe−
1
0791
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
Article
N−C-800, respectivelywhich are in good agreement with
those estimated based on HAADF-STEM images. Moreover,
the elemental mapping of Fe−N−C-700 (Figure 1d) indicates
that the signals of C, N, O and Fe are superimposed on each
other, suggesting the homogeneous distribution of these
elements at the nanoscale.
The atomic dispersion of Fe−N−C catalysts is further
approved by EXAFS analysis. Figure 2a shows the Fourier-
Figure 3. XPS Fe 2p (a) and N 1s (b) spectra of Fe−N−C samples.
results of the three samples are similar, showing two peaks at
binding energies of 711.2 eV (Fe 2p3/2) and 724.5 eV (Fe
2
p1/2), respectively. These binding energy values and the
3+
45,46
associated shakeup structure point to Fe species.
The
0
absence of Fe in the XPS of Fe−N−C-800 appears to be
contradictive with the result of HAADF-STEM. Nevertheless, if
one considers the fact that the small particles of Fe are
encapsulated with thick carbon layers (Figure S5c,d, SI), it is
reasonable that these embedded Fe/Fe C particles are not
x
detectable by the XPS technique. Therefore, only the surface or
subsurface single-atom Fe species are detected by XPS, and
3+
they are assigned to Fe . The XPS spectra of N 1s in the three
samples can be deconvoluted into four peaks with binding
energies of 398.4, 399.8, 401.1, and 405.2 eV, which are
assigned to pyridinic N, pyrrolic N, graphitic N, and N-oxide
+
−
16,39
(
N −O ), respectively.
The relative concentrations of
different N species are listed in Table S1 (SI). According to
previous studies, in Fe−N−C materials the Fe atoms
coordinate with pyridinic N to form catalytically active Fe−
1
5,47,48
N sites.
For our three samples, the concentration of
x
pyridinic N species follows the order of Fe−N−C-600 > Fe−
N−C-700 > Fe−N−C-800, well agreeing with surface density
of Fe single atoms.
The predominance of Fe3+ in the three Fe−N−C samples is
corroborated by the XANES spectra at the Fe K-edge. As
shown in Figure 2b, the pre-edge peak at 7117 eV in FePc
(
navy line), which arises from the 1s → 4p transition with
z
simultaneous ligand-to-metal charge transfer, is regarded as the
2
fingerprint of square-planar Fe−N moieties (D symme-
4
4h
Figure 2. (a) The k -weighted Fourier transform spectra of Fe−N−C
22,23,39,49
try).
It is noted that a similar pre-edge peak also
catalyst as well as reference samples. (b) The normalized XANES
spectra at the Fe K-edge of different samples. The dotted ellipse shows
the pre-edge peak at 7117 eV.
appears in the three Fe−N−C samples, whereas it is absent
from the Fe(phen) precursor, suggesting that the Fe−N−C
x
samples contain a similar Fe−N planar structure, which is
4
quite different from the Co−N−C structure we previously
2
39
transformed k -weighted EXAFS spectra at the Fe K edge. In
reported. On the other hand, a comparison of E values (the
0
contrast to the reference samples of Fe foil and Fe O , neither
first inflection point) shows that Fe−N−C-600 has an E value
2
3
0
Fe−N−C-700 nor Fe−N−C-600 presents any prominent peak
at the position of Fe−Fe coordination, in agreement with the
above HAADF-STEM result that all the Fe species are
atomically dispersed without any aggregation. However, a
small peak of Fe−Fe coordination can be clearly observed in
Fe−N−C-800, suggesting the formation of Fe aggregates.
Moreover, all three Fe−N−C samples show a main peak at 1.47
Å, which is due to Fe−N coordination in the first shell,
of 7124.3 eV, which is close to that of the Fe O reference
3
4
2
+
3+
(Table S2, SI), suggesting the copresence of Fe and Fe in
this sample. In comparison, both Fe−N−C-700 and Fe−N−C-
800 have an E value of 7126.7 eV, very close to those of ferric
0
3+
samples, indicating the predominance of Fe in the two
samples. It is also interesting to note that the Fe(II)Pc
reference sample presents an E value similar to that of Fe−N−
0
2
+
C-600, which is probably caused by the partial oxidation of Fe
resembling the FePc and Fe(phen) reference samples.
to Fe3+ upon exposure to ambient atmosphere. This will be
validated by the Mossbauer spectroscopy below.
Mossbauer spectroscopy is a powerful technique to
discriminate different Fe species. Figure 4 shows the
Mossbauer spectra of the three Fe−N−C samples, and Table
1 lists the Mossbauer parameters as well as the relative areas of
x
Chemical State and Coordination Environment of Fe
Single Atoms. The chemical state of single-atom Fe as well as
its local environment is the key to the catalytic performances.
Therefore, we subsequently employed X-ray photoelectron
spectroscopy (XPS), X-ray absorption near edge structure
̈
̈
2
4
̈
̈
(
XANES), and Mo
̈
ssbauer spectroscopy to determine the exact
different Fe species. For both Fe−N−C-600 and Fe−N−C-700
samples, the spectra can be well-fitted with three doublets. The
chemical state of Fe. As shown in Figure 3, the XPS Fe 2p
1
0792
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
Article
Fe−N−C-700, while they may change into oxygen molecules
(
O ) in Fe−N−C-800 due to the breaking of axial Fe−N bonds
2
at higher temperatures and the weak adsorption of ambient
oxygen molecules on the central Fe.
23,54
In contrast to D2, D3
has a similarly low IS value (0.2−0.3) but a much larger QS
III
value (1.3−1.6) and can be assigned to the N−(Fe N )−N
4
24
low-spin structure, in which the central Fe atom is strongly
coordinated with four nitrogen atoms in graphitic layers and the
other two nitrogen atoms in the axial direction perpendicular to
the Fe−N4 plane. Such a perfect saturated coordination
structure is robust and therefore cannot be acting as the
24,55
catalytically active site.
On the other hand, the D4 species,
which featured the smallest IS value (0.12−0.13) and the
largest QS value (2.8−3.3) among the four doublets, can be
Figure 4. 57Fe Mo
samples.
ssbauer spectra of the Fe−N−C-600/700/800
̈
III
assigned to N−(Fe N ) medium-spin species with a penta-
4
58,59
coordinated rhombic monopyramidal structure.
Its smaller
IS value and larger QS value than the Fe−N structure in FePc
4
̈
Table 1. Summary of the Mossbauer Parameters and
(
Figure S8, Table S3, SI) is caused by the additional N
Assignments to Different Iron Species in Fe−N−C-600/
coordinating with the central Fe in the apical position. The
unsaturated coordination structure of D4 will enable it to be
catalytically active in chemical reactions. This will be validated
in the subsequent catalytic tests.
700/800 Catalysts
IS/mm s‑ QS/mm s‑1 area %
1
a
assignment
b
Fe species
II
6
00-D1
0.61
0.21
2.62
0.82
14.2
66.3
Fe N , like FePc, MS
4
III
It is also noted that the relative amount of each Fe species is
6
00-D2
X−Fe N −Y, (X, Y
4
ligands like O/N), HS
critically dependent on the pyrolysis temperature (Table 1). D1
III
II
6
00-D3
00-D2
0.31
0.25
1.60
0.78
19.5
28.3
N−(Fe N )−N, LS
(Fe N ) only exists in the Fe−N−C-600 sample, suggesting
4
4
III
7
X−Fe N −Y, (X, Y
that this structure is not stable at higher temperatures. In
4
ligands like O/N), HS
III
comparison, D2 (high-spin X−Fe N −Y) exists in all the three
4
III
7
7
8
00-D3
00-D4
00-D2
0.24
0.13
0.19
1.37
2.81
0.71
53.8
17.9
59.0
N−(Fe N )−N, LS
N−(Fe N ), MS
4
samples and even becomes the predominant Fe species in both
Fe−N−C-600 and Fe−N−C-800 samples, while it appears to
be the least populated species in Fe−N−C-700. The reason for
such an irregular change with pyrolysis temperature is not clear
III
4
III
X−Fe N −Y, (X, Y
4
ligands like: O/N), HS
III
8
8
8
00-D4
0.12
−0.12
0.13
3.28
7.8
10.0
19.5
N−(Fe N ), MS
γ-Fe
4
yet, but the decreased QS value with increasing pyrolysis
00-Singlet
00-Sext
III
temperature indicates that the X−Fe N −Y structure becomes
0
Fe C (H = 27.9 T)
x
4
0
a
more symmetric. On the other hand, the concentration of D3
The relative absorption area of each iron species in Fe−N−C
III
b
[low-spin N−(Fe N )−N] increases with the pyrolysis
4
samples. LS, MS, and HS denote low-spin, medium-spin, and high-
temperature, until it completely disappears in Fe−N−C-800;
spin, respectively.
concurrently, nanoparticles of γ-Fe and Fe C are formed in Fe−
x
N−C-800. This result seems to suggest that the D3 structure is
absence of sextet and singlet indicates the absence of Fe0
species. By contrast, in the Fe−N−C-800 sample, besides two
doublets, one singlet (S1) and one sextet (Sext1) correspond-
0
destroyed and then aggregation of Fe occurs as a consequence
of pyrolysis at higher temperatures (800 °C or above). It is also
interesting to note that the concentration of D4 [medium-spin
5
0
III
ing to γ-Fe and Fe C species appear in the spectrum. This
N−(Fe N
4
)] gets maximized in Fe−N−C-700, which might
x
result is well-consistent with the HAADF-STEM and EXAFS
bring about activity enhancement.
results that both Fe−N−C-600 and Fe−N−C-700 exclusively
Catalytic Performance of Fe−N−C for Selective
Oxidation of the C−H Bond. The selective oxidation of
the C−H bond is one of the most important organic
transformations for construction of high-value-added carbonyl
compounds, such as aldehydes, ketones, and carboxylic acid and
esters, which are widely used as key intermediates in the fine
0
consist of Fe single atoms whereas aggregates of Fe particles
are present in Fe−N−C-800.
According to the isomer shift (IS) and quadrupole splitting
(
QS) values (Table 1), four different doublets (D1−D4) can be
identified in the above three samples. D1, which has relatively
larger values of IS and QS and appears only in the Fe−N−C-
60,61
chemical industry.
However, due to the high dissociation
II
51−53
6
00 sample, can be assigned to FePc-like Fe N species.
energy of the C−H bond and the overoxidation side reactions,
4
D2, with an IS value of ∼0.2 and a QS value of 0.7−0.8, appears
the direct oxidation of the C−H bond has been a long-standing
in all the three samples and can be assigned to ferric X−
challenge accompanied by low conversion and poor selectiv-
III
62−65
Fe N −Y high-spin species (X and Y refer to O or N ligand),
ity.
Considering that the Fe−N−C catalysts possess the
4
in which the Fe atom is slightly pulled out of the N plane,
resulting in a deformed octahedral structure.
unique feature of single-atom dispersion and high oxidation
state of Fe, we expect them to be efficient catalyst in the
selective oxidation reactions.
4
54,55
This assign-
ment is further confirmed by EPR (electron paramagnetic
resonance) characterization results. As shown in Figure S7 (SI),
an isotropic signal at g = 4.24 is clearly observed, which can be
assigned to high-spin ferric species in a rhombic symme-
To study the catalytic performances of the atomically
dispersed Fe−N−C catalysts for the selective oxidation of
C−H bond, ethylbenzene was first chosen as a substrate to
explore the efficiency. First, in a blank experiment, 1,10-
5
2,56,57
try.
It is also noted that the apical atoms (X and Y) may
differ with the calcination temperature. For example, both X
and Y are most probably nitrogen atoms in Fe−N−C-600 and
phenanthroline and nonpyrolyzed Fe(phen) −MgO precursor
x
were tested; however, only negligible conversion of ethyl-
1
0793
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
Article
a
Table 2. Catalytic Activity of Different M−N-C Samples for the Oxidation of Ethylbenzene in Water
selectivity (%)
2c
entry
catalyst
time (h)
conv (%)
2a
2b
2d
1
2
3
4
5
6
7
none
phen
8
8
8
5
5
5
7
7
7
0.7
1.1
1.5
94
77
55
71
96
97
94
99
95
85
0
23
32
20
0
0
b
13
Fe(phen) −MgO
8.6
0.7
0.4
2.1
0.3
0.3
1.7
trace
0
x
Fe−N−C-600
Fe−N−C-700
Fe−N−C-800
Fe−N−C-700
Fe−N−C-700
Co−N−C-700
2.6
98
2.5
4.5
0.3
4.7
0
81
0
99
0
c
8
99
0
9
74
13
0
a
Reaction conditions: 10.0 mg of catalyst, 0.5 mmol of ethylbenzene, room temperature, 500 μL of TBHP as a 70 wt % aqueous solution (7 equiv)
diluted with 6.5 mL of H O. The conversion and selectivity were determined by GC−MS and GC analysis with dodecane as an internal standard.
2
b
c
phen = 1,10-phenanthroline. Two equivalents of TBHP (1 mmol) was used.
benzene was observed (Table 2, entries 1−3). In remarkable
contrast, the three Fe−N−C catalysts afforded high activity and
excellent selectivity; among them, Fe−N−C-700 gave the best
performance with 98% ethylbenzene conversion and 97%
acetophenone selectivity (Table 2, entries 4−6). When the
reaction time was prolonged to 7 h, both the ethylbenzene
conversion and the acetophenone selectivity reached 99% at the
ambient temperature (Table 2, entry 7). More interesting, even
when the tert-butyl hydroperoxide (TBHP) was reduced to a
stoichiometric amount (2 equiv), the conversion of ethyl-
benzene could still reach 99% with an acetophenone selectivity
of 95% (entry 8), indicating a nearly 100% TBHP efficiency. It
is also noted that, under identical conditions, the atomically
dispersed Co−N−C catalyst we previously developed gave only
moderate conversion and selectivity (Table 2, entry 9),
demonstrating the superiority of the Fe−N−C single-atom
catalysts. Indeed, when compared with reports in the literature,
it is found that our Fe−N−C-700 catalyst has the highest
activity at room temperature (Table S4, SI).
Table 3. Selective Oxidation of Hydrocarbons in Water with
Fe−N−C-700 as Catalyst
a
On the basis of the conversion−time profiles (Figure S9, SI)
and the total Fe contents of the three Fe−N−C catalysts, the
initial turnover frequencies (TOFs) were calculated to be 682.5,
−
1
1
932, and 850.5 h for the Fe−N−C-600, Fe−N−C-700, and
Fe−N−C-800 catalysts, respectively. Evidently, the pyrolysis at
different temperatures resulted in different Fe-centered
5
7
structures, as identified with Fe Mo
and Table 1), which consequently led to remarkable
difference in intrinsic activities. The Fe−N−C-700 sample is
−3 times more active than the other two samples because it
contains a greater number of unsaturated coordination sites
̈
ssbauer spectra (Figure
4
2
(
Fe−N ). This will be discussed in the following section.
5
The excellent performance of Fe−N−C-700 catalyst was also
a
Reaction conditions: 10.0 mg of catalyst, 0.5 mmol of substrates,
tested for the selective oxidation of C−H bonds of a broad
scope of substrates at room temperature. As shown in Table 3,
various aromatic hydrocarbons with an electron-donating group
room temperature for 7 h, 500 μL of TBHP as a 70 wt % aqueous
solution (7 equiv) diluted with 6.5 mL of H O. Conversion and
2
selectivity are determined by GC−MS and GC analysis with dodecane
(
−OMe) or electron-withdrawing group (−NO ) were trans-
2
as an internal standard.
formed smoothly into the corresponding ketone compounds
with >98% selectivities at high conversions (Table 3, entries 2
and 3). Bulky and sterically hindered substrates, such as
cumene, diphenylmethane, and fluorene, could also be
smoothly activated by the atomically dispersed Fe−N−C-700
catalyst with high conversion and selectivity (Table 3, entries
4−7). Moreover, the heterocyclic substrate could also be
tolerated, and the corresponding product was obtained with
excellent selectivity (Table 3, entry 8). Notably, the selective
oxidation of aliphatic hydrocarbon was also tested. For example,
1
0794
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
2% conversion of cyclohexane was obtained after 7 h reaction
at room temperature, with 99% selectivity to cyclohexone and
cyclohexanol (Table 3, entry 9).
Article
5
7
6
On the other hand, based on the Fe Mo
̈
ssbauer
spectroscopy results (Figure 4 and Table 1), there are four
types of Fe−N species (D1−D4) depending on the pyrolysis
x
III
Understanding the Origin of Catalytic Activity. It has
temperature. We assume that D3 species with a N−(Fe N )−
4
been reported that heteroatom-doped carbon materials could
N low-spin structure is catalytically inactive due to the saturated
and strong coordination of central Fe with six N atoms, and in
this case, we can calculate the intrinsic activities of the other
three Fe species (D1, D2, and D4) on the basis of the TOFs
6
4,66
catalyze the oxidation of the C−H bond.
In our Fe−N−C
catalysts, there are also such N-doped carbon sites in addition
to Fe−N sites, as shown by the high N/Fe atom ratios in XPS
x
result (N/Fe = 30−60). To discriminate the role played by the
̈
per Fe atom as well as their relative areas in Mossbauer spectra
N-doped carbon and the Fe−N sites, we conducted titration
(for calculation details, see the Supporting Information). The
x
result shows that the TOFs are 2063, 588, and 6455 h− for D1,
1
experiments by using KSCN to poison the Fe−N sites,
x
−
because SCN could form a stable chelate complex with Fe
D2, and D4 sites, respectively. Clearly, as we expect, the D4
species with a N−(Fe N ) medium-spin structure is the most
catalytically active species, and it is 1 order of magnitude more
active than the high-spin 6-fold D2 species and 3 times more
active than the D1 FePc-like Fe N species. This result is in
1
5
III
cations. The titration experiment was performed with Fe−
N−C-700 catalyst within a kinetically controlled region, that is,
the catalyst amount was reduced to 2 mg while the reaction
time was shortened to 5 min. As shown in Figure 5, with an
4
II
4
good agreement with the KSCN titration result, where the first
poisoned species should be D4 with an estimated TOF of 5350
−
1
−1
h , followed by the second (D2, 533 h ) and third (D3, 160
−1
h ). The slight difference in the exact TOF values should be
caused by the fitting errors in both methods, but the trend is
well-consistent. On the other hand, to approve the catalytic
activity of D1 species, we tested FePc, which has a similar
structure as D1, for the same reaction. As expected, the FePc is
active for the reaction and the TOF based on Fe content is 373
−1
h . The lower activity of FePc than the D1 species is probably
due to the poor accessibility of FePc.
In summary, both the KSCN titration method and the
̈
Mossbauer spectroscopy method demonstrate independently
that there are different Fe−N sites on the Fe−N−C catalysts,
x
III
and the D4 site with a N−(Fe N ) medium-spin structure is
4
the most catalytically active species. For the best-performing
catalyst Fe−N−C-700, the intrinsic activity follows the order of
D4 ≫ D2 ≫ D3; in particular, D4 is 33 times more active than
D3 and 1 order of magnitude more active than D2 on the basis
of the KSCN titration method, confirming the reasonability of
Figure 5. Titration experiment of active sites in Fe−N−C-700 with
KSCN. Before the reaction, a specified equivalent of KSCN aqueous
solution (1 mg/mL) was injected into the reaction system. Reaction
conditions: 2 mg of Fe−N−C-700 catalyst (Fe 1.6 wt %, containing
̈
our assumption in the Mossbauer spectroscopy method that D3
0.0006 mmol of Fe), 0.5 mmol of substrates, room temperature, 500
is inactive due to the saturated and strong coordination of the
central Fe with six N atoms. Meanwhile, it is noted that the
most active D4 site is unfortunately the least abundant in the
catalysts (Table 1), suggesting that there is much room to
enhance the activity by increasing the number of D4 species.
To further shed light on the reaction mechanism, DMPO
(5,5-dimethyl-1-pyrroline N-oxide) spin-trapping EPR experi-
ments were conducted. As shown in Figure 6, the •OH radical
μL of TBHP as a 70 wt % aqueous solution diluted with 6.5 mL of
H O, 5 min.
2
−
increase of the SCN amount that was introduced into the
reaction system, one can clearly observe a three-stage decrease
in the ethylbenzene conversion. At the first stage, ethylbenzene
conversion drastically decreased from 14.5% to 3.8% upon
(a = a = 15.1 G), alkyl free radical (a = 16.3, a = 23 G),
N H N H
−
introducing 0.2 equiv of SCN . At the second stage,
oxidized DMPO radical, and t-Bu-OO• radical are all
67
ethylbenzene conversion decreased less fast, from 3.8% to
detected; meanwhile, t-BuOH, 1-phenylethyl alcohol, and
−
2
.2% after introducing an additional 0.3 equiv of SCN . At the
benzaldehyde are detected as intermediates of the reaction by
third stage, the conversion decreased at a much slower rate,
GC−MS analysis. On the basis of these experimental results
−
68,69
until it remained constant (1.4%) at 1.0 equiv of SCN . Such a
and the reports in the literature,
a plausible mechanism is
nonlinear decrease behavior strongly suggests that there are
three types of active sites on the catalyst surface, which is in
proposed, as shown in Figure 7. First, TBHP is adsorbed and
activated at Fe−N sites via the interaction between the lone
5
good agreement with the Mo
D3, and D4, Table 1). On the basis of the amount of SCN , we
̈
ssbauer spectroscopy results (D2,
pair of electrons of the distant oxygen in TBHP and the
electron-deficient Fe(III) center as well as the hydrogen
bonding between N and H atoms, leading to homolytic
cleavage of TBHP to produce tert-butyl oxygen radical (t-
BuO•) and hydroxyl radical (•OH). At the same time,
ethylbenzene can be adsorbed on N/O-containing functional
−
can calculate the intrinsic activity (TOFs) of different sites: the
−1
most active site has a turnover frequency (TOF) of 5350 h ,
−1
while the less active site has a TOF of 533 h , and the least
−1
active one has a TOF of only 160 h (for calculation details,
see the Supporting Information). Moreover, the titration result
indicates that the contribution of Fe-free sites accounts for only
groups adjacent to the Fe−N active site via hydrogen bonding
x
69
(e.g., in the form of O/N···H). The presence of such
functional groups on the catalyst surface is revealed by the N 1s
(Figure 3b) and O 1s (Figure S10, SI) XPS spectra. Then, the t-
9
.7%. Therefore, Fe−N sites should play a predominant role in
x
catalysis.
1
0795
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
Article
is composed of a central Fe ion coordinating with four N atoms
in the square plane and with the fifth N atom in the imidazole
ring of histidine residue in the axial direction (Figure S11, SI).
Such a Fe−N under-coordinated moiety can reversely bind
5
oxygen and thus can transport O for our life activities. In this
2
regard, our present atomically dispersed Fe−N structure acts
x
as a mimic of the enzyme in organic transformations,
demonstrating the great potential of a single-atom catalyst in
bridging heterogeneous and homogeneous catalysis.
Stability of the Fe−N Structure. The main advantage of
x
heterogeneous catalyst lies in its good stability, recyclability and
ease of separation. For this purpose, the stability and
recyclability of the Fe−N−C-700 catalyst was investigated. As
shown in Figure 8a, the catalyst could be reused at least five
Figure 6. DMPO (5,5-dimethyl-1-pyrroline N-oxide) spin-trapping
EPR experiment. The spectrum was recorded after 1 min upon
introducing DMPO to the reaction system. Ethylbenzene was used as
the substrate.
Figure 8. (a) Reusability of the Fe−N−C-700 catalyst. (b) HAADF-
STEM image of the Fe−N−C-700 catalyst after five cycles of reaction.
(
c) The 57Fe Mo
ssbauer spectra of Fe−N−C-700 before and after
̈
different cycles of reaction. (d) The relative absorption area of each
iron species in the Fe−N−C-700 catalyst before and after different
cycles of reaction.
times without decay in either activity or selectivity. The ICP
analysis of the reaction medium after five runs shows no loss of
Fe. The BET surface areas remain almost the same before and
2
after the reaction (835 vs 818 cm /g). Moreover, XRD (Figure
S12, SI), XAS (Figure S13, SI), and HAADF-STEM (Figure
Figure 7. Proposed reaction mechanism of ethylbenzene oxidation on
the Fe−N−C-700 catalyst.
8
b) characterizations show that the atomic dispersion of Fe
species is well-preserved after the five repetitive runs. To probe
BuO• abstracts one α-H of ethylbenzene to yield the
corresponding α-ethylbenzene radical, which reacts quickly
with •OH to produce 1-phenylethyl alcohol. Subsequently, t-
BuO• captures H from the hydroxyl group of 1-phenylethyl
alcohol to form 1-ethylbenzene oxygen radical, which then
reacts with hydroxyl radical (•OH) to form the 1-hydro-
peroxyethylbenzene intermediate. Finally, acetophenone as the
desired product and benzaldehyde as a byproduct were
produced via carbonyl-forming elimination of 1-hydroperox-
yethylbenzene. Among these elementary reaction steps, the
dissociation of the C−H bond of methylene in ethylbenzene
any changes of the Fe−N sites after different cycles of reaction,
x
̈
we further performed Mossbauer spectroscopy character-
izations for the three used catalysts (after the first, third, and
fifth reaction cycles). As shown in Figure 8c, the Mossbauer
̈
spectra of all three used catalysts show similar peak profiles to
that of the fresh catalyst, which can be well-fitted with three
III
doublets assignable to the D2 site (ferric X−Fe N −Y high-
4
III
spin species), D3 site (ferric N−Fe N −N low-spin species),
4
III
and D4 site (ferric N−Fe N medium spin species),
4
respectively. Further quantification reveals that the relative
content of each Fe−N species remains the same before and
x
(
step 3 in Figure 7) is believed to be the rate-determining step
after the reaction (Figure 8d), indicating that all the Fe−N
x
due to the high bond energy of α-H in ethylbenzene (412 kJ
sites (including D2, D3, and D4) are rather stable during cycles
of reaction. These results unambiguously demonstrate the
excellent stability and recycling ability of the single-atom
dispersed Fe−N−C catalyst.
−1
mol ).
The above reaction model is reminiscent of the activation of
70
oxygen by the hemoglobin molecule, in which the active site
1
0796
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
Article
CONCLUSIONS
REFERENCES
■
■
By using nano-MgO as a sacrificial template we have
successfully prepared atomically dispersed Fe−N−C catalysts
that exhibited promising activity, selectivity, and stability for the
selective oxidation of the C−H bond of a broad scope of
substrates at room temperature. Mossbauer spectroscopy and
̈
KSCN titration experiments reveal that there are four different
(1) Grirrane, A.; Corma, A.; García, H. Science 2008, 322, 1661−
664.
2) Lin, J.; Wang, A.; Qiao, B.; Liu, X.; Yang, X.; Wang, X.; Liang, J.;
Li, J.; Liu, J.; Zhang, T. J. Am. Chem. Soc. 2013, 135, 15314−15317.
3) Kang, J.; Cheng, K.; Zhang, L.; Zhang, Q.; Ding, J.; Hua, W.; Lou,
1
(
(
Y.; Zhai, Q.; Wang, Y. Angew. Chem., Int. Ed. 2011, 50, 5200−5203.
(
(
4) Chen, J. G. Chem. Rev. 1996, 96, 1477−1498.
FeN species (x = 4−6) in the atomically dispersed Fe−N−C
x
5) Tran, P. D.; Morozan, A.; Archambault, S.; Heidkamp, J.;
catalysts, the relative concentrations of which are critically
dependent on the pyrolysis temperature. Among them, the
most active Fe−N−C-700 catalyst is comprised of high-spin
FeN (28.3%), low-spin FeN (53.8%), and medium-spin FeN
5
Chenevier, P.; Dau, H.; Fontecave, M.; Martinent, A.; Jousselme, B.;
Artero, V. Chem. Sci. 2015, 6, 2050−2053.
(6) Ren, Y.; Wei, H.; Yin, G.; Zhang, L.; Wang, A.; Zhang, T. Chem.
Commun. 2017, 53, 1969−1972.
6
6
(
17.9%) species, where the FeN5 is at least 1 order of
(7) Sun, J.; Cai, Q.; Wan, Y.; Wan, S.; Wang, L.; Lin, J.; Mei, D.;
Wang, Y. ACS Catal. 2016, 6, 5771−5785.
magnitude more active than the other two species, although it
accounts for only 28.3% of the total Fe species. Pyrolysis at a
higher temperature (Fe−N−C-800) resulted in a further
(
4
(
8) Lin, L.; Yang, Z. K.; Jiang, Y.-F.; Xu, A.-W. ACS Catal. 2016, 6,
449−4454.
9) Wang, Y. C.; Lai, Y. J.; Song, L.; Zhou, Z. Y.; Liu, J. G.; Wang, Q.;
decrease in the concentration of medium-spin FeN species
5
Yang, X. D.; Chen, C.; Shi, W.; Zheng, Y. P.; Rauf, M.; Sun, S. G.
Angew. Chem., Int. Ed. 2015, 54, 9907−9910.
10) Wu, G.; More, K. L.; Johnston, C. M.; Zelenay, P. Science 2011,
32, 443−447.
11) Zhao, Y.; Watanabe, K.; Hashimoto, K. J. Am. Chem. Soc. 2012,
34, 19528−19531.
(12) Huan, T. N.; Ranjbar, N.; Rousse, G.; Sougrati, M.; Zitolo, A.;
Mougel, V.; Jaouen, F.; Fontecave, M. ACS Catal. 2017, 7, 1520−
525.
13) Zeng, M.; Liu, Y.; Zhao, F.; Nie, K.; Han, N.; Wang, X.; Huang,
W.; Song, X.; Zhong, J.; Li, Y. Adv. Funct. Mater. 2016, 26, 4397−4404.
(14) Jin, H.; Wang, J.; Su, D.; Wei, Z.; Pang, Z.; Wang, Y. J. Am.
Chem. Soc. 2015, 137, 2688−2694.
15) Liang, H. W.; Bruller, S.; Dong, R.; Zhang, J.; Feng, X.; Mullen,
K. Nat. Commun. 2015, 6, 7992.
(16) Fei, H.; Dong, J.; Arellano-Jimenez, M. J.; Ye, G.; Dong Kim, N.;
Samuel, E. L.; Peng, Z.; Zhu, Z.; Qin, F.; Bao, J.; Yacaman, M. J.;
Ajayan, P. M.; Chen, D.; Tour, J. M. Nat. Commun. 2015, 6, 8668.
17) Jagadeesh, R. V.; Surkus, A. E.; Junge, H.; Pohl, M. M.; Radnik,
J.; Rabeah, J.; Huan, H.; Schunemann, V.; Bruckner, A.; Beller, M.
to less than 10%, thus leading to significant loss in the activity
based on total Fe atoms. This work demonstrates clearly the
heterogeneity of the Fe−N−C catalysts, although all the Fe
species are atomically dispersed, and more importantly,
provides a guide to fabricate single-atom Fe−N−C catalyst
(
3
(
1
with more medium-spin Fe−N sites.
5
1
(
ASSOCIATED CONTENT
■
*
S
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.7b05130.
(
Details of the catalyst preparation and characterization
results (PDF)
AUTHOR INFORMATION
(
■
*
*
Corresponding Authors
aqwang@dicp.ac.cn
taozhang@dicp.ac.cn
Science 2013, 342, 1073−1076.
18) Jagadeesh, R. V.; Stemmler, T.; Surkus, A.-E.; Bauer, M.; Pohl,
M.-M.; Radnik, J.; Junge, K.; Junge, H.; Bru
Protoc. 2015, 10, 916−926.
19) Westerhaus, F. A.; Jagadeesh, R. V.; Wienhofer, G.; Pohl, M. M.;
Radnik, J.; Surkus, A. E.; Rabeah, J.; Junge, K.; Junge, H.; Nielsen, M.;
Bruckner, A.; Beller, M. Nat. Chem. 2013, 5, 537−543.
20) Deng, D.; Chen, X.; Yu, L.; Wu, X.; Liu, Q.; Liu, Y.; Yang, H.;
(
̈
ckner, A.; Beller, M. Nat.
ORCID
(
Xiaoyan Liu: 0000-0003-2694-2306
Aiqin Wang: 0000-0003-4552-0360
Tao Zhang: 0000-0001-9470-7215
(
Tian, H.; Hu, Y.; Du, P.; Si, R.; Wang, J.; Cui, X.; Li, H.; Xiao, J.; Xu,
T.; Deng, J.; Yang, F.; Duchesne, P. N.; Zhang, P.; Zhou, J.; Sun, L.; Li,
J.; Pan, X.; Bao, X. Sci. Adv. 2015, 1, e1500462.
Author Contributions
W.L. and L.Z. contributed equally to this work.
†
Notes
(
21) Zhang, L.; Wang, A.; Wang, W.; Huang, Y.; Liu, X.; Miao, S.;
Liu, J.; Zhang, T. ACS Catal. 2015, 5, 6563−6572.
22) Ramaswamy, N.; Tylus, U.; Jia, Q.; Mukerjee, S. J. Am. Chem.
The authors declare no competing financial interest.
(
Soc. 2013, 135, 15443−15449.
ACKNOWLEDGMENTS
■
(23) Zitolo, A.; Goellner, V.; Armel, V.; Sougrati, M. T.; Mineva, T.;
Stievano, L.; Fonda, E.; Jaouen, F. Nat. Mater. 2015, 14, 937−942.
(24) Zhu, Y.; Zhang, B.; Liu, X.; Wang, D. W.; Su, D. S. Angew.
Chem., Int. Ed. 2014, 53, 10673−10677.
The authors are grateful for the support from the National
Natural Science Foundation of China (21690080, 21690084,
1373206, 21522608, 21503219, 21672210, and 21673228),
2
(25) Sa, Y. J.; Seo, D. J.; Woo, J.; Lim, J. T.; Cheon, J. Y.; Yang, S. Y.;
the Strategic Priority Research Program of the Chinese
Academy of Sciences (XDB17020100), the National Key
Projects for Fundamental Research and Development of
China (2016YFA0202801), and the BL 14W beamline at the
Shanghai Synchrotron Radiation Facility (SSRF). We also
thank Prof. Hongxian Han from Dalian Institute of Chemical
Physics and Prof. Jihu Su from The University of Science and
Technology of China, for their help in EPR analysis, and Prof.
Lee, J. M.; Kang, D.; Shin, T. J.; Shin, H. S.; Jeong, H. Y.; Kim, C. S.;
Kim, M. G.; Kim, T. Y.; Joo, S. H. J. Am. Chem. Soc. 2016, 138,
1
(
5046−15056.
26) Xie, J.; Yin, K.; Serov, A.; Artyushkova, K.; Pham, H. N.; Sang,
X.; Unocic, R. R.; Atanassov, P.; Datye, A. K.; Davis, R. J.
ChemSusChem 2017, 10, 359−362.
(27) Wu, Z. Y.; Xu, X. X.; Hu, B. C.; Liang, H. W.; Lin, Y.; Chen, L.
F.; Yu, S. H. Angew. Chem., Int. Ed. 2015, 54, 8179−8183.
(28) Zhang, B.; Asakura, H.; Zhang, J.; Zhang, J. G.; De, S.; Yan, N.
Angew. Chem., Int. Ed. 2016, 55, 8319−8323.
Junhu Wang from Mo
Institute of Chemical Physics.
̈
ssbauer Effect Data Center of Dalian
1
0797
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798

Journal of the American Chemical Society
Article
(
29) Liu, P. X.; Zhao, Y.; Qin, R. X.; Mo, S. G.; Chen, G. X.; Gu, L.;
(56) Zang, C.; Liu, Y. R.; Xu, Z. J.; Tse, C.-W.; Guan, X. G.; Wei, J.
H.; Huang, J. S.; Che, C. M. Angew. Chem., Int. Ed. 2016, 55, 10253−
10257.
Chevrier, D. M.; Zhang, P.; Guo, Q.; Zang, D. D.; Wu, B. H.; Fu, G.;
Zheng, N. F. Science 2016, 352, 797−800.
(
Zeng, R.; Li, A.; Li, H.; Wang, M.; Zheng, X.; Zhu, J.; Zhang, W.; Ma,
C.; Si, R.; Zeng, J. Nat. Commun. 2016, 7, 14036.
30) Wang, L.; Zhang, W.; Wang, S.; Gao, Z.; Luo, Z.; Wang, X.;
(57) Christoforidis, K. C.; Pantazis, D. A.; Bonilla, L. L.; Bletsa, E.;
Louloudi, M.; Deligiannakis, Y. J. Catal. 2016, 344, 768−777.
(58) Kramm, U. I.; Herranz, J.; Larouche, N.; Arruda, T. M.; Lefevre,
M.; Jaouen, F.; Bogdanoff, P.; Fiechter, S.; Abs-Wurmbach, I.;
Mukerjee, S.; Dodelet, J. P. Phys. Chem. Chem. Phys. 2012, 14,
(
31) Yang, M.; Li, S.; Wang, Y.; Herron, J. A.; Xu, Y.; Allard, L. F.;
Lee, S.; Huang, J.; Mavrikakis, M.; Flytzani-Stephanopoulos, M. Science
014, 346, 1498−1501.
32) Wei, H.; Liu, X.; Wang, A.; Zhang, L.; Qiao, B.; Yang, X.;
Huang, Y.; Miao, S.; Liu, J.; Zhang, T. Nat. Commun. 2014, 5, 5634.
33) Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu,
J.; Li, J.; Zhang, T. Nat. Chem. 2011, 3, 634−641.
34) Yang, X.; Wang, A.; Qiao, B.; Li, J.; Liu, J.; Zhang, T. Acc. Chem.
Res. 2013, 46, 1740−1748.
35) Li, Z.; Schweitzer, N. M.; League, A. B.; Bernales, V.; Peters, A.
1
(
1673−11688.
2
(
59) Jia, Q. Y.; Ramaswamy, N.; Hafiz, H.; Tylus, U.; Strickland, K.;
Wu, G.; Barbiellini, B.; Bansil, A.; Holby, E. F.; Zelenay, P.; Mukerjee,
S. ACS Nano 2015, 9, 12496−12505.
(
(
(
60) Song, G.; Wang, F.; Li, X. Chem. Soc. Rev. 2012, 41, 3651−3678.
61) Wang, X.; Leow, D.; Yu, J. Q. J. Am. Chem. Soc. 2011, 133,
(
1
3864−13867.
62) Garcia-Bosch, I. G.; Siegler, M. A. Angew. Chem., Int. Ed. 2016,
5, 12873−12876.
63) Gao, Y.; Tang, P.; Zhou, H.; Zhang, W.; Yang, H.; Yan, N.; Hu,
G.; Mei, D.; Wang, J.; Ma, D. Angew. Chem., Int. Ed. 2016, 55, 3124−
128.
64) Yang, S.; Peng, L.; Huang, P.; Wang, X.; Sun, Y.; Cao, C.; Song,
W. Angew. Chem., Int. Ed. 2016, 55, 4016−4020.
65) Wang, L.; Zhu, Y.; Wang, J. Q.; Liu, F.; Huang, J.; Meng, X.;
Basset, J. M.; Han, Y.; Xiao, F. S. Nat. Commun. 2015, 6, 6957.
66) Wang, Y.; Zhang, J.; Wang, X.; Antonietti, M.; Li, H. Angew.
Chem., Int. Ed. 2010, 49, 3356−3359.
67) Zhu, K.; Wang, J.; Wang, Y.; Jin, C.; Ganeshraja, A. S. Catal. Sci.
Technol. 2016, 6, 2296−2304.
68) Zhang, J.; Su, D.; Zhang, A.; Wang, D.; Schlogl, R.; Hebert, C.
Angew. Chem., Int. Ed. 2007, 46, 7319−7323.
69) Mestl, G.; Maksimova, N. I.; Keller, N.; Roddatis, V. V.; Schlo
R. Angew. Chem., Int. Ed. 2001, 40, 2066−2068.
70) Sahu, S.; Goldberg, D. P. J. Am. Chem. Soc. 2016, 138, 11410−
1428.
(
5
(
(
W.; Getsoian, A. B.; Wang, T. C.; Miller, J. T.; Vjunov, A.; Fulton, J. L.;
Lercher, J. A.; Cramer, C. J.; Gagliardi, L.; Hupp, J. T.; Farha, O. K. J.
Am. Chem. Soc. 2016, 138, 1977−1982.
3
(
(
36) Matsubu, J. C.; Yang, V. N.; Christopher, P. J. Am. Chem. Soc.
015, 137, 3076−3084.
37) Lin, L.; Zhou, W.; Gao, R.; Yao, S.; Zhang, X.; Xu, W.; Zheng,
S.; Jiang, Z.; Yu, Q.; Li, Y. W.; Shi, C.; Wen, X. D.; Ma, D. Nature
017, 544, 80−83.
38) Bayatsarmadi, B.; Zheng, Y.; Vasileff, A.; Qiao, S. Z. Small 2017,
3, 1700191.
39) Liu, W.; Zhang, L.; Yan, W.; Liu, X.; Yang, X.; Miao, S.; Wang,
W.; Wang, A.; Zhang, T. Chem. Sci. 2016, 7, 5758−5764.
40) Chen, P.; Zhou, T.; Xing, L.; Xu, K.; Tong, Y.; Xie, H.; Zhang,
L.; Yan, W.; Chu, W.; Wu, C.; Xie, Y. Angew. Chem., Int. Ed. 2017, 56,
10−614.
41) Zhu, C.; Fu, S.; Song, J.; Shi, Q.; Su, D.; Engelhard, M. H.; Li,
X.; Xiao, D.; Li, D.; Estevez, L.; Du, D.; Lin, Y. Small 2017, 13,
603407.
42) Fan, L.; Liu, P. F.; Yan, X.; Gu, L.; Yang, Z. Z.; Yang, H. G.; Qiu,
S.; Yao, X. Nat. Commun. 2016, 7, 10667.
43) Wu, G.; Johnston, C. M.; Mack, N. H.; Artyushkova, K.;
2
(
(
2
(
1
(
(
(
(
(
(
̈
gl,
6
(
(
1
1
(
(
Ferrandon, M.; Nelson, M.; Lezama-Pacheco, J. S.; Conradson, S. D.;
More, K. L.; Myers, D. J.; Zelenay, P. J. Mater. Chem. 2011, 21, 11392.
(
(
44) Liu, J. Microsc. Microanal. 2000, 6, 388−399.
45) Chen, G. X.; Zhao, Y.; Fu, G.; Duchesne, P. N.; Gu, L.; Zheng,
Y. P.; Weng, X. F.; Chen, M. S.; Zhang, P.; Pao, C. W.; Lee, J.-F.;
Zheng, N. Science 2014, 344, 495−499.
(
46) de Smit, E.; van Schooneveld, M. M.; Cinquini, F.; Bluhm, H.;
Sautet, P.; de Groot, F. M.; Weckhuysen, B. M. Angew. Chem., Int. Ed.
011, 50, 1584−1588.
47) Chen, Y.; Ji, S.; Wang, Y.; Dong, J.; Chen, W.; Li, Z.; Shen, R.;
Zheng, L.; Zhuang, Z.; Wang, D.; Li, Y. Angew. Chem., Int. Ed. 2017,
6, 6937−6941.
48) Li, X.; Bi, W.; Zhang, L.; Tao, S.; Chu, W.; Zhang, Q.; Luo, Y.;
Wu, C.; Xie, Y. Adv. Mater. 2016, 28, 2427−2431.
49) Zhang, H. J.; Yuan, X.; Ma, Z. F.; Wen, W.; Yang, J. J.
Electrochem. Soc. 2014, 161, H155−H160.
50) Strickland, K.; Miner, E.; Jia, Q.; Tylus, U.; Ramaswamy, N.;
Liang, W.; Sougrati, M. T.; Jaouen, F.; Mukerjee, S. Nat. Commun.
015, 6, 7343.
51) Kuzmann, E.; Homonnay, Z.; Mylonakis, A.; Yin, H.; Wei, Y.;
Kovacs, K.; Kubuki, S.; Klencsar, Z.; Vertes, A.; Nath, A. J. Phys.: Conf.
Ser. 2010, 217, 012029.
52) Kramm, U. I.; Herrmann-Geppert, I.; Behrends, J.; Lips, K.;
2
(
5
(
(
(
2
(
́
́
́
(
Fiechter, S.; Bogdanoff, P. J. Am. Chem. Soc. 2016, 138, 635−640.
(53) Moss, T. H.; Robinson, A. B. Inorg. Chem. 1968, 7, 1692−1694.
(54) Sahraie, N. R.; Kramm, U. I.; Steinberg, J.; Zhang, Y.; Thomas,
A.; Reier, T.; Paraknowitsch, J. P.; Strasser, P. Nat. Commun. 2015, 6,
618.
55) Li, J.; Ghoshal, S.; Liang, W.; Sougrati, M.-T.; Jaouen, F.; Halevi,
8
(
B.; McKinney, S.; McCool, G.; Ma, C.; Yuan, X.; Ma, Z.-F.; Mukerjee,
S.; Jia, Q. Energy Environ. Sci. 2016, 9, 2418−2432.
1
0798
DOI: 10.1021/jacs.7b05130
J. Am. Chem. Soc. 2017, 139, 10790−10798