Defective Tungsten Oxide Hydrate Nanosheets for Boosting Aerobic Coupling of Amines: Synergistic Catalysis by Oxygen Vacancies and Br?nsted Acid Sites

Article abstract of DOI:10.1002/smll.201701354

Adsorption and activation of molecules on a surface holds the key to heterogeneous catalysis toward aerobic oxidative reactions. To achieve high catalytic activities, a catalyst surface should be rationally tailored to interact with both organic substrate

Full text of DOI:10.1002/smll.201701354

www.advancedsciencenews.com
Catalysis
Defective Tungsten Oxide Hydrate Nanosheets for
Boosting Aerobic Coupling of Amines: Synergistic
Catalysis by Oxygen Vacancies and Brønsted Acid Sites
Ning Zhang, Xiyu Li, Yifei Liu, Ran Long, Mengqiao Li, Shuangming Chen,
Zeming Qi, Chengming Wang, Li Song, Jun Jiang, and Yujie Xiong*
Adsorption and activation of molecules on a surface holds the key to heterogeneous
catalysis toward aerobic oxidative reactions. To achieve high catalytic activities, a
catalyst surface should be rationally tailored to interact with both organic substrates
and oxygen molecules. Here, a facile bottom-up approach to defective tungsten
oxide hydrate (WO3·H2O) nanosheets that contain both surface defects and lattice
water is reported. The defective WO3·H2O nanosheets exhibit excellent catalytic
activity for aerobic coupling of amines to imines. The investigation indicates that
the oxygen vacancies derived from surface defects supply coordinatively unsaturated
sites to adsorb and activate oxygen molecules, producing superoxide radicals. More
importantly, the Brønsted acid sites from lattice water can contribute to enhancing
the adsorption and activation of alkaline amine molecules. The synergistic effect
of oxygen vacancies and Brønsted acid sites eventually boosts the catalytic activity,
which achieves a kinetic rate constant of 0.455 h−1 and a turnover frequency of
0.85 h−1 at 2 h, with the activation energy reduced to ≈35 kJ mol−1.This work provides
a different angle for metal oxide catalyst design by maneuvering subtle structural
features, and highlights the importance of synergistic effects to heterogeneous
catalysts.
1. Introduction
N. Zhang, X. Li, Y. Liu, Dr. R. Long, M. Li, Dr. S. Chen,
Z. Qi, Dr. C. Wang, Prof. L. Song, Prof. J. Jiang,
Heterogeneous catalysis plays an important role in organic
Prof. Y. Xiong
synthesis, as it can offer the relative ease of catalyst separa-
Hefei National Laboratory for Physical Sciences
tion from product stream for continuous chemical processes,
as well as more tolerance of extreme operating conditions
than homogeneous analogues.[1–13] Aerobic oxidative reac-
tions are a large group of organic reactions with the addi-
tion of oxygen to (and/or the removal of hydrogen from)
functional groups using molecular oxygen, which may further
trigger coupling of organic molecules.[1–7] The catalysts for
this class of reactions generally contain noble metals such as
Pd,[1–3] Pt,[4] and Au,[5,6] whose low abundance and high cost
would hinder large-scale applications. For this reason, non-
noble-metal oxides have received much attention toward het-
erogeneous catalysis owing to their natural abundance and
at the Microscale
iChEM (Collaborative Innovation Center
of Chemistry for Energy Materials)
School of Chemistry and Materials Science
Hefei Science Center (CAS) and National Synchrotron Radiation
Laboratory
University of Science and Technology of China
Hefei, Anhui 230026, P. R. China
E-mail: yjxiong@ustc.edu.cn
The ORCID identification number(s) for the author(s) of this arti-
cle can be found under https://doi.org/10.1002/smll.201701354.
DOI: 10.1002/smll.201701354
(1 of 10) 1701354
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

full papers
www.advancedsciencenews.com
promising prospect.[8–10] To achieve highly efficient oxidative triggering the oxidation of the adsorbed amine molecules.
reactions, tremendous efforts have been made for the activa- The locations of lattice water and oxygen vacancies, which
tion of O2 molecules—a vital step in the reactions.[11–13] The serve as the sites for the adsorption of amine molecules and
activation of O2 molecules on a metal oxide surface relies on the activation of oxygen molecules, respectively, are neigh-
the creation of coordinatively unsaturated (CUS) sites, which bored to enable a synergistic effect. As a result, the syner-
can be accomplished through defect engineering.[14–16] Most gistic effect—a powerful approach to catalysis which acts
recently, we have demonstrated that the chemisorption of on two reaction substrates simultaneously to allow efficient
O2 molecules at the oxygen vacancies enables the transfer reactions[24–27]—is well achieved by oxygen vacancies and
of photoexcited electrons to O2 molecules, producing super- Brønsted acid sites in defective WO3·H2O nanosheets. This
oxide radicals in a photocatalytic system.[11]
synergistic catalysis induces significant enhancement on cata-
Nevertheless, not only oxygen species but also organic lytic performance for the aerobic couplings of amines to cor-
molecules should be adsorbed from solvent onto solid surface responding imine with a reduced apparent activation energy,
in such a reaction system. Thus, the adsorption and activation as compared with other tungsten oxide (WO3) and WO3·H2O
of organic molecules should also be elementary steps in an counterparts.
aerobic oxidative reaction, which can impact on the overall
catalytic performance, but have not received sufficient atten-
tion. As a heterogeneous reaction takes place at the catalyst
surface, material chemists typically maneuver the adsorption
2. Results and Discussion
of organic molecules by modifying the surface conditions 2.1. Material Characterization
of catalysts.[17–20] From the viewpoint of electronic or steric
effects, this surface modification can be generally performed The defective WO3·H2O (namely, WOH-D) nanosheets are
by integrating the catalysts with additional inorganic mate- synthesized by a previous protocol with slight modification
rials or organic ligands. Despite the great progress, it would using glucose as a reductant.[11] The reduction by glucose can
be fundamentally important to explore if the concepts of remove some lattice oxygen atoms to create the defects of
traditional organic chemistry can be borrowed to maneuver oxygen vacancies. The methods for synthesis and characteri-
the adsorption of organic molecules on heterogeneous cata- zation are described in detail in the Experimental Section and
lysts. For instance, acid/base catalysts have been widely used Supporting Information. X-ray diffraction (XRD, Figure S1,
to tune the interactions between reaction molecules and Supporting Information) reveals the phase of product as
catalysts in homogeneous organic synthesis, which can be orthorhombic WO3·H2O (JCPDS No. 43-0679). As indicated
extended to heterogeneous catalysis particularly with solid by scanning and transmission electron microscopy (SEM
acid-catalysts. The solid acid-catalysts (e.g., zeolites, metal and TEM, Figure 1a,b), the WOH-D nanosheets have uni-
oxides, and heteropolyacids), which provide surface Lewis form morphology with the average edge length of 145.3
and Brønsted acid sites for adsorption and catalysis, can effi- 10.8 nm and thickness of 15.1 1.8 nm (Figure S2, Supporting
ciently facilitate various reactions including aromatization, Information). The lattice structure is further examined
isomerization, esterification, reforming,
and cracking.[21–23] Thus, creating acid sites
on solid surface should be another prom-
ising strategy for enhancing molecular
adsorption in heterogeneous catalysis.
Herein, we report a facile bottom-
up approach to defective tungsten oxide
hydrate (WO3·H2O) nanosheets. Our fine-
structure characterizations show that the
defective WO3·H2O nanosheets contain
surface oxygen vacancies and Brønsted
acid sites, derived from the surface defects
and lattice water, respectively. The defec-
tive WO3·H2O nanosheets can serve as an
excellent catalyst for the thermal-based
aerobic couplings for amines to imines.
The amines are considered as Lewis bases
due to the electron pairs in p-orbitals
of N atoms. As such, the adsorption of
amine molecules on a catalyst surface is
greatly enhanced by the lattice water that
serves as the Brønsted acid sites through
hydrogen bonds. Meanwhile, oxygen
molecules are activated to superoxide rad-
icals at the CUS sites of oxygen vacancies, a) SEM, b) TEM, and c,d) HRTEM images.
Figure 1. Electron microscopy characterizations of defective WO3·H2O (WOH-D) nanosheets:
1701354 (2 of 10)
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

www.advancedsciencenews.com
Figure 2. Characterizations for oxygen vacancies and lattice water. a) Room-temperature ESR spectra of WOH-D and nondefective WO3·H2O (WOH-
ND) nanosheets. The same molar number of the samples was used to normalize the ESR spectra. b) High-resolution W 4f XPS spectra of WOH-D and
WOH-ND nanosheets. c) High-resolution O 1s XPS spectra of WOH-D and defective WO3 (WO3-D) nanosheets. d) Solid-state MAS 1H-NMR spectra
of WOH-D and WO3-D nanosheets.
by high-resolution TEM (HRTEM, Figure 1c). The lattice W atoms.[29,30] Despite their difference in oxygen vacancies,
fringes with spacing of 2.5 and 2.6 Å can be assigned to the the WOH-D and WOH-ND samples show the same phase
(002) and (200) planes of WO3·H2O, respectively. Meanwhile, and comparable morphology (Figures S3–S5, Supporting
the HRTEM image of a cross section (Figure 1d) shows the Information), indicating that the existence of defects does
(020) lattice spacing of 5.3 Å, suggesting that the flat surfaces not cause significant changes in long-distance order. Another
of nanosheets are (010) planes.
evidence for the presence of defects is the change in color
To identify the defects in nanosheets, room-temperature and light absorption. As depicted in Figure S6 (Supporting
electron spin resonance (ESR) spectroscopy and X-ray Information), the defects in WOH-D nanosheets enable
photoelectron spectroscopy (XPS) are both used to charac- remarkable near-infrared light harvesting, which is ascribed
terize the sample, referenced against nondefective WO3·H2O to the upshifting of valence band maximum (VBM) and the
(namely, WOH-ND) nanosheets (characterizations are shown emergence of defect states below the conduction band min-
in Figures S3–S5, Supporting Information). The WOH-ND imum (CBM).[11] As a result, the color of WOH nanosheets is
nanosheets are prepared through a similar protocol except altered from bright yellow to deep green by the presence of
the absence of reductant glucose. As compared with oxygen vacancies.
WOH-ND nanosheets, WOH-D nanosheets exhibit a signifi-
Upon identifying the existence of oxygen vacancies, we
cant ESR signal at g = 2.002 (shown in Figure 2a) that can be further employ high-resolution O 1s XPS and solid-state
attributed to the electron trapping at oxygen vacancies.[28] In magic-angle-spinning (MAS) 1H nuclear magnetic reso-
addition, two characteristic peaks at the binding energies of nance (1H-NMR) spectroscopy to analyze the lattice water
35.8 and 38.0 eV can be well assigned to the 4f7/2 and 4f5/2 in samples. To make reliable assignments on spectroscopic
for W6+, respectively, in the high-resolution W 4f XPS spectra information, we remove the lattice water from WO3·H2O
of WOH-ND nanosheets (Figure 2b).[29] In comparison, the samples (i.e., WOH-D and WOH-ND) through calcination at
peaks of W 4f7/2 and 4f5/2 for WOH-D nanosheets are obvi- 673 K in Ar atmosphere, which produces two WO3 samples
ously shifted to 35.4 and 37.6 eV, respectively. The lower as references—defective WO3 (WO3-D) and nondefective
binding energies indicate the increase of electron density WO3 (WO3-ND) nanosheets. The details for the calcination
in W atoms, which further verifies the existence of oxygen method and the related sample characterizations are shown
vacancies.[30] According to the peak fitting (Figure 2b), no in the Experimental Section (Supporting Information) and
peaks for lower valence W (W5+ and/or W4+) can be dis- Figures S7–S13 (Supporting Information), respectively. The
tinguished from the spectra, indicating that the existence O 1s XPS spectra of WOH-D nanosheets (Figure 2c) display
of oxygen vacancies insufficiently generates lower-valence two peaks at about 530.0 and 532.5 eV, which can be assigned
(3 of 10) 1701354
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

full papers
www.advancedsciencenews.com
to the lattice oxygen atoms (Peak A) in metal oxides and chemical shift of 6.08 ppm originates from lattice water. The
the oxygen atoms (Peak B) in lattice water and/or adsorbed shoulder peak at 4.69 ppm of chemical shift is attributed to
oxygen species, respectively.[31] In comparison, the intensity the water molecules[34] (Figure S16, Supporting Information)
of Peak B has been substantially reduced for the WO3-D mainly adsorbed at defect sites, so this peak is almost unrec-
nanosheets that possess no lattice water but can adsorb ognizable for the WOH-ND and WO3-ND samples without
oxygen species at surface defects (see the quantitative anal- surface defects (Figure S16, Supporting Information).
ysis of O 1s peaks in Table S1, Supporting Information). As a
matter of fact, the Peak B almost disappears for the WO3-ND
nanosheets without surface defects (Figure S14, Supporting 2.2. Catalytic Performance
Information), indicating that the weak Peak B for WO3-D
is mainly attributed to the adsorbed oxygen species at sur- The information gleaned above demonstrates that WOH-D
face defects.[28,32] The assignment of lattice water in XPS is nanosheets contain both defect-rich surface and lattice water.
further confirmed by the comparison between WO3-ND and We are now in a position to investigate their synergistic effect
WOH-ND (Figure S14, Supporting Information); the addi- on catalytic performance. Here, we take aerobic couplings of
tion of lattice water significantly promotes the Peak B.
amines as a model reaction to reflect the functions of syner-
Another evidence for lattice water is provided by solid- gistic sites on the interactions with O2 and organic substrate
state MAS 1H-NMR spectroscopy. As displayed in Figure 2d, molecules. The adsorption and activation of O2 molecules can
the 1H-NMR spectrum of WOH-D exhibits a strong signal at be enhanced at oxygen vacancy sites,[11,15] while the amine

6.08 ppm of chemical shift,corresponding to theW OH bonds molecules are the Lewis bases that well interact with Brønsted
for lattice water, which can generally serve as a Brønsted acid sites—lattice water. The design is anticipated to enable a
acid.[33] In sharp contrast, the peak at 6.08 ppm almost dis- synergistic catalysis in aerobic couplings of amines.
appears for WO3-D. Similarly, the 6.08 ppm signal has been
The initial catalytic assessments are performed at 353 K
observed for WOH-ND rather than WO3-ND nanosheets using oxygen molecules as an oxidant. As a model sub-
(Figure S15, Supporting Information), confirming that the strate (illustrated in Figure 3a), 4-methyl-benzylamine (1A)
Figure 3. Catalytic performance in aerobic coupling of 4-methyl-benzylamine. a) Schematic illustration for the aerobic coupling. b) Time-dependent
catalytic conversion. c) The kinetic rate plots and calculated kinetic rate constants (k). d) The calculated turnover frequencies (TOFs) based on the
data at 2 h. e) The Arrhenius plots and calculated apparent activation energies (Ea). Reaction conditions: catalyst (0.4 mmol), 4-methyl-benzylamine
(1 mmol), solvent (CH3CN, 5 mL), O2 (0.2 MPa), temperature (353 K). The conversion is determined by GC-MS using benzotrifluoride as an internal
standard. The kinetic rate constants (k) are calculated through first-order reaction equation. The apparent activation energies (Ea) are obtained by
Arrhenius equation at various temperatures (313, 323, 333, and 353 K).
1701354 (4 of 10)
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

www.advancedsciencenews.com
is first oxidized to 4-methyl-benzaldehyde (2A) and even- conversion and selectivity. To demonstrate the niche of our
tually coupled into the corresponding imine (2B) as a final WOH-D nanosheets, we have examined the aerobic coup-
product.[13,35] We perform the catalytic performance evalua- lings of various amines under the optimized conditions
tion using various catalysts, while no products are detected (Table 1). Both electronic effect (Table 1, Entry 1–5) and
in the absence of catalysts (Table S2, Supporting Informa- steric effect (Table 1, Entry 3, 6 and 7) induced by the sub-
tion, Entry 1). Figure 3b shows the time-dependent catalytic stituents can be well tolerated. Meanwhile, our WOH-D
conversion over WOH-D and WO3-D nanosheets. Appar- catalyst also exhibits excellent conversion and selectivity
ently WOH-D nanosheets can offer higher catalytic activity for heterocyclic amines to imines (Table 1, Entry 8 and 9).
in the total conversion of 1A to 2B, and the conversion can Remarkably, the WOH-D nanosheets can catalyze the aer-
be completed within 8 h. WOH-D nanosheets convert 1A up obic couplings of aliphatic amines with high conversion and
to 69% at 2 h and 92% at 6 h, whereas only 24% and 46% selectivity (see Table 1, Entry 10 and 11) in addition to aro-
conversion is achieved by WO3-D nanosheets, respectively matic amines, which is generally considered more challenging
(shown in Table S2, Supporting Information, Entry 2–5). As due to the inactive α-H of aliphatic amine.[13] Another impor-
both WOH-D and WO3-D nanosheets possess surface oxygen tant parameter for catalyst is durability. Our durability tests
vacancies, the lattice water in WOH-D nanosheets turns out (Figures S17–S20, Supporting Information) show that the
to be the key to boosting the catalytic activity. The catalytic catalytic performance decreases in catalytic cycles due to the
activities of both WOH-D and WO3-D are strongly dependent consumption of surface defects, while the lattice water can be
on reaction temperatures (Table S2, Supporting Information, well maintained. Nevertheless, the consumed surface defects
Entry 6–11). Notably, the reaction can be well triggered by can be readily recovered to boost the catalytic performance
our WOH-D nanosheets with a 43% conversion at 6 h as the back up through a simple reduction process.[36]
temperature is reduced as low as 313 K (Table S2, Supporting
Information, Entry 6), demonstrating the excellent catalytic
activity. Moreover, nondefective samples (i.e., WOH-ND and 2.3. Mechanism Investigation
WO3-ND nanosheets) are also employed to assess the activity
for catalyzing aerobic coupling of 1A. However, both can The catalytic results above indicate that the samples con-
hardly trigger the reaction, and only trace amounts of 2B are taining surface defects can efficiently tune on the reactions,
detected (Table S2, Supporting Information, Entry 12 and 13), while the nondefective samples cannot (Table S2, Sup-
which indicates the necessity of surface oxygen vacancies.
porting Information, Entry 3, 5, 12, and 13). In literature,
Regarded as a pseudo-first-order reaction,[35] the aerobic the activation of O2 molecules is generally considered as a
coupling of amines is quantified with the kinetic rate plots as key step to limit the catalytic performance of aerobic oxida-
illustrated in Figure 3c. The calculated kinetic rate constant (k) tive reactions.[11–13] Previous studies have revealed that the
for WOH-D nanosheets reaches 0.455 0.032 h−1, six times O2 molecules can be chemisorbed and activated at the CUS
higher than that of WO3-D nanosheets (0.072 0.004 h−1). sites created by oxygen vacancies.[15] To correlate O2 activa-
Moreover, the turnover frequencies (TOFs) of the catalysts tion with oxygen vacancies, we employ ESR to resolve the
are calculated at the reaction time of 2 h, as demonstrated activated oxygen species—superoxide radicals (O2•−) using
in Figure 3d. The TOF of WOH-D nanosheets reach 0.85 h−1, 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin-trap-
substantially higher than that of WO3-D nanosheets (0.30 h−1). ping agent.[37] As illustrated in Figure 4a,b, when WOH-D or
Table S3 (Supporting Information) compares the TOF of our WO3-D nanosheets are used as catalysts (blue lines), the ESR
WOH-D nanosheets with the recently reported data for het- signals show a nearly 1:1:1:1 quartet patterns—the character-
•−
erogeneous catalysts in the aerobic couplings of amines to istic fingerprints of spin adducts derived from DMPO-O2
,
imines by both thermal-based and light-driven catalysis. The demonstrating the generation of O2•−. In comparison, both
TOF by our WOH-D catalyst is comparable to those of most WOH-ND and WO3-ND nanosheets are incapable of gener-
efficient catalysts that can achieve similar selectivity. To further ating O2•−, as no signals are detected in ESR spectra (green
•−
understand the reaction process, apparent activation energies lines). As such, the activation of O2 molecules into O2 spe-
(Ea) are calculated using the Arrhenius equation based on the cies should rely on the surface oxygen vacancies in our case.
•−
kinetic rate constants at various temperatures (Equation (1)):
To confirm the formation of O2 species, superoxide dis-
•−
mutase (SOD) is used as a scavenger to quench O2 during
the ESR detection.[37] As shown in Figure S21 (Supporting
Information), the intensity of ESR signals dramatically
decreases for both WOH-D and WO3-D nanosheets after
Ea
lnk = ln A−
(1)
RT
where Ea is apparent activation energy, k is kinetic rate con- adding SOD, suggesting that O2•− is the activated oxygen spe-
stant, T is absolute temperature, A is pre-exponential factor, cies formed in our system. Furthermore, as the spin adduct of
and R is molar gas constant. As shown in Figure 3e, the Ea for DMPO-O2•− has a relatively short lifetime, we employ a more
WOH-D nanosheets is dramatically lower (35.01 4.67 kJ mol−1 stable spin trapping agent 5-(diethylphosphono)-5-methyl-
vs 52.21 4.82 kJ mol−1 for WO3-D nanosheets), indicating that 1-pyrroline N-oxide (DEPMPO) to examine the formation of
less energy is required to surmount the energy barrier.
O2 .
•− [38] As depicted in Figure S22 (Supporting Information),
The synergistic catalysis by Brønsted acid and oxygen the ESR spectra for both WOH-D and WO3-D nanosheets
•−
vacancy sites can provide a generic approach to the aerobic show the characteristic pattern for DEPMPO-O2 with
couplings of various benzylamine derivatives with high similar intensities, further proving their comparable ability
(5 of 10) 1701354
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

full papers
www.advancedsciencenews.com
Table 1. Catalytic performance in the aerobic coupling of various amines to imines using WOH-D nanosheets as a catalyst.
Entrya)
1
Substrate
Product
Conv. [%]
93
Select.b) [%]
98
2
3
99
99
97
98
4
5
6
98
88
99
99
98
93
7
99
96
8
94
76
77
89
99
99
90
81
9
10
11
a)Reaction conditions: WOH-D nanosheets (0.4 mmol), substrate (1 mmol), solvent (CH3CN, 5 mL), O2 (0.2 MPa), temperature (353 K), reaction time (8 h). The conversion (Conv.) and selectivity
(Select.) are determined by GC-MS using benzotrifluoride as an internal standard; b)Main by-products: corresponding aldehyde, nitrile, and oxime.
for O2 activation. In addition, we also verify the key role of of O2 on defective surface, which would induce efficient O2
•−
O2 species in the aerobic coupling of amines by employing activation during the reaction process.
p-benzoquinone (p-BQ) and 4-hydroxy-2,2,6,6-tetramethyl-
The results mentioned above clearly demonstrate the
piperidinyloxy (TEMPOL) as scavengers.[39,40] As shown in role of oxygen vacancies in the catalytic reactions. Never-
Figure S23 (Supporting Information), the catalytic activity in theless, it still remains elusive why WOH-D nanosheets can
the aerobic coupling of 4-methyl-benzylamine is significantly offer the significantly enhanced catalytic activity as compared
•−
suppressed by quenching O2 with p-BQ or TEMPOL, indi- with WO3-D. Although WOH-D nanosheets possess an infe-
cating that the O2•− species indeed participates in the aerobic rior ability for O2 adsorption than WO3-D nanosheets as
coupling of amines.
indicated by adsorption isotherms (Figure S26, Supporting
To further look into the role of surface oxygen vacan- Information), the two catalysts have a comparable capa-
•−
cies in O2 activation, we examine the adsorption configura- bility of activating O2 molecules into O2 species based on
tions of O2 molecule on the surface of perfect and defective the intensity of ESR signals (Figure 4a,b, blue lines). The
•−
WO3·H2O (see the surface structural models in Figure S24, amount of generated O2 is the key parameter to catalytic
Supporting Information) by first-principles simulations. As performance associated with O2 molecules.[11–13] In addition,
illustrated in Figure S25a (Supporting Information), O2 mole- we have excluded the possible influence of surface areas and
cule cannot be chemisorbed on the perfect surface due to the crystal phases on catalytic performance: 1) the WOH-D and
lack of adsorption sites. In sharp contrast, the surface oxygen WO3-D nanosheets possess comparable surface areas, as
vacancies can perfectly serve as the chemisorption sites for shown in Figure S27 (Supporting Information); 2) although
O2 molecule through an end-on chemisorption configuration, the WOH-D and WO3-D possess different crystal phases
and the surface H atoms can hardly alter the O2 adsorption due to the loss of lattice water, both (010) surfaces consist
(Figure S25b, Supporting Information). The high simulated of distorted WO6 octahedrons by sharing corners with analo-
adsorption energy (Ead) of −0.82 eV (vs −0.01 eV for perfect gous atomic arrangements (Figure S7, Supporting Informa-
surface) suggests the substantially enhanced chemisorption tion).[41] For this reason, we then examine the other reaction
1701354 (6 of 10)
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

www.advancedsciencenews.com
Figure 4. ESR detection of superoxide radicals using a DMPO spin-trapping agent with a) WO3·H2O samples and b) WO3 samples. The * labels a
nearly 1:1:1:1 quartet pattern corresponding to the characteristic fingerprints for the spin adducts derived from DMPO-O2•−. DRIFTS spectra for the
chemisorption of c) pyridine and d) benzylamine molecules on WOH-D and WO3-D nanosheets. The background spectra have been subtracted from
the DRIFTS spectra of the adsorbed molecules. The adsorption configurations of benzylamine molecule on the surface of e) defective WO3·H2O (i.e.,
WOH-D) and f) defective WO3 (i.e., WO3-D) with the computed adsorption energy.
component—the adsorption of amine molecules at the cata- 1637 cm−1 can be attributed to Brønsted acid sites. Mean-
lyst surface.
while, the 1496 cm−1 signal is induced by both Lewis acid
The interaction between substrate molecules and cata- and Brønsted acid sites.[43,44] The Lewis acid sites originate
lyst surface generally plays a vital role in tuning catalytic from the unsaturated tungsten atoms created by defect engi-
performance.[18–20] As mentioned above, the lattice water in neering.[45,46] Notably, WOH-D nanosheets exhibit distinct

WO3·H2O supplies W OH bonds, whose H atoms at surface Brønsted acid sites as compared with WO3-D nanosheets.
1
can serve as a typical Brønsted acid.[33] The H-NMR signal Given WOH-D and WO3-D nanosheets possess comparable
at the chemical shift of 6.08 ppm by WOH-D nanosheets morphologies and defect concentrations, the Brønsted acid
(Figure 2d) indicates that the catalyst surface possesses sites on WOH-D nanosheets should be designated by the
strong Brønsted acidity.[42] To further verify the acidity, we W OH bonds associated with lattice water. As a further

employ diffuse reflectance infrared Fourier-transform spec- evidence, the Brønsted acid sites have also been well iden-
troscopy (DRIFTS) to examine the catalyst using pyridine as tified for WOH-ND nanosheets (Figure S28, Supporting
a probe molecule. As shown in Figure 4c, the peaks at 1455, Information).
1575, and 1608 cm−1 are ascribed to the chemisorption of pyr-
The Brønsted acidity of WO3·H2O would enable better
idine at surface Lewis acid sites, while the peak at 1538 and adsorption of Lewis-base amines through hydrogen bonds.
(7 of 10) 1701354
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

full papers
www.advancedsciencenews.com
Figure S29 (Supporting Information) depicts the adsorption
isotherms of benzylamine on catalysts. The fitting results
based on the Langmuir equation (Table S4, Supporting
Information) demonstrate that WOH-ND nanosheets
can adsorb substantially more benzylamine molecules
than WO3-D under the catalytic condition. Furthermore,
DRIFTS spectra (Figure 4d) clearly reveal the chemisorp-
tion of benzylamine on the surface of WOH-D and WO3-D
catalyst. After adsorbing benzylamine molecules, WOH-D
nanosheets exhibit recognizable peaks at 1454, 1495, 1590,
and 1603 cm−1, which are assigned to the adsorbed ben-
zylamine. Specifically, the peaks at 1454 and 1603 cm−1 are


indexed to the bending vibrations of C H and N H bonds,
respectively,[47,48] while the bands at 1495 and 1590 cm−1 are
associated with the stretching vibrations of benzene ring.[49]
In comparison with WOH-D, the corresponding peaks are
much weaker for WO3-D nanosheets. This indicates that the
Scheme 1. Schematic illustrating the synergistic catalysis for aerobic
couplings of amines to imines by oxygen vacancies and Brønsted acid
sites on the surface of defective WO3·H2O nanosheets.
interaction between benzylamine and WOH-D nanosheets vacancies and Brønsted acid sites in defective WO3·H2O
is substantially stronger than WO3-D nanosheets. Moreover, nanosheets. As illustrated in Scheme 1, the lattice water pro-
as compared with free benzylamine molecules (1606 cm−1 in vides Brønsted acid sites for efficiently adsorbing amine mole-
Figure S30, Supporting Information), the bending vibration cules via strong hydrogen bonds. The enhanced adsorption

of N H bonds is slightly red-shifted, which suggests that the can facilitate the oxidation of amine molecules by the super-

formation of hydrogen bonds can alter the strength of N H oxide species whose generation relies on the CUS sites cre-
bonds and facilitate their activation. Our control experiments ated by oxygen vacancies. This synergistic effect eventually
(Figure S31, Supporting Information) reveal that WOH-ND boosts the catalytic activity of WOH-D nanosheets for the
nanosheets can also well adsorb benzylamine molecules aerobic coupling of amines. The aerobic coupling reactions
whereas WO3-ND hardly works. This further confirms that undergo the pathway below (Equations (2)–(5)):[13,35]
lattice water instead of surface defects is responsible for the
e−
enhanced adsorption of benzylamine on catalyst surface.
O2 →O•2−
(2)
To further gain insight into the amine–catalyst inter-
action, we conduct first-principles simulations to examine
the adsorption of benzylamine to the surface of defective
WO3·H2O (i.e., WOH-D, Figure S24b, Supporting Informa-
tion) and defective WO3 (i.e., WO3-D, Figure S32, Supporting
Information). As shown in Figure 4e, the lattice water of
WOH-D supplies Brønsted acid sites so that benzylamine
molecules can be efficiently adsorbed via the hydrogen bonds
between the Brønsted acid sites and N atoms of benzylamine.
In contrast, the adsorption of benzylamine to WO3-D surface
can only take place through the bonding of H atoms (ben-
zylamine) to surface O atoms (WO3-D) (Figure 4f), more
(3)
(4)
(5)
In this process, the activated O2 molecules oxidize amine
weakly than that of Brønsted acid sites. As a result, the Ead molecules (chemisorbed at the neighbored lattice water)
of benzylamine on the surface of WOH-D and WO3-D are and combine H atoms to generate H2O2 as an intermediate
−1.27 and −1.13 eV, respectively. This suggests the stronger product. This process does not involve the dissociation of

interaction of benzylamine with WOH-D surface, which O O bonds. The H2O2 molecules can spontaneously decom-
would facilitate the activation of substrate molecules for pose to H2O and O2. To verify this process, we employ a
enhanced catalytic efficiency. To further correlate catalytic well-established N,N-diethyl-p-phenylenediamine (DPD)/
performance with the adsorption and activation of substrate horseradish peroxidase (POD) method to detect the pro-
molecules, we take benzyl alcohol as the substrate to carry out duced H2O2.[50] As shown in Figure S34 (Supporting Infor-
oxidative reactions under the same condition. As hydroxyl mation), the formation of H2O2 intermediate has indeed

group ( OH) is not as a strong Lewis base as amino group been observed.

( NH2), the Brønsted acid sites on WOH-D nanosheets
cannot efficiently facilitate the adsorption and activation of
benzyl alcohol. As a result, the addition of lattice water can
hardly alter the catalytic activity (see Figure S33, Supporting
3. Conclusion
Information). This control experiment further unravels the In summary, we have developed a facile bottom-up approach
role of Brønsted acid sites in boosting catalytic performance. to the defective WO3·H2O nanosheets that contain surface
Taken together, the experimental and theoretical results oxygen vacancies and Brønsted acid sites. The oxygen vacan-
above clearly depict the synergistic catalysis by oxygen cies can provide coordinatively unsaturated sites for activating
1701354 (8 of 10)
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

www.advancedsciencenews.com
oxygen molecules into superoxide radicals—a key component (GC-MS, 7890A and 5975C, Agilent) using benzotrifluoride as an
for triggering reactions. More importantly, the existence of internal standard. The TOF was calculated using Equation (6):
Brønsted acid sites induces significant changes in catalytic
n0 ×C × S
performance. The strong interaction between Lewis-base
amine molecules and Brønsted-acid catalyst surface facili-
tates the adsorption and oxidation of amine molecules. As a
proof of concept, the synergistic effect of oxygen vacancies
and Brønsted acid sites at the surface of WOH-D nanosheets
enables sixfold enhancement on aerobic couplings of amines
against the catalysts in the absence of lattice water, in terms of
kinetic rate constants. The WOH-D catalyst achieves a calcu-
lated TOF of 0.85 h−1 at 2 h and a reduced activation energy
down to ≈35 kJ mol−1. This strategy can work for not only aro-
matic amines but also aliphatic amines with inactive α-H. The
work reported here calls for future efforts on catalyst design
at molecular level, and demonstrates the necessity of looking
into subtle surface features of catalytic materials.
TOF =
(6)
ncat × t
where n0 is the initial molar number of substrate, C and S are the
conversion and selectivity at the reaction of t, ncat is the molar
number of catalyst, and t is the reaction time.
DRIFTS Spectroscopy Characterizations: The DRIFTS measure-
ments for pyridine and benzylamine chemisorption were performed
using a Bruker IFS 66v Fourier transformation spectrometer with a
Harrick diffuse reflectance accessory at the Infrared Spectroscopy
and Microspectroscopy Endstation (BL01B) in National Synchro-
tron Radiation Laboratory (NSRL), Hefei. The catalyst samples were
first pretreated at 353 K for 2 h under vacuum, and a background
spectrum was obtained from 1700 to 1400 cm−1 before adsorbing
the pyridine or benzylamine molecules. Then 10 µL of pyridine or
benzylamine was injected into the samples. After reaching the
adsorption equilibrium, the samples were retreated at 353 K for
4 h under vacuum to completely remove the physically adsorbed
molecules. Finally, the DRIFTS spectra for pyridine or benzylamine
chemisorption were collected. The background spectra were sub-
tracted from the DRIFTS spectra of adsorbed species.
First-Principles Simulations: The simulations were performed
based on spin-polarized density functional theory (DFT) using the
VASP package.[51,52] Perdew–Burke–Ernzerhof (PBE) exchange-
correlation functional within a generalized gradient approximation
(GGA) and the projector augmented-wave (PAW) potential were
employed.[53] The Brillouin zone was sampled with Monkhorst–Pack
generated sets of k-points, 10 × 5 × 10 k-point mesh was chosen for
bulk computations, and 5 × 5 × 1 k-point mesh for surface computa-
tion. All geometric structures were fully relaxed until energy and forces
were converged to 10−5 eV and 0.02 eV Å−1, respectively. The calcu-
lated lattice parameters of the WO3·H2O unit cell are a = 5.25 Å, b =
10.71 Å, and c = 5.13 Å. A system of 2 × 2 slab consisting of seven
layers was employed to model the surface of WO3·H2O (010) with
oxygen atom and H2O terminated, in which three top atom layers are
relaxed and four atom layers are fixed to the bulk positions during the
geometry optimizing simulations. The periodic boundary condition
was set with a 18 Å vacuum region above the plane of WO3·H2O (010).
For the WO3 simulation, the lattice structure was used by a previous
model.[11] The monoclinic structure of WO3 with 8 W and 24 O atoms
in a primary unit cell was built, and the calculated lattice parameters
are a = 7.49 Å, b = 7.66 Å, c = 7.84 Å, and β = 90.3°. A system of 2 × 2
slab with four layers was employed to model the WO3 (010) surface,
with oxygen atom terminated. Three top layers were relaxed from the
bulk positions during the geometry optimizing. The periodic boundary
condition was set with a 18 Å vacuum region above surface. All the
atom sites of adsorbates (i.e., benzylamine and O2 molecules) on
WO3·H2O (010) or WO3 (010) were relaxed to investigate the interac-
tion between adsorbates and WO3·H2O (010) or WO3 (010).
4. Experimental Section
Materials: Sodium tungstate dehydrate (Na2WO4·2H2O,
99.5%), citric acid (CA, 99.5%), SOD, p-BQ (97%), TEMPOL (98%),
DPD (98%), and POD were purchased from Aladdin. DMPO (98%)
was purchased from Adamas-beta. DEPMPO (98%) was obtained
from TCI. d(+)-glucose (AR), hydrochloric acid (HCl, 36–38%),
sodium borohydride (NaBH4, AR), acetonitrile (CH3CN, AR), and
anhydrous ethanol (EtOH, AR) were obtained from Sinopharm
Chemical Reagent Co., Ltd. All amines and benzyl alcohol for cata-
lytic experiments were obtained from Energy Chemical. The water
used in all experiments was deionized (DI). All of the chemical rea-
gents were used as received without further purification.
Synthesis of WO3·H2O Nanosheets: The defective WO3·H2O
nanosheets were synthesized through a facile hydrothermal
method. Briefly, 1.0 mmol of Na2WO4·2H2O and 1.5 mmol of CA
were added into 30 mL of H2O to produce a transparent solu-
tion. Then 10 mmol of glucose was added into the solution as a
reductant. After vigorous stirring for 10 min, 3 mL of HCl (6 m) was
added into the mixed solution, followed by another 30 min stir-
ring. The mixture was transferred into a 50 mL Teflon-lined auto-
clave, sealed, and heated at 393 K for 24 h. After the autoclave
had cooled down to room temperature, the product was separated
by centrifugation, washed with H2O and EtOH for several times
until the organics were completely removed, and then dried at
333 K in a vacuum oven for 24 h for further use and characteri-
zation. The defects in the samples can be tailored by changing
the amount of glucose, and nondefective WO3·H2O nanosheets
were synthesized using the same method except the absence of
glucose.
Catalytic Measurements for Aerobic Couplings of Amines: The
catalytic reactions were carried out in CH3CN solution in O2 atmos-
phere at various temperatures. Typically, 1 mmol of amine and
0.4 mmol of catalyst were dispersed in 5 mL of CH3CN in a pressure
bottle, which was then filled with 0.2 MPa of O2 and sealed. The
reaction was performed in an oil bath for different time at desired
temperature with magnetic stirring. After the completion of reac-
tion, the catalyst was removed by filtration. 0.5 mL of the resultant
solution was taken and diluted to 5 mL by CH3CN. The final solu-
Supporting Information
Supporting Information is available from the Wiley Online Library
tion was analyzed by gas chromatography-mass spectrometry or from the author.
(9 of 10) 1701354
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com

full papers
www.advancedsciencenews.com
[20] M. Zhao, K. Yuan, Y. Wang, G. Li, J. Guo, L. Gu, W. Hu, H. Zhao,
Z. Tang, Nature 2016, 539, 76.
Acknowledgements
[21] B. M. Reddy, M. K. Patil, Chem. Rev. 2009, 109, 2185.
[22] A. Corma, Chem. Rev. 1995, 95, 559.
[23] A. Primo, H. Garcia, Chem. Soc. Rev. 2014, 43, 7548.
[24] I. X. Green, W. Tang, M. Neurock, J. T. Yates Jr., Science 2011,
333, 736.
[25] A. E. Allen, D. W. MacMillan, Chem. Sci. 2012, 3, 633.
[26] D. Uraguchi, N. Kinoshita, T. Kizu, T. Ooi, J. Am. Chem. Soc. 2015,
137, 13768.
[27] D. Wang, H. L. Xin, R. Hovden, H. Wang, Y. Yu, D. A. Muller,
F. J. DiSalvo, H. D. Abruña, Nat. Mater. 2013, 12, 81.
[28] F. Lei, Y. Sun, K. Liu, S. Gao, L. Liang, B. Pan, Y. Xie, J. Am. Chem.
Soc. 2014, 136, 6826.
N.Z. and X.L. contributed equally to this work. This work was finan-
cially supported in part by 973 Program (No. 2014CB848900),
NSFC (Nos. 21471141, U1532135, 21601173, and 21573212),
CAS Key Research Program of Frontier Sciences (QYZDB-SSW-
SLH018), CAS Interdisciplinary Innovation Team, Recruitment Pro-
gram of Global Experts, and CAS Hundred Talent Program. The
authors thank Prof. Bo Liu and Prof. Jiafu Chen for their help on O2
sorption isotherm and ESR characterizations, respectively. DRIFTS
measurements were performed at the Infrared Spectroscopy and
Microspectroscopy Endstation (BL01B) in National Synchrotron
Radiation Laboratory (NSRL), Hefei.
[29] G. Liu, J. Han, X. Zhou, L. Huang, F. Zhang, X. Wang, C. Ding,
X. Zheng, H. Han, C. Li, J. Catal. 2013, 307, 148.
[30] M. Ma, K. Zhang, P. Li, M. S. Jung, M. J. Jeong, J. H. Park, Angew.
Chem. Int. Ed. 2016, 55, 11819.
Conflict of Interest
[31] M. Breedon, P. Spizzirri, M. Taylor, J. du Plessis, D. McCulloch,
J. Zhu, L. Yu, Z. Hu, C. Rix, W. Wlodarski, K. Kalantar-zadeh, Cryst.
Growth Des. 2010, 10, 430.
[32] L. Li, J. Yan, T. Wang, Z.-J. Zhao, J. Zhang, J. Gong, N. Guan, Nat.
Commun. 2015, 6, 5881.
The authors declare no conflict of interest.
[33] S. M. Kanan, Z. Lu, J. K. Cox, G. Bernhardt, C. P. Tripp, Langmuir
2002, 18, 1707.
[34] S. Hayashi, M. Yanagisawa, K. Hayamizu, Anal. Sci. 1991,
7, 955.
[35] S. Biswas, B. Dutta, K. Mullick, C.-H. Kuo, A. S. Poyraz, S. L. Suib,
ACS Catal. 2015, 5, 4394.
[36] G. Xi, S. Ouyang, P. Li, J. Ye, Q. Ma, N. Su, H. Bai, C. Wang, Angew.
Chem. Int. Ed. 2012, 51, 2395.
[1] L. Kesavan, R. Tiruvalam, M. H. Ab Rahim, M. I. bin Saiman,
D. I. Enache, R. L. Jenkins, N. Dimitratos, J. A. Lopez-Sanchez,
S. H. Taylor, D. W. Knight, C. J. Kiely, G. J. Hutchings, Science
2011, 331, 195.
[2] R. Long, K. Mao, X. Ye, W. Yan, Y. Huang, J. Wang, Y. Fu, X. Wang,
X. Wu, Y. Xie, Y. Xiong, J. Am. Chem. Soc. 2013, 135, 3200.
[3] P. Zhang, Y. Gong, H. Li, Z. Chen, Y. Wang, Nat. Commun. 2013, 4,
1593.
[4] Y. Shiraishi, H. Sakamoto, Y. Sugano, S. Ichikawa, T. Hirai, ACS
Nano 2013, 7, 9287.
[5] L. Wang, Y. Zhu, J.-Q. Wang, F. Liu, J. Huang, X. Meng, J.-M. Basset,
Y. Han, F.-S. Xiao, Nat. Commun. 2015, 6, 6957.
[6] M. Turner, V. B. Golovko, O. P. Vaughan, P. Abdulkin,
A. Berenguer-Murcia, M. S. Tikhov, B. F. Johnson, R. M. Lambert,
Nature 2008, 454, 981.
[7] B. Dutta, S. Biswas, V. Sharma, N. O. Savage, S. Alpay, S. L. Suib,
Angew. Chem. Int. Ed. 2016, 55, 2171.
[8] R. V. Jagadeesh, A.-E. Surkus, H. Junge, M.-M. Pohl, J. Radnik,
J. Rabeah, H. Huan, V. Schünemann, A. Brückner, M. Beller, Sci-
ence 2013, 342, 1073.
[9] X. Xie, Y. Li, Z.-Q. Liu, M. Haruta, W. Shen, Nature 2009, 458, 746.
[10] W. Huang, Acc. Chem. Res. 2016, 49, 520.
[11] N. Zhang, X. Li, H. Ye, S. Chen, H. Ju, D. Liu, Y. Lin, W. Ye, C. Wang,
Q. Xu, J. Zhu, L. Song, J. Jiang, Y. Xiong, J. Am. Chem. Soc. 2016,
138, 8928.
[12] H. Su, K.-X. Zhang, B. Zhang, H.-H. Wang, Q.-Y. Yu, X.-H. Li,
M. Antonietti, J.-S. Chen, J. Am. Chem. Soc. 2017, 139, 811.
[13] C. Su, M. Acik, K. Takai, J. Lu, S.-J. Hao, Y. Zheng, P. Wu, T. Enoki,
Y. J. Chabal, K. P. Loh, Nat. Commun. 2012, 3, 1298.
[14] Y. Sun, Q. Liu, S. Gao, H. Cheng, F. Lei, Z. Sun, Y. Jiang, H. Su,
S. Wei, Y. Xie, Nat. Commun. 2013, 4, 2899.
[37] H. Wang, X. Sun, D. Li, X. Zhang, S. Chen, W. Shao, Y. Tian, Y. Xie,
J. Am. Chem. Soc. 2017, 139, 2468.
[38] E. L. G. Samuel, D. C. Marcano, V. Berka, B. R. Bitner, G. Wu,
A. Potter, R. H. Fabian, R. G. Pautler, T. A. Kent, A.-L. Tsai,
J. M. Tour, Proc. Natl. Acad. Sci. USA 2015, 112, 2343.
[39] F. Raza, J. H. Park, H.-R. Lee, H.-I. Kim, S.-J. Jeon, J.-H. Kim, ACS
Catal. 2016, 6, 2754.
[40] N. Zhang, C. Han, Y.-J. Xu, J. J. Foley, D. Zhang, J. Codrington,
S. K. Gray, Y. Sun, Nat. Photonics 2016, 10, 473.
[41] H. Zheng, J. Z. Ou, M. S. Strano, R. B. Kaner, A. Mitchell,
K. Kalantar-zadeh, Adv. Funct. Mater. 2011, 21, 2175.
[42] A. Takagaki, M. Sugisawa, D. Lu, J. N. Kondo, M. Hara, K. Domen,
S. Hayash, J. Am. Chem. Soc. 2003, 125, 5479.
[43] T. Barzetti, E. Selli, D. Mosotti, L. Forni, J. Chem. Soc., Faraday
Trans. 1996, 92, 1401.
[44] C. Martin, G. Solana, V. Rives, Catal. Lett. 1997, 49, 235.
[45] K. K. Akurati, A. Vital, J.-P. Dellemann, K. Michalow, T. Graule,
D. Ferri, A. Baiker, Appl. Catal., B 2008, 79, 53.
[46] S. Horike, M. Dinca, K. Tamaki, J. R. Long, J. Am. Chem. Soc. 2008,
130, 5854.
[47] S. Liang, L. Wen, S. Lin, J. Bi, P. Feng, X. Fu, L. Wu, Angew. Chem.
Int. Ed. 2014, 53, 2951.
[48] N. H. Cromwell, F. A. Miller, A. R. Johnson, R. L. Frank, D. J. Wallace,
J. Am. Chem. Soc. 1949, 71, 3337.
[49] R. Ma, L. Wang, B. Zhang, X. Yi, A. Zheng, F. Deng, X. Yan, S. Pan,
X. Wei, K.-X. Wang, D. S. Su, F.-S. Xiao, ChemSusChem 2016, 9,
2759.
[15] D.-N. Pei, L. Gong, A.-Y. Zhang, X. Zhang, J.-J. Chen, Y. Mu,
H.-Q. Yu, Nat. Commun. 2015, 6, 8696.
[16] J. Suntivich, H. A. Gasteiger, N. Yabuuchi, H. Nakanishi,
J. B. Goodenough, Y. Shao-Horn, Nat. Chem. 2011, 3, 546.
[17] Y. Bai, W. Zhang, Z. Zhang, J. Zhou, X. Wang, C. Wang, W. Huang,
J. Jiang, Y. Xiong, J. Am. Chem. Soc. 2014, 136, 14650.
[18] G. Chen, C. Xu, X. Huang, J. Ye, L. Gu, G. Li, Z. Tang, B. Wu,
H. Yang, Z. Zhao, Z. Zhou, G. Fu, N. Zheng, Nat. Mater. 2016, 15,
564.
[50] H. Bader, V. Sturzenegger, J. Hoigne, Water Res. 1988,
22, 1109.
[51] G. Kresse, J. Furthmüller, Phys. Rev. B 1996, 54, 11169.
[52] G. Kresse, J. Furthmüller, Comput. Mater. Sci. 1996, 6, 15.
[53] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
3865.
[19] S. T. Marshall, M. O’Brien, B. Oetter, A. Corpuz, R. M. Richards,
D. K. Schwartz, J. W. Medlin, Nat. Mater. 2010, 9, 853.
Received: April 27, 2017
Published online:
1701354 (10 of 10)
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2017, 1701354
www.small-journal.com