Activating Inert Alkali-Metal Ions by Electron Transfer from Manganese Oxide for Formaldehyde Abatement

Article abstract of DOI:10.1002/chem.201704398

Alkali-metal ions often act as promoters rather than active components due to their stable outermost electronic configurations and their inert properties in heterogeneous catalysis. Herein, inert alkali-metal ions, such as K⁺ and Rb⁺, are activated by electron transfer from a Hollandite-type manganese oxide (HMO) support for HCHO oxidation. Results from synchrotron X-ray diffraction, absorption, and photoelectron spectroscopies demonstrate that the electronic density of states of single alkali-metal adatoms is much higher than that of K⁺ or Rb⁺, because electrons transfer from manganese to the alkali-metal adatoms through bridging lattice oxygen atoms. Electron transfer originates from the interactions of alkali metal d–sp frontier orbitals with lattice oxygen sp³ orbitals occupied by lone-pair electrons. Reaction kinetics data of HCHO oxidation reveal that the high electronic density of states of single alkali-metal adatoms is favorable for the activation of molecular oxygen. Mn L₃-edge and O K-edge soft-X-ray absorption spectra demonstrate that lattice oxygen partially gains electrons from the Mn e_g orbitals, which leads to the upshift in energy of lattice oxygen orbitals. Therefore, the facile activation of molecular oxygen by the electron-abundant alkali-metal adatoms and active lattice oxygen are responsible for the high catalytic activity in complete oxidation of HCHO. This work could assist the design of efficient and cheap catalysts by tuning the electronic states of active components.

Full text of DOI:10.1002/chem.201704398

A Journal of
Accepted Article
Title: Activating Inert Alkali Metal Ions by Electron Transfer from
Manganese Oxide for Formaldehyde Abatement
Authors: Jiayi Gao, Zhiwei Huang, Yaxin Chen, Jing Wan, Xiao Gu,
Zhen Ma, Jianmin Chen, and Xingfu Tang
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201704398
Link to VoR: http://dx.doi.org/10.1002/chem.201704398
Supported by

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
Activating Inert Alkali Metal Ions by Electron Transfer from
Manganese Oxide for Formaldehyde Abatement
[
a]
[a]
[a]
[b]
[b]
[a]
[a]
Jiayi Gao, Zhiwei Huang, Yaxin Chen, Jing Wan, Xiao Gu, Zhen Ma, Jianmin Chen, and
[a, c]
Xingfu Tang*
[
2]
Abstract: Alkali metal ions often act as promoters rather than active
components due to their stable outermost electronic configurations
and their inert properties in heterogeneous catalysis. Herein, we
pollution, because it can lead to eye, nose, throat, and skin
[
3]
irritation. More seriously, long-term exposure to HCHO may
[4,5]
lead to cancer.
HCHO has been classified as a probable
+
+
activate inert alkali metal ions such as K and Rb by electron
transfer from hollandite-type manganese oxide (HMO) support for
HCHO oxidation. Results from synchrotron X-ray diffraction,
absorption, and photoelectron spectroscopies demonstrate that the
electronic density of states of single alkali metal adatoms is much
human carcinogen by the Environmental Protection Agency of
USA and the World Health Organization. Furthermore, HCHO
is also one of the typical atmospheric pollutants in the outdoor
air, because it induces photochemical pollution.
is of vital importance to control the HCHO emission.
Catalytic oxidation of HCHO to CO and H O by using solid
catalysts in the presence of molecular oxygen is a promising
technology.
excellent catalytic performance,
expensive. Some transition metal oxides can show catalytic
[
6]
[3]
[
7]
[8,9]
Therefore, it
+
+
higher than that of K or Rb , because electrons transfer from
manganese to the alkali metal adatoms through bridging lattice
oxygen. The electron transfer originates from the interactions of
2
2
[
10]
Supported noble metal catalysts often show
3
[11−13]
alkali d-sp frontier orbitals with lattice oxygen sp orbitals occupied
but they are rather
by lone-pair electrons. Reaction kinetics data of HCHO oxidation
reveal that the high electronic density of states of single alkali metal
[
14,15]
activity as high as noble metal catalysts.
In particular,
adatoms is favorable for the activation of molecular oxygen. Mn L
edge and O K-edge soft-X-ray absorption spectra demonstrate that
lattice oxygen gains part of electrons from the Mn e orbitals, leading
3
-
hollandite-type manganese oxides (HMO) have high catalytic
activity in HCHO oxidation because the tunnel sizes are suitable
[
16]
g
for the adsorption and activation of HCHO
and for the
[
17]
to the upshift of lattice oxygen orbitals in energy. Therefore, the
facile activation of molecular oxygen by the electron-abundant alkali
metal adatoms and active lattice oxygen are responsible for the high
catalytic activity in complete oxidation of HCHO. This work could
assist the design of efficient and cheap catalysts by tuning the
electronic states of active components.
activation of oxygen
performance, noble metals such as Pt,
. To further enhance the catalytic
[
18]
[19]
[20−22]
Au,
and Ag
were highly dispersed or even atomically dispersed/deposited
onto HMO surfaces. However, noble metal catalysts can only be
applied in some special environments such as aircrafts due to
cost reasons.
Occasionally, alkali metals and alkaline earth metals such as
+
2+
K and Ca , atomically dispersed and with high electron density
of states, have been proved to be highly active for catalytic
Introduction
[23,24]
oxidation of HCHO.
For instance, Han et al. found that
2
+
isolated Ca in the tunnels of hydroxyapatite showed high
catalytic activity in the complete oxidation of HCHO because
HCHO has been widely used in making building and furnishing
materials, consumer products, etc. The worldwide production
capacity for HCHO is over 30 million tons per year. However,
HCHO emitted from the above materials may cause indoor air
2+
[24]
Ca can activate molecular oxygen.
Our recent work also
[
1]
+
showed that single K of a single-site catalyst with hybridized d-
sp orbitals has high electronic density of states and possesses
excellent ability towards activating both molecular oxygen and
lattice oxygen, and thus the catalyst showed high catalytic
[
23]
[
a]
J. Gao, Dr. Z. Huang, Dr. Y. Chen, Prof. Z. Ma, Prof. J. Chen, Prof.
X. Tang
activity for HCHO oxidation.
It should be mentioned that
previous studies have found that potassium exhibited promotion
Institute of Atmospheric Sciences, Shanghai Key Laboratory of
Atmospheric Particle Pollution & Prevention (LAP )
Department of Environmental Science & Engineering
Fudan University
[25]
[26]
3
effects in Fisher-Tropsch reaction and ammonia synthesis,
and the promoting function of potassium mainly originates from
the electron-donating effects of potassium. Likewise, electron-
2
00433 Shanghai (P. R. China)
abundant alkalis can facilitate the dissociation of O
2
by charge
, thus
E-mail: tangxf@fudan.edu.cn
J. Wan, Prof. X. Gu
Department of Applied Physics
Chongqing University
transfer from potassium to the antibonding orbitals of O
2
[
b]
c]
[
27−29]
showing significant activity in oxidation reactions.
In this work, a series of alkali metal ions were atomically
dispersed on HMO surfaces, and the catalytic performance of
these catalysts in HCHO oxidation were studied at low
temperatures. The crystal structures and morphologies of the
catalysts were determined by using synchrotron X-ray diffraction
4
00044 Chongqing (P.R.China)
[
Prof. X. Tang
Jiangsu Collaborative Innovation Center of Atmospheric
Environment & Equipment Technology (CICAEET)
Nanjing University of Information Science & Technology
2
10044 Nanjing (P. R. China)
(
SXRD) and transition electron microscopy (TEM). The specific
surface areas and pore volumes of the catalysts were analyzed
by N adsorption-desorption measurements, and calculated by
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.2017xxxxx.
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
the Brunauer-Emmett-Teller (BET) and the t-plot method,
respectively. The local structures of single alkali adatoms were
determined by extended X-ray absorption fine structure (EXAFS)
spectra. A combination of X-ray photoelectron spectra (XPS)
and X-ray absorption near-edge structure (XANES) spectra
revealed the electronic states and the outermost orbital
configuration of the isolated alkali ions. Electronic interactions
3
between Mn and O orbitals were also studied by Mn L -edge
and O K-edge soft-X-ray absorption spectra. The catalytic
performance of catalysts in HCHO oxidation is correlated with
the catalysts’ structures and characteristics mainly in terms of
the activation of molecular oxygen and active lattice oxygen.
This work could open an avenue to develop efficient and
inexpensive single-atom catalysts for abating volatile organic
compounds such as HCHO.
Results and Discussion
The SXRD patterns of the samples are shown in Figure 1.
Rietveld refinement analyses were performed to investigate the
effect of alkalis’ loading on HMO. The corresponding lattice
parameters and partial atom coordinates are listed in Tables S1
and S2, respectively, from which the structural model can be
constructed (Figure 1, insets). HMO can be indexed to a
tetragonal structure with an I4/m space group (Figure S1), and
thus HMO, constructed by MnO
dimensional square tunnels of 4.7 Å × 4.7 Å along the [001]
direction. The tunnel walls consist of lattice oxygen with sp
6
octahedra, has one-
1 1
Figure 1. SXRD patterns of (a) K /HMO and (b) Rb /HMO including the
Rietveld refinements. The blue curves are the differential SXRD patterns. The
short vertical lines below the SXRD patterns mark the peak positions of all of
the possible Bragg reflections. Insets: oblique perspective models showing
alkali ions at the Wyckoff 2b sites. Red, blue, green, and yellow balls represent
3
3
hybridization, and these lattice oxygen ions use three sp
4
+
3
orbitals to bond to three Mn , and another sp orbital occupied
6
O, Mn, K, and Rb atoms, respectively, and blue octahedron represents MnO .
by lone-pair electrons can donate electrons to unoccupied
orbitals of anchored metals such as Ag and K .
+
[23,30,31]
After loading alkalis, the tetragonal structures are preserved
regardless of the types of alkalis, consistent with the results in
our recent report. One K or Rb ion coordinates to eight O sp
The local structures of the alkali atoms were investigated by
EXAFS, and the related spectra at the K-edge are shown in
Figure 2. Tables S4 and S5 summarize the structural
parameters obtained by fitting the spectra with theoretical
models, and Figures S2 and S3 illustrate the curve-fitting of R-
space and inverse Fourier transform (FT) spectra. The FT
amplitudes of the EXAFS data of K /HMO and Rb /HMO are
1 1
similar to each other. The results indicate that the immediate
structures of K and Rb are almost the same, i.e., they are both
[
32]
3
orbitals to form bulk-centered tetragonal prism structures
according to the Rietveld refinement analyses, and the
reconstructed models are shown in the insets of Figure 1. The
K-O and Rb-O bond lengths are 2.88 and 2.96 Å, respectively,
equal to the sum of the corresponding ionic radius of alkali ions
[31]
[
30]
and lattice oxygen.
6
The MnO octahedra are also preserved
after loading K or Rb. In Table S1, the lattice parameter a
slightly increases with the atomic number of alkalis, and the
lattice parameter c almost remains unchanged. Thus, the cell
volume also increases when these alkali ions are loaded,
consistent with the insertion of alkalis into the HMO tunnels in
the interval stacking model because of Coulomb charge
located at the Wyckoff 2b sites. The first shells of K
Rb /HMO are assigned to the A-O bonds with an average bond
length of ~2.88 and 2.96 Å, respectively, and a coordination
number of 8. Thus, a coordination configuration of an AO
1
/HMO and
1
8
polyhedron is a D4h symmetry (Figure 2, insets), as evidenced by
the SXRD refinements. The result agrees with the conclusion
from the SXRD refinement analyses that the alkali ions are
arranged at an alternate stacking model along the HMO tunnel
direction. Thus, the alkali ions are located at isolated states and
the single alkali metal ions are exposed on the HMO tunnel
[
31]
repulsion.
Furthermore, the specific surface areas of HMO
2
-1
and alkali-doped HMO are in a range of 54-64 m g , and the
pore volumes calculated from the t-plot method are about 1.4-
-3
3
-1
2
.4 × 10 cm g , as listed in Table S3.
[23]
openings to form "single-atom" alkali catalysts.
Furthermore,
the EXAFS data of the single-atom alkali catalysts are distinctly
different from those of KCl and RbOH. The EXAFS spectrum of
+
KCl reveals that the local structure of K is a KCl
6
octahedron
with an average K-Cl bond length of ~3.13 Å and an octahedral
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
[
33]
(
O
h
) symmetry (Figure 2a, inset). The EXAFS spectrum of
single alkali metal ions are anchored on oxygen cavity sites
consisting of four surface sp oxygen ions.
+
3
RbOH reveals that the local structure of Rb is a RbO
tetragonal pyramid with an average Rb-O bond length of ~3.02 Å
and a C4v symmetry (Figure 2b, inset).
5
[34]
(a)
(c)
(e)
(b)
(300)
×
-
200 nm
200 nm
200 nm
2 nm
[
011]
(d)
K
10 nm
(f)
Rb
10 nm
Figure 3. TEM images of (a) HMO, (c) K
1
1
/HMO, and (e) Rb /HMO. (b) HRTEM
image of HMO. STEM images, EDX mapping data, and atomic arrangements
on the HMO(001) planes of (d) K /HMO and (f) Rb /HMO. Red, blue, green,
and yellow balls represent O, Mn, K, and Rb atoms, respectively. Blue
octahedron stands for the MnO structural motif.
1
1
8
HMO keeps the same morphology and similar structure after
loading alkalis onto it, which is advantageous for comparatively
studying the catalytic activity of HMO and the single-atom alkali
catalysts. Catalytic oxidation of HCHO was conducted over
HMO and the single-atom alkali catalysts at low temperatures,
1 1
Figure 2. FT EXAFS spectra of (a) K /HMO and (b) Rb /HMO at the K and Rb
K-edge with the k weight. Red, green, purple, and yellow balls represent O, K,
Cl, and Rb atoms, respectively.
2
The morphologies of HMO, K
1
/HMO, and Rb
1
/HMO were
and the results in term of HCHO conversions (XHCHO) and CO
selectivity as a function of reaction temperature are shown in
2
observed by using TEM. As shown in Figure 3, HMO has a rod-
shaped morphology with an average length of ~800 nm and a
Figure 4a. All catalysts show 100% CO selectivity at the whole
tested temperature window. The XHCHO over HMO are
2
[
32]
width of ~14 nm.
The HMO rods grow along the [001] axis
according to the high resolution TEM (HRTEM) image: (i) the
fringe distances of 3.26 and 2.4 Å can be assigned to the lattice
spacing of the (300) and (211) planes of HMO, respectively, and
the (300) plane is parallel to the side line; (ii) A cross angle
significantly lower than that over the single-atom alkali catalysts.
/HMO is slightly more active than Rb /HMO, and Rb /HMO is
the same active as Na /HMO prepared using a similar method
K
1
1
1
1
(Figure S4). Over 95% XHCHO are obtained on these single-atom
o
o
between the (211) plane and the side line is ~60 . After loading
alkali catalysts at 110 C, whereas HMO only gives 20% XHCHO
o
alkalis, the one-dimensional morphologies of HMO remain
unchanged. The appearance of smooth surfaces in the TEM
images indicates the absence of KCl and RbOH nanoparticles
on HMO surfaces, while the presence of K and Rb species is
confirmed by energy dispersive X-ray spectrometer (EDX)
mapping data. The K/Mn or Rb/Mn molar ratios obtained from
at 110 C. The single-atom alkali catalysts can even exhibit
o
significant catalytic activities below 90 C, but HMO is almost not
o
active below 90 C, implying that the single alkali metal ions
anchoring on HMO surfaces serve as catalytically active sites for
the low-temperature oxidation of HCHO. Similar results using
alkalis as catalytic sites were also reported in the oxidation of
[
35,36]
EDX determine the formula of K0.9Mn
8
O
16 and Rb1.0Mn
8
O16, in
soot and volatile organic compounds.
a 1 1
The apparent activation energy (E ) over K /HMO or Rb /HMO
line with the results obtained by X-ray Fluorescence
Spectrometer (XRF) analyses. Consequently, the HMO tunnels
are filled with the alkali ions, and alkali ions are exposed on the
-
1
-1
is ~84 kJ mol in HCHO oxidation, approximately 5 kJ mol
+
lower than that of HMO (Figure 4b). Moreover, when Rb in the
[
23]
HMO(001) planes, as modelled in insets of Figure 3d,f. These
form of RbOH was supported on TiO
HCHO oxidation (Figure S4). Considering the decrease of E
after alkali loadings and the inert alkali ions when loaded on TiO
2 2
, RbOH/TiO is inactive to
a
2
,
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
it is convincing that the catalytic performance of the single alkali
adatoms on HMO is triggered by the interactions among alkalis
and HMO.
alkali ions are very active for activating O
atom alkali catalysts have significantly strong abilities for
activating both molecular O and surface lattice oxygen.
2
. Therefore, the single-
2
Figure 4. (a) XHCHO and CO
2
selectivity as a function of reaction temperature
/HMO, or HMO. (b) Arrhenius plots for the reaction rates
) with the corresponding activation energy (E ) of the HCHO oxidation over
/HMO, Rb /HMO, or HMO.
Figure 5. Reaction kinetics data of the catalytic oxidation of HCHO. Reaction
o
(T) over K
1
/HMO, Rb
1
orders (k
r
) for (a) HCHO and (b) O
/HMO.
2
at the reaction temperature of 90 C over
(R
r
a
HMO, K /HMO, or Rb
1
1
K
1
1
The oxygen activation of the catalysts is essentially
associated with the nature of the active sites, and thus we used
X-ray absorption spectroscopies (XAS) to investigate the
electronic structures of the single alkali ions. Figure 6 shows the
The mechanism of catalytic oxidation of HCHO at low
[
21,37]
temperatures often follows the Mars-van Krevelen model,
and thus the dissociation of O and the oxidation of adsorbed
formate intermediate are considered as two critical steps.
2
[37]
To
1 1
XANES data of K /HMO and Rb /HMO together with two
determine the rate-determining step, we studied the surface
reaction kinetics of HCHO oxidation at low temperatures over
references at the K and Rb K-edges. Since the 1s  3d pre-
edge features are sensitive to the coordination geometry of
[
38]
HMO, K
shown in Figure 5a, the reaction orders (k
samples are equal (k = -0.3), indicating that the oxidation of
adsorbed formate intermediate is not the rate-determining step.
The k of O is only 0.2 for HMO, and it dramatically increases to
.9 for K /HMO and 0.8 for Rb /HMO (Figure 5b). Because of
the negative k for HCHO over both HMO and the single-atom
alkali catalysts, the activation of O will be inhibited by surface
adsorbed HCHO until surface lattice oxygen species are active
enough to react with the adsorbed HCHO to produce CO and
O, concomitant with the generation of oxygen defects. The
almost first-order dependence on O concentration for the
1
/HMO, and Rb
1
/HMO at the XHCHO less than 30%. As
excited atoms, the alkali ions in KCl or RbOH do not have the
r
) of HCHO over three
pre-edge absorption peaks because of the 1s  3d dipole-
[
39]
r
forbidden transitions, indicating the absence of the d-sp orbital
hybridization, consistent with their stable outermost electronic
r
2
configurations. Particularly, the alkali ions of K /HMO and
1
0
1
1
1
Rb /HMO with a D4h symmetry have relatively intense pre-edge
peaks, indicating the presence of the 1s  3d dipole-allowed
transition. The results demonstrate that the alkali ions have d-sp
r
2
[
38]
hybridized orbitals.
The alkali d-sp hybridizations may
3
2
originate from the electron transfer from O sp orbitals with lone-
pair electrons to unoccupied alkali d-sp frontier orbitals, resulting
in electron-rich alkali atoms, analogous to metallic alkali atoms
H
2
2
[40]
single-atom alkali catalysts implies that the HMO lattice oxygen
species in contact with the single alkali ions are highly active for
with the outermost hybridized orbitals.
2 2
converting adsorbed HCHO to CO and H O, and that the single
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
states of the single alkali adatoms is mainly responsible for the
strong ability towards activating O .
2
1 1
Figure 6. XANES spectra of (a) K /HMO and KCl, and (b) Rb /HMO and
RbOH/TiO at the K and Rb K-edges. Insets: spectral fitting edges, the pink
2
shadows showing the 1s  3d transition, and models of (a) KO and (b) RbO
8 8
structure motifs.
Figure 7. (a) K 2p XPS of K
1
/HMO and KCl showing the shift of binding energy.
(
b) Rb 3d XPS of Rb /HMO and RbOH showing the shift of binding energy.
1
The oxidation states of the alkali atoms of the single-atom
alkali catalysts were determined by using K 2p and Rb 3d XPS.
The electron densities of states of single alkali adatoms are
much higher than the ideal alkali ions with a normal oxidation
state of +1 in KCl or RbOH. As shown in Figure 7a and
The activation of oxygen also includes the activation of lattice
oxygen besides molecular O . Figure 8 depicts the O 1s XPS of
HMO and K /HMO and Rb /HMO together with their differences.
Obviously, the O 1s XPS of K /HMO and Rb /HMO slightly shifts
to lower BE compared to HMO, illustrating that the O species
gains electrons after the loading of alkalis on HMO. The main
peak in the BE range of 529.6-529.8 eV is assigned to the
surface lattice oxygen (Osl) species of HMO, and the smaller
shoulder peak in a BE range of 530.9-531.3 eV is ascribed to
2
1
1
[23]
according to our recent report, it is reasonably deduced that
the potassium atoms of K /HMO have the electron density of
1
1
1
+
states higher than that of K , and the same results are also
obtained from density functional theoretical (DFT) calculation
shown in Figure S5. Likewise, the binding energy (BE) of the Rb
3d5/2 peak of RbOH is centered at 109.3 eV, characteristic of
+
[41]
Rb ,
whereas the BE of Rb 3d5/2 of Rb
1
/HMO decreases to
[42]
surface defect oxygen (Osd) species of HMO. In particular, the
O
~108.2 eV by a downshift of ~1.1 eV with respect to that of
sd species decrease while the Osl species increase in single-
RbOH (Figure 7b), indicative of the high electronic density of
states of Rb ions. The results are in line with the above XANES
data. The single alkali adatoms with the hybridized d-sp orbitals
and the high electronic density of states are energetically
atom alkali catalysts with respect to those of HMO, implying that
the concentration of the Osl species of HMO increases after the
loading of alkalis. On the basis of the Mars-van Krevelen model,
the Osl species are of great importance for low-temperature
oxidation of HCHO because Osl with nucleophilic property can
attack the electrophilic carbonyl group of HCHO readily.
Therefore, the enhancement of the Osl species concentration
with alkali loading could partially accounts for the activity
2
favorable for the adsorption and subsequent dissociation of O :
(
i) The orbital hybridization of single alkali adatoms by local
[43]
2
transfer of sp electrons to empty d shell allows O to approach
alkali atoms until chemisorption occurs because of the Pauli
repulsion weakening; (ii) the abundant electrons of single alkali
adatoms are favorable for charge transfer to antibonding *
1 1
improvement of K /HMO and Rb /HMO. Therefore, according to
the evidence and the related discussion, the enhancement of the
single-atom alkali catalysts mainly originates from the facile
orbitals of O
2
to weaken the O-O bond until dissociation
[
27]
occurs.
As a consequence, the high electronic density of
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
3
1 [47]
activation of molecular O
2
by electron-abundant alkali ions and
electrons in a high-spin electron configuration t2g
e
g
.
Possibly,
active lattice oxygen due to the alkali loading.
electron transfer from Mn to K via the bridging oxygen of K-O-
Mn occurs when K ions are inserted into HMO tunnels.
K /HMO
1
HMO
K /HMO-HMO
1
HMO-HMO
Osl
Osd
Rb /HMO
1
HMO
Rb /HMO-HMO
1
HMO-HMO
Osl
Osd
535
533
531
529
527
Binding Energy / eV
1 1
Figure 8. O 1s XPS of HMO and K /HMO and Rb /HMO together with their
differences (dotted lines). The shades in different colors show curve-fitting
results, and the shades in yellow and green represent Osd and Osl species,
respectively.
Figure 9. (a) Mn L
HMO and K /HMO. Inset: MnO
and Mn, respectively.
3
-edge and (b) O K-edge soft-X-ray absorption spectra of
The significant change of the catalytic performance of the inert
alkali metal ions when they are loaded on HMO may essentially
originate from the electronic metal-support interactions (EMSI).
Thus, we obtained soft-X-ray adsorption spectra of the samples
to shed light on such interactions. According to the forthcoming
1
6
octahedron. Red and blue balls represent O
To substantiate the above hypothesis, we recorded the O K-
edge soft-X-ray adsorption spectra of HMO and K /HMO. As
shown in Figure 9b, a distinct feature of two peaks with an
energy difference of 2.4 eV evidences that the O 2p orbitals
[
44,45]
1
DFT calculations and the previous reports,
the insulating
HMO will become conducting after alkalis are inserted into the
tunnels. Because the electronegativity of manganese is much
smaller than that of oxygen, the electrons transferred to alkalis
should mainly originate from manganese. Accordingly, we
studied the changes of the Mn L -edge soft-X-ray absorption
3
spectra of HMO after the insertion of K ions into the tunnels. The
results are shown in Figure 9a.
[46]
have hybridized with the Mn 3d orbitals. The high energy peak
intensity of K /HMO with the “e ” feature is slightly weaker with
respect to that of HMO, indicating that the occupied states of the
O 2p orbitals with the “e ” feature increase. By comparing the
peak intensity due to the Mn e orbitals with the intensity of the
O 2p orbitals with the “e ” feature, we conclude that electrons on
the Mn e orbitals have transferred into the unoccupied O 2p
1
g
g
g
g
According to the O
theory, two peaks in the Mn L
spectra of HMO and K /HMO are assigned to the 2p3/2  t2g and
transitions, and thus an energy difference between t2g
orbitals, a so-called octahedral ligand-field splitting
parameter 10Dq, is calculated to be 2.4 eV, consistent with the
h
symmetry of Mn and the crystal field
g
3
-edge soft-X-ray absorption
orbitals through hybridization. Therefore, the electron transfer
from manganese to alkalis through bridging oxygen allows
alkalis to be at the high electronic density of states.
In Figure 9b, another important characteristic of the O K-edge
soft-X-ray absorption spectra is the upshifted spectrum of
1
2p3/2  e
g
and e
g
[
46]
reported value.
the 2p3/2  e
HMO, elucidating that the unoccupied states increase for
/HMO, i.e., Mn loses part of e electrons after the K insertion
because the pristine HMO has a minority of Mn with 3d
In particular, the intensity of the peak due to
K
1
/HMO with respect to that of HMO, indicating that electron
g
transition is slightly stronger for K /HMO than
1
transfer from the Mn e orbitals to the unoccupied O 2p orbitals
g
leads to the electronic states of O to shift up to a higher energy
by 0.2 eV with comparison to that of HMO. As discussed above,
K
1
g
3
+
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
the electronic density of states of the high-energy O 2p orbitals
alkali ions, indicating that the number of the active Osl increases
and the redox ability is improved, consistent with the upshift in
+
with the “e
g
” feature increases due to the insertion of K into
HMO tunnels. Hence, the upshifted O 2p orbitals with the high
electronic density of states should improve the redox ability of
the single-atom alkali catalysts, which possibly accounts for the
1
energy of O K-edge soft-X-ray absorption spectra of K /HMO in
Figure 9b. Moreover, the abundance of active Osl species is also
favorable for improving the catalytic activity in HCHO oxidation.
production of the highly active lattice oxygen after loading alkalis. Note that after alkali loading, the high-temperature reduction
Figure 10 shows the H
2
temperature-programmed reduction
/HMO. As Figure 10a shows,
peak of HMO drastically shifts down to the low temperature
range, and the peaks due to the reduction of Osl are in a narrow
temperature window, implying that the alkali loading accelerates
the transfer rates of electrons produced in the redox reaction of
(
H
2
-TPR) profiles of HMO and K
1
HMO exhibits three reduction peaks. The first and second peaks
are located at ~269 and 279 C, respectively, while the third
o
o
o
weak broad peak starts at 300 C with a maximum at ~445 C.
The area ratio of H consumption (A 2) between the sum of
2
H with Obl. Therefore, the electron transfer between alkalis and
2
H
HMO enhances the redox ability, thus leading to the high
catalytic activity of the single-atom alkali catalysts in HCHO
oxidation.
former two peaks and the third peak is ~1.6:1, from which the
oxidation state of Mn can be calculated to be ~3.7, indicating the
3
+
4+
presence of Mn and Mn , and thus the formula of HMO can be
simplified to be MnO1.9. The first peak is due to the reduction of
the surface oxygen (Osd and Osl) species. The second and third
peaks can be assigned to the reduction of the bulk lattice
oxygen (Obl) species, i.e., the reduction of MnO1.9 to Mn
3
O
4
, and
to MnO. After alkali loading, two typical
-TPR profile of K /HMO can be observed: (i)
[
16]
the reduction of Mn
3 4
O
features of the H
2
1
the surface O species increase, (ii) the temperature required for
the reduction of the Obl species shifts down to the low-
temperature window.
Figure 11. DOS of the (a) Osp3 and (b) Mn 3dx2-y2 of HMO (black) and K
red). (c) Total DOS of HMO (black) and K /HMO (red).
1
/HMO
(
1
To investigate the EMSI in detail, we carried out DFT
calculations based on HMO and K /HMO. Figure 11a,b depicts
the electronic densities of states (DOS) of HMO and K /HMO.
1
1
+
The DOS of the Osp3 orbital bound to K shifts downward in
energy with respect to that of HMO, making the DOS of Mn 3dx2-
y2 orbitals shift down below the Fermi level (EFermi). The DOS
downshifts of both O and Mn demonstrate that the strong EMSI
occurrs along the K-O-Mn structural motif. These results agrees
fairly with the soft-X-ray absorption spectroscopy results in
Figure 9, which proved electron transfer from Mn to K via the
bridging O. Figure 11c shows the total DOS of HMO and
Figure 10. (a) H
curves by subtracting the H
represents the H consumption by active Osl at the vicinity of alkali ions.
2
-TPR profiles of HMO and K
1
/HMO, and (b) the difference
2
-TPR curve of HMO. The green shade in panel (b)
2
To shed light on the discrepancy after the alkali loading, the
difference curves are obtained by subtracting the H -TPR curve
of HMO, as shown in Figure 10b, and thus a peak centered at
2
K
1
/HMO. Interestingly, the EMSI induces an insulator-to-
conductor transition (HMO is insulating, but /HMO is
conducting), which is favorable for facilitating electron transfer in
K
1
o
~
230 C can be attributed to the reduction of Osl at the vicinity of
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
catalytic reactions, thus leading to the enhancement of redox
(2.000 g, 25.0 mmol) with RbOH (0.297 g, 2.9 mmol) and without further
annealing.
[
27]
ability (Figure 10) and catalytic activity (Figure 4).
alkali earth metal ions such as Ca were proved to be able to
Likewise,
2
+
[
48]
Catalyst characterizations
improve the oxidizing ability of manganese-oxido clusters
or
[
49]
iron-oxo complexs
by accelerating electron transfer in
The SXRD were performed at BL14B of the Shanghai Synchrotron
oxidation reactions. Therefore, the EMSI allows the DOS of the
alkali ions to increase by the electron transfer from Mn to K,
which improves the activation of O and lattice oxygen, thus
2
Radiation Facility (SSRF) at
a wavelength of 0.6887 Å. Rietveld
refinements of the diffraction data were performed with the FULLPROF
software package on the basis of the space group I4/m.
The specific surface area and pore volume were analyzed on a
leading to the high catalytic activity in HCHO oxidation.
Micromeritics Tristar II 3020 surface area and pore size analyzer by N
adsorption-desorption measurements at -196 C, and calculated by the
2
o
Brunauer-Emmett-Teller (BET) and the t-plot method, respectively.
The XAS of the K /HMO including the XANES spectra and EXAFS
1
Conclusions
spectra at the K K-edge were collected at BL4B7A of the Beijing
Synchrotron Radiation Facility (BSRF). The Rb K-edge XANES and
EXAFS spectra were collected at BL14W of the SSRF. The raw data
were analyzed using the IFEFFIT 1.2.11 software package. The soft-X-
ray absorption spectra at the Mn L-edge and O K-edge were obtained in
a total electron yield (TEY) mode at BL08U1A of the SSRF.
In summary, single-atom alkali catalysts were synthesized by
anchoring individual alkali atoms on the tunnel openings of HMO.
The structures were precisely determined by using SXRD, X-ray
absorption, and TEM. These single-atom alkali catalysts
exhibited high catalytic activity in the complete oxidation of
HCHO in the presence of
O
2
at low temperatures. The
TEM and HRTEM images were obtained on a JEM 2100F transmission
electron microscope with EDX for determination of elemental contents.
The XPS was obtained on Kratos Axis Ultra-DLD system with a charge
neutralizer and a 150 W Al (Mono) X-ray gun (1486.6 eV). The samples
were dried in vacuum before XPS analysis. The XPS were referenced to
the C 1s peak at the BE of 284.6 eV. Data analysis and processing were
undertaken using XPSPeak 4.1 with the Shirley type background.
characterization results demonstrated that the single alkali
atoms have the frontier d-sp hybridized orbitals via the strong
interaction with lattice oxygen, and that the electron transfer
g
from the Mn e orbitals to the alkali d-sp hybridized orbitals
through the alkali-O-Mn bridging O makes alkali atoms have
high electronic density of states. Hence, the electron-rich alkali
atoms with hybridized orbitals can facilitate the sticking and
The H -TPR profiles were measured by a thermal conductivity detector
2
(TCD) on a Micromeritics 2720 adsorption instrument (Micromeritics,
activation of O
absorption spectra and H
2
. The combination of O K-edge soft-X-ray
-TPR profiles elucidated that the
USA). A sample (50 mg) was loaded and reduced by 10.0 vol.% H
2
/Ar
-
1
o
-1
(
50 mL min ) at a heating rate of 10 C min . The A was calculated
H
2
2
and calibrated by pure CuO samples.
electronic states of the single-atom alkali catalysts shift up in
energy, so that the lattice oxygen shows strong redox ability.
Therefore, the high electron density of states of the single alkali
atoms and the high activity of lattice oxygen are responsible for
the high catalytic activity of the single-atom alkali catalysts in
HCHO oxidation.
Catalytic evaluations
The catalytic oxidation of HCHO was conducted in a fixed-bed quartz
reactor (inner diameter = 8 mm) at atmospheric pressure. A fresh catalyst
(
50 mg, 40–60 mesh) was loaded for each run. Gaseous HCHO was
generated by passing N over paraformaldehyde (96%, Acros) in a
container, which was immersed in a water bath to keep the temperature
2
o
Experimental Section
at 45 C. The HCHO/N
total gas stream of 140 ppm HCHO/10.0 vol.% O
gas hourly space velocity is 60,000 h . The gas analysis was undertaken
using an on-line Agilent 7890A gas chromatograph equipped with a TCD.
2
gas was mixed with another O
2
flow to give a
-
1
2
/N (100 mL min ). The
2
-
1
Catalyst preparation
The CO
890B gas chromatograph equipped with FID and Ni catalyst converter,
which could convert carbon oxides quantitatively into methane in the
presence of before the detector. No other carbon-containing
compounds except CO were detected in the effluents for the whole
tested temperature window and all tested catalysts. The kinetics of
HCHO oxidation was studied at 90 C over K /HMO, Rb /HMO, and HMO
1 1
at the XHCHO less than 30% to achieve the data of the intrinsic reaction
kinetics by excluding the external and internal diffusion. The HCHO
2
selectivity was correctly measured by using an on-line Agilent
[
32]
HMO was prepared by a refluxing route: MnSO
mol), (NH (34.200 g, 0.15 mol) and (NH
4
·H
2
O (25.350 g, 0.15
(99.000 g, 0.75
7
4
)
S O
2 2 8
4
)
2
SO
4
mol) were dissolved in deionized water (0.4 L) to give an aqueous
solution. Then the solution was heated and kept at 100 C for 12 h under
refluxing. The resulting slurry was filtered, washed with deionized water,
dried at 110 C for 12h and the obtained black solid was calcined in air at
H
2
o
2
o
o
o
4
00 C for 4h.
Single-atom alkali catalysts were prepared by an impregnation method.
In a typical synthesis, KCl (0.216 g, 2.9 mmol) or RbOH (0.297 g, 2.9
mmol) was added to a suspension (50 mL) containing HMO (2.000 g,
3.0 mmol) with vigorously stirring for 15 min. Then the suspension was
transfered to a rotary evaporator and kept the temperature at 80 C to
remove the excess water. The obtained dry samples were labeled as
KCl/HMO and RbOH/HMO, respectively. A /HMO (A = K or Rb) sample
1
was obtained after annealing KCl/HMO or RbOH/HMO at 400 C for 12 h
and the final powder was washed by deionized water at room
temperature. The alkali loadings (4.8 wt.% for K
Rb /HMO) were determined by using XRF on a Bruker-AXS S4 Explorer.
For comparison, RbOH/TiO was also prepared by impregnation of TiO
concentration ranged from 140 to 560 ppm, and the corresponding O
2
concentration ranged from 700 to 2100 ppm. In typical runs, the data
were recorded at the steady state.
2
o
Theoretical calculations
o
All of the features were based on the density functional theory as
implemented in the Vienna Ab-initio Simulation Package (VASP). The
Local Density Approximation (LDA) for the exchange and correlation
effects was used in the calculations. To describe the interactions
1
/HMO and 10.4 wt.% for
1
2
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
between the core and valence electrons, the pseudo potentials of
projector augmented wave method were used. The plane wave energy
was equal to 500 eV, and a 2 × 2 × 6 monkhorst-pack grid was used for
[18] X. Yu, J. He, D. Wang, Y. Hu, H. Tian, Z. He, J. Phys. Chem. C 2012,
116, 851–860.
[19] X. Yu, J. He, D. Wang, Y. Hu, H. Tian, T. Dong, Z. He, J. Nanopart.
Res. 2012, 14, 1–14.
ƙ-space sampling in the calculation of HMO or K
1
/HMO. For the
geometry optimization, the force and energy cutoff were set as 0.01 eV/Å
and 10 eV. The lattice constants and wyckoff sites informations for a
conventional tetragonal cell were listed in the Table S1 and S2. A
[20] H. Tian, J. He, L. Liu, D. Wang, Ceram. Int. 2013, 39, 315–321.
[21] Z. Huang, X. Gu, Q. Cao, P. Hu, J. Hao, J. Li, X. Tang, Angew. Chem.
Int. Ed. 2012, 51, 4198−4203.
-
6
supercell containing 8 formula units of MnO
2
was employed in the
[22] P. Hu, Z. Amghouz, Z. Huang, F. Xu, Y. Chen, X. Tang, Environ. Sci.
Technol. 2015, 49, 2384−2390.
+
calculations for K in the terminal the tunnels. For the DOS calculation of
the system, we have used LDA with Hubbard U corrections (LDA+U)
method, and the U and J values were set as 5.6 eV and 1.0 eV,
[23] F. Xu, Z. Huang, P. Hu, Y. Chen, L. Zheng, J. Gao, X. Tang, Chem.
Commun. 2015, 51, 9888−9891.
[
50,51]
respectively.
[24] J. Xu, T. White, P. Li, C. He, Y. F. Han, J. Am. Chem. Soc. 2010, 132,
13172−13173.
[
[
[
25] D. J. Dwyer, J. H. Hardenbergh, Appl. Surf. Sci. 1984, 19, 14−27.
26] G. Ertl, M. Weiss, S. B. Lee, Chem. Phys. Lett. 1979, 60, 391−394.
27] C. Janiak, R. Hoffmann, P. Sjövall, B. Kasemo, Langmuir 1993, 9,
Acknowledgements
3427−3440.
This work was financially supported by the NSFC (21477023
and 21777030). The SXRD and XAS measurements were
conducted at the SSRF and BSRF, Chinese Academy of
Sciences.
[28] J. Harris, S. Andersson, Phys. Rev. Lett. 1985, 55, 1583−1586.
[29] J. E. Post, R. B. Von Dreele, P. R. Buseck, Acta Cryst., B 1982, 38,
1056−1065.
[
[
30] R. D. Shannon, C. T. Prewitt, Acta Cryst., B 1969, 25, 925−946.
31] J. Vicat, E. Fanchon, P. Strobel, D. Tran Qui, Acta Cryst., B 1986, 42,
162−167.
Keywords: single-atom alkali catalysts • heterogeneous
catalysis • structure elucidation • electron transfer • oxygen
activation
[
32] P. Hu, M. E. Schuster, Z. Huang, F. Xu, S. Jin, Y. Chen, W. Hua, D. S.
Su, X. Tang, Chem. Eur. J. 2015, 21, 9619−9623.
[
[
33] J. Vogt, H. Weiss, Surf. Sci. 2001, 491, 155–168.
34] S. Odde, H. M. Lee, M. Kołaski, B. J. Mhin, K. S. Kim, J. Chem. Phys.
2
004, 121, 4665–4670.
[
1] X. Tang, Y. Bai, A. Duong, M. T. Smith, L. Li, L. Zhang, Environ. Int.
[
[
35] Q. Li, X. Wang, Y. Xin, Z. Zhang, Y. Zhang, C. Hao, M. Meng, L. Zheng,
L. Zheng, Sci. Rep. 2014, 4, 4725.
2009, 35, 1210−1224.
[
[
2] O. Wittmann, Eur. J. Wood Wood Prod. 1962, 20, 221−224.
36] Y. Chen, G. Tian, M. Zhou, Z. Huang, C. Lu, P. Hu, J. Gao, Z. Zhang, X.
Tang, Environ. Sci. Technol. 2016, 50, 5825–5831.
3] WHO Guidelines for Indoor Air Quality: Selected Pollutants, World
Health Organization Regional Office for Europe, Copenhagen, Denmark,
[
[
37] C. F. Mao, M. A. Vannice, J. Catal. 1995, 154, 230−244.
38] S. Bordiga, E. Groppo, G. Agostini, J. A. van Bokhoven, C. Lamberti,
Chem. Rev. 2013, 113, 1736−1850.
2010.
[
[
[
4] W. D. Kerns, K. L. Pavkov, D. J. Donofrio, E. J. Gralla, J. A. Swenberg,
Cancer Res. 1983, 43, 4382−4392.
[
[
[
39] L. G. Parratt, Rev. Mod. Phys. 1959, 31, 616−645.
40] L. Pauling, General Chemistry, BN Publishing, 2011.
41] R. Itti, I. Tomeno, K. Ikeda, K. Tai, N. Koshizuka, S. Tanaka, Phys. Rev.
B 1991, 43, 435–440.
5] J. A. Swenberg, W. D. Kerns, R. I. Mitchell, E. J. Gralla, K. L. Pavkov,
Cancer Res. 1980, 40, 3398−3402.
6] Integrated Risk Information System (IRIS): Formaldehyde, U.S.
Environmental Protection Agency, 2006.
[
[
[
42] Q. Tang, L. Jiang, J. Liu, S. Wang, G. Sun, ACS Catal. 2014, 4, 457–
[
[
7] T. Salthammer, Angew. Chem. Int. Ed. 2013, 52, 3320−3327.
8] Y. Kuwabara, G. V. Alexeeff, R. Broadwin, A. G. Salmon, Environ.
Health Persp. 2007, 115, 1609−1616.
4
63.
43] X. Tang, J. Chen, Y. Li, Y. Xu, W. Shen, Chem. Eng. J. 2006, 118,
19–125.
44] M. Toupin, T. Brousse, D. Bélanger, Chem. Mater. 2004, 16,
184−3190.
1
[
9] J. P. Pinto, G. R. Gladstone, Y. L. Yung, Science 1980, 210, 183−184.
[
[
[
[
[
[
[
[
10] J. Q. Torres, S. Royer, J. P. Bellat, J. M. Giraudon, J. F. Lamonier,
ChemSusChem 2013, 6, 578−592.
3
[
[
45] W. Xiao, H. Xia, J. Y. Fuh, L. Lu, J. Power Sources 2009, 193, 935−938.
46] M. Abbate, F. M. F. de Groot, J. C. Fuggle, A. Fujimori, O. Strebel, F.
Lopez, M. Domke, G. Kaindl, G. A. Sawatzky, M. Takano, Y. Takeda, H.
Eisaki, S. Uchida, Phys. Rev. B 1992, 46, 4511−4519.
11] X. Tang, J. Chen, X. Huang, Y. Xu, W. Shen, Appl. Catal. B: Environ.
2
008, 81, 115−121.
12] L. Ma, D. Wang, J. Li, B. Bai, L. Fu, Y. Li, Appl. Catal. B: Environ. 2014,
48−149, 36−43.
13] D. W. Kwon, P. W. Seo, G. J. Kim, S. C. Hong, Appl. Catal. B: Environ.
015,163, 436−443.
1
[
47] P. Umek, A. Gloter, M. Pregelj, R. Dominko, M. Jagodič, Z. Jagličić, A.
Zimina, M. Brzhezinskaya, A. Potočnik, C. Filipič, A. Levstik, D.
Arčon, J. Phys. Chem. C 2009, 113, 14798−14803.
2
14] J. Wang, J. Li, C. Jiang, P. Zhou, P. Zhang, J Yu, Appl. Catal. B:
Environ. 2017, 204, 147–155.
[
[
[
[
48] E. Y. Tsui, R. Tran, J. Yano, T. Agapie, Nature Chem. 2013, 5,
293−299.
15] J. Zhou, L. Qin, W. Xiao, C. Zeng, N. Li, T. Lv, H. Zhu, Appl. Catal. B:
Environ. 2017, 207, 233–243.
49] Y. Morimoto, H. Kotani, J. Park, Y. Lee, W. Nam, S. Fukuzumi, J. Am.
Chem. Soc. 2011, 133, 403−405.
16] T. Chen, H. Dou, X. Li, X. Tang, J. Li, J. Hao, Micropor. Mesopor. Mater.
50] F. Zhou, M. Cococcioni, C. A. Marianetti, D. Morgan, G. Ceder, Phys.
Rev. B 2004, 70, 235121.
2
009, 122, 270−274.
17] S. Liang, F. Teng, G. Bulgan, R. Zong, Y. Zhu, J. Phys. Chem. C 2008,
12, 5307–5315.
51] D. A. Tompsett, M. S. Islam, Chem. Mater. 2013, 25, 2515−2526.
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201704398
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Inert alkali metal ions were
successfully activated in completely
oxidation of HCHO at low
Jiayi Gao, Zhiwei Huang, Yaxin Chen,
Jing Wan, Xiao Gu, Zhen Ma, Jianmin
Chen, Xingfu Tang*
temperatures by electron transfer from
Mn to K via the bridging oxygen of K-
O-Mn when anchored onto hollandite
manganese oxide surfaces.
Page No. 1 – Page No. 9
Activating Inert Alkali Metal Ions by
Electron Transfer from Manganese
Oxide for Formaldehyde Abatement
Layout 2:
FULL PAPER
Author(s), Corresponding Author(s)*
Page No. – Page No.
((Insert TOC Graphic here; max. width: 11.5 cm; max. height: 2.5 cm))
Title
Text for Table of Contents
This article is protected by copyright. All rights reserved.