Communication
www.advmat.de
Electron-State Confinement of Polysulfides for Highly
Stable Sodium–Sulfur Batteries
Chao Ye, Yan Jiao, Dongliang Chao, Tao Ling, Jieqiong Shan, Binwei Zhang, Qinfen Gu,
Kenneth Davey, Haihui Wang,* and Shi-Zhang Qiao*
Confinement of polysulfides in sulfur cathodes is pivotal for eliminating the
“shuttle effect” in metal–sulfur batteries, which represent promising solu-
tions for large-scale and sustainable energy storage. However, mechanistic
exploration and in-depth understanding for the confinement of polysulfides
remain limited. Consequently, it is a critical challenge to achieve highly
stable metal–sulfur batteries. Here, based on a 2D metal–organic framework
(2D MOF), a new mechanism to realize effective confinement of polysulfides
is proposed. A combination of in situ synchrotron X-ray diffraction, elec-
trochemical measurements, and theoretical computations reveal that the
dynamic electron states of the Ni centers in the 2D MOF enable the interac-
tion between polysulfides and the MOF in the discharge/charge process to
be tuned, resulting in both strong adsorption and fast conversion kinetics
of polysulfides. The resultant room-temperature sodium–sulfur batteries are
amongst the most stable reported so far, thus demonstrating that the new
mechanism opens a promising avenue for the development of high-perfor-
mance metal–sulfur batteries.
As a promising electrode material, sulfur benefits from low cost
and high theoretical specific capacity of ≈1675 mAh g−1.[1] Addi-
tionally, sulfur cathodes can be conjugated with a range of metal
anodes in metal–sulfur batteries with high energy density of
≈350 W h kg−1 (based on devices). This holds promise for prac-
tical energy-storage applications.[2] However, a major drawback
C. Ye, Dr. Y. Jiao, Dr. D. Chao, J. Shan, Dr. B. Zhang, Dr. K. Davey,
Prof. S.-Z. Qiao
School of Chemical Engineering and Advanced Materials
The University of Adelaide
Adelaide, SA 5005, Australia
E-mail: s.qiao@adelaide.edu.au
Prof. T. Ling, Prof. S.-Z. Qiao
School of Materials Science and Engineering
Tianjin University
Tianjin 300072, China
Dr. Q. Gu
Australian Synchrotron (ANSTO)
800 Blackburn Rd, Clayton, VIC 3168, Australia
Prof. H. Wang
School of Chemistry and Chemical Engineering
South China University of Technology
Guangzhou 510640, China
E-mail: hhwang@scut.edu.cn
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adma.201907557.
DOI: 10.1002/adma.201907557
Adv. Mater. 2020, 1907557 1907557  (1 of 8)
with metal–sulfur batteries is the “shuttle
effect.” This arises from solubility of the
intermediate metal polysulfides, together
with a parasitic reaction between the poly-
sulfides and the metal anodes. This leads
to a low Coulombic efficiency (CE) and a
rapid capacity decay in metal–sulfur bat-
teries.[3] For example, room-temperature
sodium–sulfur (RT Na–S) batteries benefit
from low cost of Na compared with the
other metal–sulfur batteries. However, its
practical application is severely hindered
by the shuttle effect because of the high
reactivity of Na.[4] Therefore confinement
of polysulfides in sulfur cathodes is critical
to long-term stability and practical applica-
tion of metal–sulfur batteries.[2a,5]
Currently, polysulfides confinement
with weak physical adsorption on carbon-
based materials would inevitably inhibit
their conversion kinetics, resulting in
a compromise between high discharge
capacity and low capacity decay of the sulfur cathodes.[6] On
the other hand, too strong polysulfides adsorption would cause
decomposition of the polysulfides, leading to blockage of the
adsorption sites.[7] The keys for efficient polysulfides confine-
ment lie in appropriate adsorption of the polysulfides and their
facilitated conversion kinetics, which are largely determined
by local electronic state of a sulfur cathode material.[8] For
example, on N-doped graphene, pyridinic N with an extra pair
of electrons interacts with terminal Li in lithium polysulfides
to form Li bond, which greatly promotes lithium polysulfides
confinement.[9] However, systematic investigation to correlate
local electron state with adsorption behaviors of a series of
polysulfides remains very limited, which is due to the com-
plexity of the polysulfide conversion process. For example, in
the case of the RT Na–S batteries, at least four sodium poly-
sulfides (NaPoSs) including Na2S5, Na2S4, Na2S2, and Na2S are
involved.[10] Consequently, realizing efficient polysulfides con-
finement in sulfur cathodes has been a challenging topic in
high-energy-density metal–sulfur batteries.
Although polysulfides are too sensitive to be detected
in air, advances in in situ synchrotron-based characteriza-
tions permit the identification of specific polysulfides and the
tracking of dynamic macroscopic polysulfides conversions.
In-depth comprehension can thus be offered for macroscopic
polysulfides conversion kinetics.[11] Nevertheless, atomic-level
understanding for the polysulfides adsorption behaviors is
imperative and remains difficult to achieve experimentally.[12]
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Extraordinary progress in density functional theory (DFT)
computations that take into account the surface chemistry and
local electron state of the sulfur cathode materials is essential
for investigating the polysulfides adsorption behaviors.[13] For
instance, d-orbital electron numbers of the transitional metal
sulfides strongly determine adsorption energies of lithium
polysulfides.[14] Therefore, combination of the advanced in situ
synchrotron characterizations and computational quantum
chemistry can reveal the mechanism of polysulfides confine-
ment on cathode materials. One can engineer potential sulfur
cathode materials with efficient polysulfides confinement by
tailoring their electron states. It is thus urgently required for
introducing the aforementioned advanced methodology to the
field of metal–sulfur batteries.
Here, we report a new mechanism for polysulfides con-
finement based on a 2D metal–organic framework (2D
MOF). Using the MOF as a model together with a judicial
combination of in situ synchrotron X-ray diffraction (XRD),
electrochemical tests, and DFT computations, we demonstrate
that dynamic electron states of Ni centers in the MOF enable
tuning of interaction between polysulfides and the MOF in
the discharge/charge process, facilitating strong polysulfide
adsorption and rapid polysulfides conversion kinetics. The
resulting performance of the RT Na–S batteries is amongst the
best ones reported so far. We report for the first time the cor-
relation between the local electron states of the sulfur cathode
materials and facilitated polysulfides conversion kinetics. Our
findings offer a practical means to address the shuttle effect in
metal–sulfur batteries and guide the design of sulfur cathode
materials for highly stable metal–sulfur batteries.
A hydrothermal method was adopted to fabricate a single
crystal bulk Ni2(PymS)4 MOFs (Ni-MOF-bulk, PymSH = 2-mer-
captopyrimidine).[15] Morphology of the Ni-MOF-bulk was
characterized by scanning electron microscopy (SEM) to reveal
its layered structure (Figure S1, Supporting Information).
High-angle annular dark-field scanning transmission-elec-
tron microscopy (HAADF-STEM) images and corresponding
energy-dispersive spectroscopy (EDX) elemental maps revealed
a uniform distribution of C, N, S, and Ni (Figure S2, Sup-
porting Information). Mechanical exfoliation was used on the
Ni-MOF-bulk via wet ball-milling in acetone. Because of weak
interaction between its layers, the Ni-MOF-bulk was readily
exfoliated into 2D MOF nanosheets (Ni-MOF-2D).[16] This was
evidenced by a broad signal from 3100 to 3150 cm−1 in the
Raman spectrum that is assigned to N···H hydrogen bonds
between layers of the Ni-MOF-bulk (Figure S3, Supporting
Information). By contrast, a negligible corresponding signal in
the Ni-MOF-2D would imply breaking of the N···H hydrogen
bonds as a result of exfoliation.[17] The schematic shows this
layered structure of Ni-MOF-2D with a (002) interlayer-space
of 0.84 nm (Figure  1a). XRD patterns demonstrate that the
Ni-MOF-2D maintains the original crystalline phase of the Ni-
MOF-bulk (Figure 1b). The inset high-resolution transmission
electron microscopy (HRTEM) image shows an outstanding
(002) peak in Ni-MOF-2D pattern to evidence successful exfo-
liation of the bulk MOF.[15] Representative SEM and trans-
mission electron microscopy (TEM) measurements were per-
formed to aid visualization of the morphology of Ni-MOF-2D.
These showed a 2D nanosheet structure (inset of Figure 1c;
Figures S4a and S5a, Supporting Information). Atomic force

Figure 1.  Structural characterization and electron-state analyses of Ni-MOF-2D nanosheets. a) Schematic of Ni-MOF-2D with few-layer structure from
side-view and top-view, in which the blue, light-gray, yellow, purple, and white spheres represent, respectively, N, C, S, Ni, and H. b) XRD patterns of
Ni-MOF-bulk and Ni-MOF-2D. Inset is HRTEM image of the few-layer Ni-MOF-2D. c) HRTEM and d) HAADF-STEM image and corresponding EDX
mapping images of Ni-MOF-2D. The inset of (c) is a representative TEM image of Ni-MOF-2D. e) Ni L-edge, f) S K-edge, g) C K-edge, and h) N K-edge
NEXAFS spectra of Ni-MOF-2D and Ni-MOF-bulk. Insets show the enlarged regions for charge transfer information.

Adv. Mater. 2020, 1907557 1907557  (2 of 8) © 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
microscopy images confirmed this 2D morphology with a
thickness of ≈8.6 nm (Figure S4b, Supporting Information).
HRTEM images (Figure 1c) together with the selected-area-
electron-diffraction pattern (Figure S5b, Supporting Informa-
tion) gave measured lattice distances of 0.23 and 0.27 nm,
corresponding, respectively, to the (244) and (241) facets of
Ni2(PymS)4 MOFs. To provide a more detailed composition of
Ni-MOF-2D, HAADF-STEM image with corresponding EDX
elemental maps is presented as Figure 1d. This shows that the
C, N, S, and Ni elements are uniformly distributed and con-
firms that the Ni-MOF-2D maintains the physical phase of pris-
tine Ni-MOF-bulk.
Because it is reported that an optimized electron-state of
cathode materials is favorable for improving discharge capacity
and cycling stability of sulfur cathodes.[18] We performed DFT
calculations and conducted Bader charge analysis to predict the
charge redistribution from the Ni-MOF-bulk to the Ni-MOF-
2D.[19] Findings show that ≈10% of electrons from S transfer to
Ni in processing of Ni-MOF-bulk to Ni-MOF-2D (Figure S6 and
Table S1, Supporting Information). Synchrotron-based near-
edge X-ray absorption fine structure (NEXAFS) characteriza-
tions were performed on both Ni-MOF-bulk and Ni-MOF-2D to
investigate any structural impact on electron transfer behavior.
NEXAFS is a powerful tool to provide reliable information on
local electron-states. According to the white-line adsorption
energies, Figure 1e and inset present a clear view of the Ni
L-edge spectra in which Ni-MOF-2D exhibits increased electron
density following exfoliation of Ni-MOF-bulk. For the S K-edge,
the peak at around 2473.0 eV can be assigned to S–Ni coordi-
nation species.[20] The increased intensity of the signal in Ni-
MOF-2D compared with that in Ni-MOF-bulk demonstrate a
decreased electron density of S following exfoliation (Figure 1f).
X-ray photoelectron spectra measurements were performed
on these two elements of the Ni-MOF-bulk and Ni-MOF-2D to
substantiate electron transfer behavior (Figure S7, Supporting
Information).[21] Compared with Ni-MOF-bulk, Ni-MOF-2D
showed a lower Ni valence state and a higher S valence state,
which is consistent with the computational results and indicates
the electron transfer from S to Ni. For the C and N K-edge, the
peak at around 287.0 and 399.2 eV can be assigned to CNC
species (Figure 1g,h). The increased intensity of this feature in
C K-edge and decreased intensity of the feature in N K-edge
implies an electron transfer from C to N.[22] Overall, these find-
ings confirm a charge redistribution during the exfoliation
from Ni-MOF-bulk to Ni-MOF-2D. This can be attributed to the
breaking of weak interlayer interactions (e.g., N***H bonds),
which is suggested by the decrease of corresponding Raman
signal after exfoliation (Figure S3, Supporting Information).[23]
To investigate the electrochemical performance of the as-pre-
pared Ni-MOFs in RT Na–S batteries we assembled three sulfur
electrodes of S/Ni-MOF-2D, S/Ni-MOF-bulk and S/conductive
carbon using active materials with a sulfur mass ratio of 48.6%
(Figure S8, Supporting Information). Long-term cycling experi-
ments at a high rate of 1 C and 2 C were conducted to investi-
gate electrochemical performance of S/Ni-MOF-2D (Figure 2a;
Figures S9–S11, Supporting Information). High capacities of
347 and 241 mAh g−1 were maintained after 1000 continuous
cycles under 1 C and 2 C that refer respectively to low capacity
decay of 0.042% and 0.052% per cycle, together with a stabilized
Adv. Mater. 2020, 1907557 1907557  (3 of 8)
www.advmat.de
CE held at around 100%. As is seen in Figure 2b, S/Ni-MOF-2D
exhibited a series of advantageous discharge capacities of 516,
416, 372, 331, and 284 mAh g−1 when cycled at, respectively,
0.1, 0.2, 0.5, 1, and 2 C (1 C = 1675 mA g−1). When current
density was switched back to 0.2 C a high discharge capacity of
406 mAh g−1 was maintained. By contrast, S/Ni-MOF-bulk and
S/conductive carbon cathodes showed poor rate performance
with limited capacities under high rate (Figure S12, Supporting
Information). As is shown in Figures S13 and S14 following
further cycling at 0.2 C, the S/Ni-MOF-2D exhibited an ultralow
capacity decay of 0.024% per cycle (406 to 313 mAh g−1 in
970 cycles). This was significantly better than that for either
S/Ni-MOF-bulk (0.074%) or S/conductive carbon (0.095%).
The intact electrode structure of the S/Ni-MOF-2D compared
with the other two electrodes after cycling further confirms the
best cycling stability among the three electrodes (Figure S15,
Supporting Information). This excellent cycling stability of
S/Ni-MOF-2D is superior to those of the metal oxides, metal
sulfides, or hybrid sulfur cathodes reported recently (Figure 2c;
Table S2, Supporting Information). In addition, we prepared a
sulfur electrode with higher sulfur loading of ≈2.6 mg, which
exhibits capacity of 244 mAh g−1 and CE of around 100% after
200 cycles under 0.2 C, indicating the high promise of this
strategy to promote the future practical application of Na–S bat-
teries (Figure S16, Supporting Information).
Cyclic voltammetry (CV) tests under a low scan rate of
0.1 mV s−1 between 0.5 and 2.8 V were performed to investigate
electrochemical behavior of sulfur and electrochemical evolu-
tion of NaPoSs from S/Ni-MOF-2D cathode (Figure 3a). During
the initial cathodic scan, two prominent peaks centered at
around 2.08 and 0.73 V were observed. These corresponded to
the solid–liquid transition from elemental sulfur to Na2Sx (long-
chain NaPoSs, 6 ≤ x ≤ 8) and a further reduction to less-soluble
Na2S2 and Na2S. In the ensuing cathodic scan, two strong
cathodic peaks at 1.59 and 0.94 V were observed that were
highly reproducible. These are related to reduction of Na2Sx.
For the anodic scan, only one reproducible peak at 2.04 V was
observed. This corresponds to oxidation of Na2S2 and Na2S into
Na2Sx. The lower potential gap between first reduction and oxi-
dation peaks in S/Ni-MOF-2D suggests lower polarization and
faster conversion kinetics of the sodium polysulfides compared
with those for S/Ni-MOF-bulk (Figure S17, Supporting Infor-
mation).[24] Additionally, we estimated the voltage gaps (ΔE) at
50% depth of discharge. These are closely related to electrode
polarization (Table S3, Supporting Information). S/Ni-MOF-2D
exhibited a lower ΔE2D than either S/Ni-MOF-bulk or S/con-
ductive carbon. This finding indicates significant enhancement
of NaPoSs redox kinetics on the Ni-MOF-2D and is consistent
with the CV observations.[25]
To further investigate the NaPoSs’ electrochemical and
adsorption behavior on Ni-MOF-2D, in situ synchrotron XRD
measurements were conducted in transmission mode. A modi-
fied 2032-type coin cell of S/Ni-MOF-2D was performed using
an in-house design (Figure S18, Supporting Information).[26]
As can be seen in Figure 3b during the initial discharge from
≈2.2 V, two peaks at 20.2° and 12.9° were observed in the XRD
patterns. These are assigned to organized layers of soluble
Na2Sx with relative ordered structure.[27] When the battery was
discharged to ≈1.6 V, three new peaks emerged at 11.9°, 20.1°,
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

a 1000

100
800

Colulombic efficiency (%)

S/Ni-MOF-2D
Capacity (mAh g-1) S/conductive carbon 80
600 S/Ni-MOF-bulk

60
400
40

200
20

0 0
0 100 200 300 400 500 600 700 800 900 1000
Cycle
b 1000 S/Ni-MOF-2D
c 0.02
S/conductive carbon
S/Ni-MOF-bulk This work
800

Capacity decay (cycle-1, %)

0.04
[S5]
Capacity (mAh g )

0.06
-1

0.1C

[S10]
600 0.08 [S11]
0.1
0.2C

[S12]
0.2C

0.5C

[S13]
400
1C

0.2
2C

[S14]
[S8]
[S15] [S9]
200 0.4 [S7]
[S16]
0.6 [S6]
0.8
0 1
5 10 15 20 25 30 0 300 600 900 1200
Cycle Cycle number

Figure 2. Electrochemical performance of S/Ni-MOF-2D in RT Na–S batteries. a) Cycling performances and Coulombic efficiencies (CE) of the
S/Ni-MOF-2D, S/Ni-MOF-bulk, and S/conductive carbon at 1 C after 0.1 C activation. b) Rating capacities of the three sulfur electrodes. c) Comparison
of cycle numbers and capacity retentions of recently reported RT Na–S batteries with the current work in which deeper color refers to greater current
density. References for [S5] etc. are given in the Supporting Information.

a b 12
charge

0.1 2.04 V
10

0.0 8
Current (mA)

Time (h)

1.59 V 2.08 V 6
-0.1
0.94 V
4

-0.2
2
0.73 V
discharge
0
0.5 1.0 1.5 2.0 2.5 3.0 3 2 1 0 -110 11 12 13 20 21
Voltage (V, vs Na +/Na) Voltage (V) 2 Theta

Figure 3.  Electrochemical behavior of NaPoSs on Ni-MOF-2D. a) CV curves of S/Ni-MOF-2D at scan rate of 0.1 mV S−1. b) Initial galvanostatic dis-
charge/charge curve and corresponding in situ synchrotron XRD patterns of S/Ni-MOF-2D in which the colored patterns indicate major changes along
the discharge/charge process; the yellow and cyan spheres represent S and Na atoms, respectively.

Adv. Mater. 2020, 1907557 1907557  (4 of 8) © 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
and 20.6°, corresponding, respectively, to (131), (341), and (302)
facets of Na2S5 (JCPDF no. 77–0294). This finding reveals that
the peak in the CV scan at 1.59 V is a result of the reduction of
Na2Sx to Na2S5. The peak at 11.6° is derived from the (220) facet
of Na2S4 (JCPDF no. 71–0516) and was generated when the bat-
tery was discharged further to ≈1.0 V. This indicates that the
peak at 0.94 V in the CV curves can be reliably assigned to
the formation of Na2S4. This peak for Na2S4 faded into a new
(200) peak for Na2S2 (JCPDF no. 81–1771) at 12.1° when the
battery was further discharged to 0.8 V. The final discharge to
≈0.7 V revealed a peak at 10.6°. This is attributed to the (111)
facet of Na2S (JCPDF no. 77–2149). In the charge from 0.5 to
2.8 V, most of the Na2S in the battery was oxidized to Na2Sx. It
is notable that broader peaks are related to Na2S2, Na2S4, and
Na2S5 when compared with narrow peaks for Na2Sx. This is
attributed to the formation of an organized Na2S4 and Na2S5
layer that is the result of strong adsorption on Ni-MOF-2D.[28]
Importantly, overall results from the combination of the in situ
synchrotron XRD and CV measurements underscore the facili-
tated conversion kinetics of all NaPoSs on Ni-MOF-2D.
To study the origin of the electrochemical performance of
Ni-MOF-2D, we used PymSH and another Ni-based MOFs
(Ni-BDC MOFs) as controls and investigated the individual
roles of the heterocyclic linker and the Ni center (Figures S19
and S20, Supporting Information). Results showed that both
PymSH (without the Ni) and Ni-BDC MOFs (without the
PymSH ligand) are inactive in sulfur cathodes. To provide an
objective view of electrochemical performance of Ni-MOF-2D,
we evaluated capacities normalized by electrochemically active
surface areas (ECSAs) on S/Ni-MOF-2D, S/Ni-MOF-bulk, and
S/conductive carbon cathodes (Figure S21, Supporting Infor-
mation).[29] S/Ni-MOF-2D exhibted the greatest ECSA together
with greatest normalized capacity under 0.2 C, 1 C, and 2 C
(Table S4, Supporting Information). The former finding is
attributed to the more exposed active sites of the 2D mor-
phology. The latter demonstrates a highly efficient NaPoSs
confinement on Ni-MOF-2D from strong chemical adsorption
and fast redox kinetics toward NaPoSs.[30] It is concluded there-
fore that the linkers and Ni centers synergistically contribute
to NaPoSs adsorption and conversion kinetics on Ni-MOF-2D.
To explore the mechanism of NaPoSs adsorption on Ni-MOF-
2D, a series of spectra surveys were carried out. Raman spectra
of Ni-MOF-2D and Na2S5-treated Ni-MOF-2D (Ni-MOF-2D-
Na2S5) were investigated to study the change in the chemical
environment of the heterocyclic linker. As is shown Figure 4a,
right inset, stretching bands (ν ring) of heterocyclic linker ring
were observed in the range of 1450–1650 cm−1. These are good
indicators of the electron density of the linker.[31] The bands cen-
tered at 1565 and 1541 cm−1 in the Ni-MOF-2D actually down-
shift to 1563 and 1533 cm−1 following treatment with Na2S5.
This finding suggests that the heterocyclic linker especially
electron-rich N interacts with Na2S5 via electron transfer.[32]
Importantly, the left inset shows that the CS bond stretching
vibration band (ν CS) at 1167 cm−1 in Ni-MOF-2D upshifts
to 1179 cm−1 in Ni-MOF-2D-Na2S5. This finding indicates a
decrease in electron density on S following interaction with
Na2S5. NEXAFS characterizations were carried out on Ni-MOF-
2D-Na2S5 to investigate the change in local electron density. As
is shown in the regions of Ni L-edge, a decreased Ni electron
Adv. Mater. 2020, 1907557 1907557  (5 of 8)
www.advmat.de
density for Ni-MOF-2D-Na2S5 is demonstrated by the increased
white-line adsorption energy of Ni L-edge (Figure 4b). In the
C K-edge, no electron gain or loss is observed (Figure 4c). It
is therefore concluded that this interaction leads to decreased
electron density on S and N and is responsible for the conju-
gation effect in the heterocyclic linker rings. This is different
to reported open-metal sites in 3D MOFs in which electron
density would be expected to increase due to interaction with
S in the lithium polysulfides.[33] Our results imply therefore
that electrons of Ni transferred to NaPoSs through the linker.[34]
More importantly, even lower Ni electron densities in Ni-MOF-
2D-Na2S2 and Ni-MOF-2D-Na2S are demonstrated by higher
white-line adsorption energy of Ni L-edge compared with that in
Ni-MOF-2D-Na2S5 (Figure S22, Supporting Information). This
indicates that the electron state of the Ni centers is dynamic
during discharge/charge processes. By contrast, Ni-MOF-bulk
exhibited weak interaction with Na2S5 as is evidenced in the
Raman and NEXAFS spectra (Figure S23, Supporting Informa-
tion). To confirm NaPoSs adsorption capabilities of Ni-MOF-
2D, Ni-MOF-bulk, and the linker PymSH, samples were treated
with a 5 × 10−3 m Na2S5 solution for 1 h. As is illustrated in
Figure S24 of the Supporting Information, the Ni-MOF-2D pos-
sesses the highest adsorption capabilities based on same weight
or same surface area in comparison with Ni-MOF-bulk, con-
ductive carbon, and PymSH. Overall, these results underscore
that NaPoSs are adsorbed on Ni-MOF-2D through interaction
with the heterocyclic linker as an electron donor and NaPoSs
as an electron acceptor. The N/S sites of Ni-MOF-2D provide
effective polysulfides adsorption sites which are lacking on the
conductive carbon. This is the reason that the S/Ni-MOF-2D
significantly outperforms S/conductive carbon. Notably, elec-
tron transfer behaviors are also observed on Ni centers, which
might contribute to the interaction between the polysulfides
and Ni-MOF-2D.
To further gain atomic-level insights into the polysulfides
confinement on Ni-MOF-2D, we performed theoretical calcula-
tions. The (002) facet was determined as an exposed surface of
the Ni-MOF-2D based on the HRTEM images. Ab initio mole-
cular dynamics (AIMD) calculations were applied to investigate
NaPoSs adsorption energies on Ni-MOF-2D. The energy pro-
file (Ea) of all five Na2Sn molecules (n = 1, 2, …, 5) adsorbed on
the Ni-MOF-2D surface (Ni-MOF-2D-Na2Sn) were evaluated for
4000 fs (Figure  5a). AIMD simulations predicted Ea increases
with the length of the Na2Sn chain as ≈−4.4, −4.0, −3.9, −2.6,
and −2.5 eV, respectively. These high adsorption energies indi-
cate strong adsorption of the NaPoS on Ni-MOF-2D when com-
pared with those of reported materials in sulfur cathodes for
lithium/sodium polysulfides adsorption.[6] DFT calculations
were carried out to optimize the Ni-MOF-2D-Na2Sn configura-
tions for analyses of electron transfer induced by the adsorption
(Figure S25, Supporting Information). The results implies that
N and S can function as a dominant adsorption site to interact
with Na in Na2Sn and the N/S can interact with Na in Na2Sn
as electron donors and acceptors through Na–N/S interaction.
This finding agrees well with both the Raman and NEXAFS
results. Importantly, as is shown in Figure 5b, with Na2Sn
adsorption energies increasing from Ni-MOF-2D-Na2S5 to Ni-
MOF-2D-Na2S, electron states of the Ni centers tend to change
from an electron accummulation state (in Ni-MOF-2D-Na2S5)
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

a
Ni-MOF-2D C-S ring
Ni-MOF-2D-Na2S5
1167 1563
1565
1179
Intersity (a.u.) 1541
1533

1140 1170 1200 1230 1500 1530 1560 1590

600 800 1000 1200 1400 1600 1800

Raman shift (cm-1)
b c Ni-MOF-2D
Ni L-edge Ni-MOF-2D C K-edge
Ni-MOF-2D-Na2S5 Ni-MOF-2D-Na2S5

Normalizaed adsorption (a.u.)

Normalizaed adsorption (a.u.)

852.5 853.0 853.5

845 850 855 860 865 870 875 280 290 300 310 320
Photon energy (eV) Photon energy (eV)

Figure 4.  Analysis of electron transfer between Ni-MOF-2D and Na2S5. a) Raman spectra, b) Ni L-edge NEXAFS spectra, and c) C K-edge NEXAFS
spectra of Ni-MOF-2D and Ni-MOF-2D-Na2S5.

to an electron depletion state (in Ni-MOF-2D-Na2S). This capacities and capacity retention. By contrast, the Ni-MOF-bulk
indicates that electron transfer from Ni to N correlates with suffers from weak NaPoSs confinement (Figure S26, Sup-
the strengthened Na–N/S interaction and increased adsorp- porting Information). Density of states analysis revealed a sem-
tion energy from Na2S5 to Na2S on Ni-MOF-2D, which veri- iconducting property of Ni-MOF-2D with a smaller band gap
fies the NEXAFS results of Ni L-edge (Figure S22, Supporting of ≈0.56 eV compared with those of reported MOFs in sulfur
Information). cathodes (Figure S27, Supporting Information). The semicon-
The schematic illustrates effect of the Ni centers’ electron ducting property of Ni-MOF-2D is believed responsible for the
states on NaPoSs confinement (Figure 5c). Basically, the Ni enhanced electron transfer between Ni centers and the linkers.
centers with high redox capability are understood to strengthen All these results demonstrate that the Ni centers of S/Ni-MOF-
the Na–N/S interaction through electron transfer from the Ni- 2D exhibited dynamic electron states during charge/discharge
MOF-2D to the NaPoSs. This strengthened Na–N/S interac- process, which evokes tuned Na (polysulfide)-N/S (Ni MOF)
tion should facilitate the sodiation process from Na2S5 to Na2S interaction. Therefore, strong polysulfides adsorption and fast
during discharge. During charge, the weakened Na–N/S inter- polysulfides conversion kinetics can be realized and efficient
actions also facilitate the desodiation process of Na2S to Na2S5 polysulfides confinement was promoted on the S/Ni-MOF-2D.
due to electron transfer from the NaPoSs to the Ni-MOF-2D. We conclude that our work demonstrates a new mechanism
Therefore, this dynamic electron transfer between Ni-MOF-2D for effective polysulfides confinement on a 2D Ni (II) MOF
and NaPoSs facilitates NaPoSs sodiation–desodiation kinetics and its chemical origin. The electrochemical performances of
on Ni-MOF-2D. The Ni centers with high redox capability the resulting sulfur cathodes are shown to be superior to those
appear to result from the charge redistribution of the Ni-MOF- of all sulfur cathode materials in RT Na–S batteries. On the
2D.[35] As a result, enhanced NaPoSs adsorption and conver- basis of synchrotron-based in situ XRD, NEXAFS characteriza-
sion kinetics on Ni-MOF-2D significantly boosts the discharge tions, electrochemical tests, and systematic DFT computations,

Adv. Mater. 2020, 1907557 1907557  (6 of 8) © 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

a 4 b Ni-MOF-2D-Na2Sn
-5 conductive carbon-Na2Sn

Adsorption energy (eV)

-4
Adsorption energy (eV)

N site
-3

-4
-2
Ni center
S site
-1
-8
Ni-MOF-2D-Na2S
conductive carbon-Na2S
0
0 1000 2000 3000 4000 Na2S5 Na2S4 Na2S3 Na2S2 Na2S
MD time (fs)

c Discharge process Charge process

Stronger Na-N/S interaction
Strengthened Na-N/S interaction Weakened Na-N/S interaction
Weaker Na-N/S interaction

Solvated Na+

NaPoSs

Electron paths

Ni-MOF-2D

e- e-
e-

e-

Figure 5.  Computational investigation of NaPoSs confinement on Ni-MOF-2D. a) Adsorption energies of Na2Sn on graphene and Ni-MOF-2D as a
function of ab initio molecular dynamics (AIMD) simulation time (fs). b) Adsorption energies of Na2Sn with insets of charge difference analyses, in
which yellow and cyan isosurface represent electron accumulation and electron depletion. The color code is the same as for Figure 1a and the cyan
spheres represent Na atoms. The isosurface value is 0.005 e Å−3. c) Schematic of NaPoSs confinement on Ni-MOF-2D.

we demonstrated that strong polysulfides adsorption and fast Supporting Information

polysulfides conversion kinetics can be realized by tuning
Na–N/S interaction via dynamic electron states of Ni centers,
thus results in efficient polysulfides confinement on the Ni
MOF. We conclude that our findings offer a practical means for
production of highly stable metal–sulfur batteries and for the
design of electrode materials for broader applications in energy
storage and conversion.
Experimental Section
Experimental details can be found in the Supporting Information.
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
This work was supported financially by the Australian Research Council
(ARC) through Discovery Project and Linkage Project Programs
(DP160104866, LP160100927, and FL170100154) and by the National
Natural Science Foundation of China (51722103). C.Y. was supported
by the Chinese CSC Scholarship Program. The authors thank Dr. Bruce
Cowie at the Australian Synchrotron for help with NEXAFS.

Adv. Mater. 2020, 1907557 1907557  (7 of 8) © 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com
Conflict of Interest
The authors declare no conflict of interest.
Keywords
confinement of polysulfides, metal–organic frameworks, sodium–sulfur
batteries, sulfur cathodes, 2D materials
Received: November 18, 2019
Revised: January 3, 2020
Published online:
[1] Y. Yang, G. Zheng, Y. Cui, Chem. Soc. Rev. 2013, 42, 3018.
[2] a) M. Salama, Rosy, R. Attias, R. Yemini, Y. Gofer, D. Aurbach,
M. Noked, ACS Energy Lett. 2019, 4, 436; b) X. Hong, J. Mei,
L. Wen, Y. Tong, A. J. Vasileff, L. Wang, J. Liang, Z. Sun, S. X. Dou,
Adv. Mater. 2019, 31, 1802822.
[3] A. Manthiram, Y. Fu, S.-H. Chung, C. Zu, Y.-S. Su, Chem. Rev. 2014,
114, 11751.
[4] a) X. Xu, D. Zhou, X. Qin, K. Lin, F. Kang, B. Li, D. Shanmukaraj,
T. Rojo, M. Armand, G. Wang, Nat. Commun. 2018, 9, 3870;
b) B.-W. Zhang, T. Sheng, Y.-X. Wang, S. Chou, K. Davey, S.-X. Dou,
S.-Z. Qiao, Angew. Chem., Int. Ed. 2019, 58, 1484.
[5] Z. W. Seh, Y. Sun, Q. Zhang, Y. Cui, Chem. Soc. Rev. 2016, 45, 5605.
[6] Q. Pang, X. Liang, C. Y. Kwok, L. F. Nazar, Nat. Energy 2016, 1,
16132.
[7] X. Chen, T. Hou, K. A. Persson, Q. Zhang, Mater. Today 2019, 22,
142.
[8] a) Z. Li, B. Y. Guan, J. Zhang, X. W. Lou, Joule 2017, 1, 576;
b) J. Zhou, R. Li, X. Fan, Y. Chen, R. Han, W. Li, J. Zheng, B. Wang,
X. Li, Energy Environ. Sci. 2014, 7, 2715; c) B.-W. Zhang, T. Sheng,
Y.-D. Liu, Y.-X. Wang, L. Zhang, W.-H. Lai, L. Wang, J. Yang,
Q.-F. Gu, S.-L. Chou, Nat. Commun. 2018, 9, 4082.
[9] T.-Z. Hou, W.-T. Xu, X. Chen, H.-J. Peng, J.-Q. Huang, Q. Zhang,
Angew. Chem., Int. Ed. 2017, 56, 8178.
[10] a) X. Yu, A. Manthiram, ChemElectroChem 2014, 1, 1275; b) I. Kim,
J.-Y. Park, C. Kim, J.-W. Park, J.-P. Ahn, J.-H. Ahn, K.-W. Kim,
H.-J. Ahn, J. Electrochem. Soc. 2016, 163, A611.
[11] a) Y. Yan, C. Cheng, L. Zhang, Y. Li, J. Lu, Adv. Energy Mater. 2019, 9,
1900148; b) X. Chen, W. He, L.-X. Ding, S. Wang, H. Wang, Energy
Environ. Sci. 2019, 12, 938.
[12] Q. Zhang, Y. Wang, Z. W. Seh, Z. Fu, R. Zhang, Y. Cui, Nano Lett.
2015, 15, 3780.
[13] T. Z. Hou, X. Chen, H. J. Peng, J. Q. Huang, B. Q. Li, Q. Zhang,
B. Li, Small 2016, 12, 3283.
Adv. Mater. 2020, 1907557 1907557  (8 of 8)
www.advmat.de
[14] X. Chen, H.-J. Peng, R. Zhang, T.-Z. Hou, J.-Q. Huang, B. Li,
Q. Zhang, ACS Energy Lett. 2017, 2, 795.
[15] Y. Zhao, M. Hong, Y. Liang, R. Cao, W. Li, J. Weng, S. Lu, Chem.
Commun. 2001, 11, 1020.
[16] Y. Yao, Z. Lin, Z. Li, X. Song, K.-S. Moon, C.-p. Wong, J. Mater.
Chem. 2012, 22, 13494.
[17] R. Yadav, D. Swain, P. P. Kundu, H. S. Nair, C. Narayana,
S. Elizabeth, Phys. Chem. Chem. Phys. 2015, 17, 12207.
[18] D. Liu, C. Zhang, G. Zhou, W. Lv, G. Ling, L. Zhi, Q. H. Yang, Adv.
Sci. 2018, 5, 1700270.
[19] W. Tang, E. Sanville, G. Henkelman, J. Phys.: Condens. Matter 2009,
21, 084204.
[20] A. V. Soldatov, A. N. Kravtsova, M. E. Fleet, S. L. Harmer, J. Phys.:
Condens. Matter 2004, 16, 7545.
[21] a) F. Pei, L. Lin, D. Ou, Z. Zheng, S. Mo, X. Fang, N. Zheng, Nat.
Commun. 2017, 8, 482; b) H.-L. Zhang, S. D. Evans, J. R. Henderson,
R. E. Miles, T. Shen, J. Phys. Chem. B 2003, 107, 6087; c) Z. Jiang,
H. Xie, S. Wang, X. Song, X. Yao, H. Wang, Adv. Energy Mater. 2018,
8, 1801433; d) H. Xiang, J. Chen, Z. Li, H. Wang, J. Power Sources
2011, 196, 8651.
[22] Y. Jiao, Y. Zheng, K. Davey, S.-Z. Qiao, Nat. Energy 2016, 1, 16130.
[23] S. Leikin, V. A. Parsegian, W. H. Yang, G. E. Walrafen, Proc. Natl.
Acad. Sci. USA 1997, 94, 11312.
[24] D. Chao, C. Zhu, M. Song, P. Liang, X. Zhang, N. H. Tiep, H. Zhao,
J. Wang, R. Wang, H. Zhang, H. J. Fan, Adv. Mater. 2018, 30,
1803181.
[25] J. Zhang, Z. Li, Y. Chen, S. Gao, X. W. Lou, Angew. Chem., Int. Ed.
2018, 57, 10944.
[26] Q. Gu, J. A. Kimpton, H. E. Brand, Z. Wang, S. Chou, Adv. Energy
Mater. 2017, 7, 1602831.
[27] Y.-X. Wang, J. Yang, W. Lai, S.-L. Chou, Q.-F. Gu, H. K. Liu, D. Zhao,
S. X. Dou, J. Am. Chem. Soc. 2016, 138, 16576.
[28] J. Conder, R. Bouchet, S. Trabesinger, C. Marino, L. Gubler,
C. Villevieille, Nat. Energy 2017, 2, 17069.
[29] a) B. You, N. Jiang, M. Sheng, M. W. Bhushan, Y. Sun, ACS Catal.
2015, 6, 714; b) H. Li, Y. Tao, C. Zhang, D. Liu, J. Luo, W. Fan,
Y. Xu, Y. Li, C. You, Z.-Z. Pan, M. Ye, Z. Chen, Z. Dong, D.-W. Wang,
F. Kang, J. Lu, Q.-H. Yang, Adv. Energy Mater. 2018, 8, 1703438.
[30] H. Jin, C. Guo, X. Liu, J. Liu, A. Vasileff, Y. Jiao, Y. Zheng, S.-Z. Qiao,
Chem. Rev. 2018, 118, 6337.
[31] Y. S. Pang, H. J. Hwang, M. S. Kim, J. Mol. Struct. 1998, 441, 63.
[32] G. N. R. Tripathi, M. Clements, J. Phys. Chem. B 2003, 107, 11125.
[33] J. Zheng, J. Tian, D. Wu, M. Gu, W. Xu, C. Wang, F. Gao,
M. H. Engelhard, J.-G. Zhang, J. Liu, Nano Lett. 2014, 14, 2345.
[34] T. Ling, P. Da, X. Zheng, B. Ge, Z. Hu, M. Wu, X.-W. Du, W.-B. Hu,
M. Jaroniec, S.-Z. Qiao, Sci. Adv. 2018, 4, eaau6261.
[35] X. Liang, C. Y. Kwok, F. Lodi-Marzano, Q. Pang, M. Cuisinier,
H. Huang, C. J. Hart, D. Houtarde, K. Kaup, H. Sommer, Adv.
Energy Mater. 2015, 6, 1501636.
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim