Communication
www.advmat.de
Toward Practical High-Areal-Capacity Aqueous Zinc-Metal
Batteries: Quantifying Hydrogen Evolution and a Solid-Ion
Conductor for Stable Zinc Anodes
Longtao Ma, Qing Li, Yiran Ying, Feixiang Ma, Shengmei Chen, Yangyang Li,*
Haitao Huang,* and Chunyi Zhi*
The hydrogen evolution in Zn metal battery is accurately quantified by
in situ battery–gas chromatography–mass analysis. The hydrogen fluxes
reach 3.76 mmol h−1 cm−2 in a Zn//Zn symmetric cell in each segment, and
7.70 mmol h−1 cm−2 in a Zn//MnO2 full cell. Then, a highly electronically
insulating (0.11 mS cm−1) but highly Zn2+ ion conductive (80.2 mS cm−1) ZnF2
solid ion conductor with high Zn2+ transfer number (0.65) is constructed to
isolate Zn metal from liquid electrolyte, which not only prohibits over
99.2% parasitic hydrogen evolution but also guides uniform Zn electrodeposi-
tion. Precisely quantitated, the Zn@ZnF2//Zn@ZnF2 cell only produces
0.02 mmol h−1 cm−2 of hydrogen (0.53% of the Zn//Zn cell). Encouragingly, a
high-areal-capacity Zn@ZnF2//MnO2 (≈3.2 mAh cm−2) full cell only produces
maximum hydrogen flux of 0.06 mmol h−1 cm−2 (0.78% of the Zn//Zn cell)
at the fully charging state. Meanwhile, Zn@ZnF2//Zn@ZnF2 symmetric cell
exhibits excellent stability under ultrahigh current density and areal capacity
(10 mA cm−2, 10 mAh cm−2) over 590 h (285 cycles), which far outperforms all
reported Zn metal anodes in aqueous systems. In light of the superior
Zn@ZnF2 anode, the high-areal-capacity aqueous Zn@ZnF2//MnO2 batteries
(≈3.2 mAh cm−2) shows remarkable cycling stability over 1000 cycles with
93.63% capacity retained at ≈100% Coulombic efficiency.
Zinc (Zn) metal anode is promising for aqueous batteries
with both intrinsic safety and low-cost metrics.[1–6] How-
ever, metallic Zn is thermodynamically unstable in aqueous
Dr. L. T. Ma, Q. Li, Dr. F. X. Ma, S. Chen, Dr. Y. Y. Li, Prof. C.Y. Zhi
Department of Materials Science and Engineering
City University of Hong Kong
83 Tat Chee Avenue, Kowloon, Hong Kong 999077, P. R. China
E-mail: yangli@cityu.edu.hk; cy.zhi@cityu.edu.hk
Y. R. Ying, Prof. H. T. Huang
Department of Applied Physics and Research Institute for Smart Energy
The Hong Kong Polytechnic University
Hung Hom, Kowloon, Hong Kong 999077, P. R. China
E-mail: aphhuang@polyu.edu.hk
Prof. C.Y. Zhi
Center for Advanced Nuclear Safety and Sustainable Development
City University of Hong Kong
Kowloon, Hong Kong 999077, P. R. China
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adma.202007406.
DOI: 10.1002/adma.202007406
Adv. Mater. 2021, 33, 2007406 2007406  (1 of 9)
environments,[7–12] where Zn deposition
even during shelf time is continuously
interfered by competitive hydrogen evolu-
tion through H2O decomposition (Zn +
2H2O → Zn(OH)2  + H2) and by con-
suming both the electrolyte and active Zn
metal. This long-neglected problem will
remarkably affect calendar life of batteries,
which is equivalently important with the
intensively studied cycling life of batteries.
As revealed in our preliminary quantita-
tive study shown in Figure 1a, immersing
Zn in 2 m ZnSO4 electrolyte will induce
an ≈0.25 mmol h–1 cm–2 hydrogen flux.
The continuous hydrogen evolution
will cause local pH changes, which fur-
ther induce the formation of loose and
brittle Zn4SO4(OH)6·xH2O (3Zn(OH)2  +
ZnSO4·xH2O → Zn4SO4(OH)6·xH2O),
as revealed in Figure  1b.[13–18] It is gener-
ally assumed that these by-products aug-
ment the tortuosity and irregularity at
the electrode/electrolyte interface with
physical contact surface being increased
(Figure 1c–h), which will further accelerate
hydrogen evolution reaction. The above issues necessitate an
effective method to detect hydrogen evolution. Nevertheless,
for a Zn-based battery, these assumption and hypotheses are
only based on oversimplified observations of battery swelling,
gas bubbles, or conducting polarization curve in the absence
of Zn2+ condition, not reflecting physical truth of battery.[19–21]
Hydrogen production during electrochemical procedure and
even shelf time has not been precisely quantified. Conse-
quently, efforts to make Zn metal a valid anode material may be
misdirected. Quantifying the hydrogen production on the elec-
trode during Zn deposition is key to understanding the mecha-
nisms leading to capacity loss and battery failure.
On the other hand, these Zn protrusions caused by hydrogen
evolution reaction will attract more Zn2+ flux (“tip” effect)[22]
under concentrated electric field during electrochemical
cycling, thus accelerating the vertical growth of Zn dendrites
instead of planar growth and hydrogen production further
flourish (Figure  1g). Up to now, various strategies have been
evolved to prohibit the Zn dendrite growth, such as electrolyte
optimization,[8,23–25] Zn surface coating,[9,21,26–29] and alloying.[30]
To some extent, these strategies stabilize Zn metal, but they do
© 2021 Wiley-VCH GmbH
www.advancedsciencenews.com www.advmat.de

Figure 1.  a) In situ monitor of hydrogen evolution of bare Zn metal immersed in 2 m ZnSO4 aqueous solution by a battery–gas chromatography–mass
spectrometry (GC–MS) quantitative analysis method. b) The in situ XRD patterns of bare Zn immersed in 2 m ZnSO4 electrolyte, revealing the appearance
of Zn4SO4(OH)6·xH2O and Zn(OH)2. c–f) SEM images of bare Zn (c) and soaked Zn in 2 m ZnSO4 aqueous electrode for 2 days (d), 5 days (e), and 10 days
(f). Scale bar: 5 µm. g) Schematic illustration of Zn deposition on bare Zn foil. The corrosion, abundant “dead” Zn, and dendrite are observed due to 2D
Zn2+ diffusion and thermodynamically instability of Zn metal in aqueous electrolyte. h) Schematic illustration of Zn deposition on Zn@ZnF2 foil. The ZnF2
layer endows a dense and dendrite-free Zn deposition by regulating Zn2+ diffusion, controlling nucleation, and prohibiting the permeation of H2O and O2.

not follow the original nature of Zn dendrite growth in aqueous
environment, and batteries are operated with low current den-
sity and loading mass of cathode materials, far from practical
implementation. Meanwhile, most Zn surface coating layers
with low Zn2+ conductivity or without Zn2+ are inhomogeneous
and uncompacted, being impossible to completely isolate metal
Zn from bulk electrolyte. Therefore, there is a requisite to find
a way to improve inherent H2O-resistant properties of Zn metal
anode. Building artificial solid Zn2+-ion conductor is a good
alternative to this challenge. However, most artificial solid Zn2+-
ion conductor interlayer suffers from low Zn2+ conductivity due
to a large energy barrier of ion diffusion caused by a much high
electric charge density of Zn2+ ion.[31–34]
In this work, we demonstrate a compact and homogeneous
zinc fluoride (ZnF2) layer with high Zn2+ conductivity, utilizing
an in situ ion metathesis method. At the Zn/ZnF2 interface,
the F atoms of ZnF2 layer are tightly bound with Zn atoms of
Zn metal by electrovalent bonds, giving the credit to the charge
migration between Zn and F atoms and charge redistribution.
Meanwhile, the Zn@ZnF2 interface exhibits a much higher
ΔGH value for hydrogen evolution side reaction and experimen-
tally suppresses over 99.2% hydrogen production. As a result,
the obtained structure (denoted as Zn@ZnF2) enables long-
term Zn dendrite-free plating/stripping and prevents the fresh
Zn metal from contacting with H2O molecular in the bulk elec-
trolyte, avoiding side reactions of hydrogen evolution and Zn
corrosion (Figure 1h). Consequently, the dendrite-free and side-
reaction-free Zn anode enables Zn@ZnF2//Zn@ZnF2 sym-
metric cell cycling over 2500 h even at a high current density
of 5 mA cm–2 with an areal capacity of 1 mAh cm–2, and high-
areal-capacity Zn@ZnF2//MnO2 full cell operates 1000 cycles
with 93.63% capacity retained even at a high areal capacity
of 3.2 mAh cm–2. Finally, an 850 mAh Zn@ZnF2//MnO2
large-capacity battery exhibits excellent cycling stability over
160 cycles with 93.17% capacity retention, far outperforming
Zn//MnO2 (failed after only 15 cycles) in same condition.

Adv. Mater. 2021, 33, 2007406 2007406  (2 of 9) © 2021 Wiley-VCH GmbH

www.advancedsciencenews.com www.advmat.de

Figure 2.  a) XRD patterns of bare Zn and Zn@ZnF2 foil. b) SEM image of Zn@ZnF2 foil. c) High-resolution Zn 2p spectra of Zn@ZnF2 foil.
d) High-resolution Zn 2p spectra of bare Zn foil. e) F 1s spectra of Zn@ZnF2 foil. f ) Charge density difference distribution of Zn@ZnF2 interface and
g) the corresponding sliced 2D contour map showing the ZnF bonds. Zn and F atoms are shown in gray and orange, respectively. The isosurface
value is set as 0.02e Bohr−1. h) Diffusion energy barrier diagram for Zn2+ on ZnF2 (002) surface and in ZnF2 bulk. The corresponding initial state
(IS), transition state (TS), and final state (FS) structures are shown in the inset and see Figure S1 (Supporting Information) for amplification (top:
ZnF2 bulk; bottom: ZnF2 (002) surface; Zn and F atoms are shown in gray, orange, respectively, and the intercalated Zn atom is shown in purple
for bulk ZnF2 and red ZnF2 (002) surfaces). i) Illustration of the insertion energy barrier for single Zn atom on bulk ZnF2, Zn@ZnF2 surface, and
Zn (002) surface.

The X-ray diffraction (XRD) patterns confirm the successful
fabrication of Zn@ZnF2 composite. The peaks at 20.76°,
29.33°, 32.56°, and 51.93° are well-indexed to (111), (110), (101),
and (211) planes of ZnF2, respectively (Figure 2a). The atomic
force microscopy (AFM) image displays that the Zn surface is
evenly and homogeneously covered by ZnF2 (Figure  2b). The
high-resolution X-ray photoelectron spectroscopy (XPS) Zn 2p
spectra of both Zn@ZnF2 and bare Zn foil exhibit two dis-
tinct peaks, ascribed to Zn 2p1/2 and Zn 2p3/2, respectively
(Figure  2c,d). Compared with bare Zn foil, the formation
of ZnF2 layer leads to the appearance of components with
slightly different binding energies, which can be resolved
into minimum two signals. This can be explained by ZnF
polar bonds at the interphase of ZnF2 and Zn metal, exhib-
iting dissimilar coordination environments.[14,20] The F
signal of F 1s at 683.99  eV further confirms the formation of
ZnF polar bonds, which can enhance the adhesion of the
ZnF2 layer to the Zn metal (Figure  2e). In theory, the cor-
responding differential charge density distribution of Zn@
ZnF2 is calculated by the density functional theory (DFT) to
elucidate mechanism for the interaction between F atoms
of ZnF2 layer and Zn atoms of Zn metal. Figure  2f displays
that the geometrically optimized F element can distinctly
induce charge transfer and redistribution at the Zn/ZnF2 inter-
face. We further slice the charge density difference distribution
with 2D contour map showing the ZnF bonds in Figure  2g.
The unbalanced charge distribution can be clearly observed,
with positive charge being accumulated around the Zn layer
and negative charge distributed around ZnF bonds. This can
promote the formation of substantial Zn@ZnF2 interface and

Adv. Mater. 2021, 33, 2007406 2007406  (3 of 9) © 2021 Wiley-VCH GmbH

www.advancedsciencenews.com
fast Zn2+ diffusion on the surface. Thus, the F atoms of ZnF2
layer can tightly bond with Zn atoms of Zn metal.
The Zn2+ deposition mechanism is revealed by calculating
the diffusion energy barrier of single Zn2+ on ZnF2(002) sur-
face and ZnF2 bulk by the climbing image nudged elastic band
method. As shown in Figure 2h, the energy barrier for Zn2+ dif-
fusion on ZnF2 (002) surface is calculated to be 0.27 eV, while
Zn2+ hopping in bulk ZnF2 represents a barrier as high as
0.76  eV determined by the same methods (the corresponding
atomic structures are shown in Figure  2h (inset)), indicating
that exposed ZnF2 (002) surface can greatly enhance the Zn2+
diffusion. This analysis indicates that the surface of ZnF2 could
be highly conductive if the surfaces of Zn2+ ions are not pinned
to any position on the ZnF2 lattice. Moreover, the lower energy
barrier of Zn2+ inserted into Zn@ZnF2 can be validated by
the decrease in insertion energy barrier, where the Zn inser-
tion barrier on Zn@ZnF2 is 0.52  eV, lower than that of bulk
ZnF2 (4.08  eV) and metallic Zn surface (0.67  eV) (Figure  2i).
Accordingly, the Zn@ZnF2 preferentially provides electrostatic
attraction toward Zn2+, accelerating the kinetics “through”
reducing the deposition barrier.
As the ZnF2 layer covers the Zn metal electrode surface, the
front interface/phase between the Zn metal and the electro-
lyte plays a pivotal role in directing the growth mode. Prior to
the study of Zn growth, the electrochemical properties of Zn
layer are conducted including the ionic conductivity capability,
electronic resistivity, and Zn2+ transfer number. As expected,
the ZnF2 layer shows significant higher ionic conductivity of
80.2 mS cm–1 than that of a routine glass fiber (GF) separator
(9.8 mS cm–1), and even all reported Zn foil coating layers
(Figure 3a; Table S1, Supporting Information), calculated from
AC impedance spectra (Figure S2, Supporting Information). In
comparison, the electronic conductivity of the ZnF2 layer is only
0.11 mS cm–1, tested by four-point probe, which is near four
orders of magnitude lower than that of the pure Zn metal foil,
970.87 mS cm–1 (Figure  3b). During electrochemical process,
the Zn deposition only comes up at a place, where the Zn2+
meets charges. Therefore, the ZnF2 plays a role of regulating
the transport of Zn2+ cations toward the electrode surface,
without any Zn deposition on ZnF2 surface. The Zn2+ trans-
port flux could be dominated by ZnF2 solid Zn2+-ion conductor
during the electrodeposition. Moreover, the Zn@ZnF2 electrode
shows a very low charge-transfer resistance of 39.99  Ω, a tenth
lower than that of the bare Zn electrode (412.62 Ω) (Figure S3,
Supporting Information). The Zn2+ transfer number of the
ZnF2 reaches to 0.65, which is much larger than that of the GF
separator coupled with 2 m ZnSO4 aqueous electrolyte (0.28;
Figure S4, Supporting Information). The above results suggest
outstanding Zn2+ transportation capability of the ZnF2 layer.
To prove the efficacy of Zn@ZnF2 in suppressing dendrite
formation during different rate cycling in liquid electrolytes,
we choose the widely used 2 m ZnSO4 aqueous electrolyte to
test. We perform a series of plating/stripping measurements in
symmetric Zn//Zn and Zn@ZnF2//Zn@ZnF2 cells under gal-
vanostatic condition. The cells are cycled at 1, 2, and 5 mA cm–2
with 1 mAh cm–2 of the Zn reversibly cycled. The Zn@ZnF2//
Zn@ZnF2 cells show superior performance with a cycle life of
>2500  h, where 6250 mAh cm–2 cumulative capacity is cycled
(Figure  3c). Cycling at elevated current densities (from 0.5 to
Adv. Mater. 2021, 33, 2007406 2007406  (4 of 9)
www.advmat.de
10 mAh cm–2) reveals a much more stable voltage profile with
a smaller voltage hysteresis (Figure  3d). Even at a high rate
of 10 mA cm–2 with 10 mAh cm–2 of the Zn reversibly cycled,
Zn@ZnF2//Zn@ZnF2 cells can be operated steadily for ≈590 h
without apparent irreversible voltage observed (Figure  3e).
In contrast, Zn//Zn cells last only 600 h at a low current of
1 mA cm–2 and <250 h at 2 and 5 mA cm–2, meanwhile suc-
cumbing to substantial voltage fluctuation just after a few
cycles. The difference in cycling stability is pronounced at
high current densities. The plating/stripping overpotential
for Zn@ZnF2 is only 56.4  mV, significantly lower than that
of bare Zn (238.9  mV). The excellent reversibility and cyclic
lifespan remarkably outperform all of its counterparts with
dendrite precaution-oriented and hydrogen evolution suppres-
sion designs (Figure  3f). The outstanding reversibility of Zn
deposition/dissolution low overpotential originates from high
ionic conductivity and fast ion diffusion of ZnF2 solid in Zn2+-
ion conductor as well as good stability of ZnF2 layer in aqueous
environment (Figure S5a,b, Supporting Information).
Surprisingly, the Zn@ZnF2//Zn@ZnF2 cell can cycle over
400 h at 2 mA cm–2 and 140 h at 5 mA cm–2 with 1 mAh cm–2 of
Zn reversibly cycled in an alkaline electrolyte of (6 m KOH + 0.2 m
Zn(AC)2). In sharp contrast, after only <10 h plating/strip-
ping cycles with Zn//Zn cell in the same condition, a sudden
polarization occurs due to intensified Zn-dendrite formation
(Figure S6, Supporting Information). It demonstrates that this
design achieves superior reversible cycling performance at both
mild and alkaline electrolytes. It should be mentioned that the
ZnF2 protective layer should be uniform in terms of composi-
tion and thickness, which dominates the cation transport. As a
result, it can homogenize the transport of Zn2+ cations toward
the electrode surface. In contrast, lumpy coating layer will accel-
erate the dendrite growth. As shown in Figure S7 (Supporting
Information), Zn@ZnF2 electrode with commercial ZnF2 pow-
ders being coated induces fast dendrite formation and growth,
leading to only ≈290 h cycling stability of symmetric cells in
mild electrolyte, much lower than bare Zn electrode-based cells
(600 h, purple line in Figure 3c).
The morphologies of Zn@ZnF2 and bare Zn electrodes are
analyzed after 50 cycles of stripping/plating in symmetric cells
at 2 mA cm–2 with 1 mAh cm–2 of the Zn reversibly cycled.
The scanning electron microscopy (SEM) image confirms that
the smooth and dense surface of the Zn deposition with the
ZnF2 layer remains essentially flat with no obvious dendrites
(Figure  3g). Inversely, for bare Zn electrode, the deposited Zn
is highly porous and consists of thin platelets (Figure 3h). From
the cross-sectional SEM image, a clear gap is observed beneath
the ZnF2 layer for the Zn@ZnF2 electrode at the stripped state
(Figure 3i). For the electrode at the plated state, no empty space
is observed, and the ZnF2 layer is in intimate contact with the
Zn metal (Figure 3j). The outcome originates that the Zn depo-
sition only comes up at a place, where the Zn2+ meets charges.
The electronically insulating ZnF2 layer has no relish for Zn
nucleation, and the Zn tends to grow on the Zn@ZnF2 inter-
face with lower inserting energy. Consequently, it prefers to
bring Zn2+ ions back into the empty region beneath the ZnF2
layer and the Zn plating reaction can be well confined under
ZnF2 layer, in which uneven Zn layer can be deposited on Zn
anode and thus dendrite growth is suppressed.
© 2021 Wiley-VCH GmbH
www.advancedsciencenews.com www.advmat.de

Figure 3.  a) Ionic conductivity of ZnF2 layer and routine GF separator coupled with a 2 m ZnSO4 electrolyte. b) Electronic conductivity of ZnF2 layer and Zn
metal. c) Comparison of cycling stability of bare Zn//Zn symmetric cells and Zn@ZnF2//Zn@ZnF2 symmetric cells at 1, 2, and 5 mA cm–2 with a capacity
of 1 mAh cm–2. d) Rate performance. e) Cycling stability of bare Zn//Zn symmetric cells and Zn@ZnF2//Zn@ZnF2 symmetric cells at 10 mA cm–2 with a
capacity of 10 mAh cm–2. f) Comparison of cumulative capacity recycled and current density between this work and the previous ones. g,h) SEM images
of Zn@ZnF2 (g) and bare Zn (h) after 50 cycle tests. Scale bar: 5 µm. i,j) Cross-sectional SEM images of Zn dissolution (i) and Zn deposition (j) on Zn@
ZnF2 at 2 mA cm–2 with a capacity of 1 mAh cm–2. Scale bar: 20 µm. k) Coulombic efficiency of Zn plating/stripping in bare Zn//Cu cell and Zn@ZnF2//
Cu cell at 1 mA cm–2 with a capacity of 1 mAh cm–2. l) Voltage profiles of Zn@ZnF2//Cu cell at the 1st, 5th, 10th, 100th, 500th, and 1000th cycles.

The Zn plating/stripping Coulombic efficiencies are quan-
tified using asymmetric Zn//Cu, Zn@ZnF2//Cu, and Zn//
ZnF2@Cu cells with 1 mAh cm–2 of the Zn reversibly cycled.
The Coulombic efficiency of Zn plating/stripping is calculated
from the ratio of charge passed in each segment (Figure  3k;
Figure S8, Supporting Information). The discharge/charge
Coulombic efficiency of Zn@ZnF2//Cu electrodes reaches 99%
within only 40 cycles, increases to 99.5% within 90 cycles, and
stably maintains for 1000 cycles without any decay. Likewise, the
discharge/charge Coulombic efficiency of Zn//ZnF2@Cu elec-
trodes reaches to 99.3% within 50 cycles, and stably maintains
for 440 cycles without any decay. With regard to bare Zn, the
Coulombic efficiency is lower in the same condition and the cell
quickly fails after only ≈100 plating/stripping cycles (Figure S9a,
Supporting Information). The observed Coulombic efficiency’s
fluctuations are consistent with failure either as a result of
localized short-circuiting or by reconnection of fragmentary
Zn deposits that broke away from the electrode (“dead” Zn).
Figure  3l exhibits the corresponding voltage profiles for
Zn@ZnF2//Cu cell, which consists of the typical plateau for
Zn plating/stripping on the Cu surface at a very low overpoten-
tial of ≈52.7 mV, revealed by the voltage hysteresis between Zn
plating and stripping curves (155.5 mV; Figure S9b, Supporting
Information). Encouragingly, the voltage profiles remain
stable over 1000 cycles, suggesting exceptional reversibility of
Zn@ZnF2 electrode.
The nucleation overpotential of Zn//Cu and Zn@ZnF2//
Cu asymmetric cells is investigated for insight into the effect of
ZnF2 layer on the Zn deposition. The Zn@ZnF2 displays only
42.5  mV nucleation overpotential at 1 mA cm–2, much lower

Adv. Mater. 2021, 33, 2007406 2007406  (5 of 9) © 2021 Wiley-VCH GmbH

www.advancedsciencenews.com
than that of bare Zn (79.4  mV), which propones a dramati-
cally reduced Zn deposition barrier (Figure S10a, Supporting
Information). Likewise, the same rules are appropriate for the
current densities of 2 and 5 mA cm–2 on Cu foil (Figure S10b–d,
Supporting Information). The reduced nucleation overpoten-
tial is attributed to the lower local current density that forces
Zn2+ migration on the surface of Zn metal enabled by the ZnF2
matrix. The ZnF2 layer establishes a solid-state diffusion bar-
rier to homogenize the interfacial Zn2+ flux and mediate the
migration of Zn2+ special to Cu substrate surface. The chrono-
amperometry (CA) is further conducted to reveal the nucleation
procedure and surface fashion. The current curve of Zn@ZnF2
dictates an stable and constant 3D diffusion procedure at
≈9.3 mA cm–2 after short 2D diffusion (4 s) under an overpo-
tential of −150 mV, revealing that adhered underlying Zn2+ ions
move laterally and are reduced to Zn0 locally in very close prox-
imity to a place where the initial adsorption occurs (Figure S11,
Supporting Information). The deposition ultimately evolves
into a smooth and compact Zn layer. In sharp contrast, the cur-
rent density continuously elevates beyond 100 s, depicting a
long and fast 2D diffusion process and rough deposition propa-
gation. The Zn2+ tends aggregate and uneven dendrite growth
to minimize surface energy and exposed area (“tip” effect).[35,36]
The exchange current density is utilized to precisely evaluate
the kinetics of Zn deposition, which reflects the direct interac-
tion of ZnF2 layer with the Zn metal surface, by the following
equation[37,38]
i = i0 (F /RT )(ηtot /2) (1)
where η depicts the deposition overpotential, and i0 repre-
sents the exchange current density. Generally, lower values of
the exchange current density give information that the surface
of the Zn metal is inhibited or passivated, representing the
degree to which additive Zn2+ is absorbed to the active metal
surface and suppresses the reduction reaction. The Zn@ZnF2
electrode offers a low exchange current of 8.9 mA cm–2, much
lower than that of bare Zn (16.7 mA cm–2), suggesting reduced
redox reaction liveness of Zn@ZnF2 electrode under equilib-
rium potential (Figure S12, Supporting Information).
The electronically insulated ZnF2 layer can segregate active Zn
metal from bulk liquid electrolyte and turn off charges transfer
from Zn metal to H2O molecules of electrolyte, thus restraining
chemical oxidation and electrochemical hydrogen evolution reac-
tion on Zn@ZnF2 electrode. The direct proofs for suppressed
side reactions and modulated Zn plating/stripping enabled by
the ZnF2 layer are obtained from in situ optical visualization
observations of Zn deposition. Quite uneven and odd Zn deposits
and abundant gas bubbles are observed as early of 15 min after
the commencement of deposition with an applied constant cur-
rent density of 10 mA cm–2 (Figure 4a). For Zn@ZnF2 electrode
under the same condition, uniform and compact Zn deposits
without any bubbles being observed can be obtained over 45 min
process (Figure 4b). It should be mentioned that Zn metal coated
with commercial ZnF2 particles displays worse side reaction
suppression effect due to loosen architecture of coating layer
(Figures S13 and S14, Supporting Information).
We further achieve the hydrogen evolution polarization
curves of both electrodes in sodium acid phthalate buffer
Adv. Mater. 2021, 33, 2007406 2007406  (6 of 9)
www.advmat.de
solution with a pH value of 5.0, close to that for 2 m ZnSO4
electrolyte (pH = 4.6), which can separate hydrogen evolution
reaction with Zn precipitation procedure. The Zn@ZnF2 elec-
trode shows much lower hydrogen evolution current density
in a wide voltage range from −0.72 to −1.31  V versus standard
hydrogen electrode (SHE) in compared to that of bare Zn
electrode (e.g., 0.024 mAh cm–2 for Zn@ZnF2 electrode and
5.85 mAh cm–2 (almost 139 time higher than Zn@ZnF2) for
bare Zn electrode at −1 V vs Zn/Zn2+) (Figure S15a, Supporting
Information). The corresponding Tafel slope of 425.8 mV dec–1
for Zn@ZnF2 electrode is much higher than that of bare Zn
electrode (87.5 mV dec–1), suggesting slow hydrogen evolution
reaction kinetics after ZnF2 protection (Figure S15b, Supporting
Information). The Zn corrosion is another side reaction of
oxygen-induced passivation on the interface of Zn metal and
electrolyte with oxygen dissolved. With regard to the bare Zn,
the corrosion potential of the coated Zn@ZnF2 increases
from −0.665 to −0.650  V versus Zn/Zn2+ (Figure S16, Sup-
porting Information). Meanwhile, the ZnF2 protective layer also
reduces the corrosion current by 105.3 mA cm–2. The results
elucidate less tendency of corrosion reactions and a lower cor-
rosion of Zn@ZnF2 electrode.
To accurately quantitate gas production, we set up a bat-
tery–gas chromatography–mass spectrometry (GC-MS) appli-
ance to in situ monitor hydrogen flux by measuring the evolved
hydrogen utilizing GC during the Zn charging/discharging pro-
cess in routine 2 m ZnSO4 aqueous electrolyte at 10 mA cm–2
with 5 mAh cm–2 of Zn reversibly cycled. As shown in Figure 4c,
the hydrogen evolution rate during symmetric Zn//Zn cell
operation is conducted to 3.18 mmol h–1 cm–2 and increases to
3.76 mmol h–1 cm–2. The anabatic side hydrogen evolution origi-
nates from the increase of contact area between Zn metal and
electrolyte due to uneven distribution of Zn deposition. Prior to
the electrochemical test, the cell rests for 1 h to obtain adsorp-
tion/desorption balance and sufficient wetting. The hydrogen
evolved during the resting process (0.56 mmol h–1 cm–2) is
attributed to chemical oxidation due to the unstable thermo-
dynamic properties in aqueous solution. Inversely, there was
almost no H2 (0.02 mmol h–1 cm–2) detected in symmetric
Zn@ZnF2//Zn@ZnF2 cell (Figure  4d). To study rechargeable
batteries in practical application, we construct electrochem-
ical cells using a conventional α-MnO2 as cathode and bare
Zn or Zn@ZnF2 as anode. During the process of charging,
hydrogen evolution increases and reaches a maximum value
of 7.7 mmol h–1 cm–2, which is attributed to the decomposition
of water in Zn anode surface (Figure 4e). Assisted by the ZnF2
protective layer, the hydrogen evolution on Zn anode surface is
almost completely suppressed during the whole charge period
(maximum 0.06 mmol h–1 cm–2) (Figure 4f).
In theory, to investigate the mechanism of suppressing side
of hydrogen evolution reaction, we calculate hydrogen adsorp-
tion Gibbs free energy of ΔGH for Zn (002) and Zn@ZnF2
(ZnF2 side). Theoretically, the optimized hydrogen evolution
reaction activity can be achieved when ΔGH is close to zero,
and if ΔGH is larger (smaller) than zero, hydrogen binding
is too weak (strong), accordingly adsorption (desorption) is
suppressed.[39] As shown in Figure  4g, Zn@ZnF2 interface
exhibits a much higher ΔGH value of 1.31 eV compared with Zn
(1.00  eV) and Pt(111) benchmarks (−0.04  eV), confirming that
© 2021 Wiley-VCH GmbH
www.advancedsciencenews.com www.advmat.de

Figure 4.  a,b) In situ optical visualization observations of Zn plating behavior of bare Zn foil (a) and Zn@ZnF2 foil (b). Scale bar: 200 µm. c,d) In
situ monitoring of hydrogen evolution flux for a Zn//Zn symmetric cell (c) and a Zn@ZnF2//Zn@ZnF2 symmetric cell (d). e,f) In situ monitoring of
hydrogen evolution flux on a high-areal-capacity Zn//MnO2 full cell (e) and a Zn@ZnF2//MnO2 full cell (f). The hydrogen evolution flux is quantitated
by a battery–GC–MS quantitative analysis method. g) Gibbs free energy diagram for hydrogen evolution reaction on Zn@ZnF2 interface and Zn (002).

hydrogen evolution reaction can be greatly suppressed with the
interface structure.
The fully prototype pouch-type cells with high areal capacity
are assembled to prove the practical application of our strategy.
We choose full-fledged and promising α-MnO2 as cathodes
coupled with bare Zn and Zn@ZnF2 anodes for aqueous Zn
batteries, and the cell configuration consists of a single whole
set of anode–separator–cathode. The cells possess a high areal
capacity of ≈3.2 mAh cm–2 and are cycled at a current rate of 1 C.
As expected, both cells of Zn//MnO2 and Zn@ZnF2//MnO2
display similar galvanostatic charge/discharge profiles at the
first cycle (Figure 5a,b). The Zn@ZnF2//MnO2 cell can well
maintain their voltage profiles and 93.63% capacity even after
1000 cycles with 3.31 mAh cm–2 of Zn reversibly cycled in each
segment. In contrast, the Zn//MnO2 cell presents a similar
initial areal capacity of 3.20 mAh cm–2. However, the capacity
dramatically decays to 2.15 mAh cm–2 only after 100 cycles
(with 67.15% initial capacity retained) and to 1.18 mAh cm–2
(with 36.88% initial capacity retained) (Figure 5c). To the best of
our knowledge, currently, there are barely Zn anodes at depth
of discharge (DOD) higher than 1 mAh cm–2 reported. The high
areal capacity of ≈3.2 mAh cm–2  Zn reversibly cycled at each
cycle demonstrates a superhigh cycling stability of Zn@ZnF2
anodes for 1000 charge/discharge cycles.
To investigate the essence of attenuation, the surface
appearance of Zn anode after the full cell test is studied.
Not surprisingly, the Zn@ZnF2 anode remains smooth and
intact after 1000 cycles test (Figure  5d). Inversely, the bare Zn
anode becomes porous and consists of thin platelets after only
300 cycles (Figure  5e). In order to accurately quantify the gas
production of Zn//MnO2 and Zn@ZnF2//MnO2 full battery, the
4 × 4 cm2 pouch-type cells are fabricated and used to measure the
volume of pouch cell before and after the cycling test. As depicted
in Figure  5f,g, the Zn//MnO2 pouch cell is remarkably swollen
while there is no change of Zn@ZnF2//MnO2 cell. Detailed
measurement of the produced gas volume reveals that the Zn//
MnO2 cell produces about 126.4 times higher gas volume than
that of Zn@ZnF2//MnO2 cell (≈0.5 cm3 for Zn@ZnF2//MnO2
cell vs ≈63.2 cm3 for Zn//MnO2 cell) after 15 cycles’ test. The
hydrogen evolution rate for Zn//MnO2 is about ≈0.016 cm3
(mAh)–1, which is 133.3 times higher than that of Zn@ZnF2//
MnO2 cell (0.00012 cm3 (mAh)–1). The dramatically suppressed
hydrogen evolution during cycling suggests the successful inhi-
bition of side reaction by ZnF2 layer, which thus elevates the
reversibility of electrochemical plating/stripping of metallic Zn.
The exceptional performance of single-layer Zn@ZnF2//
MnO2 pouch cell inspires us to construct high-energy multi-
layer pouch-type batteries. The battery configuration consists of
ten whole sets of anode–separator–cathode stacks with a size
of 4 × 6 cm2, which is fabricated in ambient air environment
without any complicated procedures and protections. As shown
in Figure 5h, the 850 mAh Zn@ZnF2//MnO2 battery presents

Adv. Mater. 2021, 33, 2007406 2007406  (7 of 9) © 2021 Wiley-VCH GmbH

www.advancedsciencenews.com www.advmat.de

Figure 5.  a) Galvanostatic charge/discharge profiles of Zn@ZnF2//MnO2 full cell at 1st, 100th, 500th, and 1000th cycles. b) Galvanostatic charge/
discharge profiles of Zn/MnO2 full cell at 1st, 100th, 200th, and 300th cycles. c) Corresponding cyclic stability of Zn//MnO2 and Zn@ZnF2//MnO2
full cells. d,e) SEM images of Zn@ZnF2 foil after Zn@ZnF2//MnO2 full cell test (d) and Zn foil after Zn//MnO2 full cell test (e). Scale bar: 5 µm.
f,g) Optical photos of Zn//MnO2 (f) and Zn@ZnF2//MnO2 (g) full cells after cycling tests. h) Cyclic stability of 850 mAh large-capacity Zn//MnO2 and
Zn@ZnF2//MnO2 full cells.

excellent cycling performance with a 93.17% initial capacity
retained after 160 cycles at 0.2 C. In comparison, the Zn//
MnO2 battery rapidly decays to 49.78% and fails (short-circuit)
only after 15 cycles. The results demonstrate excellent large-
scale and practical use of our strategies.
In summary, we quantitatively assessed the parasitic
hydrogen evolution flux during Zn deposition process
including a Zn//Zn asymmetric cell and Zn//MnO2 full
cell by in situ battery–gas chromatography–mass spec-
trometry analysis. The hydrogen evolution flux reaches
3.76 mmol h–1 cm–2 in each segment of the Zn//Zn cell, and
maximum 7.70 mmol h–1 cm–2 in Zn//MnO2 (≈3.2 mAh cm–2)
full cell at full charged state. The Zn deposition is loose and
causes dendrite growth due to undesirable side reaction in
aqueous system. To overcome these challenges, we create
a solid Zn2+-ion conductor of ZnF2 to inhibit side reaction
and enhance electrochemical reversibility of Zn metal elec-
trode. The designed Zn@ZnF2 electrode shows remarkably
suppressed hydrogen evolution in comparison to the based
Zn electrode (3.76  vs 0.02 mmol h–1 cm–2 in symmetric cell
and 7.70  vs 0.06 mmol h–1 cm–2 in full cell). Meanwhile, the
Zn@ZnF2 architecture can regulate Zn electrodeposition and
promote the Zn2+ diffusion. Accordingly, the Zn@ZnF2//
Zn@ZnF2 symmetry exhibits highly reversible Zn cycling over
2500 h at 5 mA cm–2 and deep cycling at a high current den-
sity of 10 mA cm–2 and a high areal capacity of 10 mAh cm–2
with an low total over potential of 56.4  mV. As a result, the
Zn@ZnF2//MnO2 full cell with a high areal capacity of
≈3.2 mAh cm–2 can operate over 1000 cycles with 93.63% ini-
tial capacity retained at ≈100% Coulombic efficiency. Finally,
an 850 mAh large-capacity Zn@ZnF2//MnO2 full cell with a
size of 4 × 6 cm2 can reversibly work over 160 cycles (800 h)
with 93.17% initial capacity being retained. This strategy of
advanced Zn metal anode with high reversibility and signifi-
cantly suppressed side reaction opens a way to achieve high-
areal-capacity aqueous Zn metal batteries.

Adv. Mater. 2021, 33, 2007406 2007406  (8 of 9) © 2021 Wiley-VCH GmbH

www.advancedsciencenews.com
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
L.T.M., Q.L., and Y.R.Y. contributed equally to this work. This research
was supported by the National Key R&D Program of China under Project
2019YFA0705104.
Conflict of Interest
The authors declare no conflict of interest.
Keywords
hydrogen evolution suppression, practical-level Zn batteries, quantifying
hydrogen evolution, solid Zn2+-ion conductors, Zn deposition regulation
Received: October 29, 2020
Revised: December 6, 2020
Published online: February 19, 2021
[1] D. Kundu, B. D. Adams, V. Duffort, S. H. Vajargah, L. F. Nazar, Nat.
Energy 2016, 1, 16119.
[2] H.  Pan, Y.  Shao, P.  Yan, Y.  Cheng, K. S.  Han, Z.  Nie, C.  Wang,
J. Yang, X. Li, P. Bhattacharya, K. T. Mueller, J. Liu, Nat. Energy 2016,
1, 16039.
[3] L. Ma, S. Chen, H. Li, Z. Ruan, Z. Tang, Z. Liu, Z. Wang, Y. Huang,
Z. Pei, J. A. Zapien, C. Zhi, Energy Environ. Sci. 2018, 11, 2521.
[4] T. Li, X. Q. Zhang, P. Shi, Q. Zhang, Joule 2019, 3, 2647.
[5] J. Liu, H. Yuan, X. B. Cheng, W. J. Chen, M. M. Titirici, J. Q. Huang,
T. Q. Yuan, Q. Zhang, Mater. Today Nano 2019, 8, 100049.
[6] L. Ma, Y. Ying, S. Chen, Z. Huang, X. Li, H. Huang, C. Zhi, Angew.
Chem., Int. Ed. 2021, 60, 3791.
[7] W.  Yang, X.  Du, J.  Zhao, Z.  Chen, J.  Li, J.  Xie, Y.  Zhang, Z.  Cui,
Q.  Kong, Z.  Zhao, C.  Wang, Q.  Zhang, G.  Cui, Joule 2020, 4,
1557.
[8] F.  Wang, O.  Borodin, T.  Gao, X.  Fan, W.  Sun, F.  Han, A.  Faraone,
J. A. Dura, K. Xu, C. Wang, Nat. Mater. 2018, 17, 543.
[9] L.  Ma, M. A.  Schroeder, O.  Borodin, T. P.  Pollard, M. S.  Ding,
C. Wang, K. Xu, Nat. Energy 2020, 5, 743.
[10] T. Zhang, Y. Tang, S. Guo, X. Cao, A. Pan, G. Fang, J. Zhou, S. Liang,
Energy Environ. Sci. 2020, 13, 4625.
[11] C.  Deng, X.  Xie, J.  Han, Y.  Tang, J.  Gao, C.  Liu, X.  Shi, J.  Zhou,
S. Liang, Adv. Funct. Mater. 2020, 30, 2000599.
Adv. Mater. 2021, 33, 2007406 2007406  (9 of 9)
www.advmat.de
[12] J. Liu, C. Xu, Z. Chen, S. Ni, Z. X. Shen, Green Energy Environ. 2018,
3, 20.
[13] N. Zhang, F. Cheng, J. Liu, L. Wang, X. Long, X. Liu, F. Li, J. Chen,
Nat. Commun. 2017, 8, 405.
[14] D. Kundu, S. Hosseini Vajargah, L. Wan, B. Adams, D. Prendergast,
L. F. Nazar, Energy Environ. Sci. 2018, 11, 881.
[15] L. E. Blanc, D. Kundu, L. F. Nazar, Joule 2020, 4, 771.
[16] H.  Kim, G.  Jeong, Y. U.  Kim, J. H.  Kim, C. M.  Park, H. J.  Sohn,
Chem. Soc. Rev. 2013, 42, 9011.
[17] Z.  Wang, J.  Huang, Z.  Guo, X.  Dong, Y.  Liu, Y.  Wang, Y.  Xia, Joule
2019, 3, 1289.
[18] F. Cheng, J. Chen, Chem. Soc. Rev. 2012, 41, 2172.
[19] Q.  Wang, Z.  Yao, C.  Zhao, T.  Verhallen, D. P.  Tabor, M.  Liu,
F. Ooms, F. Kang, A. Aspuru-Guzik, Y. S. Hu, M. Wagemaker, B. Li,
Nat. Commun. 2020, 11, 4188.
[20] J. Hao, B. Li, X. Li, X. Zeng, S. Zhang, F. Yang, S. Liu, D. Li, C. Wu,
Z. Guo, Adv. Mater. 2020, 32, 2003021.
[21] X.  Xie, S.  Liang, J.  Gao, S.  Guo, J.  Guo, C.  Wang, G.  Xu, X.  Wu,
G. Chen, J. Zhou, Energy Environ. Sci. 2020, 13, 503.
[22] Q. Yang, G. Liang, Y. Guo, Z. Liu, B. Yan, D. Wang, Z. Huang, X. Li,
J. Fan, C. Zhi, Adv. Mater. 2019, 31, 1903778.
[23] L. Ma, S. Chen, N. Li, Z. Liu, Z. Tang, J. A. Zapien, S. Chen, J. Fan,
C. Zhi, Adv. Mater. 2020, 32, 1908121.
[24] L. Ma, S. Chen, X. Li, A. Chen, B. Dong, C. Zhi, Angew. Chem., Int.
Ed. 2020, 59, 23836.
[25] S. Xu, M. Sun, Q. Wang, C. Wang, J. Semicond. 2020, 41, 091704.
[26] Z. Zhao, J. Zhao, Z. Hu, J. Li, J. Li, Y. Zhang, C. Wang, G. Cui, Energy
Environ. Sci. 2019, 12, 1938.
[27] Q.  Yang, Q.  Li, Z.  Liu, D.  Wang, Y.  Guo, X.  Li, Y.  Tang, H.  Li,
B. Dong, C. Zhi, Adv. Mater. 2020, 32, 2001854.
[28] F.  Zhang, F.  Shen, Z.-Y.  Fan, X.  Ji, B.  Zhao, Z.-T.  Sun, Y.-Y.  Xuan,
X.-G. Han, Rare Met. 2018, 37, 510.
[29] W. Guo, Z. Cong, Z. Guo, C. Chang, X. Liang, Y. Liu, W. Hu, X. Pu,
Energy Storage Mater. 2020, 30, 104.
[30] S. B. Wang, Q. Ran, R. Q. Yao, H. Shi, Z. Wen, M. Zhao, X. Y. Lang,
Q. Jiang, Nat. Commun. 2020, 11, 1634.
[31] L. Ma, N. Li, C. Long, B. Dong, D. Fang, Z. Liu, Y. Zhao, X. Li, J. Fan,
S. Chen, S. Zhang, C. Zhi, Adv. Funct. Mater. 2019, 29, 1906142.
[32] L. Ma, S. Chen, C. Long, X. Li, Y. Zhao, Z. Liu, Z. Huang, B. Dong,
J. A. Zapien, C. Zhi, Adv. Energy Mater. 2019, 9, 1902446.
[33] D. Chao, C. Ye, F. Xie, W. Zhou, Q. Zhang, Q. Gu, K. Davey, L. Gu,
S. Z. Qiao, Adv. Mater. 2020, 32, 2001894.
[34] Q.  Li, Y.  Zhao, F.  Mo, D.  Wang, Q.  Yang, Z.  Huang, G.  Liang,
A. Chen, C. Zhi, EcoMat 2020, 2, e12035.
[35] J. C. Ballesteros, P. Díaz-Arista, Y. Meas, R. Ortega, G. Trejo, Electro-
chim. Acta 2007, 52, 3686.
[36] G. Trejo, H. Ruiz, R. O. Borges, Y. Meas, J. Appl. Electrochem. 2001,
31, 685.
[37] J. Lopez, A. Pei, J. Y. Oh, G.-J. N. Wang, Y. Cui, Z. Bao, J. Am. Chem.
Soc. 2018, 140, 11735.
[38] A. Pei, G. Zheng, F. Shi, Y. Li, Y. Cui, Nano Lett. 2017, 17, 1132.
[39] Z. W. Seh, J. Kibsgaard, C. F. Dickens, I. Chorkendorff, J. K. Norskov,
T. F. Jaramillo, Science 2017, 355, eaad4998.
© 2021 Wiley-VCH GmbH