ll

OPEN ACCESS

Article
A High-Performance Magnesium Triflate-
based Electrolyte for Rechargeable
Magnesium Batteries
Dan-Thien Nguyen, Alex Yong
Sheng Eng, Man-Fai Ng, ...,
Albertus D. Handoko, Gomathy
Sandhya Subramanian, Zhi Wei
Seh

sehzw@imre.a-star.edu.sg

HIGHLIGHTS
A simple yet effective electrolyte,
comprising Mg(OTf)2 and MgCl2
in monoglyme

Highly reversible, conditioning-

free, and homogeneous
magnesium deposition

Excellent performance at high

current densities and areal
capacities

Formation of a robust solid

electrolyte interphase at the Mg
anode

The development of rechargeable magnesium batteries is limited by the lack of

simple, commercially available, high-performance electrolytes that enable
reversible plating and stripping of magnesium metal. Nguyen et al. report a
promising magnesium-triflate-based electrolyte, along with investigation of the
solid electrolyte interphase and possible electroactive species.

Nguyen et al., Cell Reports Physical Science 1,

100265
December 23, 2020 ª 2020 The Authors.
https://doi.org/10.1016/j.xcrp.2020.100265
 ll
OPEN ACCESS

Article
A High-Performance Magnesium Triflate-based
Electrolyte for Rechargeable Magnesium Batteries
Dan-Thien Nguyen,1 Alex Yong Sheng Eng,1 Man-Fai Ng,2 Vipin Kumar,1 Zdenek Sofer,3
Albertus D. Handoko,1 Gomathy Sandhya Subramanian,1 and Zhi Wei Seh1,4,*

SUMMARY
The quest for a suitable electrolyte formulation is pivotal to the suc-
cess of rechargeable magnesium batteries. A simple conventional
electrolyte having high compatibility with magnesium anode and
cathode material is in great demand. Herein, we report a simple
yet effective electrolyte formulation, comprising magnesium triflate
(Mg(OTf)2) and magnesium chloride in monoglyme, that can enable
highly reversible, conditioning-free, and homogeneous magnesium
deposition. Galvanostatic Mg plating/stripping demonstrates an
average Coulombic efficiency of 99.4% over 1,000 cycles. The cells
show excellent performance at current densities up to 3 mA cm
2
and areal capacities up to 5 mAh cm
2. Post mortem analysis unveils
the formation of a robust solid electrolyte interphase, which leads to
improved kinetics at the magnesium electrode. A prototype Mg
battery with Mo6S8 cathode demonstrates stable cycling perfor-
mance over 100 cycles. This study shows that rational design of
Mg(OTf)2-based electrolytes is a promising direction toward the
realization of high-performance magnesium batteries.

INTRODUCTION
The integration of renewable energy sources and energy-storage devices is the key
to decarbonize economic growth. Energy-storage devices play a pivotal role in tran-
sitioning away from our dependence on fossil fuels to clean energy sources, such as
wind and solar power, as well as in the evolution of modern telecommunication and
automobile industries. However, the heavy reliance on lithium (Li)-ion battery tech-
nology has placed its supply chain under huge pressure. Therefore, numerous efforts
have been made to develop more-sustainable batteries based on earth-abundant
elements. Magnesium is earth abundant, has a high volumetric capacity (3,833
mAh mL
1) twice that of Li metal, and a low reduction potential of
2.4 V versus stan-
dard hydrogen electrode (SHE).1–3 In addition, Mg deposition tends to form uniform
1Institute
of Materials Research and Engineering,
structures and smooth surfaces rather than highly dendritic structures as seen in
Agency for Science, Technology and Research
lithium and sodium deposition.4,5 Using density functional theory (DFT) calculations, (A*STAR), 2 Fusionopolis Way, Innovis, Singapore
Jäckle and Groß6 showed that Mg exhibits a tendency toward the growth of smooth 138634, Singapore
2Institute
of High Performance Computing,
surfaces as it exhibits lower ion diffusion barriers than lithium and sodium, and as a
Agency for Science, Technology and Research
hexagonal close-packed metal, Mg favors higher coordinated configurations in (A*STAR), 1 Fusionopolis Way, Connexis,
contrast to the body-centered cubic metals, e.g., lithium and sodium. These unique Singapore 138632, Singapore
properties make Mg batteries attractive candidates for safer energy-storage 3Department of Inorganic Chemistry, University
of Chemistry and Technology Prague, Technická
devices.
5, 166 28 Prague 6, Czech Republic
4Lead Contact
Finding an appropriate electrolyte formulation that is uniquely suitable for recharge-
*Correspondence: sehzw@imre.a-star.edu.sg
able magnesium batteries is one of the major technological challenges. Unlike Li-ion
https://doi.org/10.1016/j.xcrp.2020.100265

Cell Reports Physical Science 1, 100265, December 23, 2020 ª 2020 The Authors. 1
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 ll
OPEN ACCESS
Article

chemistry, in which the solid electrolyte interphase (SEI) is highly conductive to Li+
ions, spontaneous reduction of the electrolyte components on the Mg metal surface
results in the formation of a passivating surface layer, which is less conductive toward
Mg2+ ions. Therefore, the search for electrolyte components, particularly Mg salts,
which have high reductive stability against Mg metal, is crucial for improving the
performance of current Mg batteries. To date, a few Mg salts allow Mg plating
and stripping to occur at various degrees of reversibility. The most notable electro-
lyte solutions for Mg batteries are based on the organohaloaluminate complexes,7
magnesium aluminum chloride complex (MACC),8 magnesium monocarborane
(Mg(CB11H12)2),9 amidomagnesium halides,10,11 and magnesium borate.12–15
However, achieving high areal capacity and high current density for Mg plating/
stripping over long cycling process, which is essential for the practical use of Mg
metal anode, is still challenging.16 In addition, new electrolyte systems with a simple
synthetic route, based on commercially available salts and solvents, are in high
demand.

Magnesium trifluoromethanesulfonate, also known as magnesium triflate

(Mg(OTf)2), is thermally stable (with a melting point above 300 C), non-toxic, insen-
sitive to ambient moisture, and commercially available with high purities. To
evaluate its anodic stability, we performed DFT calculations of the highest occupied
molecular orbital (HOMO)/lowest unoccupied molecular orbital (LUMO) levels and
ionization potential (IP) of the triflate anion (OTf
; Figure S1; Table S1). The DFT
results show that OTf
anion has similar anodic stability as bis(trifluoromethanesul-
fonyl)imide (TFSI
) and perchlorate (ClO4
) anions and much higher than that of
hexamethyldisilazide (HMDS
) anion. The ionization potential of the OTf
anion is
4.81 V, which is close to that of TFSI
(4.66 V) and ClO4
(5.04 V) anions. This sug-
gests that Mg(OTf)2 salt is also a promising salt for high-voltage electrolyte system.
An initial study on Mg(OTf)2-based electrolyte conducted by Lossius and Emmeneg-
ger17 successfully observed the deposition of Mg from an electrolyte solution of
Mg(OTf)2 in dimethylacetamide. NuLi et al.18 reported reversible deposition and
dissolution of magnesium from an ionic liquid-based electrolyte consisting of
Mg(OTf)2 and 1-n-butyl-3-methylimidazolium tetrafluoroborate. However, magne-
sium metal passivation was reported to take place due to the negative potential limit
of the ionic liquid cation.19,20 In addition, it is well established that only ethereal
solvents are compatible with Mg metal anode due to their high reductive stability.
However, Mg(OTf)2 has a low solubility in ethereal solvents, such as glymes, resulting
in their incompatibility.21 This is considered to be the biggest challenge for the
development of Mg(OTf)2-based electrolyte for Mg batteries.

Recent studies show that reversible Mg plating/stripping is obtainable in mixed

ether-based electrolytes containing Mg(OTf)2 salt with the addition of multiple
electrolyte additives.21,22 In particular, the solubility of Mg(OTf)2 was improved
by using a mixture of ether solvents, such as tetrahydrofuran (THF) and tetraethy-
lene glycol dimethyl ether (tetraglyme), or by adding a mixture of electrolyte addi-
tives, such as MgCl2, AlCl3, and anthracene.21,22 In addition to being complicated
and consisting of multiple components, the electrochemical performance of the
electrolyte was also unsatisfactory, as the average Coulombic efficiency (CE) was
low,21 and electrochemical conditioning (indicated by no Mg deposition in early
cycles) was required.22 It is worth noting that electrolyte solutions containing
AlCl3 generally exhibit conditioning behavior, which limits the ease of practical
application.23,24 To date, improving the solubility of Mg(OTf)2 in ethereal solvents
using facile means is the key to the success of Mg(OTf)2-based electrolytes for Mg
batteries.

2 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

In this work, we carefully design and conduct a systematic study on the combination
of Mg(OTf)2 and MgCl2 salt in monoglyme (1,2-dimethoxyethane [DME]).
Intriguingly, we find that, by tuning the [Mg2+]:[Cl
] molar ratio above 1, a high
concentration of Mg(OTf)2-MgCl2 can be completely dissolved into the electrolyte.
The resulting electrolyte solution exhibits a superior electrochemical performance in
Mg plating and stripping, showing dendrite-free deposition, high CE, and long cycle
life (99.4% over 1,000 cycles at 0.5 mA cm
2). Toward the practical use of Mg metal
anode, the experiments are conducted on a conventional cell configuration using
thin Celgard separator and low electrolyte volume and tested at high current
densities and high areal capacities. In addition, the deposition morphology of Mg
under various current densities, the surface chemistry of the SEI formed on Mg elec-
trode, and the electrolyte’s compatibility with Mo6S8 cathode are also carefully
investigated.

RESULTS AND DISCUSSION

Homogeneous and Reversible Mg Deposition and Dissolution
One of the greatest challenges for the utilization of Mg(OTf)2 as a salt in Mg battery
electrolytes is its low solubility in ethereal solvents, especially monoglyme and di-
glyme, which are commonly used as solvents for Mg electrolytes due to their high
reductive stability against Mg metal anode. Despite its low solubility, the galvano-
static plating/stripping of a saturated Mg(OTf)2 in DME electrolyte showed revers-
ible plating/stripping at 0.05 mA cm
2 and 0.05 mAh cm
2 with an average CE of
74.3% over 50 cycles and low overpotential of Mg plating (
0.4 V; Figure S2). On
the contrary, Mg(TFSI)2 alone in DME demonstrated poor Mg plating (Figure S3).
This result suggests that Mg(OTf)2 is highly compatible with Mg metal anode and
improving its solubility in DME may lead to a high-performance electrolyte solution.

In this study, we carefully designed and conducted a systematic investigation to

improve the solubility of Mg(OTf)2 in DME solvent. MgCl2 is frequently used in Mg
electrolytes and demonstrates effectiveness in improving the compatibility of the
electrolyte with the Mg metal anode.25–27 It should be noted here that the solubility
of MgCl2 alone in DME is very low due to low chemical polarity of DME.28 Various
combinations of Mg(OTf)2 and MgCl2 in DME were prepared (Figure S4), with
[Mg(OTf)2]:[MgCl2] ratios ranging from 0.25 to 4, although the total concentration
of Mg2+ in solution was fixed at 0.5 M. The results showed that clear and stable elec-
trolyte solutions (i.e., complete dissolution) could be obtained at [Mg(OTf)2]:[MgCl2]
ratios greater than 1, corresponding to [Mg2+]:[Cl
] ratios greater than 1 as well (Fig-
ures S4A–S4C). When the ratio of [Mg(OTf)2]:[MgCl2] = [Mg2+]:[Cl
] = 1, the initially
clear solution (Figure S4D) was instead unstable and crystals were formed after a few
days at room temperature (Figure S4G). In mixtures where both molar ratios are less
than 1, the salts were found to be insoluble (Figures S4E and S4F). Interestingly, a
clear electrolyte can be obtained with the total concentration of Mg2+ up to 2.5
M, corresponding to the electrolyte formula of 1.5 M Mg(OTf)2 + 1 M MgCl2 in
DME. Previously, Yang et al.21 attempted to combine Mg(OTf)2 and MgCl2 in
DME at a molar ratio of 0.5, but a clear electrolyte solution was not obtainable.
This result is consistent with our solubility findings above, which suggests that the
molar ratio between Mg(OTf)2 and MgCl2 plays an unexpectedly important role in
improving the solubility of their combination. At 0.5 M Mg2+, all soluble electrolyte
combinations exhibited an ionic conductivity of approximately 0.3 mS cm
1 (Table
S2). The low ionic conductivity of the electrolyte solution is due to strong interaction
between OTf
anion and Mg2+, similar to that observed in LiOTf-based electro-
lyte.29 By increasing Mg2+ concentration to 1 M, a higher ionic conductivity was

Cell Reports Physical Science 1, 100265, December 23, 2020 3

 ll
OPEN ACCESS
Article

obtained at 0.53 mS cm
1. However, upon further increasing the Mg2+ concentra-
tion above 1 M, we noticed the increased viscosity of the electrolyte solution, which
may reduce the mobility of ions in the electrolyte solution.

Highly reversible plating/stripping of the Mg anode is an essential factor to evaluate

the electrolyte performance, especially in metal anode batteries. Hence, Mg//Al
asymmetric cells were assembled using 2032-type coin cells to measure the Mg
plating/stripping CE of various Mg(OTf)2-MgCl2 combinations (Figure 1) to deter-
mine the optimal composition. In each cycle, a fixed areal capacity (0.1 mAh
cm
2) of Mg was deposited onto the carbon-coated aluminum (Al) electrode at a cur-
rent density of 0.5 mA cm
2, which was then stripped by charging the cell to 1.2 V
versus Mg/Mg2+ at the same current density. The CE for each cycle was calculated
based on the ratio of the capacity of stripped Mg to that of plated Mg. Among
the three stable electrolyte combinations, 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME
demonstrated the highest average CE of 99.4% for 1,000 cycles. The cell also shows
an initial CE (ICE) of 89.9% and reached 95.8% in the second cycle. In addition, it also
exhibited the lowest Mg plating and stripping overpotential at
0.17 V and +0.17 V
(Figure S5), respectively. It should be noted that all electrolyte combinations with
Mg2+ concentration of 0.5 M have similar ionic conductivity, implying that the
improved cycling performance of 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME is related
to the concentration of chloride in the electrolyte. It is well established that MgCl2
plays a crucial role in enhancing Mg plating/stripping: it can inhibit Mg surface
passivation through the formation of adsorbed Cl
and/or MgCl2 on the Mg surface
as well as through a dynamic competition with adventitious H2O impurity in the dou-
ble layer.30

To evaluate the effect of doubling the total salt concentration on the reversibility of
the anode plating/stripping process, 0.6 M Mg(OTf)2 + 0.4 M MgCl2 in DME electro-
lyte was also prepared for comparison (Figure 1A). The asymmetric cell showed
lower ICE (75.6%), but the CE rapidly reached 98.8% after the second cycle. Over
986 cycles, the cell exhibited an average CE of 99.2%. However, a slight decrease
of CE after 500 cycles was observed, and the Mg plating and stripping overpotential
also increased to
0.22 V and 0.23 V (Figure S5D), respectively. This is due to the
increase of contact ion pairs in Mg electrolyte at a higher salt concentration as re-
ported in the literature.31 Herein, we conclude that 0.3 M Mg(OTf)2 + 0.2 M
MgCl2 in DME is the optimal electrolyte formula for highly reversible Mg plating/
stripping among the solutions tested. For comparison, a similar experiment was
performed on 0.25 M Mg(TFSI)2 + 0.5 MgCl2 in DME electrolyte, which is a common
formula used in Mg(TFSI)2-based system (Figures S6A and S6B).27 The asymmetric
cell showed an average CE of 94.0% and a cycle life of around 100 cycles, lower
than our developed electrolytes, which indicates the superior performance of
Mg(OTf)2 salt to Mg(TFSI)2 (Figure S6C).

In order to obtain a thorough understanding of Mg(OTf)2-MgCl2 electrolyte perfor-

mance, we conducted various electrochemical characterizations as shown in Figures
1B–1E. A typical cyclic voltammogram measured with Pt working electrode in 0.3 M
Mg(OTf)2 + 0.2 M MgCl2 in DME (Figure 1B) exhibited highly reversible and condi-
tioning-free Mg plating/stripping. The first cycle demonstrated a large Mg deposi-
tion overpotential at
1.0 V versus Mg/Mg2+, but it quickly reduced to
0.3 V in the
second cycle, which is well consistent with voltage profiles of the galvanostatic
cycling in Figure 1D. The large overpotential in the first cycle was probably due to
the elimination of thin MgO oxide on Mg anode, which was formed by the reduction
of trace moisture from the glovebox and/or the reduction of impurities during the

4 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

Figure 1. Electrochemical Characterization of Mg(OTf)2
MgCl2 in DME Electrolyte

(A) Plating/stripping Coulombic efficiency of Mg//Al asymmetric cell using various compositions of Mg(OTf) 2
MgCl 2 in DME electrolyte at 0.5 mA
cm
2 and 0.1 mAh cm
2 .
(B–E) Electrochemical characterizations of Mg//Al asymmetric cell using 0.3 M Mg(OTf) 2 + 0.2 M MgCl2 in DME.
(B) Cyclic voltammetry at 5 mV s
1 between
1.5 V and 2.2 V.
(C) Linear sweep voltammetry on various metal electrodes using three-electrode flood cell at 5 mV s
1 .
(D and E) Voltage profile (D) and Coulombic efficiency (E) of galvanostatic cycling of Mg//Al asymmetric cell at 0.5 mA cm
2 and 0.5 mAh cm
2 .

first cathodic scan. The first positive scan exhibited a large cathodic peak, beginning
near 0 V versus Mg/Mg2+. In subsequent cycles, the cyclic voltammogram showed
highly reversible Mg deposition and dissolution with low Mg deposition overpoten-
tial (
0.3 V) and high cathodic peak in the positive scan. The anodic stability of 0.3 M
Mg(OTf)2 + 0.2 M MgCl2 in DME was further studied by linear sweep voltammetry
using inert Pt electrode and non-noble metal electrodes (Figure 1C). On the inert

Cell Reports Physical Science 1, 100265, December 23, 2020 5

 ll
OPEN ACCESS
Article

Pt electrode, the onset of the electrolyte oxidation appeared at 3.0 V versus Mg/
Mg2+, which is close to that of Mg(TFSI)2-MgCl2 system reported by Aurbach’s
group.27 On non-noble metal electrodes, including Al, stainless steel (SS), nickel
(Ni), and Mo, the oxidation current began to increase at 1.9 V, 2.0 V, 2.4 V, and
3.1 V, respectively, due to chloride corrosion.

Figures 1D and 1E show the galvanostatic cycling performance of the Mg//Al asym-
metric cell at an areal capacity of 0.5 mAh cm
2 using a current density of 0.5 mA
cm
2. During initial plating, the asymmetric cell shows an overpotential of
0.4 V
versus Mg/Mg2+ (Figure 1D), which is probably due to the elimination of the thin
MgO layer on Mg metal anode. The first Mg dissolution showed a low overpotential
at 0.2 V versus Mg/Mg2+ and delivers a CE of 84.8%. The irreversible capacity cor-
responds to the reduction of trace moisture and impurities presented in electrolyte
and cell components. In subsequent cycles, the Mg deposition and dissolution over-
potentials were reduced and well maintained near G0.2 V over 100 cycles. The CE
was 97.3% in the second cycle, reached 99.1% by the 7th cycle, and stabilized at 99%
over 100 cycles, indicating a typical behavior of a conditioning-free electrolyte sys-
tem. The cell delivered an average CE of 99.1% over 100 cycles. At 1 mA cm
2 and 1
mAh cm
2, the asymmetric cell also demonstrated an excellent Mg plating/stripping
with average CE of 99.2% (Figure S7). This electrolyte system demonstrated a signif-
icant performance enhancement compared to other Mg(OTf)2 electrolyte systems
reported in literature and is highly competitive with other reported systems (Table
S3), especially in terms of CE, areal capacity, and current density.21,22 It should be
noted here that the excellent cycling performance of the asymmetric cell using
0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME electrolyte was achieved using an experi-
mental setup that is close to the practical requirement of batteries, such as the
use of thin Celgard separators, unlike the majority of other works, which use glass
fiber separators.13,15,16,22,32 In addition, we also demonstrate low electrolyte vol-
ume (25 mL per coin cell), relatively high current density (0.5–1 mA cm
2), and areal
capacity of 0.5–1 mAh cm
2 as indicated in Table S3.

In addition to the CE, the morphology and composition of the deposited Mg film are
also crucial to the cycle life of Mg batteries. Figure 2 demonstrates various charac-
terizations conducted on Mg film deposited on Al foil at 0.5 mA cm
2. The Mg
film was uniform, crystalline, dense, and free of dendritic structures (Figures 2A
and 2B), which is very important to deliver high energy density and safe operation
of Mg batteries. X-ray diffractometer (XRD) measurement was also performed on
the Mg deposits (Figure 2C) to confirm its crystal structure. The measurement was
conducted on Mg film plated at 0.5 mA cm
2 and 3 mAh cm
2. The XRD pattern
clearly confirmed that the agglomerated deposits are pure Mg polycrystalline in na-
ture (PDF card no. 00-001-1141). Energy-dispersive X-ray (EDX) spectrum (Figure 2D)
and elemental mapping (Figure S8) of the sample confirms that the Mg deposits
were of high purity with Mg content of 91.6%. Other elements, such as C, O, F, S,
Cl, and Al, were also present but at very low percentages (Figure S8).

The galvanostatic cycling performance of asymmetric cells was also examined under
various cycling conditions to provide a useful benchmark for the evaluation of elec-
trolyte performance. Toward practical application of Mg batteries, the reversibility
of Mg plating/stripping at high current density and high areal capacity is crucial
for achieving high power and high energy densities, which has received great atten-
tion from recent studies.15,16,33,34 Figures 3A and 3B demonstrate the cycling perfor-
mance of Mg//Al cells at various current densities from 0.5 mA cm
2 to 3 mA cm
2
while keeping the areal capacity constant at 0.5 mAh cm
2. It is well known that the

6 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

Figure 2. Analysis of Mg Deposition on Al Foil Using 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME
Electrolyte
(A and B) SEM images of Mg deposits at 0.5 mA cm
2 , 1 mAh cm
2 .
(C) XRD pattern of Mg deposits at 0.5 mA cm
2 , 3 mAh cm
2 .
(D) EDX elemental analysis on Mg deposits.

electrodeposition of metals at high current density will accelerate dendritic growth,

which will lead to decreased CE and internal short circuiting. As the current densities
increased (Figure 3A), the Mg plating and stripping overpotential increased from
0.15 V at 0.5 mA cm
2 to 0.46 V at 3 mA cm
2. Despite the increasing overpotential,
the average CE remained high. At 0.5, 1, 1.5, 2, 2.5, and 3 mA cm
2, average CEs of
96.7%, 97.0%, 96.6%, 97.6%, 98.4%, and 98.4% could be attained, respectively
(Figure 3B). This result suggests that the 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME elec-
trolyte is promising for high-power applications. In addition, toward high-energy ap-
plications, we also examined Mg deposition at high areal capacity while keeping the
current density at 0.5 mA cm
2 (Figures 3C and 3D). At areal capacities of 0.25, 0.5,
1, 1.5, 2, 2.5, 3, and 5 mAh cm
2, the cell demonstrated high average CE of 96.0%,
99.1%, 99.3%, 99.2%, 99.3%, 99.3%, 99.4%, and 99.4%, respectively. This result
suggests that Mg deposition and dissolution occurred at very high reversibility
with similar overpotentials even at high areal capacity up to 5 mAh cm
2.

Cell Reports Physical Science 1, 100265, December 23, 2020 7

 ll
OPEN ACCESS
Article

Figure 3. Galvanostatic Cycling of Mg//Al at Different Current Densities and Areal Capacities Using 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME
(A and B) Voltage profile (A) and Coulombic efficiency (B) of cell cycled with current densities from 0.5 to 3.0 mA cm
2 at 0.5 mAh cm
2 .
(C and D) Voltage profile (C) and Coulombic efficiency (D) of cell cycled with a current density of 0.5 mA cm
2 at areal capacities between 0.25 and 5 mAh
cm
2 . The voltage profile is selected from the last cycle of each cycling step.

In consideration of practical battery applications, the electrochemical behavior and

morphology of the Mg metal anode are crucial to the full-cell performance. There-
fore, we carefully investigated the performance of Mg//Mg symmetric cells, in which
the cell performance is completely dependent on Mg plating and stripping
occurring at the Mg electrode surface. Figure 4 presents various electrochemical
characterization of Mg//Mg symmetric cells. It should be noted here that the cell
configuration in this study used 2 layers of Celgard separators (25 mm thick) filled
with 25 mL of electrolyte solution per coin cell. In our separate experiment with a
thicker glass fiber separator (680 mm thick), the cell showed much longer cycle life
due to increased distance between the two Mg electrodes that mitigates short cir-
cuiting, but the use of such a thick separator is less practical (Figure S9). Figures
4A and 4B present the stable cycling performance of Mg//Mg symmetric cell over
500 h at 0.5 mAh cm
2 and 1 mAh cm
2, corresponding to 250 cycles and 125 cycles,
respectively, both at current density of 0.5 mA cm
2. Figures 4C–4E provide a
detailed view of the voltage change of Figure 4A at the first 10 h, between 250
and 260 h and 490 and 500 h, respectively. In the first cycle (Figure 4C), high over-
potential is observed in both charge and discharge profiles, which is probably due
to the presence of a thin MgO film on the Mg metal electrode. In subsequent cycles,
the overpotential gradually reduced to 0.17 V versus Mg/Mg2+ (Figure 4D). After 500
cycles, a low overpotential was observed at 0.3 V versus Mg/Mg2+ (Figure 4E). To
understand the change of overpotential with cycle number, electrochemical imped-
ance spectroscopy (EIS) was conducted on Mg//Mg symmetric cells at various states

8 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

Cell Reports Physical Science 1, 100265, December 23, 2020 9

 ll
OPEN ACCESS
Article

of the cycling process (Figures 4F and 4G). Large cell impedance was observed after
resting the cell at open-circuit voltage (OCV) for 3 h before cycling, which corre-
sponds to the large overpotential in initial cycle. In subsequent cycles, the symmetric
cell shows a gradual decrease of the cell impedance with cycle number, which is due
to the elimination of surface oxide and formation of fresh Mg metal by repeated
deposition/dissolution. A similar observation was also reported by Zhao-Karger
et al.15 in the magnesium tetrakis(hexafluoroisopropyloxy) borate electrolyte.

Due to the inherent nature of the electrochemical deposition process, dendritic struc-
tures typically form when a high current density is applied to deposit metals. In the case
of Mg, the maximum current density limit that allows for dendrite-free Mg plating is
highly dependent on the electrolyte solution used. Thus, it is important to determine
this current density limit to avoid the dendritic growth of Mg with our electrolyte sys-
tem.35 Figure 4H presents the voltage versus time curve of an Mg//Mg symmetric
cell cycled at 0.5 mAh cm
2 with current density gradually increased from 0.5 mA
cm
2 to 3 mA cm
2. The overpotential increased from 0.17 V at 0.5 mA cm
2 to
0.3 V at 1 mA cm
2 and 0.5 V at 2 mA cm
2. Even at high current densities, the sym-
metric cell showed flat potential curves during plating/stripping, implying a smooth
Mg cycling process. At a current density of 3 mA cm
2, the overpotential increased
significantly to 0.8 V and a short circuit occurred in the 3rd cycle. The scanning electron
microscope (SEM) images show a significant change in the shape of Mg deposits to a
dendritic structure at 3 mA cm
2 (Figure S10). This result suggests that the limiting cur-
rent density for dendrite-free deposition in our electrolyte is 2 mA cm
2. It is worth
noting that this limiting current could be improved by enhancing the ionic conductivity
of the electrolyte solution, which will be reported in future work.

To gain insight into the Mg plating/stripping process at the Mg metal anode, we

conducted a post mortem SEM analysis of Mg//Mg symmetric cells. Figure 5 pre-
sents SEM images of Mg electrodes from a symmetric cell after 3 cycles (0.5 mA
cm
2 and 0.5 mAh cm
2), followed by a 6-h plating/stripping at 0.5 mA cm
2, cor-
responding to an areal capacity of 3 mAh cm
2. In the plated state (Figures 5A
and 5B), the uniform and crystalline structure of Mg deposition were clearly ob-
tained, even at a high areal capacity of 3 mAh cm
2. For the electrode disassembled
in the stripped state (Figures 5C and 5D), the Mg deposits from previous cycles were
completely removed and a clean surface was observed. In both cases, the electrode
surface became uneven but shows no dendritic structure formation as illustrated in
Figure 5E. To further analyze the Mg deposition morphology, we acquired a cross-
section SEM image of an Mg electrode after Mg deposition (Figure 5F). The film ap-
pears to be dense and uniform with an average thickness of 13.3 mm at 3 mAh cm
2.

Formation of a Robust Solid Electrolyte Interphase on Mg Anode

Because the Mg anode-electrolyte interface is crucial to the reversibility of the Mg
plating/stripping process, we carried out X-ray photoelectron spectroscopy (XPS)
in depth profiling mode to examine the surface physiochemistry of Mg metal anode
from the top surface to the bulk of the SEI. Figure 6 presents the XPS spectra
obtained from the Mg metal electrode from a symmetric cell after 20 cycles at the

Figure 4. Electrochemical Characterization of Mg//Mg Symmetric Cell Using 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME
(A and B) Voltage profile of symmetric cell with a current density of 0.5 mA cm
2 at an areal capacity of (A) 0.5 mAh cm
2 and (B) 1 mAh cm
2 for 500 h.
(C–E) Expanded voltage profile of (A) at (C) 0–10 h, (D) 250–260 h, and (E) 490–500 h.
(F) Nyquist plots of Mg//Mg symmetric cell at various points during cycling.
(G) Expanded Nyquist plots at high-frequency region.
(H) Voltage profile of Mg//Mg symmetric cell at 0.5, 1, 2, and 3 mA cm
2 at areal capacity of 0.5 mAh cm
2 .

10 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

Figure 5. Morphology of Mg Electrodes from a Mg//Mg Symmetric Cell after Cycling in 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME after Plating or
Stripping at 0.5 mA cm
2 for 6 h (3 mAh cm
2)
(A and B) SEM images of Mg electrode after plating.
(C and D) SEM image of Mg electrode after stripping.
(E) A schematic illustrates the morphology change of the Mg electrodes in a symmetric cell.
(F) Cross-section image of the Mg electrode after plating where the arrows refer to the Mg deposited.

current density of 0.5 mA cm
2 and areal capacity of 0.5 mAh cm
2. The electrode
was removed from a coin cell in the plated state to be examined under XPS analysis,
carefully avoiding any ambient contamination or decomposition of surface species
as described in the experimental section. Due to the weak C 1 s signal, all spectra
herein were calibrated using the Mg 2p peak of Mg0 at 49.8 eV.36

Figure 6 and Table S4 present the XPS spectra collected at the top surface (0 min)
after Ar etching for 30 s and for 5 min. Before etching (0 min), the C 1 s spectrum
contains peaks at 284.7 eV, 286.2 eV, and 288 eV, which is attributed to the C-H/
C-C, C-O-C, and C = O signals, respectively.36,37 These peaks originate from the
ether and ester moieties. The presence of a strong peak at 533.7 eV in O 1 s spec-
trum further supports this assignment.38,39 These organic species originated from
the reduction of the electrolyte solvent DME at the Mg surface. The C 1 s peak
with low binding energy of 283.8 eV is assigned to the Mg-C bonding.38,40

The Mg 2p spectrum contains characteristic peaks corresponding to Mg0 (49.8 eV),

Mg(OH)2 (50.3 eV), MgO/MgF2 (51 eV), and MgCl2/MgCO3 at 51.8 eV.30,36,41,42 We
also noticed the presence of a broad peak at a low binding energy of 48.9 eV. This
peak is attributed to Mg dangling bonds,43 which refers to Mg atoms residing on the
surface of crystals with less coordination compared to that of Mg atoms in the bulk
crystal. The existence of MgO and Mg(OH)2 in the XPS spectrum is due to the reduc-
tion of trace amount of moisture in the electrolyte solvent and/or electrolyte decom-
position.30 The MgO and Mg(OH)2 are also detected in the O 1 s spectrum with two
strong signals at 529.1 eV and 531.4 eV, respectively.38,44,45 The Mg 2p peaks at 51
eV together with F 1 s peak at 685.5 eV (C-F) confirms the presence of MgF2 in the

Cell Reports Physical Science 1, 100265, December 23, 2020 11

 ll
OPEN ACCESS
Article

Figure 6. Investigation of SEI on Mg Anode

XPS surface analysis of Mg electrode cycled in 0.3 M Mg(OTf) 2 + 0.2 M MgCl 2 in DME electrolyte for 20 cycles.

top surface of Mg metal anode. MgF2 originates from the decomposition of
CF3
from OTf
anion, which is the only source of fluorine in the electrolyte solution. In
addition, a strong signal at 688 eV (CF3) in F 1 s spectrum occurs due to the presence
of residual OTf
, which is further confirmed by an S 2p peak at 168.6 eV.46,47 The O
1 s peak at 533.7 eV is assigned to O = C-O*.48,49 The high-resolution spectrum of Cl
2p shows a broad peak between 204 eV and 196 eV, which can be fitted using 2 spin-
orbit split doublets with the binding energy of Cl 2p3/2 at 200.4 eV and 198 eV. The

12 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

Cl 2p3/2 peak at 200.4 eV is a characteristic of Cl atoms in stoichiometric

MgCl2.44,50,51 The residual MgCl2 on the cycled electrode is due to its low solubility
in DME. The Cl 2p3/2 peak appearing at low binding energy (198 eV) is assigned to
organic chloride salts and/or chloride in metal complexes.36,41,44,50 The result sug-
gests the formation of different chloride-containing species at the Mg anode-elec-
trolyte interface, confirming the two major roles of MgCl2: (1) it interacts with
Mg(OTf)2 to improve the solubility of the two salts and (2) the adsorption of various
chloride-containing species at the Mg metal-electrolyte interface improves Mg
plating/stripping and prevents the Mg metal surface from being passivated by elec-
trochemical reduction products.30 In the S 2p spectrum, MgS is also detected with a
strong signal at 162.2 eV in addition to the presence of OTf
anion, which originates
from the reduction of OTf
anion.50,52 Overall, the top of the surface film contained
both organic species (ether and ester moieties), inorganic components (MgO,
Mg(OH)2, MgF2, MgS, and MgCl2), and adsorbed chloride-containing components.

After conducting Ar etching on Mg metal anode, we saw remarkable changes in the

composition of its interior. First, after a short time of etching (30 s), the C 1 s and O 1 s
peaks corresponding to organic species are significantly weakened and became
indistinct. In addition, all peaks related to OTf
in F 1 s and S 2p also decreased
or disappeared. This result suggests that the top layer containing reduced organic
species was very thin. Second, the peaks corresponding to Mg metal (Mg 2p) and
other inorganic components, such as MgO, Mg(OH)2, MgF2, and MgS intensified,
which is due to the removal of the top organic layer. The Cl 2p peaks corresponding
to chloride species remained strong and distinct. Upon further etching up to 5 min,
we observed that the intensity of the Mg metal peak (Mg0) increased at the expense
of other components, such as MgO, Mg(OH)2, and MgF2. Chloride-containing spe-
cies and MgS were mostly diminished after 5 min etching. Further etching up to
15 min yielded the same trend. The atomic concentration of various elements as a
function of etching time for the surface film on Mg metal anode after cycling is pre-
sented in Figure S11A. This result indicates that the interior of the surface film con-
sisted largely of inorganic components, including MgO, Mg(OH)2, MgCl2, MgS, and
MgF2, with very little organic species arising from the reduction of electrolyte
solvent. Note that the concentration of oxygen remains high even with increased
etching time, which is also observed in other metal anodes, such as Li and
Na.30,53,54 Connell et al.30 suggested that the presence of MgO/Mg(OH)2 even after
a deep etching is due to the heterogeneous nature of deposits. To explain this phe-
nomenon, we propose the structure of Mg metal anode in Figure S11B. The Mg de-
posits on Mg anode (Figures 5A and 5B) consist of agglomerated Mg polycrystalline
with high surface roughness. The high concentration of MgO/Mg(OH)2 after Ar
etching is possibly due to the high surface roughness and/or their presence within
the interior surface of polycrystalline Mg deposit layers. Overall, we conclude here
that the surface film on Mg metal anode consists of organic and adsorbed chlo-
ride-containing species on the top surface and mostly inorganic components, such
as MgO, Mg(OH)2, MgCO3, MgS, MgCl2, and MgF2, within the interior (Figure S11B).
This largely inorganic surface film plays an important role in enabling the reversibility
of Mg deposition and dissolution. It is well known that solid phases, such as MgF2 or
MgO, are not conductive to Mg2+ due to its high charge density. Because high
reversibility is still observed at the Mg anode, we propose that the inorganic layer
is not entirely compact; instead, it contains pinholes that allow Mg2+ diffusion,
similar to that of hexafluorophosphate-based solutions reported by Aurbach’s
group.55 The chloride is expected to play an important role in engineering the struc-
ture and composition of this layer. Further investigation on the structure and Mg
mobility through this SEI will be reported in future work.

Cell Reports Physical Science 1, 100265, December 23, 2020 13

 ll
OPEN ACCESS
Article

Figure 7. Cycling of Mg//Mo6S8 Full Cell Using 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in DME at 0.1 C and
Voltage Window of 0.2–2.2 V
(A) Voltage profile.
(B) Capacities versus cycle number.

Electrochemical Performance of Mg//Mo6S8 Full Cell

Toward the practical application of this electrolyte, its compatibility with cathode
materials and its stability in the working voltage window of the cathode also needs
to be characterized. We examined the ability of 0.3 M Mg(OTf)2 + 0.2 M MgCl2 in
DME to support reversible Mg intercalation into a Chevrel phase (Mo6S8) cathode
(Figures S12 and S13). In this experiment, the full-cell test was performed using
2032-type coin cell with Celgard separator, and the cathode was prepared by a con-
ventional slurry route using polyvinylidene fluoride (PVDF) as the binder and Ni foil
as the current collector. This experimental setup allows us to examine our electrolyte
in a condition that is close to that of practical cells. Figure 7 presents the voltage pro-
file and cycling performance of Mg//Mo6S8 after one formation cycle. The full cell
showed high overpotential in the early cycles, which is due to the high overpotential
of Mg plating as seen in the symmetric and asymmetric cells. In subsequent cycles,
the overpotential reduced gradually with a discharge plateau at 1 V and charge
plateau at 1.3 V. This is consistent with previous reports of Mo6S8 cathode in Mg bat-
teries.9,32,56–59 The cell demonstrated a reversible capacity of 75 mAh g
1 and stable
cycling performance over 100 cycles with a capacity retention of 90.8% at 0.1 C, indi-
cating high compatibility with cathode materials up to 2.2 V. The electrolyte appears

14 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

Figure 8. Structural Analysis of Possible Electroactive Species

Refined SCXRD structure for Mg(DME) 2 (OTf) 2 single crystal, recrystallized from 1.5 M Mg(OTf) 2 +
1 M MgCl2 in DME solution.

to be promising for future development of Mg battery by combining with high ca-

pacity and fast kinetic cathode materials, such as metal chalcogenide cathodes.60,61

Reaction Mechanism of Mg(OTf)2 and MgCl2 in DME

In order to explain the improved solubility of Mg(OTf)2 and MgCl2 combination in
DME, we conducted a structural analysis of the electrolyte solution using single-crys-
tal XRD (SCXRD) (Figure 8) and Raman spectroscopy (Figure 9). To induce crystal
precipitation, an electrolyte with a high concentration of 1.5 M Mg(OTf)2 + 1 M
MgCl2 in DME solution was prepared to grow the single crystal. The single-crystal
structure refinement yielded the structure shown in Figure 8 and Table S5. The
single-crystal structure with the formula Mg(DME)2(OTf)2 clearly shows an Mg2+
cation encaged by two DME molecules and two OTf
anions without any coordina-
tion to chloride ions. Mg(OTf)2 acquires a V shape with an equilibrium angle of
91.51 between the oxygen atoms of the triflate group (Figure 8). This structure is
similar to that of MgCl2 in DME, which forms MgCl2(DME)2 solution species.62,63

Raman spectra of various triflate anion-containing salts in DME solvent and Mg(OTf)2
powder are presented in Figure 9. The DME spectrum (Figure 9A) shows strong sig-
nals at 367, 397, 821, 849, 994, and 1,025 cm
1. In all Mg(OTf)2-containing electro-
lyte solutions (Figures 9E–9H), we observed three major trends. First, the absence of
the peak in the 200–300 cm
1 region suggests that cation complexes, such as
[Mg2(m-Cl)2(DME)4]2+ and [Mg3(m-Cl)4(DME)5]2+, do not exist in these solutions,63
at least at the detection level of these measurements. Second, a new peak appears
at 876 cm
1, which is assigned to DME cage in Mg(DME)2(OTf)2. It should be noted
that the DME cage in [Mg(DME)3]2+ complex, which is well studied in Mg(TFSI)2-
based electrolyte, shows a signal at 881 cm
1,63 confirmed by the spectrum of
Mg(TFSI)2/DME (Figure 9B). Third, the existence of the Mg-OTf ion pair is also evi-
denced by the red shift of vibrational mode at 773 cm
1 to 762 cm
1 and the blue
shift of vibrational mode at 1,038 cm
1 to 1,049 cm
1 of triflate anion (Figure 9C).
In addition, we also observed the presence of free OTf
anion, evidenced by the sig-
nals at 311 and 346 cm
1. The Raman spectrum of 1-butyl-3-methylimidazolium tri-
fluoromethanesulfonate ([BMIM][OTf]) and DME (1:5 vol. ratio) solution (Figure 9D)
supported this assignment. The absence of chloride-containing species in Raman
spectra suggest that Cl
might form cation that is not detectable by Raman spectros-
copy, such as [MgCl]+. 35Cl nuclear magnetic resonance (NMR) measurements

Cell Reports Physical Science 1, 100265, December 23, 2020 15

 ll
OPEN ACCESS
Article

Figure 9. Raman Spectra of Various Mg(OTf)2-Containing Solutions and Reference Samples

(A) DME solvent only.
(B) 0.75 M Mg(TFSI) 2 in DME.
(C) Mg(OTf) 2 powder.
(D) [BMIM][OTf]-DME mixture solution (1:5 vol. ratio).
(E) Saturated solution of Mg(OTf) 2 in DME.
(F) 0.4 M Mg(OTf) 2 + 0.1 M MgCl 2 in DME.
(G) 0.3 M Mg(OTf) 2 + 0.2 MgCl2 in DME.
(H) 1.5 M Mg(OTf) 2 + 1 M MgCl 2 in DME.

showed that these solutions contain no free chloride ions (Figure S14). Characteriza-
tion of the electroactive complex, especially chloride-containing species, would
require extensive research and will be the subject of future studies.

The formation of Mg(DME)2(OTf)2 crystal structure is explained as follows: (1) the

composition of electroactive species reported in the literature is generally based on
the [Mg2+]:[Cl
] ratio in the electrolyte solution.63 In the present Mg(OTf)2-MgCl2 elec-
trolyte, the [Mg2+]:[Cl
] molar ratio must be larger than 1 (Figure S4; Table S2). Thus,
electroactive species, such as [Mg3(m-Cl)4(DME)4]2+ or even [Mg2(m-Cl)2(DME)4]2+, are
unlikely to form as the main species, especially when the [Mg2+]:[Cl
] molar ratio is as

16 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

high as 2.5, as evidenced by Raman spectroscopy (Figure 9). This result suggests that
MgCl2 does not stoichiometrically react with Mg(OTf)2 to form the electroactive spe-
cies seen in the low [Mg2+]:[Cl
] ratio systems, such as [Mg2(m-Cl)2(DME)4](TFSI)2,
[Mg3(m-Cl)4(DME)4](TFSI)2, and [Mg2(m-Cl)2(DME)4](AlCl4)2.63 (2) With regard to the
OTf
anion, the strong electron-withdrawing effect of the terminal
CF3 group helps
to delocalize the negative charge over the
SO3 group. However, the charge density
of the OTf
anion is expected to be higher than that of TFSI
anion, where all the elec-
tronegative sulfur, nitrogen, and oxygen atoms can participate in dispersing the nega-
tive charge, leading to good charge delocalization.29 In other words, the formation of
the Mg-OTf ion pair in Mg(OTf)2-based electrolyte is more significant than that of
Mg(TFSI)2 electrolytes, especially in low dielectric media, such as glyme solvents.64
(3) The smaller size of the OTf
than that of TFSI
anion also allows it to coordinate
into the octahedral structure of the Mg complex. Herein, we propose the following
possible reaction schemes of Mg(OTf)2 and MgCl2 in DME solution (Equations 1, 2,
3, 4, and 5):

MgðOTfÞ2 + 2DME/MgðOTfÞ2 ðDMEÞ2 ; (Equation 1)

 2 +

MgðOTfÞ2 ðDMEÞ2 + DME% MgðDMEÞ3 + 2OTf ; (Equation 2)

MgCl2 + 2DME%MgCl2 ðDMEÞ2 ; (Equation 3)

 2 +  2 +
MgðDMEÞ3 + MgCl2 ðDMEÞ2 % Mg2 Cl2 ðDMEÞ4 + DME; (Equation 4)

 2 +  +
Mg2 Cl2 ðDMEÞ4 %2 MgClðDMEÞx + ð4
2xÞDME (Equation 5)

Equations 1 and 2 illustrate the interactions between Mg(OTf)2 and DME, forming
Mg(OTf)2(DME)2 and possibly [Mg(DME)3]2+ cation. In the electrolyte solution con-
taining only Mg(OTf)2, Equation 2 might be negligible due to the strong interaction
between Mg2+ and OTf
anion as explained above, corresponding to the low solu-
bility of Mg(OTf)2 alone in DME. The presence of MgCl2 leads to the occurrence of
Equations 3, 4, and 5, as suggested by Aurbach’s group.62,63 These reactions effec-
tively increase the dissociation of Mg(OTf)2 (Equation 2) into more soluble species,
such as [Mg(DME)3]2+ and [MgCl(DME)x]+. The formation of complexes such as
[MgCl(OTf)(DME)2] also cannot be ruled out. The formation of these species might
well explain the enhanced solubility of Mg(OTf)2 with the addition of MgCl2. Howev-
er, because the [Mg2+]:[Cl
] ratio is above 1 in all the electrolyte solutions, a large
portion of Mg2+ ions exist as Mg(OTf)2(DME)2 and possibly [Mg(DME)3]2+. There-
fore, the crystallization of the electrolyte solution yielded only the formation of
Mg(OTf)2(DME)2, probably due to its charge-neutral nature, which made it less
soluble than other complexes. Further structural analysis of Mg complexes in this
electrolyte system will be reported in future work.

In conclusion, we report that the simple combination of Mg(OTf)2 and MgCl2 in DME
with a [Mg2+]:[Cl
] ratio larger than 1 exhibits highly reversible and conditioning-free
Mg dissolution/deposition (99.4% CE over 1,000 cycles) at room temperature. In
addition, the electrolyte is promising for practical applications at high areal capacity
(5 mAh cm
2) and high current density (3 mA cm
2) while maintaining high CE at
99.4% and 98.4%, respectively. Investigation on the surface chemistry of Mg anode
reveals the formation of a surface film, which consists of organic and adsorbed
chloride-containing species on the top surface and mostly Mg-based inorganic com-
ponents in the interior. This largely inorganic and robust SEI structure enables
reversible Mg deposition and dissolution at high CE. This work provides insights

Cell Reports Physical Science 1, 100265, December 23, 2020 17

 ll
OPEN ACCESS
Article

on the rational design of promising electrolyte systems for rechargeable magnesium

batteries, which can be extended to other energy-storage systems beyond Li-ion
batteries in the future.

EXPERIMENTAL PROCEDURES
Resource Availability
Lead Contact
Further information and requests for resources and reagents should be directed to
and will be fulfilled by the lead contact.

Materials Availability
Materials synthesized in this manuscript can be obtained by request to the lead
contact.

Data and Code Availability

All data are available from the lead contact upon reasonable request. The accession
number for the crystallographic information file (CIF) of Mg(DME)2(OTf)2 single crys-
tal reported in this paper is Cambridge Crystallographic Data Centre (CCDC):
2041129.

Electrolyte Preparation and Handling

MgCl2 (99.99%), molecular sieve (3Å beads; 4–8 mesh), DME (99.5%), N-methyl-2-
pyrrolidone (NMP) (99.5%), and PVDF (Mw 534,000) were purchased from Sigma-
Aldrich. Mg(OTf)2 (99.5%) and Mg(TFSI)2 (99.5%) were purchased from Solvionic.
Carbon black (99%) and platinum foil (99.99%) were purchased from Alfa Aesar.
Mg foil (>99%; 0.1 mm thick) was purchased from MTI corporation.

Molecular sieves were dried at 200 C for 24 h in the oven followed by vacuum
drying overnight and finally stored in an Ar-filled glovebox. DME was dried with
molecular sieves for at least 24 h prior to electrolyte preparation. A predetermined
amount of Mg(OTf)2 and MgCl2 powders were weighted directly into a glass vial, fol-
lowed by addition of DME solvent and stir bar. The mixture was stirred for 24 h at
60 C. The solution was rested overnight, and a trace amount of white solid powder
at the bottom of glass vial was removed. Finally, molecular sieves were added into
the electrolyte solution to further remove moisture from the electrolyte solution.
The moisture content in all electrolyte solutions is determined to be less than
20 ppm using Karl Fischer Coulometer C30X (Mettler Toledo). The ionic conductivity
of the electrolyte solution was measured using Fisherbrand Accumet AB200 at room
temperature.

DFT Simulation
Geometry optimizations are performed using the unrestricted B3LYP hybrid density
functional65,66 implemented in Gaussian 09 suite of programs.67 No symmetry
constraint is applied. The 6
311+G (d,p) basis sets are adopted for all atoms.68,69
‘‘Tight’’ optimization and ‘‘ultrafine’’ integration grid are specified for the DFT calcu-
lations. Monoglyme is used as the solvent within the polarization continuum model
(PCM).70 Four parameters are set for this solvent: the static dielectric constant (ε) =
7.069; the molecular radius of the solvent = 2.7904 Å; the density of the solvent =
0.005753 particles Å
1; and the molar volume of the solvent = 104.671 cm
3. The
rest of the solvent parameters are based on the program’s standard input. The ioni-
zation potential (IP) of the anions is calculated within the adiabatic approximation
and is set relative to the Mg2+/Mg redox pair.

18 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

Cell Configurations and Electrochemical Characterization

Cyclic voltammetry (CV) was performed using a three-electrode cell (EL-CELL) with
1 cm2 platinum (Pt) disk as the working electrode, 1.26 cm2 polished Mg disk as
the counter electrode (Figure S15), Mg wire as the reference electrode, 2 layers of
Celgard separators, and 50 mL of electrolyte solution. The CV measurement was con-
ducted at a scan rate of 5 mV s
1 between
1.5 and 2.2 V versus Mg/Mg2+. Linear
sweep voltammetry (LSV) was performed using 3 electrode flooded glass cells filled
with 1 mL electrolyte solution. Platinum, SS, Al, Ni, and molybdenum (Mo) were used
as the working electrode. Mg ribbons were used as the counter and reference elec-
trodes. These metal electrodes were polished with sandpaper prior to cell assem-
bling. The LSV measurements were conducted at a scan rate of 5 mV s
1 between
OCV and 6.0 V versus Mg/Mg2+. All cell assembly and voltammetry measurements
were conducted in an Ar-filled glove box (H2O < 0.1 ppm and O2 < 1 ppm).

Galvanostatic plating/stripping was performed using a 2032-type coin cell. In the asym-
metric cell configuration, the coin cell consists of a carbon-coated Al disk (Figure S16;
1 cm2) as the working electrode, polished Mg disk (1.26 cm2) as the counter electrode,
and 2 layers of Celgard separators. The cell was filled with 25 mL of the electrolyte solu-
tion. It should be noted here that, in our experiments with various working electrodes,
including carbon-coated Al, copper (Cu), SS, Ni, and titanium, the highest plating/strip-
ping CE was achieved with carbon-coated Al foil. In the symmetric cell configuration, the
Al disk was replaced with polished Mg disks (1.26 cm2). All coin cell components were
washed with isopropanol and acetone and dried overnight in 60 C oven. Celgard sep-
arators (16 mm in diameter) were punched out and dried in vacuum at 60 C for 3 days
and stored in an Ar-filled glovebox prior to cell assembly. Galvanostatic plating of Mg
was carried out at the desired current density and areal capacity, followed by stripping
of Mg by charging to a cutoff voltage of 1.2 V versus Mg/Mg2+, all performed at room
temperature. The CE for each cycle was calculated as the ratio of the capacity of Mg
stripped to that of Mg deposited. Electrochemical impedance spectroscopy was carried
out on symmetric Mg//Mg cell before cycling, after the first plating/stripping cycle, after
20 and 50 cycles using Gamry potentiostat between 1 MHz to 10 mHz. The full cell Mg//
Mo6S8 was cycled between 0.2 and 2.2 V using constant current-constant voltage mode
(CC-CV mode) at a current density of 0.1 C after a formation process, which includes one
discharge-charge cycle at 0.05 C followed by discharge to 0.2 V at 0.1 C. The C rate was
calculated based on the theoretical capacity of Mo6S8 at 129 mAh g
1 (1 C = 129 mA
g
1). The CE of Mg//MgMo6S8 cell for each cycle was calculated based on the ratio
of discharge capacity to charge capacity.

Single-Crystal XRD Analysis and Raman Spectroscopy

Single crystals were obtained by layering hexane on the top of 1.5 M Mg(OTf)2 + 1 M
MgCl2 in DME solution (0.5 mL) in 1-mL glass vial. The sample then was placed in an
Ar-filled glovebox for 7 days. Intensity data were collected at 100 K using Mo-Ka
radiation (l = 0.71073 Å) on a Bruker D8 Quest diffractometer with an Incoatec
microfocus X-ray source. The collected frames were processed with the software
SAINT.71 The data were corrected for absorption by using the SADABS program.72
The structure was solved by the direct methods and refined by full-matrix least-
squares analyses on F2 (SHELXL-2018/3).73 All non-hydrogen atoms were refined
anisotropically. The positions of all hydrogen atoms were generated geometrically
and allowed to ride on their respective parent carbon atoms before the final cycle
of least-squares refinement. A Renishaw inVia confocal microscope was employed
for Raman spectroscopy, using a 532-nm excitation laser in backscattering geome-
try. Samples (1 mL of electrolyte solution or powder) were sealed in a glass tube filled
with Ar.

Cell Reports Physical Science 1, 100265, December 23, 2020 19

 ll
OPEN ACCESS
Article

NMR Analysis
35
Cl NMR spectra of electrolyte solutions were collected using 500-MHz NMR (JEOL
JNMECA500II), operated without deuterated solvent lock centered at 450 ppm
offset and 1,200 ppm sweep (90 pulse angle; 0.1 s relaxation delay; 0.16965 s repe-
tition time). Each spectrum is an average of 800,000 consecutive scans. Background
treatment and data processing were done using MestReNova (v12.0.2-20910, Mes-
trelab S.L.). NaCl reference is set at 0 ppm to represent free Cl
. To prepare samples
for NMR analysis, 1 mL electrolyte solution was filled into an NMR tube together with
a sealed micro-capillary containing 0.1 M NaCl internal standard in D2O. The NMR
spectra of electrolyte solutions were collected with and without the internal standard
for comparison.

Post Mortem Characterizations

After cycling, coin cells were disassembled in Ar-filled glovebox. The electrodes
were washed with dried DME to remove residual electrolyte for 30 s and dried over-
night at room temperature. The Mg deposition morphology was studied using SEM
(JEOL 7600F) at an accelerating voltage of 5 kV. The electrode was mounted on the
SEM sample holder in an Ar-filled glovebox and transferred to the analysis chamber
using an airtight carrier. The crystal structure information of Mg deposits was
revealed by D8 Advance, Bruker, XRD, and the materials were scanned between
10 and 70 (2q) using Cu Ka radiation. The Mg deposited electrode was mounted
in an airtight specimen holder (Bruker).

The chemical composition of the electrode surface was revealed by XPS, Thermo
Scientific Theta Probe Angle-Resolved Spectrometer. The Mg electrode was washed
with dried DME for 30 s and dried in an Ag-filled glovebox prior to XPS measure-
ment. The sample was mounted on a specially designed air-tight XPS holder to avoid
exposing the sample to air during transfer from the Ar-filled glovebox to the XPS
analysis chamber. The sample was vacuumed overnight to remove gases and
absorbed solvent. XPS measurement was conducted within 18 h after removal
from the coin cell to minimize the decomposition of surface species. Due to the
weak signal of C 1 s spectra, herein, all spectra were calibrated using Mg 2p peak
of Mg0 at 49.8 eV. The chemical composition of the interphase as a function of its
thickness was characterized using XPS depth profiling, conducted at 3 kV (ion-
beam current of 2 mA) for 30 s, 1 min, 2 min, 5 min, 8 min, and 15 min.

Synthesis of Mo6S8 and Cathode Preparation

Chevrel phase Cu2Mo6S8 was first synthesized as described in the literature.56 The
Cu2Mo6S8 powder then underwent a chemical leaching process to yield the final
product of Mo6S8 as follows: 600 mg of Cu2Mo6S8 was added into 40 mL 8 M HCl
solution. The mixture was then stirred for 8 h at room temperature with dried air
bubbling. The crystal structure of Cu2Mo6S8 and Mo6S8 was confirmed by XRD.
Mo6S8 cathode slurry was prepared by mixing active material, carbon black, and
PVDF in the weight ratio of 80:10:10 in NMP solvent using Thinky Mixer (ARE-250)
for 20 min. The slurry was then coated on Ni foil and dried at 60 C for 12 h and
120 C for 6 h. The electrodes were then punched into 12-mm-diameter disks and
dried at 120 C in a vacuum oven prior to use. The total electrode loading mass is
8.8 mg cm2 (7.0 mg cm
2 for Mo6S8).

SUPPLEMENTAL INFORMATION
Supplemental Information can be found online at https://doi.org/10.1016/j.xcrp.
2020.100265.

20 Cell Reports Physical Science 1, 100265, December 23, 2020

 ll
OPEN ACCESS
Article

ACKNOWLEDGMENTS
This work was supported by the Singapore National Research Foundation (NRF-
NRFF2017-04). M.-F.N. acknowledges the National Supercomputing Centre
(NSCC) Singapore and A*STAR Computational Resource Centre (A*CRC) of
Singapore for the use of its high-performance computing facilities. Z.S. was sup-
ported by Czech Science Foundation (GACR no. 20-16124J).

AUTHOR CONTRIBUTIONS
Conceptualization, D.-T.N. and Z.W.S.; Data Curation, D.-T.N., A.Y.S.E., V.K., Z.S.,
and A.D.H.; Formal Analysis, D.-T.N., A.Y.S.E., V.K., Z.S., A.D.H., and G.S.S.; Inves-
tigation, D.-T.N., A.Y.S.E., M.-F.N., V.K., Z.S., A.D.H., and G.S.S.; Writing—Original
Draft, D.-T.N. and Z.W.S.; Software, M.-F.N.; Supervision, Z.W.S.

DECLARATION OF INTERESTS
The authors declare no competing interests.

Received: August 17, 2020

Revised: October 6, 2020
Accepted: November 2, 2020
Published: December 2, 2020

REFERENCES
1. Muldoon, J., Bucur, C.B., and Gregory, T.
(2014). Quest for nonaqueous multivalent
secondary batteries: magnesium and beyond.
Chem. Rev. 114, 11683–11720.
2. Canepa, P., Sai Gautam, G., Hannah, D.C.,
Malik, R., Liu, M., Gallagher, K.G., Persson,
K.A., and Ceder, G. (2017). Odyssey of
multivalent cathode materials: open questions
and future challenges. Chem. Rev. 117, 4287–
4341.
3. Choi, J.W., and Aurbach, D. (2016). Promise
and reality of post-lithium-ion batteries with
high energy densities. Nat. Rev. Mater. 1,
16013.
4. Aurbach, D., Cohen, Y., and Moshkovich, M.
(2001). The study of reversible magnesium
deposition by in situ scanning tunneling
microscopy. Electrochem. Solid-State Lett. 4,
A113–A116.
5. Kumar, V., Wang, Y., Eng, A.Y.S., Ng, M.-F., and
Seh, Z.W. (2020). A biphasic interphase design
enabling high performance in room
temperature sodium-sulfur batteries. Cell Rep.
Phys. Sci. 1, 100044.
6. Jäckle, M., and Groß, A. (2014). Microscopic
properties of lithium, sodium, and magnesium
battery anode materials related to possible
dendrite growth. J. Chem. Phys. 141, 174710.
7. Aurbach, D., Lu, Z., Schechter, A., Gofer, Y.,
Gizbar, H., Turgeman, R., Cohen, Y.,
Moshkovich, M., and Levi, E. (2000). Prototype
systems for rechargeable magnesium
batteries. Nature 407, 724–727.
8. Doe, R.E., Han, R., Hwang, J., Gmitter, A.J.,
Shterenberg, I., Yoo, H.D., Pour, N., and
Aurbach, D. (2014). Novel, electrolyte solutions
comprising fully inorganic salts with high
anodic stability for rechargeable magnesium
batteries. Chem. Commun. (Camb.) 50,
243–245.
9. Tutusaus, O., Mohtadi, R., Arthur, T.S., Mizuno,
F., Nelson, E.G., and Sevryugina, Y.V. (2015). An
efficient halogen-free electrolyte for use in
rechargeable magnesium batteries. Angew.
Chem. Int. Ed. Engl. 54, 7900–7904.
10. Liebenow, C., Yang, Z., and Lobitz, P. (2000).
The electrodeposition of magnesium using
solutions of organomagnesium halides,
amidomagnesium halides and magnesium
organoborates. Electrochem. Commun. 2,
641–645.
11. Kim, H.S., Arthur, T.S., Allred, G.D., Zajicek, J.,
Newman, J.G., Rodnyansky, A.E., Oliver, A.G.,
Boggess, W.C., and Muldoon, J. (2011).
Structure and compatibility of a magnesium
electrolyte with a sulphur cathode. Nat.
Commun. 2, 427.
12. Zhang, Z., Cui, Z., Qiao, L., Guan, J., Xu, H.,
Wang, X., Hu, P., Du, H., Li, S., Zhou, X., et al.
(2017). Novel design concepts of efficient Mg-
ion electrolytes toward high-performance
magnesium–selenium and magnesium–sulfur
batteries. Adv. Energy Mater. 7, 1602055.
13. Zhao-Karger, Z., Gil Bardaji, M.E., Fuhr, O., and
Fichtner, M. (2017). A new class of non-
corrosive, highly efficient electrolytes for
rechargeable magnesium batteries. J. Mater.
Chem. A 5, 10815–10820.
14. Du, A., Zhang, Z., Qu, H., Cui, Z., Qiao, L.,
Wang, L., Chai, J., Lu, T., Dong, S., Dong, T.,
et al. (2017). An efficient organic magnesium
borate-based electrolyte with non-nucleophilic
characteristics for magnesium–sulfur battery.
Energy Environ. Sci. 10, 2616–2625.
15. Zhao-Karger, Z., Liu, R., Dai, W., Li, Z., Diemant,
T., Vinayan, B.P., Bonatto Minella, C., Yu, X.,
Manthiram, A., Behm, R.J., et al. (2018). Toward
highly reversible magnesium–sulfur batteries
with efficient and practical Mg[B(hfip)4]2
electrolyte. ACS Energy Lett. 3, 2005–2013.
16. Tang, K., Du, A., Dong, S., Cui, Z., Liu, X., Lu, C.,
Zhao, J., Zhou, X., and Cui, G. (2020). A stable
solid electrolyte interphase for magnesium
metal anode evolved from a bulky anion lithium
salt. Adv. Mater. 32, 1904987.
17. Lossius, L.P., and Emmenegger, F. (1996).
Plating of magnesium from organic solvents.
Electrochim. Acta 41, 445–447.
18. NuLi, Y., Yang, J., and Wu, R. (2005). Reversible
deposition and dissolution of magnesium from
BMIMBF4 ionic liquid. Electrochem. Commun.
7, 1105–1110.
19. Amir, N., Vestfrid, Y., Chusid, O., Gofer, Y., and
Aurbach, D. (2007). Progress in nonaqueous
magnesium electrochemistry. J. Power Sources
174, 1234–1240.
20. Cheek, G.T., O’Grady, W.E., El Abedin, S.Z.,
Moustafa, E.M., and Endres, F. (2008). Studies
on the electrodeposition of magnesium in ionic
liquids. J. Electrochem. Soc. 155, D91–D95.
21. Yang, Y., Wang, W., Nuli, Y., Yang, J., and
Wang, J. (2019). High active magnesium
trifluoromethanesulfonate-based electrolytes
for magnesium-sulfur batteries. ACS Appl.
Mater. Interfaces 11, 9062–9072.
22. Huang, D., Tan, S., Li, M., Wang, D., Han, C.,
An, Q., and Mai, L. (2020). Highly efficient non-
nucleophilic Mg(CF3SO3)2-based electrolyte
for high-power Mg/S battery. ACS Appl. Mater.
Interfaces 12, 17474–17480.
23. Bieker, G., Salama, M., Kolek, M., Gofer, Y.,
Bieker, P., Aurbach, D., and Winter, M. (2019).
The power of stoichiometry: conditioning and
speciation of MgCl2/AlCl3 in tetraethylene

Cell Reports Physical Science 1, 100265, December 23, 2020 21

 ll
OPEN ACCESS
glycol dimethyl ether-based electrolytes. ACS
Appl. Mater. Interfaces 11, 24057–24066.
24. Canepa, P., Jayaraman, S., Cheng, L., Rajput,
N.N., Richards, W.D., Gautam, G.S., Curtiss,
L.A., Persson, K.A., and Ceder, G. (2015).
Elucidating the structure of the magnesium
aluminum chloride complex electrolyte for
magnesium-ion batteries. Energy Environ. Sci.
8, 3718–3730.
25. Liu, T., Shao, Y., Li, G., Gu, M., Hu, J., Xu, S.,
Nie, Z., Chen, X., Wang, C., and Liu, J. (2014). A
facile approach using MgCl2 to formulate high
performance Mg2+ electrolytes for
rechargeable Mg batteries. J. Mater. Chem. A
2, 3430–3438.
26. Liao, C., Sa, N., Key, B., Burrell, A.K., Cheng, L.,
Curtiss, L.A., Vaughey, J.T., Woo, J.-J., Hu, L.,
Pan, B., et al. (2015). The unexpected discovery
of the Mg(HMDS)2/MgCl2 complex as a
magnesium electrolyte for rechargeable
magnesium batteries. J. Mater. Chem. A 3,
6082–6087.
27. Shterenberg, I., Salama, M., Yoo, H.D., Gofer,
Y., Park, J.-B., Sun, Y.-K., and Aurbach, D.
(2015). Evaluation of (CF3SO2)2N
(TFSI) based
electrolyte solutions for Mg batteries.
J. Electrochem. Soc. 162, A7118–A7128.
28. Watahiki, T., Kiyosu, T., and Okamoto, K.
(2016). Electrolyte solution for electrochemical
devices. US patent US9263767B2, filed July 28,
2011, and granted February 16, 2016.
29. Webber, A. (1991). Conductivity and viscosity
of solutions of LiCF3SO3, Li(CF3SO2)2N, and
their mixtures. J. Electrochem. Soc. 138, 2586–
2590.
30. Connell, J.G., Genorio, B., Lopes, P.P.,
Strmcnik, D., Stamenkovic, V.R., and Markovic,
N.M. (2016). Tuning the reversibility of Mg
anodes via controlled surface passivation by
H2O/Cl
in organic electrolytes. Chem. Mater.
28, 8268–8277.
31. Lapidus, S.H., Rajput, N.N., Qu, X., Chapman,
K.W., Persson, K.A., and Chupas, P.J. (2014).
Solvation structure and energetics of
electrolytes for multivalent energy storage.
Phys. Chem. Chem. Phys. 16, 21941–21945.
32. Kang, S.-J., Lim, S.-C., Kim, H., Heo, J.W.,
Hwang, S., Jang, M., Yang, D., Hong, S.-T., and
Lee, H. (2017). Non-Grignard and Lewis acid-
free sulfone electrolytes for rechargeable
magnesium batteries. Chem. Mater. 29, 3174–
3180.
33. Yoo, H.D., Liang, Y., Li, Y., and Yao, Y. (2015).
High areal capacity hybrid magnesium-lithium-
ion battery with 99.9% Coulombic efficiency for
large-scale energy storage. ACS Appl. Mater.
Interfaces 7, 7001–7007.
34. He, Y., Li, Q., Yang, L., Yang, C., and Xu, D.
(2019). Electrochemical-conditioning-free and
water-resistant hybrid AlCl3/MgCl2/Mg(TFSI)2
electrolytes for rechargeable magnesium
batteries. Angew. Chem. Int. Ed. Engl. 58,
7615–7619.
35. Bonnick, P., and Muldoon, J. (2020). A trip to Oz
and a peak behind the curtain of magnesium
batteries. Adv. Funct. Mater. 30, 1910510.
36. Wanger, C., Riggs, W., Davis, L., Moulder, J.,
and Muilenberg, G. (1979). Handbook of X-Ray
Photoelectron Spectroscopy: A Reference
22 Cell Reports Physical Science 1, 100265, December 23, 2020
Book of Standard Data for Use in X-Ray
Photoelectron Spectroscopy (Physical
Electronics Division, Perkin-Elmer).
37. Nguyen, D.-T., and Song, S.-W. (2017).
Magnesium stannide as a high-capacity anode
for magnesium-ion batteries. J. Power Sources
368, 11–17.
38. Yu, Y., Baskin, A., Valero-Vidal, C., Hahn, N.T.,
Liu, Q., Zavadil, K.R., Eichhorn, B.W.,
Prendergast, D., and Crumlin, E.J. (2017).
Instability at the electrode/electrolyte interface
induced by hard cation chelation and
nucleophilic attack. Chem. Mater. 29, 8504–
8512.
39. Nguyen, C.C., and Lucht, B.L. (2014).
Comparative study of fluoroethylene
carbonate and vinylene carbonate for silicon
anodes in lithium ion batteries. J. Electrochem.
Soc. 161, A1933–A1938.
40. Wong, P.C., Li, Y.S., and Mitchell, K.A.R. (1995).
Investigations of the interface between
magnesium and polyethyleneterephthalate by
X-ray photoelectron spectroscopy. Appl. Surf.
Sci. 84, 245–252.
41. Kuwata, H., Matsui, M., and Imanishi, N. (2017).
Passivation layer formation of magnesium
metal negative electrodes for rechargeable
magnesium batteries. J. Electrochem. Soc. 164,
A3229–A3236.
42. Tie, D., Feyerabend, F., Hort, N., Willumeit, R.,
and Hoeche, D. (2010). XPS studies of
magnesium surfaces after exposure to
Dulbecco’s modified eagle medium, Hank’s
buffered salt solution, and simulated body
fluid. Adv. Eng. Mater. 12, B699–B704.
43. Kamarulzaman, N., Chayed, N.F., Badar, N.,
Kasim, M.F., Mustaffa, D.T., Elong, K., Rusdi, R.,
Oikawa, T., and Furukawa, H. (2016). Band gap
narrowing of 2-D ultra-thin MgO graphene-like
sheets. ECS J. Solid State Sci. Technol. 5,
Q3038–Q3045.
44. Zhao, X., Li, Q., Zhao-Karger, Z., Gao, P., Fink,
K., Shen, X., and Fichtner, M. (2014).
Magnesium anode for chloride ion batteries.
ACS Appl. Mater. Interfaces 6, 10997–11000.
45. Gofer, Y., Turgeman, R., Cohen, H., and
Aurbach, D. (2003). XPS investigation of surface
chemistry of magnesium electrodes in contact
with organic solutions of
organochloroaluminate complex salts.
Langmuir 19, 2344–2348.
46. Seh, Z.W., Sun, J., Sun, Y., and Cui, Y. (2015). A
highly reversible room-temperature sodium
metal anode. ACS Cent. Sci. 1, 449–455.
47. Agostini, M., Xiong, S., Matic, A., and Hassoun,
J. (2015). Polysulfide-containing glyme-based
electrolytes for lithium sulfur battery. Chem.
Mater. 27, 4604–4611.
48. Briggs, D., and Beamson, G. (1993). XPS studies
of the oxygen 1s and 2s levels in a wide range of
functional polymers. Anal. Chem. 65, 1517–
1523.
49. López, G.P., Castner, D.G., and Ratner, B.D.
(1991). XPS O 1s binding energies for polymers
containing hydroxyl, ether, ketone and ester
groups. Surf. Interface Anal. 17, 267–272.
50. Merrill, L.C., and Schaefer, J.L. (2019). The
influence of interfacial chemistry on
Article
magnesium electrodeposition in non-
nucleophilic electrolytes using sulfone-ether
mixtures. Front Chem. 7, 194.
51. Magni, E., and Somorjai, G.A. (1996).
Preparation of a model Ziegler-Natta catalyst:
electron irradiation induced titanium chloride
deposition on magnesium chloride thin films
grown on gold. Surf. Sci. 345, 1–16.
52. Ding, M.S., Diemant, T., Behm, R.J., Passerini,
S., and Giffin, G.A. (2018). Dendrite growth in
Mg metal cells containing Mg(TFSI)2/glyme
electrolytes. J. Electrochem. Soc. 165, A1983–
A1990.
53. Kumar, V., Eng, A.Y.S., Wang, Y., Nguyen,
D.-T., Ng, M.-F., and Seh, Z.W. (2020). An
artificial metal-alloy interphase for high-rate
and long-life sodium–sulfur batteries. Energy
Storage Mater. 29, 1–8.
54. Jie, Y., Liu, X., Lei, Z., Wang, S., Chen, Y.,
Huang, F., Cao, R., Zhang, G., and Jiao, S.
(2020). Enabling high-voltage lithium metal
batteries by manipulating solvation structure in
ester electrolyte. Angew. Chem. Int. Ed. Engl.
59, 3505–3510.
55. Shterenberg, I., Salama, M., Gofer, Y., and
Aurbach, D. (2017). Hexafluorophosphate-
based solutions for Mg batteries and the
importance of chlorides. Langmuir 33, 9472–
9478.
56. Lancry, E., Levi, E., Gofer, Y., Levi, M., Salitra,
G., and Aurbach, D. (2004). Leaching chemistry
and the performance of the Mo6S8 cathodes in
rechargeable Mg batteries. Chem. Mater. 16,
2832–2838.
57. Cheng, Y., Stolley, R.M., Han, K.S., Shao, Y.,
Arey, B.W., Washton, N.M., Mueller, K.T.,
Helm, M.L., Sprenkle, V.L., Liu, J., and Li, G.
(2015). Highly active electrolytes for
rechargeable Mg batteries based on a [Mg2(m-
Cl)2]2+ cation complex in dimethoxyethane.
Phys. Chem. Chem. Phys. 17, 13307–13314.
58. Muthuraj, D., and Mitra, S. (2018). Reversible
Mg insertion into chevrel phase Mo6S8
cathode: Preparation, electrochemistry and
X-ray photoelectron spectroscopy study.
Mater. Res. Bull. 101, 167–174.
59. Saha, P., Jampani, P.H., Datta, M.K., Okoli,
C.U., Manivannan, A., and Kumta, P.N. (2014). A
convenient approach to Mo6S8 chevrel phase
cathode for rechargeable magnesium battery.
J. Electrochem. Soc. 161, A593–A598.
60. Xue, X., Chen, R., Yan, C., Zhao, P., Hu, Y.,
Kong, W., Lin, H., Wang, L., and Jin, Z. (2019).
One-step synthesis of 2-ethylhexylamine
pillared vanadium disulfide nanoflowers with
ultralarge interlayer spacing for high-
performance magnesium storage. Adv. Energy
Mater. 9, 1900145.
61. Pei, C., Yin, Y., Sun, R., Xiong, F., Liao, X., Tang,
H., Tan, S., Zhao, Y., An, Q., and Mai, L. (2019).
Interchain-expanded vanadium tetrasulfide
with fast kinetics for rechargeable magnesium
batteries. ACS Appl. Mater. Interfaces 11,
31954–31961.
62. Neumüller, B., Stieglitz, G., and Dehnicke, K.
(1993). Crystal structure of MgCl2(1, 2-
dimethoxyethane)2. Z. Naturforsch. 48, 1151–
1153.

Article
63. Salama, M., Shterenberg, I., Shimon, L.J.W.,
Keinan-Adamsky, K., Afri, M., Gofer, Y., and
Aurbach, D. (2017). Structural analysis of
magnesium chloride complexes in
dimethoxyethane solutions in the context of
Mg batteries research. J. Phys. Chem. C 121,
24909–24918.
64. Saito, M., Yamada, S., Ishikawa, T., Otsuka, H.,
Ito, K., and Kubo, Y. (2017). Factors influencing
fast ion transport in glyme-based electrolytes
for rechargeable lithium–air batteries. RSC
Adv. 7, 49031–49040.
65. Becke, A.D. (1993). Density-functional
thermochemistry. III. The role of
exact exchange. J. Chem. Phys. 98, 5648–
5652.
66. Stephens, P.J., Devlin, F.J., Chabalowski, C.F.,
and Frisch, M.J. (1994). Ab initio calculation of
vibrational absorption and circular dichroism
spectra using density functional force fields.
J. Phys. Chem. 98, 11623–11627.
67. Gaussian, 09, Revision, A.02, Frisch, M.J.,
Trucks, G.W., Schlegel, H.B., Scuseria, G.E.,
Robb, M.A., Cheeseman, J.R., Scalmani, G.,
Barone, V., Petersson, G.A., Nakatsuji, H., et al.
(2016). Gaussian, Inc (CT: Wallingford).
68. McLean, A.D., and Chandler, G.S. (1980).
Contracted Gaussian basis sets for molecular
calculations. I. Second row atoms, Z=11–18.
J. Chem. Phys. 72, 5639–5648.
69. Krishnan, R., Binkley, J.S., Seeger, R., and
Pople, J.A. (1980). Self-consistent molecular
Cell Reports Physical Science 1, 100265, December 23, 2020
ll
OPEN ACCESS
orbital methods. XX. A basis set for correlated
wave functions. J. Chem. Phys. 72, 650–654.
70. Miertus, S., Scrocco, E., and Tomasi, J. (1981).
Electrostatic interaction of a solute with a
continuum. A direct utilizaion of ab initio
molecular potentials for the prevision of
solvent effects. Chem. Phys. 55, 117–129.
71. Bruker AXS (1998). Smart, Saint+ for Windows
NT, Version 6.02a (Bruker AXS).
72. Sheldrick, G.M. (1996). SADABS (University of
Göttingen).
73. Sheldrick, G.M. (2015). Crystal structure
refinement with SHELXL. Acta Crystallogr. C
Struct. Chem. 71, 3–8.
23