Inhibiting Polysulfide Shuttle in Lithium–Sulfur Batteries through Low-Ion-Pairing Salts and a Triflamide Solvent

Article abstract of DOI:10.1002/anie.201701026

The step-change in gravimetric energy density needed for electrochemical energy storage devices to power unmanned autonomous vehicles, electric vehicles, and enable low-cost clean grid storage is unlikely to be provided by conventional lithium ion batteries. Lithium–sulfur batteries comprising lightweight elements provide a promising alternative, but the associated polysulfide shuttle in typical ether-based electrolytes generates loss in capacity and low coulombic efficiency. The first new electrolyte based on a unique combination of a relatively hydrophobic sulfonamide solvent and a low ion-pairing salt, which inhibits the polysulfide shuttle, is presented. This system behaves as a sparingly solvating electrolyte at slightly elevated temperatures, where it sustains reversible capacities as high as 1200–1500 mAh g^?1 over a wide range of current density (2C–C/5, respectively) when paired with a lithium metal anode, with a coulombic efficiency of >99.7 % in the absence of LiNO₃ additive.

Full text of DOI:10.1002/anie.201701026

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201701026
German Edition: DOI: 10.1002/ange.201701026
Lithium–Sulfur Batteries
Inhibiting Polysulfide Shuttle in Lithium–Sulfur Batteries through
Low-Ion-Pairing Salts and a Triflamide Solvent
Abhinandan Shyamsunder, Witali Beichel, Petra Klose, Quan Pang, Harald Scherer,
Anke Hoffmann, Graham K. Murphy, Ingo Krossing,* and Linda F. Nazar*
Abstract: The step-change in gravimetric energy density
needed for electrochemical energy storage devices to power
unmanned autonomous vehicles, electric vehicles, and enable
low-cost clean grid storage is unlikely to be provided by
conventional lithium ion batteries. Lithium–sulfur batteries
comprising lightweight elements provide a promising alterna-
tive, but the associated polysulfide shuttle in typical ether-based
electrolytes generates loss in capacity and low coulombic
efficiency. The first new electrolyte based on a unique combi-
nation of a relatively hydrophobic sulfonamide solvent and
a low ion-pairing salt, which inhibits the polysulfide shuttle, is
presented. This system behaves as a sparingly solvating
electrolyte at slightly elevated temperatures, where it sustains
reversible capacities as high as 1200–1500 mAhgÀ1 over a wide
range of current density (2C–C/5, respectively) when paired
with a lithium metal anode, with a coulombic efficiency of
> 99.7% in the absence of LiNO3 additive.
2600 WhkgÀ1 based on the reaction S8 + 16LiQ8Li2S.[1,2]
The volumetric energy density (based on an excess of lithium
and a low electrolyte volume cell) provides a theoretical value
of 2800 WhLÀ1, which is almost five times that of commer-
cially available Li-ion batteries, and practical values that
could approach 650 WhLÀ1.[3]
These promising metrics have not yet been achieved in
practice, however, owing to numerous challenges. One is the
formation of highly soluble intermediate polysulfides during
the redox cycle, which migrate to and from the negative
electrode. This creates the well-known shuttle phenom-
enon,[4,5] which leads to poor coulombic efficiency and loss
of active material from the positive electrode on cycling.[6]
Lithium sulfide (Li2S) and S8, the ultimate discharge and
charge products of the cell, are also electronic insulators that
require the addition of electronically conductive agents.
Usually these are low density carbons that must be suitably
wetted, whereas modeling studies suggest a low target ratio of
À1
O
wing to the concern of CO2 emissions and their contri-
electrolyte:sulfur of about 1.5 mLgS is necessary for Li-S
bution to climate change, there has been increasing focus over
the last decade on developing alternative sources of energy to
conventional petroleum feedstocks. While solar and wind
power derive from renewable resources, harnessing their real
potential lies in the ability to store their intermittent energy,
and to render it dispatchable. Electrochemical energy storage
systems have emerged as one of the promising candidates to
achieve this aim, both to satisfy the needs of large and
intermediate scale storage, and also to power vehicular
transport. Of these, the Li-S battery is a particularly appeal-
ing, environmentally friendly option. Sulfur is widely abun-
dant, very low cost, and provides a high theoretical specific
capacity and energy density of 1675 mAhgÀ1 and
cells to be competitive with current Li-ion batteries.[7] To
achieve this goal, cells must be able to operate in an
electrolyte-starved mode. Nonetheless, the majority of
todayꢀs Li-S cells function with high electrolyte volume
using 1,3-dioxolane (DOL)/dimethoxyethane (DME) as a sol-
vent, which fully solubilizes the intermediate polysulfide
species and enables redox to occur in solution. Only a few
reports have emerged that deviate from this approach. They
have centered on lowering the ability of the electrolyte to
dissolve polysulfides, utilizing either solvent-in-salt electro-
lytes, or chelate ionic liquid electrolytes that complex the
lithium salt (typically LiTFSI) to a strongly solvating solvent
such as DME or acetonitrile.[8–10] While effective, such
systems suffer from the need for very high (and costly) salt
fractions, which incur their own instability problems. Here, we
report the first electrolyte system based on an entirely
different principle. It combines a relatively non-polar trifla-
mide solvent and a fluorinated lithium aluminate salt
(Scheme 1). The combination behaves as a non-solvent for
polysulfides at room temperature and transforms into a spar-
ingly solvating electrolyte at slightly elevated temperatures.
Our concept was centered on multiple design consider-
ations involving both solvent and salt. It obviates the need for
clever additives such as LiNO3[11] and LiI,[12] which are used to
prevent reduction of polysulfides that cross over to the anode
(but can be consumed on cycling) and large quantities of
expensive Li+ salts. We note that an “ideal” solvent must be
capable of dissolving a Li+ salt as well as providing good
electrochemical properties such as a wide voltage window,
low viscosity, sufficient ionic conductivity to sustain good rate
[*] A. Shyamsunder, Q. Pang, Prof. G. K. Murphy, Prof. L. F. Nazar
Department of Chemistry and Waterloo Institute of Nanotechnology
University of Waterloo
Waterloo, Ontario N2L 3G1 (Canada)
E-mail: lfnazar@uwaterloo.ca
Dr. W. Beichel, P. Klose, Prof. I. Krossing
Freiburger Materialforschungszentrum (FMF)
Albert-Ludwigs-Universitꢀt Freiburg
Stefan-Meier-Strasse 21, 79104 Freiburg im Breisgau (Germany)
E-mail: krossing@uni-freiburg.de
Dr. H. Scherer, Dr. A. Hoffmann, Prof. I. Krossing
Institut fꢁr Anorganische und Analytische Chemie
Albert-Ludwigs-Universitꢀt Freiburg
Albertstrasse 21, 79104 Freiburg im Breisgau (Germany)
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.201701026.
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
(Ohfip)4]) used herein, can be produced in the laboratory on
a large scale and are comparable in cost to the commonly used
LiTFSI.
The electrochemical stability of the new electrolyte
system was evaluated through a cyclic voltammetric tech-
nique using a simple cell with a stainless steel working
electrode and a lithium foil counter electrode. The cyclic
voltammograms of the DMTand the DPT solvents containing
0.2m dissolved Li[Al(Ohfip)4] are given in the Supporting
Information, Figure S1. In both cases, lithium plates on the
stainless steel substrate at a potential less than 0 V with no
reduction of either salt or solvent. In the anodic scan, the
current begins to rise at 0 V, signaling the stripping of the
plated lithium from the stainless steel substrate. The Li[Al-
(Ohfip)4]-DMT electrolyte exhibited a stripping and plating
current much higher than Li[Al(Ohfip)4]-DPT. This is a con-
sequence of its higher ionic conductivity and lower viscosity.
We note that DMT is stable up to 5.6 V, corresponding to the
onset of electrolyte oxidation, while the DPT system shows an
onset of oxidation at around 5.4 V (Supporting Information,
Figure S1).
Scheme 1. Molecular structures of DMT, DPT, and [Al(Ohfip)4]À. Calcu-
lated electrostatic potentials of TFSIÀ and [Al(Ohfip)4]À plotted onto an
isodensity surface with like cut off values at a PBE0/def2-TZVPP level.
Red: accumulation of negative charge; green: neutral; blue: positively
charged surfaces.
capability, stability towards the polysulfides formed during
discharge, a low vapor pressure, and high thermal stability.[13]
The ability of a solvent to dissolve polysulfides or any salt
depends on the nature of the Lewis basic sites present in the
molecule. DOL and DME both have Lewis basic oxygen
atoms, and their basicity is bolstered by electron-donating
alkyl groups. To fabricate a solvent that has deactivated Lewis
basic sites, highly electron-withdrawing groups can be intro-
duced. Taking inspiration from the LiTFSI salt, which has an
electron-withdrawing triflyl group attached to the nitrogen
atom, we investigated two new battery solvents: N,N-
dimethyl triflamide (DMT) and N,N-dipropyl triflamide
(DPT), the structures of which are shown in Scheme 1. They
fall into the broad category of sulfonamides, which are known
for their high electrochemical and thermal stability. Trifla-
mides are commonly used in fuel cells, ionic liquids, and as
conducting additives,[14] but they have not been considered as
battery electrolytes owing to their relatively non-polar
nature.[15] We have capitalized on that characteristic to inhibit
the dissolution of polar polysulfides.
To verify the inhibition of the polysulfide shuttle con-
veyed by the new triflamide electrolytes, Li-S cells utilizing
0.2m solutions of Li[Al(Ohfip)4] in DMT, DPT, and DOL
were compared (Figure 1a). The latter exhibits the classic
endless charging response associated with a polysulfide
A Li+ salt that can dissolve in the non-polar solvent is also
required to provide an ion carrier. Conventional lithium salts
[16]
such as LiTFSI, LiOTf, or LiPF6
are all comprised of
weakly coordinating anions,[17–19] enabling them to dissolve in
classical non-aqueous organic solvents. For a salt to dissolve in
a solvent with deactivated Lewis basic sites, even weaker
coordinating anions are necessary.[17] Fluorinated alkoxy
aluminates of the form [Al(ORF)4]À easily satisfy this
condition. They are well-known weakly coordinating anions
that provide sufficient ionic conductivities in solution either
as Li salts[20,21] or as ionic liquids, the latter exhibiting high
dissociation degrees.[22–25] The strong electrostatic interactions
associated with commercially available salts are negligible in
such fluorinated alkoxy aluminates, owing to the large size of
the anion and the delocalization of the negative charge
throughout the anion (see electrostatic potential plot in
comparison to TFSI in Scheme 1). Their poorly polarizing
fluorine atoms minimize the efficiency of ion-pairing and
increase stability against oxidation. These salts, such as
lithium tetrakis(hexafluoroisopropoxide) aluminum (Li[Al-
Figure 1. a) Comparison of the galvanostatic profiles for Li-S cells run
at C/10 at 258C using 0.2m Li[Al(Ohfip)4] salt in different solvents:
DOL (blue curve), DMT (red curve), and DPT (black curve). b)–
d) Photographs of separators soaked in DOL, obtained from cells that
underwent one full cycle at 258C using an electrolyte comprised of
b) 0.2m Li[Al(Ohfip)4]-DOL, c) 0.2m Li[Al(Ohfip)4]-DMT, d) 0.2m Li[Al-
(Ohfip)4]-DPT, e),f) typical solutions obtained from extracting cathodes
(harvested from a room temperature cell in the charged state) with
DMSO. The lapis blue color characteristic of the trisulfur S3CÀ radical is
observed immediately upon immersion in DMSO (e), and transforms
within minutes to the olive–green solution characteristic of medium-
length polysulfides (f).
2
www.angewandte.org
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
These are not the final page numbers!

Angewandte
Communications
Chemie
shuttle, owing to the inability of this electrolyte to function
without a LiNO3 additive to protect the Li anode, whereas
DMT and DPT both exhibit profiles associated with solvents
that do not dissolve polysulfides.[26] When the separators were
soaked in 1,3-dioxolane, the yellow color of the solution
derived from the Li[Al(Ohfip)4]-DOL cell clearly indicates
the presence of residual polysulfides that become trapped in
the separator between the positive and negative electrodes,
whereas the two triflamide cells exhibit completely colorless
solutions, indicative of essentially no polysulfides, and hence
the lack of a shuttle (Figure 1b-d). Comparison of the
electrochemical performance of the triflamide electrolytes
shows that DMT exhibits superior cycling stability compared
to DPT, which shows a sharp decrease in capacity on cycling
(Supporting Information, Figure S2). The poor performance
of DPT could be due to its longer alkyl chains, which make it
more susceptible to a polysulfide initiated radical decom-
position process.[27] However, likely more important is that
the DPT electrolyte exhibits a lower ionic conductivity than
DMT (0.24 mScmÀ1 vs. 1.25 mScmÀ1 respectively; Supporting
Information, Table S1), which limits carrier transport. This is
further confirmed from electrochemical impedance spectros-
copy studies, where the Nyquist plots (Supporting Informa-
tion, Figure S3) indicate that the DPT system suffers from
a higher electrolyte/interphase resistance due to the poorer
wetting of the electrode by the electrolyte and lower ionic
conductivity. Therefore our studies focussed on the Li[Al-
(Ohfip)4]-DMT system.
Although the Li[Al(Ohfip)4]-DMT system exhibits good
cycling performance, the capacity of 380 mAhgÀ1 obtained
after 50 cycles (Supporting Information, Figure S2) is lower
than those obtainable with conventional high-electrolyte
volume DOL-DME systems using a LiNO3 additive. Inter-
estingly, polysulfides are clearly formed in the Li[Al(Ohfip)4]-
DMT cell, as shown by postmortem analysis of a cell which
was run for 30 cycles and disassembled at the end of charge.
Dipping those cathodes in DMSO resulted in an intense blue
ultramarine color associated with the S3CÀ radical, which
immediately transformed to deep green in a matter of seconds
(Figure 1e,f). This unequivocally shows that the associated
sulfur redox process involves the formation of lithium
polysulfides, even though the process that interconverts S8
and Li2S must occur in the solid state, not in solution. We base
this on the fact that no polysulfides were found in the
separator (Figure 1b–d). The kinetic limitations of this
reaction are exacerbated in the poorly ionically conductive
DMT electrolyte, and loss of electronic contact of the active
sulfide mass with the carbon host owing to volume expansion/
contraction.
Figure 2. Operando XRD of a Li-S cell at 258C with 0.2m Li[Al-
(Ohfip)4]-DMT.
sulfides by slightly increasing the operating temperature of
the cell to 508C. Along with providing a vehicle for solution-
based chemical transformations, this has the additional
benefit of lowering the viscosity and increasing the ionic
conductivity. The positive impact of increased temperature is
clearly evident in the first discharge–charge profile of the
Li[Al(Ohfip)4]-DMT system at 508C compared to that at
room temperature (Figure 3a). While elevated temperatures
have been used in cycling Li-S batteries in conventional
DOL-DME electrolytes,[29,30] the positive effect is not so
pronounced because the solubility of polysulfides is already
sufficient at 258C and a temperature increase does not show
marked improvement. In the system herein, however, a non-
solvent system is converted into a sparingly soluble system.
The improved kinetics result in more clearly defined plateaus
and a capacity (1530 mAhgÀ1) remarkably close to theoretical
(1675 mAhgÀ1) at a moderate current rate of C/5 (C rate =
1675 mAgÀ1). This represents better performance than most
DOL-DME systems, where typically capacities of 1100–
1200 mAhgÀ1 are reported at such rates. Even at a high
current rate of 2C, an initial capacity of 1500 mAhgÀ1 was
obtained, which settled to 1200 mAhgÀ1 on the second cycle.
Very good coulombic efficiencies approaching 99.7% (Fig-
ure 3b) were observed over this wide range of current density
in the absence of a LiNO3-induced passivation film on the
negative Li electrode. While this is due in part to the
improved electrolyte properties at elevated temperature, we
The capacity decay appears to be due to the formation of
trapped polysulfide/lithium sulfide, which are inaccessible to
further redox reactions. Indeed, while operando X-ray
diffraction (XRD) measurements (Figure 2) demonstrate
the formation of Li2S on discharge, and its consumption on
charge, its oxidation is not quite complete. As in DOL/DME
electrolytes, a-S8 converts into the b-S8 polymorph on
charge.[28]
To overcome the kinetic limitations of the quasi solid state
reaction, we induced partial (very low) solubility of poly-
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
The electrochemical behavior of 0.2m solutions of DMT-
Li[Al(Ohfip)4] and DMT-LiTFSI were compared to under-
stand the effect of the dissolved salt. At room temperature,
the LiTFSI electrolyte performs slightly better than Li[Al-
(Ohfip)4]. At 508C, however, DMT-Li[Al(Ohfip)4] exhibits
better capacity (Supporting Information, Figure S7). T-de-
pendent viscosities (h) and conductivities (s) of the 0.2m
solutions were recorded between 08C and 708C (Supporting
Information, Figure S8) to better understand this behavior.
Although the viscosities of the LiTFSI solutions are lower
than those of Li[Al(Ohfip)4] (Supporting Information,
Table S1), the conductivities of the latter are higher at all
measured temperatures, exceeding the former ones by
a factor of 4.5–4.6. This suggests that the Li[Al(Ohfip)4]
based electrolyte has a higher dissociation degree than the
LiTFSI, as confirmed by the corresponding Walden plot
(Figure 4). The DMT-Li[Al(Ohfip)4] system lies in the region
Figure 3. a) Comparison of the galvanostatic profiles for Li-S cells run
at C/5 using 0.2m Li[Al(Ohfip)4] DMT for cells cycled at 258C (black)
and 508C (red) and 2C at 508C (green); b) Discharge capacity (left
axis) and coulombic efficiency (right axis) as a function of cycle
number for Li-S cells examined at 258C (black) and 508C (red);
c) photographs of Li[Al(Ohfip)4]-DMT at 258C (left) and at 508C (right)
solutions with the yellow color indicating the presence of Li2S6;
d) typical emerald green solutions obtained from extracting cathodes
(harvested from a high-temperature cell in the charged state) with
DMSO overnight.
believe it primarily results from the very sparing solubility of
polysulfides in the electrolyte at 508C. The appearance of
a pale yellow color of a typical polysulfide (Li2S6) in DMT-
Li[Al(Ohfip)4] at 508C (vs. the colorless solution at 258C)
shows the electrolyte transformation from a non-solvent to
a sparingly soluble electrolyte (Figure 3c). Most importantly,
this slight increase in polysulfide solubility does not initiate
a significant polysulfide shuttle process, even at slow rates
such as C/20, since no overcharge is observed in the electro-
chemical profile (Supporting Information, Figure S4). Fur-
thermore, no color is observed in the separator on post-
mortem analysis (Supporting Information, Figure S5). Rather
the sparing solubility provides a kinetically less hindered
pathway for the redox processes.
To further demonstrate the improved kinetics associated
with increased temperature, galvanostatic intermittent titra-
tion technique (GITT) experiments were performed for the
electrolyte systems at 258C and 508C (Supporting Informa-
tion, Figure S6). At elevated temperatures, the quasi-equilib-
rium curve is similar to that exhibited in classic DOL-DME
systems, while at room temperature the polarization is very
large indeed. This underlines the importance of achieving
sparing solubility of the polysulfides. Although an increase in
temperature provided an increase in capacity and a decrease
in the capacity fade, the cathodes recovered from fully
charged cells and soaked in DMSO still show the presence of
residual unconverted polysulfides. This is again evident from
the coloration of the electrolyte shown in Figure 3d. We note
the difference in color relative to the 258C cell (emerald
green, rather than olive-green), suggesting that the precise
speciation of polysulfides in the cathode is not the same at the
two temperatures,[31,32] a point which will be the focus of
a future operando XRD/XAS study.
Figure 4. Walden plots confirming the higher dissociation degree of
the Li[Al(Ohfip)4] salt in DMT compared to the LiTFSI salt.
between an ideal electrolyte system and the “good IL”
barrier,[22,33] while that of the LiTFSI falls quite below the
latter. Furthermore, NMR-PGSTE measurements showed
almost the same diffusion constants for the cation (D+) and
the anion (D+) in both electrolyte solutions (Supporting
Information, Table S2) being around 0.26 ꢁ 10À10 m2 sÀ1 at
258C. At 508C D+/À increase to values of around 0.44 ꢁ
10À10 m2 sÀ1 (as expected). By calculating the ionicity (I),
which is a quantitative measure of the dissociation degree[23,34]
(I = 1 fully dissociated salt, I = 0 ion pairs) one arrives at
values of 0.11 (258C) and 0.10 (508C) for the LiTFSI
electrolyte and 0.48 (258C and 508C) for the Li[Al(Ohfip)4]
electrolyte. This is in agreement with the Walden plots and
our considerations of weakly coordinating anions as suitable
conducting salts.
A further benefit of the new proposed electrolyte
formulation is the low concentration of the Li[Al(Ohfip)4]
salt. Although its maximum concentration in DMT is 0.5m,
the optimum value to optimize capacity retention is about
0.2 M. These results clearly demonstrate the effectiveness of
low ion pairing salts even at low concentrations.
A typical loss of capacity occurs when cells are cycled at
very high rates, but the capacity recovers (albeit not
completely) when the rate is decreased (Supporting Informa-
tion, Figure S9). XPS analysis of the negative electrode from
4
www.angewandte.org
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
These are not the final page numbers!

Angewandte
Communications
Chemie
the 508C cell indicates that LiF (a common component
present in electrochemical systems when fluorinated electro-
lytes are used) is formed on the surface (either from residual
electrolyte or from its reaction at the elevated temperature),
along with other materials, which may facilitate the plating of
lithium ions during charge (Supporting Information, Fig-
ure S10). In particular, the S 2p XPS spectrum confirms the
absence of Li2S on the surface of the negative electrode,
suggesting that neither polysulfides nor sulfur cross-over are
present, where they would be reduced to lithium sulfide. The
Al 2p spectrum also shows the presence of an “alumina-like”
species, unique to our electrolyte system which may arise
from Li[Al(Ohfip)4]-contamination and/or reaction to form
an aluminum oxide like species. It has recently been shown
that “alumina” helps in forming a stable interface on the
lithium anode, as once lithiated, it provides a lithium-ion
conductive pathway.[35] However, the interfacial composition
is still not conclusive, and elucidation of species formed on the
lithium surface with this electrolyte is beyond the scope of the
current study.
In conclusion, a new, stable triflamide-based low concen-
tration electrolyte combined with a low-ion pairing Li+ salt
was developed for Li-S batteries for the first time. This
electrolyte transforms from a non-solvent at room temper-
ature to a sparingly solvating electrolyte at elevated temper-
ature and inhibits the polysulfide shuttle in the system at both
low, intermediate and fast rates, yielding a coulombic effi-
ciency of > 99.7% without the use of LiNO3. These electro-
lytes are anticipated to be low cost, as the salt is both
inexpensive and comprises a small fraction of the total
electrolyte compared to a ionic liquid chelate such as
[(CH3CN)2:LiTFSI] and suggest substantial mass savings in
comparison to the highly concentrated salt-in-solvent electro-
lytes, which also exhibit reduced polysulfide shuttling. The
next challenge will be to overcome the capacity fade, which
appears to be due to gradually increasing electronic isolation
of the Li2S formed on discharge owing to the complexity of
dissolution-precipitation chemistry in poorly solvating media.
Future studies will focus on the role of soluble redox
mediators to address this problem, which have been found
indispensable in recent studies of Li-S cells.[36,37]
[1] P. G. Bruce, S. A. Freunberger, L. J. Hardwick, J.-M. Tarascon,
Nat. Mater. 2012, 11, 19.
[2] N. S. Choi, Z. Chen, S. A. Freunberger, X. Ji, Y. K. Sun, K.
Amine, G. Yushin, L. F. Nazar, J. Cho, P. G. Bruce, Angew.
Chem. Int. Ed. 2012, 51, 9994; Angew. Chem. 2012, 124, 10134.
[3] L. F. Nazar, M. Cuisinier, Q. Pang, MRS Bull. 2014, 39, 436.
[4] R. Rauh, K. Abraham, G. Pearson, J. Surprenant, S. Brummer, J.
Electrochem. Soc. 1979, 126, 523.
[5] R. Rauh, F. Shuker, J. Marston, S. Brummer, J. Inorg. Nucl.
Chem. 1977, 39, 1761.
[6] Q. Pang, X. Liang, C. Y. Kwok, L. F. Nazar, Nat. Energy 2016, 1,
16132.
[7] D. Eroglu, K. R. Zavadil, K. G. Gallagher, J. Electrochem. Soc.
2015, 162, A982.
[8] Y. Yamada, A. Yamada, J. Electrochem. Soc. 2015, 162, A2406.
[9] J.-W. Park, K. Ueno, N. Tachikawa, K. Dokko, M. Watanabe, J.
Phys. Chem. C 2013, 117, 20531.
[10] M. Cuisinier, P.-E. Cabelguen, B. Adams, A. Garsuch, M.
Balasubramanian, L. Nazar, Energy Environ. Sci. 2014, 7, 2697.
[11] D. Aurbach, E. Pollak, R. Elazari, G. Salitra, C. S. Kelley, J.
Affinito, J. Electrochem. Soc. 2009, 156, A694.
[12] F. Wu, J.-T. Lee, J. Nitta, H. Kim, O. Borodin, G. Yushin, Adv.
Mater. 2015, 27, 101.
[13] J. Gao, M. A. Lowe, Y. Kiya, H. c. D. AbruÇa, J. Phys. Chem. C
2011, 115, 25132.
[14] B. A. Shainyan, L. L. Tolstikova, Chem. Rev. 2013, 113, 699.
[15] A. S. Madsen, H. M. Kristensen, G. Lanz, C. A. Olsen, Chem-
MedChem 2014, 9, 614.
[16] R. Younesi, G. M. Veith, P. Johansson, K. Edstrçm, T. Vegge,
Energy Environ. Sci. 2015, 8, 1905.
[17] I. Krossing, I. Raabe, Angew. Chem. Int. Ed. 2004, 43, 2066;
Angew. Chem. 2004, 116, 2116.
[18] I. Krossing, Comprehensive Inorganic Chemistry II, Vol. 1,
Elsevier, Oxford, 2013, pp. 681 – 705.
[19] T. A. Engesser, M. R. Lichtenhaler, M. Schleep, I. Krossing,
Chem. Soc. Rev. 2016, 45, 789 – 899.
[20] S. Tsujioka, B. G. Nolan, H. Takase, B. P. Fauber, S. H. Strauss, J.
Electrochem. Soc. 2004, 151, A1418.
[21] H. Tokuda, S.-i. Tabata, M. A. B. H. Susan, K. Hayamizu, M.
Watanabe, J. Phys. Chem. B 2004, 108, 11995.
[22] S. Bulut, P. Klose, M. M. Huang, H. Weingꢂrtner, P. J. Dyson, G.
Laurenczy, C. Friedrich, J. Menz, K. Kꢃmmerer, I. Krossing,
Chem. Eur. J. 2010, 16, 13139.
[23] A. Rupp, N. Roznyatovskaya, H. Scherer, W. Beichel, P. Klose,
C. Sturm, A. Hoffmann, J. Tꢃbke, T. Koslowski, I. Krossing,
Chem. Eur. J. 2014, 20, 9794.
Acknowledgements
[24] A. Rupp, S. Welle, P. Klose, H. Scherer, I. Krossing, ChemPhys-
Chem 2015, 16, 1940.
[25] A. B. Rupp, I. Krossing, Acc. Chem. Res. 2015, 48, 2537.
[26] L. Suo, Y.-S. Hu, H. Li, M. Armand, L. Chen, Nat. Commun.
2013, 4, 1481.
The research was supported by the BASF International
Scientific Network for Electrochemistry and Batteries. We
thank the Waterloo Institute of Nanotechnology for awarding
a fellowship to A.S.
[27] M. Cuisinier, C. Hart, M. Balasubramanian, A. Garsuch, L. F.
Nazar, Adv. Energy Mater. 2015, 5, 1401801.
[28] S. Urbonaite, T. Poux, P. Novꢄk, Adv. Energy Mater. 2015, 5,
1500118.
[29] H. Kim, J. T. Lee, G. Yushin, J. Power Sources 2013, 226, 256 –
265.
[30] X. Li, A. Lushington, Q. Sun, W. Xiao, J. Liu, B. Wang, Y. Ye, K.
Nie, Y. Hu, Q. Xiao, R. Li, J. Guo, T. Sham, X. Sun, Nano Lett.
2016, 16, 3545 – 3549.
[31] K. H. Wujcik, D. R. Wang, A. Raghunathan, M. Drake, T. A.
Pascal, D. Prendergast, N. P. Balsara, J. Phys. Chem. C 2016, 120,
18403.
Conflict of interest
The authors declare no conflict of interest.
Keywords: lithium–sulfur batteries · low ion-pairing salts ·
polysulfide shuttle · sparingly solvating electrolyte · sulfonamide
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

Angewandte
Communications
Chemie
[32] K. H. Wujcik, T. A. Pascal, C. Pemmaraju, D. Devaux, W. C.
Stolte, N. P. Balsara, D. Prendergast, Adv. Energy Mater. 2015, 5,
1500285.
[33] M. Yoshizawa, W. Xu, C. Angell, J. Am. Chem. Soc. 2003, 125,
15411.
[36] S. Meini, R. Elazari, A. Rosenman, A. Garsuch, D. Aurbach, J.
Phys. Chem. Lett. 2014, 5, 915.
[37] P. D. Frischmann, L. C. Gerber, S. E. Doris, E. Y. Tsai, F. Y. Fan,
X. Qu, A. Jain, K. A. Persson, Y.-M. Chiang, B. A. Helms, Chem.
Mater. 2015, 27, 6765.
[34] K. Ueno, H. Tokuda, M. Watanabe, Phys. Chem. Chem. Phys.
2010, 12, 1649.
[35] X. Han, Y. Gong, K. K. Fu, X. He, G. T. Hitz, J. Dai, A. Pearse,
B. Liu, H. Wang, G. Rubloff, Y. Mo, V. Thangadurai, E. D.
Wachsman, L. Hu, Nat. Mater. 2016, https://doi.org/10.1038/
nmat4821.
Manuscript received: January 29, 2017
Revised: March 12, 2017
Final Article published: && &&, &&&&
6
www.angewandte.org
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Lithium–Sulfur Batteries
The detrimental polysulfide shuttle asso-
ciated with Li-S batteries is alleviated
(without the addition of LiNO3) through
the combined use of a non-polar sulfo-
namide solvent and low-ion-pairing salt
as the electrolyte. This combination
transforms to a sparingly solvating elec-
trolyte at 508C and sustains close to
theoretical capacity at rates of C/5 with
a coulombic efficiency of 99.7%.
A. Shyamsunder, W. Beichel, P. Klose,
Q. Pang, H. Scherer, A. Hoffmann,
G. K. Murphy, I. Krossing,*
L. F. Nazar*
&&&& — &&&&
Inhibiting Polysulfide Shuttle in Lithium–
Sulfur Batteries through Low-Ion-Pairing
Salts and a Triflamide Solvent
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7
These are not the final page numbers!