PUBLISHED: 8 SEPTEMBER 2016 | ARTICLE NUMBER: 16141 | DOI: 10.1038/NENERGY.2016.

141

comment

A solid future for battery

development
Jürgen Janek and Wolfgang G. Zeier
Solid-state batteries have recently attracted great interest as potentially safe and stable high-energy
storage systems. However, key issues remain unsolved, hindering full-scale commercialization.

L
ithium-ion batteries (LIBs) have proven metal leaching in LIB cells. SEs, on the other temperature at sufficient C-rates is one of
to be an efficient energy storage system hand, allow transfer of lithium ions only and the challenges scientists and engineers face
in terms of their energy and power act as functional separators with only minor in the forthcoming years. Clearly, polymer-
density, reliability and cyclability 1. Today self-discharge (due to negligible electronic based batteries without competitive energy
the state-of-the-art LIBs offer volumetric conductivity). Moreover, lithium ions and density and rate performance will not enter
and gravimetric energy densities up to anions are mobile in liquid electrolytes, the mass market.
770 Wh l−1 and 260 Wh kg−1, respectively, causing severe concentration gradients of The potential of inorganic SEs for
which was not expected 10 years ago. the conducting salts during current flow improving battery performance at high
Unfortunately, the energy density of and limiting the cell current, whereas only currents has recently been demonstrated4,
conventional LIBs will soon reach its limit 2 lithium ions are mobile in SEs so this bulk pointing to a misconception that SEs are
(Fig. 1). On the other hand, there is an ever- polarization cannot occur. Therefore, higher inherently poor conductors at ambient
increasing demand for batteries with even current densities and quicker charging temperature. A number of ternary and
higher energy density as well as high power times are conceivable in SSBs. Secondly, quaternary sulfides and thiophosphates
density (quick discharging and charging some inorganic solid electrolytes are stable (Li7P3S11 and Li10GeP2S12)5,7 have also
performance). Solid-state batteries (SSBs) at elevated temperatures, improving battery been reported to have room temperature
that use solid electrolytes (SEs) instead of safety. In fact, their conductivity always conductivities higher than typical
liquid ones could offer both high energy and increases with increasing temperature, liquid electrolytes. However, the major
high power density 3. While slow kinetics is so their conductivity may even reach drawback of many inorganic SEs is their
considered the Achilles’ heel of solid-state that of liquid electrolytes5. Thirdly, the low thermodynamic stability 8. Most solid
systems, a recent study has shown that SSBs mechanical rigidity of SEs may prevent electrolytes are easily reduced at low
may well offer high power densities4. the dendrite formation that is caused by potentials (for example, by lithium metal) —
Despite fast growing interest in the electrodeposition of lithium, and thus just like their liquid competitors — and
SSBs, many challenges remain in facilitate the use of lithium-metal anodes. oxidized at intermediate potentials.
both manufacturing and fundamental Solid electrolytes can be divided into Protecting interphases are therefore
understanding of the technology. Here we two major groups: inorganic solids — required to stabilize the electrolyte/electrode
aim to provide views on some of the most either crystalline, glass or glass-ceramic contact 9, as happens in conventional
relevant questions under active discussion: in nature — and organic solid polymers. lithium-ion batteries. Today’s crystalline
which type of SEs can be used; how can In addition, solid-like dispersions of inorganic SEs offer the highest lithium
we overcome the inherent mechanical nanoparticles in liquids can be considered ion conductivities in the solid state, with
stability issues of SSBs; will SSBs fulfill as ‘semi-solid’ electrolytes. Polymer- phosphates and oxides coming close. While
the expectation of safe storage systems; based electrolytes appear to be the natural thiophosphates are ductile and easily form
and will the use of SEs enable the use of choice as they can compensate for volume dense cathode composites, oxides are brittle
lithium-metal anodes for a real jump in changes of electrodes by elastic and plastic and often experience mechanical failure
energy density? deformation. However, the polymer through cracking.
electrolytes’ lithium-ion conductivity is
Solid electrolytes too low for battery operation at room Key performance measures
The vast majority of batteries use organic temperature, and the operation in electric LIB technology is already mature, and
liquid electrolytes, which are low-cost and vehicles requires temperatures above 80 °C. the fight for reduced costs (per kWh)
easy to prepare. However, safety concerns Even then, their rate capability is limited, and further performance improvement is
often arise, as these liquid electrolytes are preventing fast charging 6. Interestingly, clearly dominating the market. SSBs will
combustible. In contrast, SEs offer some at elevated temperatures lithium anodes only become a major contender if they can
unique beneficial features. Firstly, unwanted can be safely cycled with suitable polymer provide a significant performance jump in
chemical ‘cross-talk’ of the electrodes is electrolytes. Hence, the search for one or more of the key properties. Usually
overcome. In liquid electrolytes, soluble stable polymer electrolytes for use with energy density is considered the top priority,
electrode components can diffuse across the lithium-metal anodes and lithium nickel but power density is important when it
battery, which can result in the ‘polysulfide cobalt manganese oxide or lithium nickel comes to the need for quick charging. Long-
shuttle’ in lithium–sulfur cells and transition cobalt aluminum oxide cathodes at ambient term stability, both long cycle and calendar

NATURE ENERGY | VOL 1 | SEPTEMBER 2016 | www.nature.com/natureenergy 1

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
comment

a b Temperature (K)
600 500 400 300 200
1,000 500 0
wvol
Physicochemical limit Li10GeP2S12
wgrav –1
800 400 Li3.25Ge0.25P0.75S4 Li7P3S11 glass ceramic

Energy density (Wh kg–1)

–2
Energy density (Wh l–1)

log [σ (S cm–1)]
EC:DEC:LiPF6
600 300 –3
P(EO)12-
400 –4 LiTFSI EC:PC:LiPF6
200
Li3.5Zn0.25GeO4
–5 Li7La3Zr2O12
200 100
–6 LiPON
Li1.5Al0.5Ge1.5(PO4)3
0 0
1990 1995 2000 2005 2010 2015 2 3 4 5
Year Reciprocal temperature (1/1,000 K–1)

Figure 1 | Energy densities of lithium-ion batteries and ionic conductivities of lithium electrolytes. a, Energy densities of the standard cylindrical 18650 LIB
cells (with a volume of approximately 16 cm3 and a mass of 48 g) over the past 25 years are shown. The data until the year 2012 are taken from ref. 19, where
an extrapolation (represented by the dashed curves) was made for the energy density values after the year 2004. The extrapolated data in the year 2012 was
confirmed by cell producers20 and is shown as solid symbols. The data in year 2016 also comes from cell producers21. Specifically, the latest 18650 cell delivers
a 3.5 Ah charge at a mean voltage of 3.65 V, corresponding to a volumetric (gravimetric) energy density wvol (wgrav) of roughly 770 Wh l−1 (270 Wh kg−1). The
energy density of commercial LIB cells has consistently increased to this level by almost a factor of four compared to the first commercial cells in 1991, but
may soon reach a limit (represented by a bar that refers to wvol) assuming no unforeseen developments in materials technology. The exact value of the limit is
influenced by how the cell is designed and processed2. b, Ionic conductivities of solid electrolytes are shown in comparison to those of liquid electrolytes and
polymers. The data includes that of Li10GeP2S12 (ref. 5), P(EO)12-LiTFSI (ref. 22), Li3.25Ge0.25P0.75S4 (ref. 23), Li7P3S11 (ref. 24), EC:DEC:LiPF6 and EC:PC:LiPF6 (ref. 25),
LiPON (ref. 26), Li3.5Zn0.25GeO4 (ref. 27), Li7La3Zr2O12 (ref. 28) and Li1.5Al0.5Ge1.5(PO4)3 (ref. 29). The vertical grey line marks room temperature. EC:DEC:LiPF6
marks the commonly used liquid electrolyte in LIB. P(EO)12-LiTFSI marks a typical polymer electrolyte on the basis of polyethylene oxide. Other labels are
chemical formulae of the solid electrolytes.

life (lifetime of batteries in terms of number either high-voltage cathodes (for example, There is reasonable doubt whether all
of discharge/charge cycles and time after materials with a 5 V redox potential) or these problems can be solved. A successful
production), is another key requirement, as high-capacity materials that cannot be used integration of the lithium anode would offer
the volume changes of the electrodes during in combination with liquid electrolytes an increase of up to 70% in energy density
cycling of the SSB cause mechanical strain (for example, lithium metal). Oxide- or and serious attempts are surely worth the
and stability problems. While safety appears phosphate-based solid electrolytes are effort. It is worth noting that lithium-metal
to be a less critical issue with the state-of- expected to chemically withstand high electrodes operate well in thin film solid-
the-art LIBs, the further increase in energy voltages, while liquid electrolytes are not state batteries with low area-specific
and power density may still cause problems stable at these voltages. capacity, where only about 1 μm of lithium
for SSBs. Lastly, the cost per kWh of energy Solid electrolytes are often considered is cycled12.
storage capacity, highly dependent on ‘enablers’ of high-capacity lithium-metal
production technology, is also important 10. anodes, as their mechanical strength may Power density
These five challenges define the roadmap for prevent dendrite growth — yet a realistic Power density is a measure of how quickly
the successful development of SSBs. assessment of the problems in the way of batteries can deliver stored energy.
the successful integration of lithium-metal Quick charging or discharging requires
Energy density anodes is missing. Most lithium SEs are high currents, leading to faster material
Estimation of practical energy densities must thermodynamically unstable against degradation and heat release (depending on
take into account all inactive components lithium metal8,11, and protective films or a the cell impedance). As solids usually exhibit
(electrolytes, electrode additives, metal stable solid electrolyte interphase (SEI) are larger thermal conductivities than liquids,
foils and so on) and also account for required. A 20-μm-thick lithium layer is heat can be released faster and hot spots can
proper charge capacity limits of the active needed to replace the much thicker standard be prevented. If SSBs are constructed with
materials. If only the liquid electrolyte of graphite anode of a typical LIB thick-film highly conductive electrolytes, the electrode
an LIB is replaced by a SE, the volumetric cell (such as the common 18650 cell). contacts will be decisive for the cell kinetics.
energy density remains unchanged (Fig. 2). Inhomogeneous dissolution and subsequent Like all solid–solid contacts, the solid
As SEs have a higher density than liquid deposition of such a comparably thick metal electrolyte/electrode contact depends on
electrolytes, the gravimetric energy density film causes severe mechanical strain and the preparation conditions and may become
is lower in this case. Increasing the cell may lead to mechanical failure. In crystalline the kinetic bottleneck. Unfortunately,
voltage from 4.2 V to 5 V would result in solid electrolytes, lithium dendrites can there is not much information available
an increase of more than 20% in energy grow along grain boundaries and lead to on the kinetics of electrodes in SSBs yet.
density (depending on additional cathode fast short-circuits, and the cyclic dissolution As SEs with their rigid structure cannot
capacity at high voltage), and thus the use and deposition of the metal at SEs needs to de-mix (no bulk polarization as in liquids),
of a SE is favorable if it allows the use of be fast enough to allow acceptable C-rates. they may allow higher current densities

2 NATURE ENERGY | VOL 1 | SEPTEMBER 2016 | www.nature.com/natureenergy

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 comment

and higher operation temperatures4. On LiM-SSB LIB LI-SSB

the other hand, solid electrolyte/cathode – V
V + – V + – V +
contacts may form ‘rectifying’ contacts with
blocking kinetics in one direction (such
as a diode) at high potentials, due to the
formation of lithium depletion layers in wvol = +70% wvol = +0%
the SE, and additional interface protection wgrav = +40% wgrav = –10%
may be necessary. Whether fast kinetics
turns out to be one of the key advantages
of SSB is an open and inspiring question,
and the answer could offer a completely
counter-intuitive approach. Figure 2 | Typical battery architectures for the conventional lithium-ion and solid-state batteries. The
volumetric and gravimetric energy densities are represented by wvol and wgrav, respectively. Conventional
Long-term stability lithium-ion batteries (LIB, middle) contain a porous anode (negative electrode typically made of graphite,
The potential for long-term stability of grey circles) and a porous cathode (positive electrode typically made of a layered transition metal oxide,
SSBs is one of the major incentives for violet circles) as ‘active’ storage components, coated on thin copper (negative electrode) and aluminum
their development. Unfortunately, far from (positive electrode) foils, serving as current collectors. A thin separator (grey band, about 10 μm thick)
what might naively be expected, the best is placed between the much thicker electrodes (each about 100 μm thick). A liquid electrolyte infiltrates
solid electrolytes show relatively small the porous electrode and separator assembly, providing fast ion transfer between the electrodes and
thermodynamic ranges of stability 13,14 preventing electronic short-circuiting. In a lithium ion solid-state battery with a conventional anode (right,
and their electrochemical window is still LI-SSB) the liquid electrolyte in the electrodes is completely replaced by solid electrolyte (dark orange
a subject of research. Once in contact circles) in the electrodes and the electrolyte-filled separator is replaced by another or the same solid
with lithium metal, most of the sulfides electrolyte (orange circles). Using the same SE in the electrodes and the separator appears to be a natural
and thiophosphates react and form an choice, yet for stability and production reasons, a different SE might be more favourable. Only with a
interphase. If a highly conductive solid lithium-metal anode (light yellow) that has a theoretical energy density of 3,700 mA g−1, can a significant
electrolyte interphase is formed, long- gain in energy density be achieved. Changes in energy density are estimated based on the density
term operation will be secured; however, increase from liquid to solid, taking into account the high specific capacity of lithium metal (left, LiM-SSB)
mixed conducting interphases will lead to and complete replacement of the graphite and anode electrolyte.
fast deterioration of the SSB. Interfacial
instabilities may also occur on the cathode15.
While coating the cathode particles seems be required, which complicates production leads to a volume fraction of about 25%
to provide long-term stability, allowing and increases costs. Thus, safety may liquid electrolyte.
high-voltage operation and facilitating not automatically be enhanced in SSBs, Figure 3 shows a schematic architecture
high-energy cells4,16, the stability of the especially when the risk of mechanical of a potential SSB (single cell). While the
interphases remains challenging. In general, failure is considered. use of liquid electrolytes ensures good
SSBs also suffer from interfacial instabilities electrode contact and efficient ionic
and degradation, mirroring the situation Architecture and processing of SSB pathways, a solid electrolyte will typically
in the conventional LIBs. While the solid Still, high costs of LIBs are an obstacle to have a microstructure consisting of small
electrolyte prevents unwanted electrode mass production of electric vehicles. Due to grains that must be in good contact
‘cross-talk’, the unwanted products of side strong competition among cell producers the with the electrode material as well as
reactions form and stay at the interface. The cost per kWh has dropped to US$145, and other electrolyte grains. Additionally, an
kinetic instability of the lithium anode itself public company reports and forecasts expect independent network of electronically
due to the pore formation during discharge it to drop further to US$100 within the next conducting particles has to be established:
and the dendrite formation during charge few years. No reliable cost estimate for SSB solid composite electrodes ensuring
also requires further research. production is available yet. US$100 kWh−1 sufficient electronic and ionic percolation
corresponds to roughly US$30 kg−1 material have to be formed. As intercalation
Safety and manufacturing costs on the cell level, electrodes are known to exhibit volume
If only oxide ceramic components (which which will not leave much room for complex changes, the mechanical integrity of the
are stable against lithium) were used, the materials processing in SSBs. solid composite is a critical issue. In a liquid
safety of high-energy batteries would be The electrolyte in state-of-the-art LIBs electrolyte the electrode can ‘breathe’ and
guaranteed. However, the best-suited ionic is a ‘cocktail’ of different organic solvents cycle reversibly. In a solid composite, pores
conductors are unfortunately sulfides. The (mostly carbonates), a conducting salt can form and cracking and loss of particle
formation of toxic H2S in case of battery (usually LiPF6), and functional additives (for contact may occur, resulting in increased
damage or the formation of SO2 by oxidation example, molecular compounds supporting interfacial resistance and overvoltage as
of the solid electrolyte may present a the formation of protective films on the well as capacity fading. Using mechanically
potential risk17. The more stable garnet-type anode). Pre-formed electrode sheets and soft ionic conductors such as sulfide and
oxides (for example, Li7La3Zr2O12) appear the separator are wound or stacked, and the thiophosphate glasses may help. As the solid
to be promising candidates. However, electrolyte is simply infiltrated. In optimized electrolyte requires mechanical mixing and
oxides are not as mechanically soft 18 and, LIBs only a few μl of electrolyte per cm2 the particle contacts are extremely sensitive
even though initially believed otherwise, cell are used. A top-of-the-range 18650 LIB to mechanical effects, solid electrolyte-
dendrites may enter into polycrystalline cell is equipped with a very thin electrolyte based batteries will probably require
solids along grain boundaries and lead layer of only 10 μm or less (depending on a revolutionary change of production
to short-circuits and mechanical failure. the separator) and thick porous electrodes technology. With the perspective of direct
Sophisticated protection techniques may that are also filled with electrolytes, which bipolar stacking of SSB cells, that is, the

NATURE ENERGY | VOL 1 | SEPTEMBER 2016 | www.nature.com/natureenergy 3

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
comment

Jürgen Janek and Wolfgang Zeier are at the Institute

RSE/SEI of Physical Chemistry and the Center for Materials
RSE,gb φ Research, Justus Liebig University, 35392 Giessen,
Li metal

Li+-depleted SE Germany. Jürgen Janek is also scientific

RSE director of BELLA laboratory at the Institute of
Coated
Nanotechnology, KIT, Karlsruhe, Germany.
SEI

Mechanical cathodes
pressure e-mail: [email protected];
[email protected]
Dendrites
References
1. Goodenough, J. B. & Park, K.‑S. S. J. Am. Chem. Soc.
135, 1167–1176 (2013).
2. Adelhelm, P. & Janek, J. in Handbuch Lithium‑Ionen‑Batterien
(ed. Korthauer, R.) Ch. 16, 199–217 (Springer, 2013).
Figure 3 | Major challenges in developing solid-state batteries. Owing to its high specific capacity, the 3. Robinson, A. L. & Janek, J. MRS Bull. 39, 1046–1047 (2014).
use of a lithium metal anode can significantly increase the cell energy density. However, a resistive SEI 4. Kato, Y. et al. Nat. Energy 1, 16030 (2016).
5. Kamaya, N. et al. Nat. Mater. 10, 682–686 (2011).
may form between the lithium anode and the SE, represented by RSE/SEI. In addition, the grains and grain
6. Stephan, A. M. & Nahm, K. S. Polymer 47, 5952–5964 (2006).
boundaries of the SE also contribute resistance, represented by RSE and RSE,gb. Dendrite formation due 7. Minami, T., Hayashi, A. & Tatsumisago, M. Solid State Ionics
to highly inhomogeneous lithium metal deposition causes serious risk of short-circuiting. Most SEs 177, 2715–2720 (2006).
8. Zhu, Y., He, X. & Mo, Y. ACS Appl. Mater. Interface
react with cathode active materials and need to be protected by coating the active material. Potential
7, 23685–23693 (2015).
Li depletion at the cathode (space charge with rectifying effect) could also be avoided by employing 9. Luntz, A. C., Voss, J. & Reuter, K. J. Phys. Chem. Lett.
a coating. The red curve indicates schematically the drop of the electric potential φ across the space 6, 4599–4604 (2015).
charge. The blue arrows indicate that mechanical pressure is required both to avoid contact loss due to 10. Troy, S. et al. Appl. Energy 169, 757–767 (2016).
11. Zhou, Y., He, X. & Mo, Y. J. Mater. Chem. A 4, 3253–3266 (2016).
local volume changes upon lithiation/delithiation and to achieve high power density. 12. Li, J., Ma, C., Chi, M., Liang, C. & Dudney, N. J. Adv. Energy Mater.
5, 1401408 (2015).
13. Wenzel, S. et al. Solid State Ionics 286, 24–33 (2016).
serial connection of cells in the SSB, which electrolytes allow high current densities 14. Wenzel, S. et al. Chem. Mater. 28, 2400–2407 (2016).
is not possible in liquid-based LIB, further without concentration polarization, high 15. Richards, W. D., Miara, L. J., Wang, Y., Kim, J. C. & Ceder, G.
Chem. Mater. 28, 266–273 (2016).
advantages for a higher energy density power densities may be achieved as well in 16. Oh, G., Hirayama, M., Kwon, O., Suzuki, K. & Kanno, R.
are offered. SSBs with intercalation anode and cathode Chem. Mater. 28, 2634–2640 (2016).
(lithium-metal anodes probably do not offer 17. Muramatsu, H., Hayashi, A., Ohtomo, T., Hama, S.
& Tatsumisago, M. Solid State Ionics 182, 116–119 (2011).
Towards a solid future high power densities). 18. Sakuda, A., Hayashi, A., Takigawa, Y., Higashi, K.
Solid-state batteries have experienced SSBs as a concept are connected with & Tatsumisago, M. J. Ceram. Soc. Jap. 121, 946–949 (2013).
fast growing interest. Nevertheless, the prejudices and misjudgments — leading 19. Broussely, M. & Archdale, G. J. Power Sources 136, 386–394 (2004).
20. Lithium ion NCR18650B specifications (Panasonic, 2012);
understanding of interfacial reactions to overly optimistic or pessimistic views. http://go.nature.com/2b1p3SS
and cell architecture and performance are A number of key issues have not yet been 21. http://go.nature.com/2b1oKaK
still at an early stage. Three main points addressed and major efforts are required 22. Edman, L., Ferry, A. & Doeff, M. M. J. Mater. Res.
15, 1950–1954 (2000).
are clear. Firstly, optimization of poor for a better evaluation of the technological 23. Kanno, R. & Murayama, M. J. Electrochem. Soc.
interfacial kinetics between SE and active future of all-solid-state batteries. In our 148, A742–A746 (2001).
materials requires more investigation. opinion, high voltages and high power 24. Seino, Y., Ota, T., Takada, K., Hayashi, A. & Tatsumisago, M.
Energy Environ. Sci. 7, 627–631 (2014).
Secondly, mechanical pressure is required densities will be achieved in the coming 25. Stallworth, P. et al. J. Power Sources 81–82, 739–747 (1999).
to guarantee stable operation of a solid-state years, and while the employment of 26. Yu, X. et al. J. Electrochem. Soc. 144, 524–532 (1997).
cell, and quantitative studies are needed thick lithium anodes seems unlikely, 27. Bruce, P. G. & West, A. R. J. Solid State Chem. 44, 354–365 (1982).
here. Finally, only the lithium-metal anode SSB development may be prompted 28. Buschmann, H. et al. Phys. Chem. Chem. Phys.
13, 19378–19392 (2011).
really brings a major advantage to SSBs by the superior kinetics when using 29. Francisco, B. E., Stoldt, C. R. & M’Peko, J.-C. Chem. Mater.
with respect to energy density. As solid intercalation anodes. ❐ 26, 4741–4749 (2014).

4 NATURE ENERGY | VOL 1 | SEPTEMBER 2016 | www.nature.com/natureenergy

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.