R ES E A RC H
TECHNICAL RESPONSE
◥
standard conditions, the equilibrium voltage of
cell reaction 2 can be expressed as a function of
the activity (a) of the reactants and products.
!
v
v
avLi a O22 aHH22OO
2:303RT
Ee ¼ E o þ
ð3Þ
lg Li OvLiOH
aLiOH
nF
BATTERIES
Response to Comment on “Cycling
Li-O2 batteries via LiOH formation
and decomposition”
where R is the gas constant, F is the Faraday
constant, n is the number of electrons involved
in the reaction, and T is the temperature.
Because Li and LiOH are solids, and gaseous
O2 is at close to ~1 bar during the cell reaction—
i.e., they are in their respective standard states—
the above equation can be simplified.
Tao Liu,1 Gunwoo Kim,1,2 Javier Carretero-González,1 Elizabeth Castillo-Martínez,1
Paul M. Bayley,1 Zigeng Liu,1 Clare P. Grey1*
Ee ¼ E o þ 0:01481gða2H2 O Þ
W
e recently reported (1) on a highly reversible lithium-oxygen battery composed
of a reduced graphene oxide (rGO) electrode and a dimethoxyethane (DME)/
LiTFSI electrolyte. The additives (H2O
and LiI) were key in controlling the nature of
the battery reactions. LiOH, instead of the commonly reported Li2O2 phase, was the predominant
product during discharge, the protons primarily
coming from H2O in the cell. On charge, LiOH
was removed [as seen by x-ray diffraction and
proton nuclear magnetic resonance (1H NMR)],
and we proposed that this occurs via a reaction
with I3–, O2 being observed by mass spectroscopy.
In their Comment, Viswanathan et al. (2) argue
that a charge mechanism involving reaction with
I3– to produce O2 is not feasible, because the
redox potential for reaction 1
4Li+ + 4e– + O2,g + 2H2Ol ↔ 4LiOHs
(1)
is 3.34 V versus Li+/Li under standard conditions,
whereas the redox potential for 6I– ↔ 2I3– + 4e–,
is ~3.0 V in the DME-based electrolyte used here.
We discuss this issue and possible mechanisms
for LiOH removal.
LiOH formation in the presence of LiI has been
previously reported in lithium-oxygen batteries
(3), and the reversible formation/decomposition
of LiOH—e.g., using ruthenium catalysts and tetraglyme (4)/ dimethyl sulfoxide (DMSO) (5)–based
electrolytes with added water—was observed at
3.1 to 3.2 V. It is well established that the electrolyte affects the potential of a redox couple. Indeed, even the I3–/I– couple drops from 3.53 V in
water under standard conditions (6, 7) to 3.35,
3.1, and 3.0 V in acetonitrile (6), tetraglyme, and
DME (1), respectively. The O2/O22– couple varies
from ~3.0 V in DMSO electrolyte to ~3.5 V in
acetonitrile electrolyte (8), and the recent “water
in salt” work (9) showed that the redox potentials
for water reduction/oxidation shift considerably
due to the chemical potential changes of water
and Li+ in the electrolyte.
Changes in the redox potential of reaction 1
under nonstandard conditions arise from at least
two factors. First, the concentrations of the species
in the electrolyte deviate noticeably from standard
conditions. Second, the different coordination environments of the species/ions differ dramatically
between solvents. The standard Gibbs free-energy
change, DGro, of the overall cell reaction
4Lis + O2,g + 2H2Ol ↔ 4LiOHs
(2)
is –1282 kJ/mol (10) at standard conditions resulting in Eo = –DGro /nF = 3.32 V. Under non-
The activity, or the chemical potential, of
water in the batteries is given by
m−mo
aH2 O ¼ exp
RT DH
DS
¼ exp
exp −
ð5Þ
RT
R
where m is the chemical potential, and DH and
DS are the enthalpy and entropy difference of
water in LiI/LiTFSI/DME electrolyte compared
with those in a LiI/LiTFSI aqueous electrolyte at
standard conditions
H2Ol ↔ H2OLiI,LiTFSI,DME
(6)
and “H2Ol” in reaction 2 should be replaced by
“H2O(Li,LiTFSI,DME)”. It is then relevant to ask how
large a value of DG for reaction 6 is required to
shift the potential of reaction 2 so that it drops
below that of the I–/I3– couple in the same electrolyte. Although this value is currently not known
for DME, a relatively small value of –60.0 kJ/mol
is required to reduce the couple of reaction 2
down to 3.0 V. DS makes only small changes to
voltage (of –20 mV to –77 mV, where the entropy
is approximated by the concentration of water in
the system). The DH for reaction 6 is controlled
by the loss of water-water interactions and the
difference between water–Li salt and DME–Li
salt interactions. Evidence for strong water–Li+
interactions in DME comes from the hygroscopic
Fig. 1. Reaction kinetics
of a LiOH + LiI3 mixture
in DME/water solutions
followed by UV-visible
spectroscopy (with LiOH
in a 103 excess).
1
Department of Chemistry, University of Cambridge,
Lensfield Road, Cambridge CB2 1EW, UK. 2Cambridge
Graphene Centre, University of Cambridge, Cambridge CB3
0FA, UK.
*Corresponding author. Email: [email protected]
SCIENCE sciencemag.org
6 MAY 2016 • VOL 352 ISSUE 6286
667-d
Downloaded from http://science.sciencemag.org/ on June 18, 2017
Lithium-oxygen (Li-O2) batteries cycle reversibly with lithium iodide (LiI) additives in
dimethoxyethane (DME) to form lithium hydroxide (LiOH). Viswanathan et al. argue that
because the standard redox potential of the four-electron (e−) reaction, 4OH– ↔ 2H2O +
O2 + 4e–, is at 3.34 V versus Li+/Li, LiOH cannot be removed by the triiodide ion (I3–).
However, under nonaqueous conditions, this reaction will occur at a different potential.
LiOH also reacts chemically with I3– to form IO3–, further studies being required to
determine the relative rates of the two reactions on electrochemical charge.
ð4Þ
R ES E A RC H | T E C H N I C A L RE S P O N S E
nature of LiTFSI and LiI salts and their higher
solubility in water than in ethers (7, 9). Theoretical calculations (9) also suggest a very high binding energy of water with Li+ cations in an aqueous
LiTFSI electrolyte. Hydration enthalpies for LiI
and LiTFSI in water/nonaqueous solvents are an
order of magnitude larger [e.g., –828 kJ/mol for
LiI in water and –756 kJ/mol for LiTFSI in acetonitrile (11, 12)] than the –60 kJ/mol assumed
above, all suggesting that this value for DGvi is
certainly plausible, but we stress that further
measurements/calculations are required.
The chemistry is, however, more complicated
because I3– can both react with LiOH forming
I– and liberating oxygen
4LiOH + 2I3– ↔ 4Li+ + 6I– + 2H2O + O2 (7)
and also react to form metastable IO–, which then
disproportionates forming IO3– and I– (13, 14).
The low concentrations of water in our electrolytes help drive both reaction equilibria to the
right-hand side. Our Raman and ultraviolet (UV)–
visible studies of the chemical reaction of LiOH
and I3– show that reaction 8 dominates under
667-d
6 MAY 2016 • VOL 352 ISSUE 6286
discharge-charge all require further investigation. Nonetheless, the evidence for reversible
LiOH formation, in the presence of O2 (and
LiI), is compelling.
REFERENCES
1. T. Liu et al., Science 350, 530–533 (2015).
2. V. Viswanathan et al., Science 352, 667 (2016).
3. W. J. Kwak et al., J. Mater. Chem. A 3, 8855–8864
(2015).
4. S. Wu, J. Tang, F. Li, X. Liu, H. Zhou, Chem. Commun. 51,
16860–16863 (2015).
5. F. Li et al., Nat. Commun. 6, 7843 (2015).
6. G. Boschloo, A. Hagfeldt, Acc. Chem. Res. 42, 1819–1826
(2009).
7. Y. Zhao, L. Wang, H. R. Byon, Nat. Commun. 4, 1896
(2013).
8. L. Johnson et al., Nat. Chem. 6, 1091–1099 (2014).
9. L. Suo et al., Science 350, 938–943 (2015).
10. D. R. Lide, Ed., CRC Handbook of Chemistry and Physics (CRC
Press, ed. 84, 2003)
11. D. F. C. Morris, Electrochim. Acta 27, 1481–1486 (1982).
12. D. M. Seo, O. Borodin, S. D. Han, P. D. Boyle, W. A. Henderson,
J. Electrochem. Soc. 159, A1489–A1500 (2012).
13. J. Paquette, B. L. Ford, Can. J. Chem. 63, 2444–2448
(1985).
14. C. H. Li, C. F. White, J. Am. Chem. Soc. 65, 335–339
(1943).
15. J. E. Vitt, D. C. Johnson, J. Appl. Electrochem. 24, 107
(1994).
1 December 2015; accepted 25 March 2016
10.1126/science.aad8843
sciencemag.org SCIENCE
Downloaded from http://science.sciencemag.org/ on June 18, 2017
6LiOH + 3I3– ↔ 6Li+ + IO3– + 8I– + 3H2O (8)
aqueous conditions, the rate of the reaction slowing down noticeably, but still occurring in DME
solutions containing 3 to 6 weight % water (Fig.
1). Interestingly, the addition of LiTFSI speeds up
the reaction. However, electrochemical studies of
the oxidation of I3– in aqueous solutions with
carbon electrodes often observe a combination of
O2 evolution and IO3– formation (15). Thus, it is
clear that further studies are required to determine how much O2 is evolved on charge versus
being tied up as IO3–. Our 17O/1H NMR studies
of electrodes cycled in and beyond the first cycle
observe reversible LiOH formation, even when
using the thick electrodes (>0.150 mg) required
for the NMR experiments. Finally, Viswanathan
et al. suggested that parasitic reactions (3) between I–/I3– and the ether-based electrolyte represented the source of proton for LiOH and that
such irreversible processes cannot form the basis
of a practical Li-O2 battery. We agree that irreversible processes must be avoided. In our work,
DME electrolyte decomposition is not, however,
the dominant source of the protons in LiOH, and
super P carbon, rGO, and titanium carbide electrodes operate for hundreds of cycles (1). In our
current Li-O2 battery, the equilibria among water,
oxygen, and iodide; the surface functionality of
rGO; and the detailed reaction mechanism during
Response to Comment on ''Cycling Li-O2 batteries via LiOH formation and decomposition''
Tao Liu, Gunwoo Kim, Javier Carretero-González, Elizabeth Castillo-Martínez, Paul M. Bayley, Zigeng Liu and Clare P. Grey
Science 352 (6286), 667.
DOI: 10.1126/science.aad8843
http://science.sciencemag.org/content/352/6286/667.4
PERMISSIONS
http://www.sciencemag.org/help/reprints-and-permissions
Use of this article is subject to the Terms of Service
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title
Science is a registered trademark of AAAS.
Downloaded from http://science.sciencemag.org/ on June 18, 2017
ARTICLE TOOLS