|
|||||||||
|
|||||||||
Electrical Energy
Storage and Intercalation Chemistry
Abstract. The electrochemical reaction of
layered titanium disulfide with lithium giving the intercalation compound
lithium titanium disulfide is the basis of a new battery system. This
reaction occurs very rapidly and in a highly reversible manner at ambient
temperatures as a result of structural retention. Titanium disulfide
is one of a new generation of solid cathode materials.
We
have recently reported on the fundamental properties of the LixTiS2 series of intercalation compounds
(1-6). Those studies were initiated by our finding that titanium disulfide
could be used as the cathode of a high-energy-density reversible battery
with a lithium anode (7). We report here on this relationship between
intercalation chemistry and electrical energy storage.
The
recent work on high-energy batteries, which are required for electric
vehicle propulsion and for the storage of off-peak and solar power,
has been reviewed by Cairns and Shimotake (8). The high operating temperatures
of most of these batteries cause substantial corrosion problems that
are not yet solved and will pose significant barriers to their acceptance
by the general public. We therefore set out to determine whether an
alkali metal-based battery couple could be found that would operate
reversibly at ambient temperatures. For convenience, a lithium anode
and an organic electrolyte were chosen, and a search was initiated for
an electronically conductive, highly oxidizing solid that could react
readily and reversibly with lithium as the cathode. It was also necessary
that the solid be light in weight and inexpensive.
A
group of materials that are known to be highly conductive are the layered
dichalcogenides of the transition metals of groups IVB and VB of the
periodic table. They have also aroused much interest because of their
ability to intercalate a variety of molecules, such as pyridine, in
the van der Waals layers of the structure and because of the effect
of these inserted species on their superconductive properties, which
has been described by Gamble et al. (9). However, nothing was known
concerning the possibility of forming them in an electrochemical cell,
or of the energy or reversibility of the reaction. We have shown that
these intercalation complexes can be readily formed electrochemically
at ambient temperatures; thus, for example, pyridine was inserted in
a few minutes by the electrolysis of a solution of pyridinium hydrochloride
at a TaS2 electrode (10,11). The direct reaction
of pyridine with Ta S2 normally requires temperatures around
200°C and several days. The first indication that the free energy of
formation of the alkali metal complexes was substantial came from the
reaction of KOH with TaS2 (1). In this reaction potassium surrounded
by a ring of water molecules was intercalated; the water could be readily
driven out by heating to 100°C. That the potassium did not reduce the
water indicates a potential of K+ of around 2 volts or more relative
to potassium metal; this is in marked contrast to the corresponding
compound of graphite, C8K, where this potential is only 0.2 volt (12).
For
cell testing we chose to use TiS2, not only because it had the lowest
weight and
cost of all the layered dichalcogenides, but also because of its metallic
conductivity, whose origin has undergone extensive study (13, 14). In
addition, nuclear magnetic resonance studies (4) showed
that lithium self-diffusion is most rapid in TiS2 of all the layered sulfides. A simple
electrochemical cell was thus set up in which the anode was lithium,
the electrolyte LiPF6 dissolved in propylene carbonate,
and the cathode a single crystal of TiS2. The cell electromotive force (emf)
was 2.5 volts, a high value, as suggested by the hydration studies.
On discharge, initial current densities of 10ma per square centimeter
of active crystal area were found; these are about an order of magnitude
higher than those previously reported for any organic electrolyte battery
system. These measurements indicate that the lithium diffusion coefficient
is about 10-7 cm2/sec. The emf of this cell on open
circuit is shown by curve a in Fig. 1. The slope of this plot
is indicative of a single-phase reaction represented by the equation
xLi + TiS2 --> LixTiS2
X-ray
analysis of the discharge products showed that reaction proceeded by
inter- calation
of the lithium into the TiS2 lattice with a maximum expansion
of the structure
of 0.5 Å, ~ 10 percent, perpendicular to the basal planes (5). This was
confirmed by x-ray analysis of LiTiS2 (3) iformed from n-butyl lithium and
TiS2 (2).
By
starting with a cell with an LiTiS2 cathode made from n-butyl lithium
or a fully
discharged electrode, curve a in Fig. 1 could be retraced exactly, indicating
the complete reversibility of the system and the 100 percent coulombic
efficiency of the TiS2 electrode. To test the cell in a
more realistic configuration, a mixture of finely divided TiS2 and Teflon (9
to 1 by weight) was hot-pressed into a stainless steel grid of area
2 cm2, sur- rounded
by a polypropylene separator and a lithium anode, and immersed in an electrolyte
consisting of LiClO4 dissolved in a mixture of dimethoxyethane and
tetrahydrofuran (30 to 70). The discharge characteristics of this cell
at 10 and
4 ma are shown by curve b in Fig. 1. A cell was then shallow-cycled
- oper- ated
at 4 percent of its full capacity -more than 1100 times and retained
re- versibility
throughout, as shown by curve c in Fig. 1, which is a discharge curve
after these cycles. The TiS2 still maintained more than 70 percent
of its theoretical
capacity at this stage (based on the reaction Li + TiS2 --> LiTiS2), in- dicating
the remarkable reversibility of TiS2.
The
current densities measured on the single crystal were confirmed in the
powder cell configuration described above and the data obtained are
shown in Fig. I.
The
current density for high utilization optimized for; low values of y in Ti1+y because the excess titanium occupies
sites in the van der Waals layers, impeding the diffusion of lithium
by pinning the layers together (3). These currents are comparable to
those obtained in the intermediate-temperature (200°C) Na/SbClx molten salt cell (15). They are only
slightly less than those used in the high-temperature (400+°C) lithium/methanol-sulfide
cells (16, 17), where, for example
CuS was cycled at 50 ma/ cm2, FeS2 and FeS at 40 ma/ cm2, and NiS and Cr2S3 at10 ma/ cm2. Even higher current densities may
be obtained by using mol- ten salts at elevated temperatures. These high currents and the ready
reversibility of the reaction are directly associated with
the crystal structure, which remains essentially unchanged during reaction; no chemical
bonds are broken in the host TiS2 matrix during the insertion or removal
of lithium (5). Holleck et al. (18) studied these cells and found good reversibility
but reported very low current densities, 0.33 ma/ cm2.
The
energy density of the Li/TiS2 couple is found from Fig. 2 to be
480 watt-hour/kg, which is comparable to the energy densities calculated
for Na/S cycling in the single-phase region and the LiAl/FeS high-temperature cells now under development.
The values for the latter are 330 and 460 watt-hour/kg, respectively,
and are anticipated to reach 100+ watt-hour/kg in practical cell configurations.
As the TiS2 ambient-temperature cell will require
less deadweight associat- ed with heat insulation and corrosion-resistant
materials, it should also fall in this area, making it feasible for
electric vehicle propulsion. Preliminary calculations and extended high-current
operation near full capacity indicate that the required power densities
are achievable with the TiS2 cell (19).
In
conclusion, TiS2 has a high energy density and rate
capability when coupled with a lithium anode, a high electrical conductivity,
and a discharge-charge mechanism involving intercalation of lithium
between the layers of the host's crystal structure that permits extended
reversibility(7). Moreover, in contrast to most oxidants such as Cl2, TiS2 has a kinetically selective oxidizing
power, making it highly reactive to species that can be intercalated
but noncorrosive to its environment. This couple has potential as an
ambient-temperature, as well as high-temperature, battery for electric
vehicle propulsion. The Na/TiS2 couple is less interesting because
of the much greater free energy change with x (20) and the presence
of a number of crystalline phases, which places an upper limit of ~0.8
on x at 25°C (6). M. S. WHITTINGHAM Corporate
Research Laboratories, Exxon
Research and Engineering Company,
Linden, New Jersey 07036
References and Notes
1. M. S. Whittingham, Mater. Res. Bull. 9, 1981
(1974). 2. M. B. Dines, ibid. 10, 287 (1975); U.S. Patent
No. 3,933,688 (1976). 3. M. S. Whittingham and F. R. Gamble, Mater.
Res. Bull. 10, 363 (1975). 4. B. G. Silbernagel, Solid State Commun. 13,
1911 (1975). 5. M. S. Whittingham, J. Electrochem. Soc. 123,
315(1976). 6. B. G. Silbernagel and M. S. Whittingham,
Mater. Res. Bull. 11, 29 (1976). 7. M. S. Whittingham, Belgian Patent No. 819,672
(1973). 8. E. J. Cairns and H. Shimotake, Science 164,
1347 (1969). 9. F. R. Gamble, J. H. Osiecki, M. Cais, R.
Pisharody, F. J. DiSalvo, T. H. Geballe, ibid. 174, 493 (1971). 10.
M. S. Whittingham, Chem. Commun. (1974), p. 328. 11.
G. V. Subba Rao and J. C. Tsang, Mater. Res. Bull. 9,921(1974). 12.
S. Aronson, F. J. Salzano, D. Bellafiore, J. Chem. Phys. 49,434 (1968). 13.
A. H. Thompson, F. R. Gamble, C. R. Symon, Mater. Res. Bull. 10, 915
(1975). 14.
A. H. Thompson, Phys. Rev. Lett. 35, 1786 (1975). 15. J. Werth, I. Klein, R. Wylie, J. Electrochem. Soc. 122, 265C (1975). 16.
L. Heredy and L. R. McCoy, U.S. Patent No. 3,898,096(1975). 17.
D. R. Vissers, Z. Tomczuk, R. K. Steunenberg,
J. Electrochem. Soc. 121, 665 (1974). 18.
G. L. Holleck, F. S. Shuker, S. B. Brummer, in Proceedings of the 10th
Intersociety Energy Conversion Engineering Conference, Newark,
Delaware, August 1975 (Institute of Electrical and Electronics Engineers, New York, 1975). 19.
L. H. Gaines, preprint. 20.
D. A. Winn, thesis, Imperial College (1975). 21.
This electrolyte is susceptible to decomposition on overdischarge and
so is not suitable for com- mercial use. The cell was, however, deep-cycled-operated
at >58 percent of capacity-at these rates for 16 cycles before any apparent
degradation set in. 22.
I would like to thank F. R. Gamble for introducing me to the layered
sulfides and for his constant encouragement. In addition I gratefully
acknowledge the help of my colleagues, R. R. Chianelli, M. B. Dines, B. G. Silbernagel,
A. H. Thompson, R. W. Francis, L. H. Gaines, G. H. Newman, and B. M. L. Rao. 10
March 1976 |
|||||||||
|