See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/223152806

Thermal activated ("thermal") battery technology - Part IIIa: FeS(2) cathode

material

Article in Journal of Power Sources · March 2008

DOI: 10.1016/j.jpowsour.2007.11.017

CITATIONS READS

148 2,838

2 authors:

Patrick Masset Ronald Guidotti

Warsaw University of Technology / Kaziranga University Sierra Nevada Consulting
184 PUBLICATIONS 2,334 CITATIONS 165 PUBLICATIONS 1,604 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Thermodynamics View project

Carbons for Li-ion cells View project

All content following this page was uploaded by Ronald Guidotti on 11 October 2017.

The user has requested enhancement of the downloaded file.

 Available online at www.sciencedirect.com

Journal of Power Sources 177 (2008) 595–609

Review

Thermal activated (“thermal”) battery technology

Part IIIa: FeS2 cathode material
Patrick J. Masset a,∗ , Ronald A. Guidotti b
a Karl Winnacker Institut der Dechema e.V., Theodor-Heuss-Allee 25, 60486 Frankfurt am Main, Germany
b Sierra Nevada Consulting, 1536, W. High Pointe Ct., Minden, NV 89423, USA

Received 13 July 2007; received in revised form 1 November 2007; accepted 1 November 2007
Available online 17 November 2007

Abstract
This article presents an overview of the pyrite FeS2 used as cathode material in thermally activated (“thermal”) batteries. A large emphasis was
placed on the physicochemical properties and electrochemical performance of the pyrite FeS2 , including the discharge mechanisms, self-discharge
phenomena, and recent developments.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Thermal batteries; Molten salts; Cathode materials; FeS2

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 596
2. Physicochemical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 596
2.1. Basic properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 597
2.2. Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 597
2.2.1. Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 597
2.2.2. Thermal stability in inert atmosphere. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 598
2.2.3. Thermal stability in molten salts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599
2.3. Chemical stability in low-melting-point electrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
2.3.1. Halide electrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
2.3.2. Nitrate electrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
2.3.3. Organic electrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
2.4. Electrochemical behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601
2.4.1. Discharge mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601
2.4.2. Ambient temperature applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 603
2.4.3. Self-discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 603
3. Recent developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
3.1. Nanostructured FeS2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
3.2. Thermal-sprayed electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 607
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 607

∗ Corresponding author. Tel.: +49 697 564 362; fax: +49 697 564 388.
E-mail address: masset@dechema.de (P.J. Masset).

0378-7753/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jpowsour.2007.11.017
596 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609

1. Introduction - Low solubility of the cathode materials in the molten elec-

trolyte: to minimize self-discharge reactions with attendant
Thermally activated (“thermal”) batteries are mainly used loss in capacity.
for military purposes that require a high level of reliability and - Low solubility of discharge products in the molten electrolyte:
whose performance is not compromised after lengthy storage to minimize possible self-discharge reactions.
times. Applications and the electrochemistry of such power - Stable towards moisture and/or oxygen: to prevent the pro-
sources were described in detail in the first part of this review duction of oxides at the cathode surface. (This gives rise to a
dedicated to thermal batteries [1]. The properties of molten voltage peak at the beginning of discharge.)
salts (high- and low-temperature electrolytes) were thoroughly - Ability to be wetted by electrolyte: to minimize the contact
reviewed in the second part of this review [2]. The third part resistance at the electrolyte (separator)/electrode interface.
of this review of thermal batteries is devoted to the cathode - Low equivalent weigh: for higher coulombs/mole.
materials. Thermal batteries are complex chemical systems that - Good discharge kinetics (high exchange-current density): for
include electrochemical, chemical and physical properties that high-rate capability.
should be well mastered to understand the global functioning - Reasonable costs.
of these systems. To reach the high level of confidence required - Being environmentally friendly (“green”) is an additional
by such electrical generators, the physicochemical properties of desirable attribute.
the cathode materials must be well assessed and understood.
The main physicochemical properties required for the cathode This review on cathode materials is shared in two parts: this
materials to be used in thermal batteries are highlighted below: first one is dedicated to the pyrite FeS2 and the second part
concerns the other sulfides and oxide-based cathodes [160]. In
- Redox potential: it should have a discharge potential com- this paper, the properties and performances of pyrite material
patible with the electrochemical window of the electrolyte in are reviewed. Its properties were analyzed with regard to their
order to avoid its oxidation. use as cathode material in thermal batteries. The most important
- Ability to provide a fixed discharge plateau: it should undergo properties for thermal-battery applications are highlighted.
multiphase discharge and not intercalation.
- High thermal stability: to minimize thermal decomposition 2. Physicochemical properties
and associated possible chemical reactions caused by the
decomposition products (e.g., S2 in the case of FeS2 reacting The use of pyrite as a cathode in high-temperature batter-
with the anode or pyrotechnic source in the battery). These ies was first reported in a patent by Schneider and Bowser in
products can also results in increased self-discharge. 1978 [3]. In the beginning of the 1980s, the FeS2 cathode was
- Electronically conductive: to minimize the resistance of the extensively studied by Argonne National Laboratory (ANL)
cathode. for rechargeable applications [4] for replacement of FeS. At

Fig. 1. Fe–S phase diagram (from Ref. [6]).

 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609 597

Table 1
Selected values of the thermodynamic properties of FeS and FeS2 [26]
◦ ◦
T (K) Hf (kJ mol−1 ) Sf (J K−1 mol−1 ) Cp (T) (J K−1 mol−1 ) Ht (kJ mol−1 )

FeS
298 −100 60.29 −0.5 + 170.71 × 10−3 T (K) –
411 – – 72.8 2.4
598 – – 51.04 + 9.96 × 10−3 T (K) 0.5
1461 – – 50.54 32.3
FeS2
298 −178.2 52.93 68.95 + 14.1 × 10−3 T (K)–9.87 × 105 /T2 (K)

that time, pyrite was the most widely used cathode material in (1.2 eV [19], 1.14 eV [15], 1.0 eV [20], 0.92 eV [21,22], 0.92 eV
thermal batteries. Therefore the properties and performances of [23] at 300 K and 0.77 eV [24] at 550 K). It was shown that the
pyrite are covered in greater detail here. energy gap decreases with temperature [25] which enhances the
metallic character of the pyrite. This makes it perfect for use
2.1. Basic properties in thermal batteries, as it has a higher electrical conductivity at
elevated temperatures.
The thermodynamic properties of FeS2 are well established. Selected values of the thermodynamic properties of FeS and
FeS2 exists in two forms: pyrite and marcasite. Pyrite is the sta- FeS2 from Kubaschewski et al. [26] are reported in Table 1. The
bile phase of FeS2 above 700 K (see the Fe–S phase diagram heat capacity of FeS2 was measured [27–29] and recently re-
in Fig. 1). The Fe–S phase diagram has been widely investi- investigated [30]. However, some discrepancies appeared. The
gated [5,6] and recently optimized [7]. FeS2 has a cubic structure latter is approximately 10% higher in the temperature range
(group Pa3 (Th6 )) where the Fe atoms and S2 groups are located investigated (Fig. 2). The heat of vaporization of FeS2 Hvap has
on the Cl and Na positions in the NaCl-type structure, respec- been measured by several techniques. Hvap values as well as
tively. The atomic positions in the cubic structure are (u is close the techniques used are reported in Table 2. From literature data
to 0.386): [31–42] the decomposition temperature Td is given by Eq. (1).
15, 700
Fe: 0 0 0; 0 1/2 1/2; 1/2 0 1/2; 1/2 1/2 0 Td (◦ C) = − 273 (1)
[16.2 − log Pvap ]
S: +/− (u u u; u + 1/2, 1/2 − u, −u; −u, u, u + 1/2; 1/2 − u, −u,
u + 1/2) where Pvap represents the sum of the partial pressures of the
sulfur-based species Sn . The decomposition temperature is
The pyrite can deviate from the ideal stoichiometry of 2.00 lowered in presence of impurities in pyrite ores [43].
by as much as 7.5 at.% but this does not greatly affect either the
lattice perfection or the cube edge of the material [8]. 2.2. Stability
The basic properties of FeS2 vary with the ore’s origins. The
bonding in pyrite and related chalcogenides has been the sub- 2.2.1. Oxidation
ject of numerous investigations [9–13]. The band structure of Sulfates form rather easily on the pyrite surface even at low
the pyrite was determined by Ennaoui et al. [14]. Pyrite is a oxygen pressure as sketches the predominance diagram shows
good semiconductor – both n and p types being reported – with (Fig. 3). In nature, topochemical pyrite oxidation to FeSO4
electrical conductivities at room temperature ranging from 0.03 occurs readily and is depends greatly on the ambient moisture
to 333 S cm−1 [15–18]. Energy gaps range from 0.77 to 1.2 eV content of the air, being greatly accelerated above 20% relative

Table 2
Thermodynamic data relative to pyrite and marcasite decomposition
◦
Temperature range (◦ C) Hvap (kJ mol−1 FeS2 ) Reference Experimental technique

647–681 449 [32] Manometry

575–680 412 [33] Condensation temperature
590–686 383 [34] Manometry
600–690 327 [35] Manometry
324–743 304 [36]
500–552 301 [37] Manometry
436–591 297 [38] Transpiration
324–438 292 [39] Sensitive sensor
547–743 276 [40] Manometry
548–676 266 [41] Manometry
695–770 268 (marcasite to troilite) [43] Torsion–Knudsen effusion, transpiration
420–600 234 [42] Isotopic tracer
598 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609

Fig. 2. Experimental values of Cp of FeS2 versus the absolute temperature, ()

data from Ref. [29], () data from Ref. [30], (—) data from Ref. [31].

humidity [44]. Oxidation activation energy of 27.2 J mole−1 was

reported. Surface oxidation of pyrite has been reported to be
Fig. 3. Predominance diagram of the Fe–S–O system at 500 ◦ C.
dependent on the temperature and humidity as well as the par-
ticle size [45]. Experimentally, sulfate crystal growth has been
evidenced by XPS measurements on fresh pyrite surface in vac- 2.2.2. Thermal stability in inert atmosphere
uum [46]. The thermal decomposition of sulfates leads to the The thermal stability of pyrite has been widely investi-
formation of Fe2 O3 [47,48]. A large voltage transient (“spike”) gated under inert or corrosive atmosphere by non-isothermal
occurs upon activation of a thermal battery if the FeS2 contains or isothermal analyses [46,55–69,37]. The overall mechanism
impurities such as oxides, sulfates, and elemental sulfur or if the is described by Eq. (2).
activity of Li is not fixed in the cathode. This interferes with
(1 − x)FeS2(s) → Fe1−x S(s) (x = 0–0.2) + 21 (1 − 2x)S2(g) (2)
maintaining strict voltage control. This is readily remedied by
lithiation, however [49–52]. Common lithiation agents are Li2 O By isothermal analysis it was shown that the mass losses
or Li2 S, added in small quantities of 1–2 (w/o). It should be during the decomposition process were linear with time. In addi-
mentioned that pure S was also tested as anti-peak agent [53]. tion, the dependence of the kinetic constant versus temperature
The starting pyrite used in the cathode is leached with a 1:1 follows an Arrhenius-type law [68]. The activation energy var-
(v/v) solution of HCl to remove the major portion of the impuri- ied between 275 and 325 kJ mol−1 FeS2 . Literature values of
ties, leaving behind siliceous gangue, which is electrochemically the activation energy and experimental conditions are reported
inert. This gangue can also be removed by leaching with con- in Table 3. Values of the activation energy determined either
centrated HF, if a higher purity pyrite is desired or required. An by non-isothermal or by isothermal analysis were found to be
alternative, mineral beneficiation technique, such as flotation, similar [68]. During its thermal decomposition, the pyrite is
can also be used for pyrite purification [54]. progressively transformed into pyrrhotite and the FeS2 grains

Table 3
Kinetics data relative to pyrite decomposition
Temperature range (◦ C) Activation energy (kJ mol−1 ) Reference Experimental conditions

500–600 230 [46] Linear weight loss, isothermal analysis, dry He flow—50 cm3 min−1
500–600 260 [46] Linear weight loss, isothermal analysis, dry He flow—50 cm3 min−1
400–750 110 [55] Linear weight loss, isothermal analysis, vacuum and Ar atmosphere
450–690 110 [56] Linear weight loss, isothermal analysis, vacuum and N2 atmosphere
486–554 120 [57] Linear weight loss, isothermal analysis, vacuum
400–650 130 [58] Linear weight loss, isothermal analysis, air, CO2 and H2 atmosphere
390–530 140 [59] Linear weight loss, isothermal analysis, air, CO2 and mixtures
400–550 297 ± 34 [37] Linear weight loss, isothermal analysis, dry He and N2 flow 100 cm3 min−1
450–600 297 ± 15 [37] Linear weight loss, isothermal analysis flow (with 54, 108 and 1003 ppm CO)
450–600 275 ± 20 [37] Linear weight loss, isothermal analysis, CO2 atmosphere
T < 475 82 ± 9 [37] Linear weight loss, isothermal analysis, flow with 100 ppmv O2
T > 475 293 ± 52
600–653 281 [60] Linear weight loss, isothermal analysis, Ar atmosphere
427–927 286 [61] Linear weight loss, isothermal analysis, N2 atmosphere
451–476 310 [62] Linear weight loss, isothermal analysis, vacuum
594–625 – [63] Fluidized bed, linear variation of the S/Fe ratio with time
620–643 – [64] Fluidized bed, linear variation of the S/Fe ratio with time
610–750 – [65,66] Fluidized bed, linear variation of the S/Fe ratio with time
25–800 325 [67] First order-based analysis, non-isothermal analysis (v = 10 ◦ C min−1 )
 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609 599

Table 4
Expression of the solubilities of sulfur-containing species (Li2 S, FeS2 , FeS1.14 )
versus the temperature in the LiCl–KCl eutectic
ln XMx Sy (10−4 mole fractions) Reference

Li2 S 11.077 − 6.1046 × 103 /T (K) [72–77]

FeS2 10.753 − 11.882 × 103 /T (K) [71]
FeS1.14 6.4477 − 7.6622 × 103 /T (K) [71]

The pyrite as well as the pyrrhotite dissolves partially in

molten salts at high temperature. Solubility limits of FeS2 [71],
FeS1.14 [71] and Li2 S [72–77] were experimentally measured.
Analytical expressions of the solubility limits of FeS2 , FeS1.14
and Li2 S in the LiCl–KCl eutectic are reported in Table 4. In
Fig. 5, one can observed that at high temperature the solubil-
ity limits of FeS2 and FeS1.14 were equal which agrees the
progressive decomposition of FeS2 into FeS1.14 at high temper-
ature. Santarini [78] established the FeS2 –pS2 -predominance
diagram in the LiCl–KCl eutectic. It was found that the solubil-
ity limit of Li2 S was lower than in other studies. It was ascribed
to the formation of the so-called J-phase that forms between
FeS2 and the LiCl–KCl eutectic above 470 ◦ C [4,79]. The J-
phase has the djerfisherite structure and the general composition
K5.5 Li0.6 Fe24 S25.9 Cl1.0 [80,81] (Fig. 6).
As the pyrite thermally decomposes, sulfur gas is able to react
with already dissolved sulfur-based species [82] according the
following mechanism to form polysulfides [83]. Those reactions
(Eq. (4)) might enhance the dissolution process. The stability of
sulfur-based species in molten chlorides was studied by Delarue
[84] even in presence of dissolved oxide ions [85–87].
2S2− + 3S2 ⇔ 4S−
2
Fig. 4. SEM picture of decomposed FeS2 (a) grain and (b) detail. (4)
2S2− + 5S2 ⇔ 4S−
3

become porous as sulfur gas escapes (Fig. 4). The mechanism of The decomposition kinetic of FeS2 in molten salts has been
FeS2 and FeS1.14 thermal decomposition was investigated and little investigated by Barlow [88], who reported the thermal sta-
modelled by Hoare [61]. When synthetic pyrite is compared bility of pyrite in the molten LiCl–KCl and LiF–LiCl–LiBr
to natural pyrite, thermal decomposition occurs more slowly eutectics. More recently Masset et al. measured the thermal
for equivalent grain size [46]. This was attributed to the lower
intrinsic purity of the mineral material.
In contrast to pyrite, the thermal decomposition of FeS
(troilite) does not become significant until 1200 K, being only
∼0.47 kPa [42]. The vapor pressure of sulfur over troilite is given
by Eq. (3):
16, 040
ln pS2 (kPa) = 8.03 − (K) (3)
T

2.2.3. Thermal stability in molten salts

Above 580 ◦ C FeS2 thermally decomposes to form a non-
stoichiometric monosulfide (pyrrhotite) and sulfur vapor which
reacts exothermic ally with the lithium or lithium-alloy anode or
dissolved lithium in the molten electrolyte. It can also react with
the hot iron in the pyrotechnic used in the battery. This reduces
the battery capacity as well as generating more heat. This, in turn,
leads to even more thermal decomposition of FeS2 , which can Fig. 5. Evolution of the logarithm of Li2 S, FeS2 and FeS1.14 solubility in the
then destroy the battery if thermal runaway occurs. Moreover, it LiCl–KCl eutectic versus the inverse of the absolute temperature. () Li2 S data
forms solid insulating Li2 S layer in the retained electrolyte [70]. from Refs [72–77], () FeS2 from Ref. [71], () FeS1.14 from Ref. [71].
600 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609

Fig. 7. Li–Fe–S–O phase diagram [51].

our best knowledge no wetting behavior of FeS has been yet

reported.
Fig. 6. FeS2 –pS2 -predominance diagram in the LiCl–KCl eutectic [79]. One question remains unresolved: how does the molten salt
fill the empty space created by the sulfur escape and how does
decomposition kinetics in LiX–KX (X = Cl, Br, I) mixtures, it influence the wetting behavior (pyrrhotite formation at the
in the LiF–LiCl–LiI and LiF–LiCl–LiBr electrolytes [68,89]. pyrite surface). No wetting angles of FeS2 or Li3 FeS4 have been
In many cases the weight variations varied linearly with time. measured in the presence of dissolved oxides or hydroxides.
Moreover, a linear relationship was evidenced between the
kinetic constant and the radius of the anion X− [68]. The acti- 2.3. Chemical stability in low-melting-point electrolytes
vation energy of the reaction was found close to 300 kJ mol−1
FeS2 in LiX–KX mixtures (X = Cl, Br, I) whereas it was approxi- 2.3.1. Halide electrolytes
mately equal to 360 kJ mol−1 FeS2 in ternary mixtures. In molten A number of alkali halide eutectic salts have been exam-
media the kinetic constants are at least three orders of magnitude ined for potential use with pyrite cathodes for possible use in
lower than in inert gas atmosphere. The molten salt probably acts geothermal applications [2,92–97]. These batteries would not
as a physical barrier for the ready escape of the sulfur gas and use an internal pyrotechnic but would depend on the immedi-
may repress dissociation by increasing the partial pressure of ate borehole thermal environment to provide the needed energy
sulfur in the immediate vicinity of the discrete FeS2 particles. for melting of the electrolyte to allow the batteries to function.
The influence of dissolved hydroxides or oxides in the elec- Thermal batteries typically operate over a temperature range of
trolyte on the chemical stability of the pyrite has been poorly 400–550 ◦ C for standard electrolytes. In contrast, borehole tem-
investigated. The temperature of thermal decomposition of FeS2 peratures can be as low as 300–400 ◦ C. In the case of boreholes
in the LiCl–KCl eutectic was investigated by means of thermo- associated with oil and gas drilling, the maximum tempera-
gravimetric analysis (TGA) in the presence of hydroxides [30] tures are much lower—typically, only up to 250 ◦ C. This would
and oxides [30,90]. The beginning of the thermal decomposition require the use of electrolytes that are even lower melting.
is a function of the amount of hydroxides or oxides contained
in the electrolyte. According to the quaternary phase diagram 2.3.2. Nitrate electrolytes
Li–Fe–S–O (see Fig. 7) [50,91], the reaction pathway leads to There are a number of low-melting electrolytes based on
the formation of Li3 Fe2 S4 by adding Li2 O to the pyrite. The alkali nitrates that have been examined for use in high-
weight variations recorded during the TGA experiments are in temperature batteries [98]. However, molten nitrates are
good agreement with these theoretical predictions. incompatible with sulfide cathodes, oxidizing them to oxides or
The wetting properties of pyrite by molten salts may also sulfates, with the generation of NOx and SO2 and considerable
be influenced by the morphology changes during the decom- heat [99].
position step as the pyrite becomes more and more porous.
This might also modify the interface properties between the 2.3.3. Organic electrolytes
cathode material and the electrolyte. The poor wetting behav- A number of organic electrolytes have been examined
ior of the pyrite by the molten LiCl–KCl eutectic has been for use with sulfide cathodes for various battery applica-
previously reported (wetting angle: 120◦ at 500 ◦ C [30]). To tions. For organic solvents or low-melting-point electrolytes,
 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609 601

the thermal stability should be replaced by the chemical sta- The first discharge step consists in the reaction of 1.5 mole
bility, as the pyrite is stable in the temperature range of of lithium with one mole of pyrite. This reaction contrasts to
interest of these solvents. FeS2 has been successfully used the intercalation process, where electrochemical insertion of the
in room-temperature primary batteries using organic solvents lithium cation in carbon occurs (Graphite Intercalation Com-
(EC-PC-DME) [100–104]. Similar work with 1 M LiClO4 in pounds: GIC [118]). It presents the advantage that it ensures a
PC-DME and synthetic pyrite was reported by Iwakura et al. flat discharge plateau before the potential transition [119]. This
[105]. The performance of synthetic and natural pyrite in non- transition is equivalent to 1206 A s g−1 of FeS2 .
aqueous systems has also been studied [106,107]. The rate It should be mentioned that Ritchie [120] took advantage of
capabilities of the synthetic pyrite were greater than those of this reaction to produce the cathode material Li3 Fe2 S4 using the
natural pyrite because of the finer grain size of the former. spontaneous reaction of pyrite with pure LiCl and LiBr used as
Recently, Choi et al. [108] reported the use of FeS2 pyrite as solvent.
cathode material in Li/FeS2 temperature batteries using lithium-
based salts (LiTFSI) dissolved in organic solvents as electrolyte. Step 2:
(Similar cells are commercially available from Eveready [109].)
Considerable work has also been done with pyrite cathodes in (1 − x) Li3 Fe2 S4 ⇔ (1 − 2x) Li2−x Fe1−x S2 + Fe1−x S (6a)
conjunction with polymer-based electrolytes [100–104]. In Eq. (6a), x is close to 0.2.
The use of ionic liquids – the so-called room-temperature When x = 0, Eq. 6b results for the reduction of Li3 Fe2 S4 .
molten salts – has also been explored as one low-temperature
medium for use with the Li–Si/FeS2 couple. This type of elec- Li3 Fe2 S4 + Li+ + e− → Li2 FeS2 + FeS + Li2 S (6b)
trolyte has an intrinsic high-temperature stability that would
allow its use for certain borehole applications. Being liquid Step 3:
at room temperature would allow its use from ambient to the
temperatures in the borehole. While it is compatible with the Li2−x Fe1−x S2 → Li2 FeS2 (“X-phase ) (7)
cathode, such material are not compatible with high-activity Step 4:
anodes such as Li–Si, Li–Al, Ca, or Mg at temperatures above
100 ◦ C [110]. Li2 FeS2 + 2e− → Li2 S + Fe + S2− (8)
More recently, another category of electrolyte has been
examined, the tetraalkyl ammonium salts. The tetraalky- Li–Si1 /FeS2 thermal batteries are designed to use only the
lammonium salts have good thermal stability and a high first cathode transition (Eq. 5) because of rigid voltage require-
voltage-stability window. One such salt, tetramethylammonium ments associated with the use of such batteries.
bis(trifluoromethylsulfonyl)imide (TMAIm) has been studied The species actually undergoing reduction in Eq. (5) is the
with the Li–Si/FeS2 couple [111]. While it did function as polysulfide, S2 −2 , i.e., the oxidation state of Fe in FeS2 is +2
expected, it suffered from one major problem: a very low and not +4 [21,22]. This is illustrated in Eq. (9):
ionic conductivity – only 85 mS cm−1 at 250 ◦ C, which is more
than an order of magnitude less than some of the traditional S2 −2 + 2 e− → 2 S−2 (9)
thermal-battery electrolytes – which resulted in unacceptable
FeS2 begins to thermally decompose at temperatures above
polarization losses. This greatly limits its possible use.
550 ◦ C to form a non-stoichiometric monosulfide (pyrrhotite)
and sulfur vapor, as shown in Eq. (2):
2.4. Electrochemical behavior Any fugitive sulfur can react very exothermically with the
Li-alloy anodes in the battery. This reduces the battery capacity
2.4.1. Discharge mechanism as well as generating more heat. This, in turn, leads to even more
The discharge reactions that occur when a FeS2 cathode is thermal decomposition of FeS2 , which can destroy the battery
used in high-temperature batteries have been extensively stud- if thermal runaway occurs.
ied by ANL for rechargeable applications. A number of other The open-circuit potentials for a number of the major dis-
researchers had studied the various electrode processes asso- charge reactions are summarized in Table 5. Note that the first
ciated with pyrite but there was some disagreement as to the two discharge steps experience entropy changes that result in
exact nature of the sequence of discharge phases in molten salts cell cooling, while the discharge of the Li2 FeS2 phase results
[112,113]. Schmidt, for example, postulated the formation of an in cell heating [121]. These changes are very important for
“Li2 S2 ” metastable intermediate [114,115]. However, the multi- proper thermal management of high-temperature rechargeable
step discharge mechanism put forth by ANL is felt to be the most batteries.
accepted [4,116,117]. The four discharge sequences of FeS2 in The electrolyte used in pyrite-based cells can impact the
molten LiCl–KCl are described in Eqs. (5)–(8): discharge processes in a number of ways. High levels of
K+ can lead to increased formation of the J-phase material
Step 1:
1 All reference to Li–Si in this paper is for the composition 44 (w/o) Li/56
FeS2 + 23 Li+ + 23 e− → 21 Li3 Fe2 S4 (“Z-phase”) (5) (w/o) Si.
602 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609

Table 5 Table 6
Open-circuit potentials for several of the discharge steps involving pyrite at Electrical conductivities of the phases associated with the discharge of pyrite
400 ◦ C (from Ref. [116]) cathodes
Discharge reaction emf at 400 ◦ C Entropy effects Phase Temperature (◦ C) Conductivity (S cm−1 )
versus Li–Al (V)
FeS2 400 80–100 [4,25]
FeS2 → Li3 Fe2 S4 1.750 Cooling Li3 Fe2 S4 (Z-phase) 400 ∼0.1 [123]
Li3 Fe2 S4 → Li2+x Fe1−x S2 + Fe1−y S 1.645 Cooling Li2 FeS2 (X-phase) 400 4.2 [123]
Li2 FeS2 → Fe + Li2 S 1.261 Heating [121] FeS2 500 80–100 [14,26]
Li3 Fe2 S4 (Z-phase) 500 ∼0.3 (extrapolated) [123]
Li2 FeS2 (X-phase) 500 ∼6.3 [123]
(K5.5 Li0.6 Fe24 S25.9 Cl1.0 ). However, this phase appears not to
be stable above 500 ◦ C. The generation of a large Li+ flux at the
anode–separator interface under high-rate discharge can lead to phases are semiconductors, which is ideal for thermal-battery
very high concentration gradients, with the associated localized applications.
increase in melting point of the electrolyte. This is illustrated in The effects of the conductivity changes in the cathode are
Fig. 8 (from Ref. [122]) for a Li–Si/LiCl–KCl/FeS2 cell. The readily apparent with the use of a suitable reversible reference
increase in Li+ concentration at the anode is accompanied by a electrode. For short-time runs, an Ag/AgCl (0.1 M) reference
corresponding increase in the K+ concentration at the cathode. contained in a glass capillary tube sealed at one end works quite
The precipitation of solids reduces the volume of free electrolyte well [124]. However, after several hours, the reference deteri-
and increases the resistance of the separator, leading to a sub- orates and emf becomes unstable. An aluminum wire can be
stantial increase in the overall impedance of the battery. This charged with Li to form a two-phase region consisting of ␣-
is most important near the end of life of the battery when its Al + ␤-LiAl. This provides a very stable emf with long discharge
temperature is lowest. times. ANL developed a Ni/Ni3 S2 reference that works very
The conductivity of the discharge phases also influences the well for this purpose [125–127]. The area-specific impedances
performance of the cell. While the conductivity of the pyrite of typical anode and cathode for a Li–Si/LiCl–KCl/FeS2 cell
phase is very good, that of the first discharge phase, Li3 Fe2 S4 , discharged at 450 ◦ C and 50 mA cm−2 are shown in Fig. 9.
is much lower, while that of subsequent discharge phase is inter- As can readily be seen, the change in the anode impedance
mediate. Data taken from Ref. [123] for temperatures of 400 and is insignificant with depth of discharge, while there is a dra-
500 ◦ C are presented in Table 6. Both the starting and discharge matic increase for the cathode. This is similar to what was
reported by Wang [128]. Using rotating disc voltammetry, he
found that the bulk of the total cell polarization (∼75%) is asso-
ciated with the porous cathode. The bulk of the observed increase
in impedance is due to the formation of the less-conductive Z-
phase at the start of discharge. There is some contribution as
well from KCl precipitation (see Fig. 8) and J-phase formation.
At normal discharge rates associated with thermal batteries, the
Z-phase begins to discharge to form the X-phase before all of the

Fig. 8. Variation in atomic K/Cl ratio in Li–Si/LiCl–KCl/FeS2 single cell dis- Fig. 9. Area-specific impedances of half cells of a Li–Si (25%
charged at 500 ◦ C and 50 mA cm−2 [122]. (The nominal K/Cl ratio is 0.418 for electrolyte)/LiCl–KCl/FeS2 cell discharged at 450 ◦ C and 50 mA cm−2
the eutectic composition.) [122].
 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609 603

Fig. 11. Schematic representation of self-discharge reactions in

Li–Si/LiCl–KCl/FeS2 cells.

and elemental Li. Wang and Seefurth also hypothesized the pres-
ence of dissolved Fe compounds in the electrolyte that resulted
Fig. 10. Photomicrograph of FeS2 phase (white) and Li3 Fe2 S4 phase (light in loss of capacity of similar cells [113]. The discharge mecha-
gray) (650×) formed during discharge of Li–Si/LiCl–KCl/FeS2 cell at 400 ◦ C nism is illustrated schematically in Fig. 11. The reacting solution
and 50 mA cm−2 [122]. species migrate towards each other and react in the separator
matrix. The band of reaction product appears orange colored
initial pyrite phase is consumed. The discrete discharge phases under polarized light. (However, when Li2 FeS2 is substituted
are readily evident in the photomicrograph of Fig. 10. There is for FeS2 , no band formation is observed [122] because of the
a large volumetric change in the cathode during discharge due reduced solubility of this FeS-containing phase.) The rate of self-
to the lower density of the Z-phase. This can lead to mechani- discharge is very dependent upon temperature, Li activity of the
cal issues for secondary high-temperature batteries using pyrite anode, and electrolyte composition [131]. Data for the LiCl–KCl
cathodes. eutectic electrolyte are presented in Table 7 [132]. These data
are in agreement with a study by Burrow et al. where discharged
2.4.2. Ambient temperature applications Li–Si/FeS2 cells were cross-sectioned and the anode–separator
When FeS2 is discharged at temperatures of 200 ◦ C or more, and cathode–separator interfaces were examined by scanning
the reaction is quite reversible. However, when Li/FeS2 cells are electron microscopy (SEM), energy dispersive X-ray analysis
discharged at ambient temperatures using organic electrolytes, (EDX), and Auger electron spectroscopy (AES) for elemental
the discharge reaction is not completely reversible. The Z-phase distribution [133].
which is the first discharge product of pyrite cathodes in molten This proposed discharge mechanism is supported by the iden-
salts does not form in nonaqueous systems. Instead, the X-phase tification of elemental Fe and Li2 S crystals in the center of
is observed [109,105]. The reversibility in nonaqueous systems the separator of discharged Li–Si/FeS2 cells [122,134,70]. The
was also studied by Fong et al. [129,130]. amount of Li2 S that forms is found to be related to the activity of
the anode used, with high activity (e.g., pure Li) associated with
2.4.3. Self-discharge greater levels of Li2 S deposition. The Li2 S level also increases
The decrease in battery efficiency comes mainly from with increase in the Li+ content of the electrolyte. The reaction
self-discharge reactions. In this system where complex elec-
trochemical and chemical reactions take place, the knowledge Table 7
and the understanding of the source of these unwanted self- Rate of self-discharge for pure Li in LiCl–KCl eutectic electrolyte [132]
discharge reactions are of great importance. Both the pyrite and
Temperature (◦ C) Self-discharge rate (mA cm−2 )
Li–Si have finite solubilities in the molten salts. (In the case
of pure Li, a solubility of 0.038 mol l−1 has been reported in 395 1.0
LiCl–KCl eutectic at 400 ◦ C [2].) This leads to the creating of a 415 1.4
436 1.9
flux of migrating dissolved species, such as Fe+2 , polysulfides,
604 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609

Fig. 12. Equilibria among various species for FeS2 in contact with molten
electrolyte.

band tends to move closer to the cathode with increasing Li

activity of the anode [70].
The equilibria between the various species in the case of
pyrite in contact with molten electrolyte are summarized in
Fig. 12. There is competition between decomposition of FeS2 Fig. 13. Self-discharge losses for Li–Si (25% electrolyte)/FeS2 cells at 500 ◦ C
to FeS and sulfur gas, which can dissolve in the electrolyte. In as a function of time on open circuit prior to discharge at 125 mA cm−2 for
addition, the polysulfide will be in equilibrium with sulfide and LiCl–KCl and LiCl–LiBr–LiF electrolytes.
dissolved sulfur. Once the solubility limit of FeS is reached, it
can precipitate out of solution. The relative importance of these increased with increasing temperatures and with decreasing cur-
associated reactions is not known at this time. In the presence of rent densities, with open circuit being the worst condition [131].
significant K+ concentrations, formation of the J-phase can also Significant loss in capacity of Li–Si/FeS2 cells was observed
occur at temperatures above 470 ◦ C. when the cells were place on open circuit for up to 2 h. While
The composition of the electrolyte has a major influence on substantial loss was observed for the LiCl–KCl eutectic, the loss
the solubility of pyrite and self-discharge of Li–Si/FeS2 cells was greatest for the all-Li (K+ -free) LiCl–LiBr–LiF electrolyte.
(Table 8). The concentration of Li+ appears to be responsible for The response at 500 ◦ C is shown in Fig. 13. In the former case,
this behavior. With a K-rich LiCl–KCl electrolyte, a discharge the loss in capacity was almost 12% after 30 min but did not
loss rate of 0.001% h−1 was reported at 450 ◦ C and 88 mA cm−2 increase much more after 2 h. For the all-Li electrolyte, how-
[113]. This increased to 0.013% h−1 for the stoichiometric ever, the corresponding value was over 38%, for 30 min open
composition. With a K-free electrolyte of LiCl–LiBr–LiF, the circuit. After 2 h, almost all of the capacity has been exhausted
self-discharge rate increased to 0.07% h−1 . Dissolution of FeS2 (>93%). Fortunately, such conditions are not encountered dur-
has also been reported even in low-melting, tetrachloroaluminate ing normal thermal-battery operation and are more important
melts. Solubilities of 6.8 × 10−4 and 2.0 × 10−3 mole fractions for secondary applications.
at 170◦ and 270 ◦ C, respectively, were reported [135]. When FeS2 is equilibrated with various molten halide eutectic
Similar capacity-loss behavior has been observed at Sandia mixtures, an amber-colored solution is formed. (Similar colored
National Laboratories (SNL). Self-discharge of Li–Si/FeS2 cells solutions were observed in the electrochemical study of sulfur

Table 8
Self-discharge rate of FeS2 cathode (open circuit) [131]
Electrolyte Catholyte treatment T (◦ C) Capacity losses (% min−1 ) (r2 for LSF)

450 0.9889
Fused, unlithiated
550 0.450 0.98703
LiCl–KCl
450 0.0915 0.9942
Fused, lithiated
550 0.438 0.9997
450 0.130 0.9778
Fused, unlithiated
550 0.582 0.9534
LiBr–KBr–LiCl (LM#1)
450 0.0998 0.5396
Fused, lithiated
550 0.330 0.95029
450 0.344 0.99965
Fused, unlithiated
550 0.751 0.98379
LiBr–KBr–LiF (LM#2)
450 0.203 0.99467
Fused, lithiated
550 0.354 0.99467
450 0.374 0.98507
Fused, unlithiated
550 0.722 0.97499
LiCl–LiBr–LiF (all-Li)
450 0.448 0.92364
Fused, lithiated
550 0.716 0.88492
 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609 605

in the molten salt. This has been reported for a number of alkali
metal–alkali halide (M–MX) systems [138–142]. Reynolds et
al. used the Wagner polarization technique to study the contri-
bution to the electronic conductivity of LiCl–KCl electrolyte as
a function of Li activity of the anode [143]. Dissolved Li can be
represented by Eq. (12).
◦
Li(soln) → Li+ + e− (12)

They reported an electronic conductivity of 0.18 S cm−1 for

pure Li (unit activity) at 465 ◦ C and 8 × 10−5 S cm−1 at 383 ◦ C
for Li0.75 Al alloy (0.056 Li activity).

3. Recent developments

3.1. Nanostructured FeS2

Fig. 14. Soluble S and Fe as a function of temperature for a solution of LiCl–KCl
eutectic equilibrated with FeS2 for 1 h.
Nanostructured FeS2 prepared by a high-energy grinding
process and used as the cathode material in LiSi/FeS2 ther-
and sulfide in LiCl–KCl eutectic and were attributed to com-
mal batteries resulted in improved electrochemical performance
plexed sulfur species [136].) Analysis for Fe and S of filtered
[144]. With the same weight, the nanostructured cathode pellets
and quenched aliquots of these solutions show that the dissolved
were reported to be 23% thinner than conventional counterparts
S increases sharply with temperature. However, the soluble Fe
resulting in 31% increase of pellet density. With the nanostruc-
content changes little and, in some cases, actually decreases
ture, the electrode materials react more rapidly and completely.
over the same temperature interval. This is shown in Fig. 14
The nanostructured cathode pellets are more mechanically
for the LiCl–KCl eutectic. These data indicate that Fe+2 is being
robust than the conventional materials, which leads to poten-
removed by precipitation of FeS (or Lix FeSy species) by the reac-
tial high productivity and lower cost in battery manufacturing.
tions shown in Fig. 12. As a result, the atomic S/Fe ratio is not
Only data for single-cell tests at one temperature were reported.
fixed at 2 but increases rapidly with temperature above 500 ◦ C.
These limited data are in agreement with more comprehensive
Similar results were obtained with other electrolytes, with a S/Fe
data generated over a wide range of discharge conditions for
atomic ratio of 13 at 550 ◦ C for the LiCl–LiBr–LiF electrolyte.
pyrite and other disulfides synthesized by an aqueous method
The trend in solubility of FeS2 in molten salts with temperature
[145]. Using nanostructured materials, it is expected that ther-
and electrolyte composition mirrors that observed for Li2 S. The
mal batteries can be prepared that are more compact and robust
data of Table 9 show that the solubility increases considerably
with higher energy density.
with temperature with the all-Li electrolyte showing the greatest
solubility [137].
These solubility data are for FeS2 in the absence of the Li- 3.2. Thermal-sprayed electrode
alloy anodes. In a complete cell, the reaction of the migrating
soluble Li◦ species with the soluble species arising from the All thermal-battery electrodes today are formed by press-
FeS2 will also be taking place and is reflected in the self- ing powder mixes into pellets that are then stacked to construct
discharge behavior of such cells. Under moderate rate (e.g., the battery. This requires increasing larger – and expensive –
400–500 mA cm−2 ) discharge, the self-discharge reactions will presses as the pellet diameters increase. (The forming pressure
not generally be significant, as the electrochemical reactions will will increase as the square of the pellet diameter.) The large
dominate the chemical reactions that can occur in parallel during inventory of pressing dies that are necessary adds to the equip-
discharge. ment costs. Several years ago, the concept of plasma spraying
The dissolution of elemental Li from the anodes adds further of thermal-battery electrodes was evaluated [146–156]. This
complicates matters in that it gives rise to electronic conductivity approach involves using a thermal-spray process to deposit thin-
film electrodes onto a graphite-paper or stainless steel current
Table 9 collector. It was demonstrated that both FeS2 [154,156] and
Solubility of Li2 S in various molten salts [137] CoS2 [156] cathodes could be formed in this fashion.
Electrolyte composition (m/o) Temperature (◦ C) Li2 S solubility (ppm)
The main advantages of this type of cathode are:

65LiCl/35KCl 400 1100

55LiCl/45KCl (eutectic) 400 650 - better mixing of the electrolyte and pyrite in the electrode
66LiCl/34KCl 500 1800 layer;
54LiCl–46KCl 500 840 - decrease of the cathode thickness (volume gain, since one can
22LiF–31LiCl–47LiBr 466 6840 deposit only what is necessary and not what must be used
22LiF–31LiCl–47LiBr 500 8700
because of pellet mechanical issues);
606 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609

Fig. 15. Scanning electron microphotograph of plasma-sprayed pyrite cathode

co-sprayed with LiCl–LiBr–LiF electrolyte. (Bottom layer is graphite-paper
current collector.)

- reduction of the ohmic losses in the cathode because of better

particle–particle contact;
- better wetting behavior of the cathode particles by the elec-
trolyte;
- better contact (bonding) with the current collector;
- no need for expensive dies; inexpensive “cookie cutter”
dies can punch any size and shape from a painted-cathode Fig. 16. Cross-sectional view of a Li–Si/LiCl–KCl/FeS2 discharged cell. (The
separator layer shows the colored deposition band noted earlier. The cathode is
sheet.
at the top of the picture; the separator layer is ∼1 mm thick.)

However the thermal-sprayed cathode process presents sev-

sprayed Li–Si anodes [158]. This was the first report of batteries
eral drawbacks:
made with all-plasma-sprayed electrodes.
More recently, work is underway studying the use of painted
- high cost of the process as the equipment is very expensive;
(sprayed) electrodes for thermal batteries [159]. This provides
- loss of capacity for very thin electrodes due to self-discharge;
all of the advantages of plasma spraying without the disadvan-
- the need to conduct the spraying under an inert atmosphere
tages. Preliminary data for Li–Si/LiCl–KCl/FeS2 cells made
cover;
- presence of free sulfur when it is used as a thermal-barrier
coating. This requires leaching of the sulfur from the electrode
afterwards with CS2 . This issue was resolved by substitution
of electrolyte as a barrier coating [156,146];
- a change in composition of the cathode from the starting feed-
stock material caused by large FeS2 particle bouncing off the
substrate, resulting in lower FeS2 content;
- difficulty in controlling the density of the deposit. (The porous
nature of the deposit is evident in Fig. 15.) A cross-sectional
photomicrograph of a cell made with pressed-powder parts of
nominal 25% porosity is shown in Fig. 16 (note the red color
of the deposition band in the separator);
- it is a batch process, which makes it more difficult to commer-
cialize.

Composite separator–cathode deposits were also prepared

in the same manner by sequential thermal spraying of
LiCl–KCl-based separator material onto a pyrite-cathode sub- Fig. 17. Discharge at 500 ◦ C and 125 mA cm−2 of cathode-limited, Li–Si
strate [150,157]. Both single cells and batteries were tested using (25% electrolyte)/LiCl–KCl/FeS2 cells made with plasma-sprayed, painted, and
the two-layered, plasma-sprayed composites along with plasma- pressed-powder cathodes.
 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609 607

with painted cathodes is presented in Fig. 17 along with com- [30] P. Masset, Doctoral Thesis, National Polytechnic Institute of Grenoble,
parable data for cells made with plasma-sprayed cathodes and 2002.
[31] F.W. Dickson, L.D. Shields, G.C. Kennedy, Econ. Geol. 57 (7) (1962)
pressed-powder cathodes. The lack of the large “humps” in the
1021.
polarization traces for the cells with the sprayed and painted [32] E.T. Allen, R.H. Lombard, Am. J. Sci. 43 (255) (1917) 173.
electrodes for the cathode voltage transitions is attributed to bet- [33] M.G. Reader, Kgl. Norske Videnskabers Selskab 2 (43) (1929) 151.
ter particle–particle contact. Work is continuing to extend this [34] F. de Rudder, Bull. Soc. Chim. Fr. 47 (1936) 1225.
approach to include the separator and anode, as well, with the [35] P. Toulmin III, P.B. Barton Jr., Geochim. Cosmochim. Acta 28 (1964)
goal of producing a three-layered, composite—a complete cell. 641.
[36] S. Bog, T. Rosenqvist, Trans. Faraday Soc. 55 (1959) 1555.
This will greatly reduce the number of piece parts needed for [37] Y. Hong, B. Fegley Jr., Ber. Bunsenges. Phys. Chem. 101 (12) (1997)
battery assembly and will reduce the production costs of thermal 1870.
batteries, if successful. [38] E.P. Schneeberg, Econ. Geol. 68 (1973) 507.
[39] H. Rau, J. Phys. Chem. Solids 37 (1976) 425.
[40] D’Or, J. Chim. Phys. 28 (1931) 378.
4. Conclusions [41] B.J. Giletti, R.A. Yund, T.H. Lin, Econ. Geol. 57 (1968) 702.
[42] D. Ferro, V. Piacente, P. Scardala, J. Chem. Thermodyn. 21 (1989)
In this paper, we have presented an overview of the use of 483.
[43] C.J. Johnson, J.L. Kavandi, Proc. ISE (Abstract 102) (1992) 145.
pyrite (FeS2 ) as a cathode material for use in thermal batter-
[44] J.E. Pahlman, G.W. Reimers, US Bureau of Mine Report RI9059 (1986).
ies. The physical and chemical properties and the discharge [45] A.N. Spektor, A.D. Markov, A.N. Pyrikov, F.S. Umanskii, Zh. Prikladnoi
sequences and reaction mechanisms are described including Kh. 45 (8) (1972) 1713.
self-discharge processes. [46] J.P. Pemsler, R.K.F. Lam, J.K. Litchfield, S. Dallek, B.F. Larrick, B.C.
Beard, J. Electrochem. Soc. 137 (1) (1990) 1.
[47] P.S. Thomas, D. Hirschauen, R.E. White, J.P. Guerbois, A.S. Ray, J.
References Therm. Anal. Cal. 72 (2003) 769.
[48] P. Masset, J.-Y. Poinso, J.-C. Poignet, J. Therm. Anal. Cal. 83 (2) (2006)
[1] R.A. Guidotti, P. Masset, J. Power Sources 161 (2) (2006) 1443. 457.
[2] P. Masset, R.A. Guidotti, J. Power Sources 164 (1) (2006) 397. [49] R.A. Guidotti, US Patent 4,731,307 (March 15, 1988).
[3] A.A. Schneider, G.C. Bowser, US Patent 4,119,769 (October 10, 1978). [50] T.L. Aselage, E.E. Hellstrom, J. Electrochem. Soc. 134 (8) (1987) 1932.
[4] Z. Tomzcuk, B. Tani, N.C. Otto, M.F. Roche, D.R. Vissers, J. Electrochem. [51] R.A. Guidotti, F.W. Reinhardt, W.F. Hammetter, Sandia Report SAND85-
Soc. 129 (5) (1982) 925. 1737 (1988).
[5] R.C. Sharma, Y.A. Chan, Metall. Trans. B 10B (1979) 103. [52] R.A. Guidotti, F.W. Reinhardt, Proceedings of the 2nd Annual Battery
[6] W. Burgmann, G. Urbain, M.G. Frohberg, Mem. Sci. Rev. Metall. 65 Conference on Applications and Advances, 1987, p. 229.
(7–8) (1968) 567. [53] K. Grjotheim, B. Haugsdal, H. Kvande, H.G. Nebell, T.A. Utigard, J.
[7] P. Waldner, A.D. Pelton, Metall. Mater. Trans. B, Process Metall. Mater. Electrochem. Soc. 135 (1) (1988) 51.
Process. Sci. 35 (5) (2004) 897. [54] R.A. Guidotti, F.W. Reinhardt, Sandia Report SAND92-0874 (1992).
[8] S. Fiechter, M. Birkholz, A. Hartmann, P. Dulski, M. Giersig, T. Tributsch, [55] V.S. Udintseva, G.I. Chuforov, J. Appl. Chem. USSR 14 (1941) 3.
R.J.D. Tilley, J. Mater. Res. 7 (7) (1992) 1829. [56] V.M. Zhukovskii, I.A. Montillo, A.A. Babadzhan, Tr. Ural. Nauch. Issled.
[9] J.A. Tossell, D.J. Vaughan, J.K. Burdett, Phys. Chem. Miner. 7 (1981) Proek. Inst. Med. Prom. 8 (1967) 387.
177. [57] G.I. Samal, Geterogennye Khim. Reaktsii, Inst. Obshch. I Neorgan. Khim.
[10] G.L. Zhao, J. Callaway, M. Hayashibara, Phys. Rev. B 48 (21) (1994) 48. Akad. Nauk Belorussk. SSR (1965) 93.
[11] S.L. Finklea III, L. Cathey, E.L. Amma, Acta Crystallogr. A32 (1976) [58] G.M. Schwab, J. Philinis, J. Am. Chem. Soc. 69 (1947) 2588.
529. [59] B. Fegley Jr., K. Lodders, A.H. Treiman, G. Kleingelhofer, Icarus 115
[12] J.A. Morice, L.V.C. Rees, D.T. Rickard, J. Inorg. Nucl. Chem. 31 (1969) (1995) 159.
3797. [60] A.W. Coats, N.F.H. Bright, Can. J. Chem. 44 (1966) 1191.
[13] L. Pauling, Can. Mineral. 16 (3) (1978) 447. [61] I.C. Hoare, H.J. Hurst, W.I. Stuart, T.J. White, J. Chem. Soc. Faraday
[14] A. Ennaoui, S. Fiechter, W. Jaergermann, H. Tribusch, J. Electrochem. Trans. 1 (84(9)) (1988) 3071.
Soc. 133 (1) (1986) 97. [62] G. Pannetier, L. Davignon, Bull. Soc. Chim. 72 (1961) 1513.
[15] A. Sasaki, Mineral. J. 1 (5) (1955) 290. [63] J.L.F. Monteiro, Can. J. Chem. Eng. 59 (1981) 511.
[16] K.K. Mishra, K. Osseo-Asare, J. Electrochem. Soc. 135 (10) (1988) 2502. [64] S. Yamazaki, Z. Asaki, Y. Kondo, Trans. TMS-AIME 242 (1968) 896.
[17] S.B. Lalvani, A. Weston, J.T. Masden, J. Mater. Sci. 25 (1A) (1990) 107. [65] M.D. Gibbs, T.N. Smith, B. Verbaan, Trans. Inst. Min. Metall. Section C
[18] H.D. Banerjee, N. Godhaunkar, Mater. Lett. 10 (3) (1990) 99. 106 (1997) 69.
[19] J.C. Marinace, Phys. Rev. 96 (1954) 593. [66] M.D. Gibbs, T.N. Smith, B. Verbaan, Trans. Inst. Min. Metall. Section C
[20] R. Otsuka, Synopsis, Graduate School of Science and Engineering, 106 (1997) 74.
Waseda University, Japan, 1957, p. 6. [67] A.G. Ritchie, P. Carter, Proceedings of the 38th Power Sources Confer-
[21] S.L. Finklea, L. Cathey, Acta Crystallogr. A32 (1976) 529. ence, 1998, p. 215.
[22] V.P. Gupta, N.M. Ravindra, V.K. Srivastava, J. Phys. Chem. Solids 41 [68] P. Masset, V. Frotté, J.-Y. Poinso, J.-C. Poignet, J. Power Sources, to be
(1980) 145. submitted.
[23] T.A. Bither, Inorg. Chem. 7 (1968). [69] S. Dallek, B.F. Larrick, Thermochim. Acta 95 (1985) 139.
[24] H. Horita, T. Suzuki, Sci. Rep., Tohoku Univ. A-25 (4–5) (1975) 124. [70] M.C. Hash, J.A. Smaga, R.A. Guidotti, F.W. Reinhardt, Proceedings of
[25] R. Schieck, A. Hartmann, S. Fiechter, R. Konenkamp, H. Wetzel, Mater. the 8th International Symposium on Molten Salts, 1992, p. 228.
Res. 5 (7) (1990) 1567. [71] R.A. Sharma, R.N. Seefurth, J. Electrochem. Soc. 131 (5) (1984) 1084.
[26] O. Kubaschewski, C.B. Alcock, P.J. Spencer, Materials Thermochemistry, [72] C.H. Liu, A.J. Zielen, D.M. Gruen, J. Electrochem. Soc. 120 (1) (1973)
sixth ed., Pergamon Press, 1997. 14.
[27] S.C. Mraw, D.F. Naas, J. Chem. Thermodyn. 11 (1979) 567. [73] J. Phillips, H.F. Gibbard, Proceedings of the Second International Sym-
[28] F. Gronvold, E.F. Westrum Jr., J. Inorg. Chem. 1 (1961) 36. posium on Molten Salts, The Electrochemical Society, Pennington, NJ,
[29] F. Gronvold, E.F. Westrum Jr., J. Chem. Thermodyn. 8 (1976) 1039. 81-10, p. 45.
608 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609

[74] D. Warin, Z. Tomczuk, D.R. Vissers, J. Electrochem. Soc. 130 (1) (1983) [113] S.S. Wang, R.N. Seefurth, J. Electrochem. Soc. 134 (3) (1987) 530.
64. [114] J.A. Schmidt, An. Asoc. Quim. Argent. 74 (2) (1986) 149.
[75] S. Sharma, J. Electrochem. Soc. 133 (5) (1986) 859. [115] J.A. Schmidt, An. Asoc. Quim. Argent. 75 (3) (1987) 359.
[76] M.L. Saboungi, J.J. Marr, M. Blander, J. Electrochem. Soc. 125 (10) [116] Z. Tomczuk, D.R. Vissers, J. Electrochem. Soc. 133 (12) (1986) 2505.
(1978) 1567. [117] S.K. Preto, Z. Tomczuk, S. von Winbush, M.F. Roche, J. Electrochem.
[77] M.L. Saboungi, J.J. Marr, M. Blander, Metall. Trans. 10B (1979) 477. Soc. 130 (2) (1983) 264.
[78] G. Santarini, Electrochim. Acta 27 (4) (1982) 495. [118] D. Guerard, A. Herold, Carbon 3 (4) (1975) 337.
[79] Z. Tomczuk, S.K. Preto, M.F. Roche, J. Electrochem. Soc. 128 (4) (1981) [119] M. Armand, University of Grenoble, 1978 (in French).
760. [120] A. Ritchie, Proceedings of the 18th International Power Sources Sympo-
[80] F.C. Mrazek, J.E. Battles, J. Electrochem. Soc. 124 (10) (1977) 1556. sium, 1993, p. 299.
[81] B. Tani, F. Mrazek, J. Faber, R. Hitterman, J. Electrochem. Soc. 133 (12) [121] D.M. Chen, H.F. Gibbard, J. Electrochem. Soc. 130 (10) (1983) 1975.
(1986) 2644. [122] L. Redey, J.A. Smaga, J.E. Battles, R. Guidotti, Argonne National Labo-
[82] F.G. Bodewig, J.A. PLambeck, J. Electrochem. Soc. 117 (7) (1970) 904. ratory Report ANL-87-6 (1987).
[83] D.M. Gruen, R.L. McBeth, A.J. Zielen, J. Am. Ceram. Soc. 93 (24) (1971) [123] S.P.S. Badwal, R.J. Thorn, J. Sol. St. Chem. 43 (1982) 163.
6691. [124] R.A. Guidotti, unpublished data.
[84] G. Delarue, Bull. Soc. Chim. Fr. (1969) 906. [125] L. Redey, D.R. Vissers, J. Electrochem. Soc. 128 (12) (1981) 2703.
[85] G. Delarue, J. Electroanal. Chem. 1 (1959) 43. [126] Z. Tomczuk, L. Redey, D.R. Vissers, J. Electrochem. Soc. 130 (5) (1983)
[86] G. Delarue, J. Electroanal. Chem. 1 (1960) 285. 1074.
[87] G. Delarue, Bull. Soc. Chim. Fr. (1960) 1654. [127] L. Redey, D.R. Vissers, J. Electrochem. Soc. 130 (1) (1983) 231.
[88] G. Barlow, Investigation of Mechanism for Capacity Loss in Iron [128] S.S. Wang, J. Electrochem. Soc. 135 (7) (1988) 1675.
Disulfide Electrodes in High-Temperature Lithium-Alloy/Metal Sulfide [129] R. Fong, J.R. Dahn, C.H.W. Jones, J. Electrochem. Soc. 136 (11) (1989)
Batteries, Final Report to Lawrence Berkeley Laboratory, December 3206.
1986. [130] F. Fong, C.H.W. Jones, J.R. Dahn, J. Power Sources 26 (3–4) (1989) 333.
[89] P. Masset, J.-Y. Poinso, S. Schoeffert, J.-C. Poignet, Proceedings of the [131] R.A. Guidotti, F.W. Reinhardt, J.A. Smaga, Proceeding of the 34th Inter-
39th International Power Sources Conference, 2002, p. 246. national Power Sources Symposium, 1990, p. 132.
[90] P. Masset, Z. Naturforsch., in press. [132] L. Redey, Argonne National Laboratory, unpublished data.
[91] T.L. Aselage, E.E. Hellstrom, J. Electrochem. Soc. 134 (8) (1987) 1929. [133] B.J. Burrow, K.W. Nebesny, N.R. Armstrong, R.K. Quinn, D.E. Zurawski,
[92] R.A. Guidotti, F.W. Reinhardt, J. Odinek, J. Power Sources 136 (2) (2004) J. Electrochem. Soc. 128 (9) (1981) 1919.
257. [134] Argonne National Laboratory Progress Report for October
[93] R.A. Guidotti, R.A. Normann, F.W. Reinhardt, J. Odinek, Trans. Geother- 1980–September 1981, ANL-81-65 (1981).
mal Resour. Council 27 (2003) 173. [135] N. Takami, N. Koura, Proceedings of the Stationary Energy Storage Load
[94] R.A. Guidotti, F.W. Reinhardt, Proceedings of The Electrochemical Soci- Leveling and Remote Applications, PV 88-11, The Electrochemical Soci-
ety, vol. 2002-19 (Molten Salts XIII), 2002, p. 63. ety, 1988, p. 106.
[95] R.A. Guidotti, Proceedings of the 35th IECEC Meeting, vol. 2, 2000, p. [136] M.J. Weaver, D. Inman, Electrochim. Acta 20 (1975) 929.
1276. [137] Z. Tomzcuk, D.R. Vissers, J.L. Saboungi, Proceedings of the Fourth Inter-
[96] R.A. Guidotti, F.W. Reinhardt, Proceedings of the International Sympo- national Symposium on Molten Salts, PV 84-2, The Electrochemical
sium on Molten Salts, XII, vol. 99-41, 2000, p. 701. Society, 1984, p. 352.
[97] R.A. Guidotti, F.W. Reinhardt, Proceedings of the 11th International [138] J.J. Warren Jr., in: G. Mamamtov, R. Marassi (Eds.), Molten Salts Chem-
Symposium on Molten Salts, 1998, p. 316. istry, 1987, p. 237.
[98] R.A. Guidotti, F.W. Reinhardt, Proceedings of the 41st Power Sources [139] A.S. Dworkin, H.R. Bronstein, M.A. Bredig, J. Phys. Chem. 66 (1962)
Conference, 2005, p. 141. 572.
[99] R.A. Guidotti, F.W. Reinhardt, Proceeding of 201st ECS Meeting, Molten [140] R. Knödler, J. Electrochem. Soc. 130 (1) (1983) 16.
Salts, XIII, PV 2002-19, The Electrochemical Society, Pennington, NJ, [141] G.M. Haarberg, K.S. Osen, J.J. Heus, J.J. Egan, J. Electrochem. Soc. 137
2002, p. 31. (9) (1991) 2777.
[100] E. Strauss, D. Golodnitsky, E. Peled, Electrochim. Acta 45 (2000) 1519. [142] G.M. Haarberg, K.S. Osen, J.J. Egan, H. Heyer, W. Freyland, Ber. Bun-
[101] E. Strauss, D. Golodnitsky, E. Peled, Electrochem. Solid State Lett. 2 (3) senges. Phys. Chem. 92 (1988) 139.
(1999) 115. [143] G.J. Reynolds, M.C.Y. Lee, R.A. Huggins, Proceedings of the Fourth
[102] D. Golodnitsky, E. Peled, Electrochim. Acta 45 (1999) 335. International Symposium on Molten Salts, PV 84-2, The Electrochemical
[103] E. Peled, D. Golodnitsky, E. Strauss, J. Lang, Y. Lavi, Electrochim. Acta Society, 1984, p. 579.
43 (10–11) (1998) 1593. [144] M. Au, J. Power Sources 115 (2) (2003) 360.
[104] E. Peled, D. Golodnitsky, J. Lang, Y. Lavi, in: N. Doddapaneni, A.R. [145] R.A. Guidotti, P.J. Nigrey, F.W. Reinhardt, J.G. S Odinek, Batteries Pro-
Landgrebe (Eds.), Proceeding of the Symposium on Lithium Batteries, ceeding of the 40th Power Sources Conference, 2002, p. 250.
PV 94-4, The Electrochemical Society, Pennington, NJ, 1994, p. 75. [146] R.A. Guidotti, H. Ye, T.D. Xiao, D.E. Reisner, D.H. Doughty, US Patent
[105] C. Iwakura, N. Isobe, H. Tamura, Electrochim. Acta 28 (3) (1983) 277. 6,926,997 (August 9, 2005).
[106] Y. Shao-Horn, Q.C. Horn, Electrochim. Acta 46 (2001) 2613. [147] R.A. Guidotti, F.W. Reinhardt, J. Dai, J. Roth, D.E. Reisner, Proceedings
[107] Y. Shao-Horn, S. Osmialowski, Q.C. Horn, J. Electrochem. Soc. 149 (11) of the 41st Power Sources Conference, 2004, p. 153.
(2002) A1499. [148] J. Dai, R.A. Guidotti, T.D. Xiao, D.E. Reisner, US Patent 6,794,086
[108] J.-W. Choi, G. Cheruvally, H.-J. Ahn, K.-W. Kim, J.-H. Ahn, J. Power (September 21, 2004).
Sources 163 (1) (2006) 158. [149] R.A. Guidotti, F.W. Reinhardt, J. Dai, T.D. Xiao, D. Reisner, Proceedings
[109] Y. Shao-Horn, S. Osmialowski, Q.C. Horn, J. Electrochem. Soc. 149 (12) of the 35th IECEC Meeting, vol. 2, 2002, p. 976.
A1547. [150] R.A. Guidotti, F.W. Reinhardt, J. Dai, D.E. Reisner, Proceedings of the
[110] T.D.J. Dunstan, J. Caja, R.A. Guidotti, in: P.C. Truelove, H.C. Delong, 40th Power Sources Conference, 2002, p. 339.
R.A. Mantz, G.R. Stafford, M. Matsunaga (Eds.), Proceeding of the [151] R.A. Guidotti, F.W. Reinhardt, J. Dai, D.E. Reisner, Proceedings of the
International Symposium on Molten Salts XIII, vol. 2002-19, 2002, p. 36th IECEC, vol. 1, 2001, p. 897.
43. [152] R.A. Guidotti, F.W. Reinhardt, J. Dai, T.D. Xiao, D.E. Reisner, Review of
[111] R.A. Guidotti, S. Preston, D. Pickett, Z. Johnson, J. Wilkes, Proceedings US Nanocorp® —SNL Joint Development of Thermal-Sprayed Thin-Film
of the 42nd Power Sources Conference, 2006, p. 241. Cathodes for Thermal Batteries, 16th Annual Conference on Applications
[112] K. Abe, T. Chiku, J. Electrochem. Soc. 122 (10) (1975) 1322. and Advances, Long Beach, CA, January 9–12, 2001.
 P.J. Masset, R.A. Guidotti / Journal of Power Sources 177 (2008) 595–609 609

[153] J. Broadhead, J. Dai, J. Roth, R.A. Guidotti, D. Reisner, Thermal-Sprayed [157] R.A. Guidotti, F.W. Reinhardt, J. Dai, D.E. Reisner, J. Power Sources 160
Thin-Film Electrodes for Primary and Secondary Batteries, IBA-2001 (2) (2006) 1456.
Battery Materials Symposium, Pilanesberg, S. Africa, March 11–14, [158] C.J. Crowley, N.A. Elkouth, C. Lamb, Proceedings of the 39th Power
2001. Sources Conference, 2000, p. 517.
[154] R.A. Guidotti, F.W. Reinhardt, J. Dai, J. Roth, D.E. Reisner, Proceedings [159] Work underway at Electro Energy, Inc., Colorado Springs, CO, under US
of the 17th Conference on Applications and Advances, 2001, p. 97. Navy SBIR Phase II program.
[155] R.A. Guidotti, J. Dai, T.D. Xiao, D.E. Reisner, Proceedings of the 15th [160] P.J. Masset, R.A. Guidotti, J. Power Sources (2007),
Annual Conference on Applications and Advances, 2000, p. 127. doi:10.1016/j.jpowsour.2007.11.073.
[156] D.E. Reisner, T.D. Xiao, J. Dai, R.A. Guidotti, F.W. Reinhardt, J. New
Mater. Electrochem. Syst. 2 (1999) 279.

View publication stats