Comparison of the Different Anode Technologies Used

7.3
Comparison of the Different Anode Technologies Used in Thermal Batteries
J. Douglass Briscoe*, Emmanuel Durliat**, Florence Salver-Disma**, Ian Stewart***
*ASB Group – Advanced Thermal Batteries
107 Beaver Court, Cockeysville, MD 21030, USA
**ASB Group – Aérospatiale Batteries (ASB)
Allée Sainte Hélène, 18021 Bourges cedex, France
Fax: +33 2 4848 5601 Phone: +33 2 4848 5639 Email: [email protected]
*** ASB Group – Missiles and Space Batteries (MSB)
Hagmill Road, East Shawhead, Coatbridge, ML5 4UZ, UK
Abstract: This paper presents a comparison of three
anode technologies: LAN (lithium anode), lithium
aluminum and lithium silicon against iron disulfide and one
of a proprietary metal disulfide compound MS2. LAN comes
out as the best anode technology in terms of voltage, energy
and power densities, internal resistance and polarization. It
is also very safe against overheating. Used with our MS2,
LAN appears as particularly suitable for high power/high
energy applications.
ƒ
The electrolyte consists of binary salt with magnesia
binder. The cathode was either iron disulfide (FeS2) (so
called “cathode 1” studied in part 1) or our proprietary
metal disulfide compound (MS2) (so called “cathode 3”
studied in part 2) mixed with salts.
The heat powder for the heat pellets was adapted to the heat
sensitivity of each anode. LAN batteries as shown below
can be designed with higher heat input.
Keywords: Thermal batteries, LAN, Alloy, Comparison
Lithium alloy cells were designed using a 0.1mm thick iron
separator. In the case of LiSi anodes, the capacity was
determined taking into account the first and second plateaus
only, and assuming a capacity of 2000A-s/g.
Introduction
LAN (Lithium Anode) was compared to lithium aluminum
and lithium silicon in battery tests under the same
requirements:
-
capacity of 3600A-s,
-
fixed volume for the stack : 84 cubic cm with a
cell diameter of 53mm.
For the study with FeS2 as cathode, all cells had the same
adiabatic temperature evaluated using our proprietary
model (575°C @ +60°C), the same cathode, the same ratio
{thickness of electrolyte over (thickness of anode +
thickness of cathode)} and the same coulombic capacity for
the anode.
In a first phase, the anodes were evaluated with iron
disulfide as cathode. Iron disulfide is a standard cathode,
which is inexpensive and easily available. In a second
phase, the best anode was evaluated versus improved
cathodes.
Current loads
Three different baseline currents were used: 2.2A, 4.4A and
13.2A giving current densities of 0.1A, 0.2A and 0.6A/cm2.
The test loads consisted in one of these baseline currents
plus pulses of twice the baseline intensity during 100ms
every 50s.
This study aimed at complementing existing studies at
single cell test level by taking into account the thermal
effect (side thermal insulation and end heats). It enables for
example to assess and to compare the energy/power
volumetric density of the three anodes.
Discharge
Batteries were discharged after conditioning at the test
temperature during at least 4 hours. Voltage and current
were monitored.
Experimental set-up
Battery and cell definition
Part 1: Results with standard FeS2 as cathode
Results
All batteries were made using the can, header, lead
assemblies, lateral thermal insulation and end heats of an
existing battery. The volume available for the stack was
thus a cylinder with a diameter of 53mm and a length of
33.6mm. For each cell type, as many cells as possible were
fitted in the battery.
Their overall thickness was 2.98mm with LiAl, 3.19mm
with LiSi and 2.40mm with LAN.
In the volume available, it was possible to fit stacks with
either 11 LiAl cells, 10 LiSi cells or 14 LAN cells. The
following figures present the voltage vs. time at -32 (cell @
490°C) and +60°C (cell @ 575°C) (red and orange curves
for LAN, green curves for LiAl and blue curves for LiSi).
Three kinds of anodes were characterized:
ƒ
lithium metal mixed in a metallic matrix: LAN
supplied by MSB (with 15% weight of Li).
two alloys: LiAl (with 19% weight of Li) and LiSi
(with 44% weight of Li),
117
Analysis
30
The LAN anode technology gives by far the best
performance in terms of voltage in a given battery volume.
This is due to the higher cell electromotive force, the higher
weight density and the lower polarization, which is
estimated by linear regression of the 2.2A discharge curves
in the time frame 80 to 500s and in cold conditions as:
voltage (V)
25
20
LiAl @ 490°C 2.2
LiAl @ 575°C 2.2A
LiSi @ 490°C 2.2A
LiSi @ 575°C 2.2A
LAN @ 490°C 2.2A
LAN @ 575°C 2.2A
15
10
5
0
0
200
400
600
time (s)
800
1000
1200
Note that the hot LiSi discharge ended at 1000s, thus
explaining the sudden tail end
LiSi
-0.11 mV/A-s
LAN
-0.08 mV/A-s
70
Energy (kJ)
Power (W)
60
50
30
LiAl @ 575°C 4.4A
LiSi @ 575°C 4.4A
LAN @ 575°C 4.4A
25
voltage (V)
-0.14 mV/A-s
In the 2.2A discharge, the energy and average power (till
voltage drop to 75% of the initial maximum voltage) are:
Fig. 1 discharge @ 2.2A, 490 and 575°C
40
30
20
20
15
10
10
0
5
LAN
0
0
200
400
600
time (s)
800
1000
30
LiAl @ 575°C 13.2A
LiSi @ 575°C 13.2A
LAN @ 590°C 13.2A
25
LiAl
20
15
10
30
5
LAN @ 640°C 4.4A
LAN @ 640°C 4.4A
25
100
200
300
time (s)
400
500
voltage (V)
0
0
LiSi
Assuming we could take the same double layer (electrolyte +
cathode) for all kinds of anodes, one could integrate then 12
alloy cells (either LiAl or LiSi). In this case LAN still remains
the best anode followed by LiSi and then LiAl. In terms of
internal resistance, LAN gives also the best results: at 600s,
R=9.0mΩ for LAN, 9.3mΩ for LiSi and 11.6mΩ for LiAl.
Furthermore, LAN is an anode technology, which is tolerant of
a high cell temperature. It is thus a much safer technology than
the alloys. The following figure presents two discharges in the
same test conditions with a cell temperature of 640°C.
1200
Fig. 2 discharge @ 4.4A, 575°C
voltage (V)
LiAl
600
Fig. 3 discharge @ 13.2A, 575°C
20
15
10
Remark 1: the hot LiSi batteries showed some overheating
after post-mortem (Fig. 2 and 3). That explains why the
FeS2 transition from the 1st to the 2nd plateau appears earlier
for LiSi batteries than for the other ones, due to increased
self-discharge of the cells.
5
0
0
100
200
300
time (s)
400
500
600
Fig. 4 LAN cell @ 4.4A, 640°C
Remark 2: Figures 2 and 3 show that the capacity of the
LiSi anode is more of the order of 4000A-s than the
estimated 3600A-s, if both the 1st and 2nd plateaus are
considered.
Such a hot cell temperature is not possible for alloy based
cells, which would fail in thermal runaway. On the
example, we can nevertheless see that the LAN cell life is
shortened by the thermal decomposition of the cathode.
This can be mitigated by improving the cathode.
Remark 3: in Fig.1, the cold batteries are cooling down.
The performance is thermally limited.
118
Part 2: Results with improved cathodes
Battery and cell definition
30
In this part, we used the same battery design as previously.
The anode was LAN with a capacity of 3600A-s.
voltage (V)
25
We only changed the cathode. Two cathodes were
investigated:
-
One cathode (so called “cathode 2”) made with the
same FeS2 based cathode but with a chemical protector
inserted between the cathode and the heat pellet; this
weight of FeS2 was similar to the weight of the cathode
investigated in the previous part. Thus cathode 2 had
the same capacity as cathode 1 but was slightly thicker,
FeS2
FeS2 + Protector
MS2 + Protector
15
10
5
0
0
200
400
600
800
1000
time (s)
Fig. 5 LAN vs. MS2 and FeS2 @ 2.2A
One cathode (so called “cathode 3”) made with a metal
sulfide compound MS2. In this case, we also protected
the cathode from the heat pellet by insertion of a
chemical protector. To keep things comparable, we
designed the cathode so that the total thickness of
cathode + protector is equal to the thickness of cathode
1 as evaluated in the first part. Thus cathode 3 had the
same thickness as cathode 1 and cathode 2 but 25%
less capacity.
Figure 6 presents results obtained at 4.4A.
30
25
voltage (V)
-
20
MS2 + Protector
FeS2 + Protector
FeS2
20
15
10
Cathode
Cathode 1
Cathode 2
Cathode 3
5
FeS2
FeS2 +
MS2 +
0
Protector
Protector
Thickness
t
t+
0.13mm
t
Capacity
C
C
0.75 C
0
100
200
300
time (s)
400
500
600
Fig. 6 LAN vs. MS2 and FeS2 @ 4.4A
Figure 7 presents results obtained with MS2 at high cell
temperatures.
Table 1: Summary of cathode definitions
30
“Cathode 3” was tested at 4 different cell temperatures:
575, 590, 600 and 640°C. As a comparison, “cathode 1”
and “cathode 2” were tested at 575°C.
voltage (V)
25
Batteries were built with these cathodes and with 14 cells
each. Considering “cathode 2”, fitting 14 cells was only
possible by removing 2mm of packing.
20
MS2 @ 640°C
MS2 @ 600°C
MS2 @ 590°C
15
10
Current loads
5
Two different baseline currents were used: 2.2A and 4.4A
giving current densities of 0.1A/cm2 and 0.2A/cm2. The test
loads consisted in one of these baseline currents plus pulses
of twice the baseline intensity during 100ms every 50s.
0
0
100
200
300
time (s)
400
500
600
Fig. 7 LAN / MS2 at various cell temperatures
Discharge
Analysis
Batteries were discharged after conditioning at the test
temperature for at least 4 hours. Voltage and current were
monitored.
The chemical protection of a FeS2 cathode against the heat
pellet increases the performance of the battery by a factor
of about 2 in a very low discharge rate (See Fig. 5). The
slope of the discharge curve increases drastically at 400s
for the cathode without protector (red curve) whereas the
slope of the discharge curve of the cathode with protection
(purple curve) remains unchanged till about 800s. For
Results
Figure 5 presents results of MS2 and FeS2 (with or without
chemical protector) in a 2.2A discharge.
119
shorter discharges (Fig. 6), the protection of the cathode
does not bring any advantage.
30
The use of MS2 also enables the internal resistance per cell
to be reduced by about 25%, as shown on the following
table.
time (s)
200
400
500
600
voltage (V)
25
Internal Resistance per cell
(mΩ) @ 575°C
MS2 +
FeS2 +
Protector
Protector
4.5
3.6
5.5
4.1
7.5
5.0
8.9
5.9
5
0
0
0.2
0.4
0.6
Cathode Depth of Discharge
0.8
1
Fig. 9 voltage vs. DoD in the cathode (4.4A discharge
@ +575°C)
The use of LAN/MS2 is thus of high interest for high
energy density applications.
The results with a cell temperature of 640°C show that the
performances of LAN/MS2 are very stable on a wide range
of cell temperature. The use of LAN/MS2 enables to design
much safer batteries than the use of alloys, and especially
LiSi. It is also particularly suitable for the design of high
duration/high energy batteries.
The polarization is also much smaller with LAN/MS2
(about -0.038mV/A-s) than with LAN/FeS2+Protector
(about -0.090mV/A-s).
Furthermore, figure 6 shows that as the current density
increases (from 0.1A/cm2 on Fig. 5 to 0.2A/cm2 on Fig. 6),
MS2 shows increasing cell efficiency compared to FeS2.
Indeed, figure 6 shows that the performance of both
FeS2+protector and MS2 are quite the same although the
MS2 “cathode 3” had about 25% less capacity than the
protected FeS2 “cathode 2”. The use of MS2 thus increases
the efficiency of LAN by about 30%.
Nevertheless in a wide range of applications, iron disulfide
remains a good compromise between cost and
performance; additionally, its performance can be further
improved by protecting it from the heat pellet for life
duration above 400s.
Conclusion
We have compared the LAN technology to lithium
aluminum and lithium silicon. We have shown that:
Figures 8 and 9 highlight the higher efficiency of MS2 by
presenting the voltage as a function of depth of discharge
(DoD) in the cathode. They clearly show that MS2 has an
advantage over FeS2 in terms of efficiency.
30
25
voltage (V)
15
10
The use of LAN/MS2 is thus of high interest for high power
density applications.
20
FeS2
FeS2 + Protector
MS2 + Protector
15
MS2 + Protector
FeS2 + Protector
FeS2
20
-
LAN is the best technology in terms of voltage per
cell, internal resistance and polarization. It is
furthermore a very safe technology due to its
robustness to overheating.
-
Its performance can be optimized by coupling LAN to
an iron disulfide cathode with protection from the heat
pellet (life duration > 400s), or to our proprietary metal
disulfide compound (life duration > 800s).
FeS2 as shown in figures 5 and 6 gives very good
performance up to 600 to 800s duration, and is a very good
compromise between cost and performance.
10
5
MS2 is of interest for very long life batteries or special very
high power applications.
0
0
0.2
0.4
0.6
Cathode Depth of Discharge
0.8
1
Using all technologies and specially LAN, the ASB Group
(ASB, MSB, ATB) is able to cope with the very wide range
of customer requirements by designing reliable and safe
thermal batteries.
Fig. 8 voltage vs. DoD in the cathode (2.2A discharge
@ +575°C)
120