High specific energy Lithium Sulfur cell for space application

The battery energy density remains a key parameter accounting for the satellite mass budget. As illustration to this, the rechargeable battery still represents 100 to 200 of kilograms for a typical Eurostar 3000 satellite, which can represent up to 5% of the total mass, and about 100 kilograms for the next meteorological satellite program MetOp-SG. Any reduction in weight in these applications has therefore significant financial benefits, considering that the launch cost for such a satellite can be around 10k€/kg. Lithium-ion technology represented a revolution in terms of specific energy compared to Ni-Cd and is currently the most used and well suited for spacecraft. But it has also many drawbacks like price, some safety issues and its toxicity. Lithium-Sulfur (Li-S) cells are likely to become the next generation of energy storage to replace them. One of the reasons is that sulfur is an abundant element so it’s more affordable than cobalt used in Li-Ion cells. On top of that, Li-S cells are safer and more environmentally friendly. But the main advantage of this technology is the high energy density: around 5 times higher than Li-Ion cells. The major obstacle for application is due to dissolved polysulfide shuttling between anode and cathode. This phenomeno leads to permanent loss of active mass from the cathode into the electrolyte and onto the Li metal anode (passivating the Li anode with insoluble Li2S), severe self-discharge, low efficiency and fast capacity decay. Airbus DS has been testing and characterizing prototype Li-S cells manufactured by OXIS Energy Ltd. since 2014, demonstrating the potential and fast evolution of the cells performance. This paper presents the last test results on a set of different batches provided by OXIS and performed at Airbus DS premises in the frame of an ESA Innovation Triangle Initiative (ITI).


INTRODUCTION
This paper will first present a brief state of the art on Lithium-Sulfur technology, followed by the description of OXIS' cells tested during this study. After that, the test plan is described and the test results summarized and analysed to illustrate the current state of Li-S technology and its astonishing evolution.

STATE OF THE ART
A Lithium-Sulfur Battery (LSB) is a secondary battery composed of lithium metal anode and sulfur-based cathode. The trend for the secondary battery requirements is to be cheaper, safer, environmentally friendly and especially to have a high energy density.

Figure 1: Lithium-Sulfur electrochemical process
The Li-S battery works on the basis of redox reactions between lithium metal anode and sulfur cathode. The reaction in Li-S battery is a reversible conversion reaction. The sulfur cathode offers a theoretical capacity of 1675 mAh/g and a theoretical energy density of 2600 Wh/kg with the full utilization of sulfur in the following process: (1) Lithium-sulfur secondary battery technology is currently the only rechargeable battery technology that can reach 400 Wh/kg energy density and 1350 cycles [RD1]. A better understanding and further research about the problems of this new technology could allow reaching even higher energy density in the future.

Cathode materials
Elemental sulfur is an attractive choice as a cathode material for high-specific-energy rechargeable lithium cells because of its high theoretical specific capacity and the fact that it is very low cost and non-toxic. Sulfur can react with lithium ions between 1.5 and 3.0 V (vs. Li/Li + ) producing soluble and insoluble lithium polysulfides. A major drawback of sulfur is its low electronic conductivity (5 x 10 -30 S/cm at 25 °C) compared to the cathode materials used for Li-Ion batteries. In comparison to Sulphur, carbon based cathodes have several orders of magnitude higher conductivity values. For example, the conductivity of MCMB (MesoCarbon MicroBeads) is around 10 3 S/cm.

Anode materials
In the Li-S cells, dendrite deposition of Li metal is not as severe as observed in other lithium cell systems. This is attributed to the dissolved polysulfides (PS), which react with the anode and prevent the growth of Li dendrites. Actually, while facilitating cell performance, the dissolved PS causes severe redox shuttle between the sulfur cathode and Li anode, which results in low columbic efficiency for charging and a fast selfdischarge rate for storage [RD2].

Electrolyte
PS anions and anionic PS radicals are extremely reactive, they participate in many types of reactions such as basic, nucleophilic, redox, and radical reactions. PS is known to react with most of the common electrolyte solvents, such as esters, carbonates, and phosphates. It appears that the suitable solvents for the Li-S cell electrolytes are limited within the linear and cyclic ethers, such as dimethyl ether (DME) and 1, 3dioxolane (DOL). So far, Lithium trifluoromethanesulfonate (or lithium triflate) LiSO 3 CF 3 and lithium bis(trifluoromethanesulfony)amide (or LiTFSA) LiN(SO 2 CF 3 ) 2 have been reported to be the most suitable salt for the electrolyte of Li-S cells, of which LiN(SO 2 CF 3 ) 2 is superior in providing higher ionic conductivity and less corrosion to the Al substrate [RD30].

Electrochemical behaviour
Based on the phase change of sulfur species, the discharge process can be divided into four reduction regions (see Figure 2): •Region I: a solid / liquid two-phase reduction from elemental sulfur to Li 2 S 8 , which shows the first upper voltage plateau at 2.2 -2.3 V. In this region, the formed Li 2 S 8 dissolves into liquid electrolyte to become a liquid cathode. This process leaves numerous voids in the cathode.
•Region II: a liquid / liquid single-phase reduction from the dissolved Li 2 S 8 to low-order PS, during which the cell voltage steeply declines and the solution's viscosity gradually increases with a decrease in the length of S-S chain and an increase in the number (concentration) of PS anions. The solution's viscosity reaches a maximum value in the end of the discharge region. •Region III: a liquid / solid two-phase reduction from the dissolved low-order PS to insoluble Li 2 S 2 or Li 2 S, during which the following equations are competing with each other. This region forms the second lower voltage plateau at 1.9 -2.1 V, which contributes to the major capacity of a Li-S cell.
•Region IV: A solid / solid reduction from insoluble Li 2 S 2 to Li 2 S. This process is kinetically slow and generally suffers from high polarization due to the nonconductive and insoluble natures of Li 2 S 2 and Li 2 S.
Among the four regions above, Regions I and II show the highest redox shuttle, during which the cell suffers from the highest self-discharge rate and the cell's theoretical capacity can be seldom obtained. Region III contributes to the major capacity of a Li-S cell. Region IV becomes very short or even vanishes.

Safety
Li-S secondary batteries have demonstrated safe performance consistent with the technology's current state of development. Failures of early prototypes have been experienced, but they are of a magnitude and type that will yield to improvements in cell chemistry, packaging and electronic controls for cell charge / discharge management. It has to be noted that no fire or explosion occurred during the tests performed in the frame of the study.
The research and development areas specifically targeted to enhance Li-S battery safety are: •Electrode stabilization, •Electrolyte composition, •Electronic controls and packaging designs internal and external to the cell.

OXIS LITHIUM-SULFUR CELLS
OXIS provides different cells to be tested at Airbus DS facilities in order to analyse the evolution of the performances. Different batches of 5-6 cells have been specified: • Batch 1a: Ultra-light cells of 6.5A.h with an energy density of 250Wh/kg; • Batch 1b: Long-life cells of 10A.h with an energy density of 150Wh/kg; • Batch 2: Ultra-light cells of 12A.h with an energy density around 300Wh/kg; • Batch 3: Long-life cells of 20A.h with an energy density of 190Wh/kg.
The cells currently under test are described hereafter.

OXIS POA0084 6.5Ah Ultra-Light Pouch Cells
The main characteristics of the POA0084 Ultra-Light cells are listed in the following

OXIS POA0217 12Ah Ultra-Light Pouch Cells
The main characteristics of the POA0217 Ultra-Light cells are listed in the following table. Operating temperature 5°C to +30°C

TEST PLAN
The test plan is designed to evaluate the interest of this technology for space systems. For each cell batch delivery, the following tests will be performed.

Initial Reference test
The objective of this test is to confirm the characteristics shown on OXIS' datasheet. It is performed at the temperature indicated on the datasheet for the capacity measurement and performed at C/5 with internal resistance measurements at 0, 10, 20, 60, 80 and 100% of Depth of discharge (DoD). In order to assess the self-discharge, two reference tests will be performed: a first one with the discharge starting just after the end of the charge and a second one with one hour pause between the two phases.

LEO cycling test
If Li-S is meant to be used for satellite applications, it will need to face a high number of cycles representative of any LEO mission. For this test, a discharge at C/3 down to 20% of DoD has been selected with a charge rate of C/5. A hysteresis method is selected for charge management, so the voltage of the cell is always higher than end of charge voltage (EOCV) -100mV.

GEO cycling test
GEO missions based on Li-Ion technology use high discharge currents but such high currents are currently one of the weak points of Li-S. A test of cycling at 80% DoD and another one following a real GEO season profile are performed. The impact of EoCV is also assessed by performing one test at 80% DoD with the maximum voltage as EoCV and another one with a reduced EoCV, expecting to increase the life of the cell, even if slightly degrading the total capacity. After testing the first batch of cells (POA0084) at C/1.5, it was observed that the cell did not achieve the required capacity. It was then decided to characterize the discharge currents and perform a GEO cycling at C/5.

Launcher test
The wide temperature range expected for Li-S batteries make them very attractive for launcher applications since they often require starting at very low temperatures. A power profile coming from an Ariane 5 GTO DDO mission is run on one cell.

Cells characterisation tests
The last type of tests includes the characterisation of the cells at different temperatures: +20°C for comparison and minimum and maximum temperatures included in the datasheet. The discharge is performed at C/5 with internal resistance measurements as per the reference test. Current rates tests are done at Reference test temperature and at C/5, C/2 and C rates for discharge.

Initial Reference test
At the time of writing, the tests results available are those on POA0084 Ultra-Light cells and POA0122 Long-Life cells. Partial results are available for the POA0217 Ultra-light cellS.

Figure 6: Initial Reference Test on Li-S POA0084
As it can be observed, the POA0084 capacity of 6.5Ah is confirmed by all the cells. The self-discharge after one hour in open-circuit varies from 4.6% to 0%.

Figure 7: Initial Reference Test on Li-S POA0122
As it can be observed, the cell capacity shows always values higher than 10 A.h. However, in this type of cell the self-discharge is more important: between 12% and 24% of loss capacity in just one hour of open-circuit.   A new batch of Ultra-light cells shows a lower internal resistance that the first one. The resistance values decreased from a range of 10 to 40mΩ (at 1s) down to 5 to 15mΩ (at 1s) for five out of six samples.

LEO cycling test
In figure 11 the hysteresis method for charge management can be observed: once the EOCV (2.6V) is reached, the charge ends, the cell voltage goes down progressively and reaches 2.5V (EOCV-100mV) just when the discharge phase starts.

Figure 11: LEO cycling Test on Li-S POA0084
The evolution of the End Of Discharge voltage shows the progressive loss of capacity.

Figure 12: EODV evolution during LEO cycling Test on Li-S POA0084
At the time of writing, the POA0217 Ultra-Light cell has completed 55 cycles and the test is still running. The next table summarizes the performance of the other cells.

GEO cycling test
GEO cycling was initially foreseen with a discharge rate of C/1.5 but as it can be observed, the value of Ah expected (5.2Ah) was higher than the capacity that could be actually provided (3.97Ah).

Figure 13: GEO cycling Test on Li-S POA0084
Unlike LEO cycling, the cell cannot provide enough capacity. The test was stopped when the fading reached 83%.

Figure 14: Fading and capacity provided during GEO
cycling Test on Li-S POA0084 The cell following a real GEO eclipse season could provide enough capacity until eclipse number 8. It was then decided to assess higher levels of DoD at lower discharge currents: C/5 instead of C/1.5.
At the time of writing, the POA0217 Ultra-Light cell has completed 25 cycles and the test is still running. The next table summarizes the performance of the other cells.

Launcher test
As explained, a power profile coming from an Ariane 5 GTO DDO mission has been run on one cell per batch.

Figure 15: Launcher profile Test on Li-S POA0084
The launcher profile test showed that the cell could withstand without any problem this type of mission. The DoD reached at the end of the mission was 70%, letting enough margin for its application.
The two other cells passed without any problem the adapted launcher profile.

Cells characterisation tests
Characterisation tests include discharge at different current rates and temperature values. As it has already been demonstrated by the GEO cycling, the discharge current clearly impacts the capacity that can be provided by the cell. C/5 guarantees the nominal capacity, but when it is increased up to C/2 this is reduced by 22%. At 0.75C, the capacity is reduced by 75%. Figure 16: C-rate characterisation Test on Li-S POA0084 The tolerance to high C-rates has been considerably enhanced with the new batch of Ultra-light cells, reducing the loss from 75% down to 27% as it can be observed on the next figure. The following figures show that the impact at capacity and internal resistance level increases when the temperature decreases. No dramatic differences are observed for +20°C and +30°C but a huge capacity loss appears at +5°C.

Figure 18: Temperature characterisation Test on Li-S POA0084
Again, this issue has been assessed for the last batch of Ultra-Light cells, reducing the sensitivity to the temperature. Again, as it can be observed on the figure below, no major difference appears between +20°C and +30°C, but the difference with regard to the +5°C case has been considerably reduced.  • It is demonstrated by testing that Li-S is a technology with a specific energy superior to that of Li-Ion, i.e 306 Wh/kg delivered by a POA0217 Ultra-light cell at Airbus DS premises.
• The major drawback of Li-S batteries for spacecraft applications continues to be the cycling capability. Up to now, only 100 cycles have been achieved at a 20% DoD. An additional cell has been provided by OXIS for batches 2 and 3 in order to study the impact of EoCV but there are no significant results at the time of writing. However, background from previous tests performed by OXIS and Airbus DS has shown a much higher cycling capability on other Li-S cells by reducing the EoCV (up to 1400 cycles at 20% DoD with a capacity loss of 15%) by reducing for instance the voltage from 2.6V down to 2.45V. Similar results are expected to be obtained with the additional cells under test for batch 2 and 3. If the cycling capability proves to indeed be significantly improved by a 150 mV reduction in EoCV also for this latest Li-S technology, it would be worth considering for certain mission profiles. Thanks to the significant leap in specific energy for Li-S batteries compared to the Li-ion ones, despite the for-mentioned decrease of the EoCV, Li-S batteries would still deliver a higher specific energy compared to Li-ion. • Two of the main drawbacks up to now such as the dependency of the capacity on the temperature and the current rate have been addressed.
Despite the limitations to this moment, Li-S battery technology shows a great potential for many space applications like launchers or LEO satellite missions. Performances regarding cycling, rate capability and temperature need to be however further improved. Li-S technology is just at the beginning and already shows great potential and constant evolution. The results of the tests performed during this ITI are encouraging and suggest that Li-S will in fact become the next generation of energy storage.