Skip to main content

Advanced European lithium sulphur cells for automotive applications

Final Report Summary - EUROLIS (Advanced European lithium sulphur cells for automotive applications)

Executive Summary:
The research project EUROLIS started in October 2012 and ended in September 2016. The aim of EUROLIS was to develop a sustainable three generations of advanced lithium sulphur battery (LSB) prototypes in the standard 18650 battery configuration. For each reporting period the consortium of partners has prepared a new set of chemical environments (electrolyte, host matrix for cathode, separator, additives, etc.) to be integrated into the three different generations of prototypes. Based on the obtained knowledge within the project, we have improved the electrochemical performance of LBS prototypes in each prototype generation.

In the 1st reporting period we established a chain of knowledge, where each partner contributes to the certain activity related to the production of the prototype cells (Figure I). Among different types of carbon host matrixes tested during this period, a doped carbon prepared by template synthesis using glucose was selected and the synthesis was scaled up. After impregnation with sulphur, laboratory tests of the obtained composite showed its suitability for the 1st generation of LSB. A similar approach was used for the selection and preparation of a larger batch of electrolyte used by all partners working on the analysis and electrochemical characterisation in laboratory scale LSB. In the 2nd reporting period we successfully continued research and development of different parts of Li-S battery (LSB) with a focus on distributed work between partners, where each partner contributed to the certain activity related to the production of the prototype cells. Among different types of carbon host matrixes tested during this period, a doped carbon prepared by template synthesis using glucose was selected and the synthesis was scaled up. After impregnation with sulphur, laboratory tests of the obtained composite showed its suitability for the 2nd generation of LSB. A similar approach was used for the selection and preparation of a larger batch of electrolyte used by all partners working on the analysis and electrochemical characterisation of laboratory size LSB. In the last reporting period, the focus has been the preparation of 3rd generation of cathode foils with an aqueous, as well as a PIL based binder system. 12 cells of the 3rd prototype generation have been built. Initial discharged capacities measured on generation 3 at low rate were above 1100 mAh/gS. Rate capability was improved compared to GEN2 since we reached more than 600 mAh/gS at 1C.

The solid project consortium of various partners with different interests, but focuses on the same goals, enabled the project to fully cover the whole range of R&D including manufacturing and battery tests. Clustering and exchange of ideas have been carried out during the third reporting period. With dissemination activities we have raised the awareness of the project’s significant results.

Project Context and Objectives:
The aim of the EUROLIS project was to develop three generations of lithium sulphur battery (LSB) prototypes in the standard 18650 battery configuration. The timing of the work plan was organized in three major steps: (i) preparation of active components, (ii) characterisation and use of analytical tools, and (iii) prototype assembling and testing.

For each reporting period the consortium of partners has prepared a new set of chemical environments (electrolyte, host matrix for cathode, separator, additives, etc.) to be integrated into the three different generations of prototypes. Based on the obtained knowledge within the project, we have improved the electrochemical performance of LBS prototypes of each prototype generation. The final goal was to develop cells with energy density of 500 Wh/kg, power density of 1000 W/kg and charge efficiency higher than 95%. The predicted cycle life was 1000 cycles. The set goals have been reached mainly through six different work packages (WPs), where WP2 (System definition) has been acting as a supporting WP to WP3 (Cathode composite) and WP4 (Electrolytes and separators: Formulation and Modelling). WP3 and WP4 have been assisting in the selection and transmission of components between partners for each generation of prototype cells. Components have been extensively tested and analysed within WP5 (Analytical tools) and integrated into prototype cells within WP7 (Integration, scale up, testing, life cycle assessment and benchmarking). Other possible configurations of LSB have been explored within WP6 (Benchmarking of other Li-S technologies).

Within consortium many different carbons as host materials for sulphur have been prepared. At the end for each generation the one with the most optimal BET surface area, type of porosity, pore size, pore volume have been chosen:
- for the 1st generation prototype cell the MPG prepared sugar derived salt-templating carbons;
- for the 2nd generation prototype cell the salt-templating carbons using binary combinations of salts and;
- for the 3rd generation prototype cell commercial carbon from Imerys called ENSACO®.

Optimization of the electrolyte formulation has been important to improve battery performances, especially in terms of safety and life span. As electrolyte for the 1st generation of the Li-S prototype cells we decided to use 1M LiTFSI in TEGDME/DIOL (1:1 v/v). Both electrolytes were upscaled to 2 L. Best stability and excellent coulombic efficiency were achieved with TFSI family of IL in the combination DEME cation - N122(2O1)TFSI. Methoxyethyl chemical functionality grafted on the nitrogen atom of the IL cation attributed to the beneficial stability with N122(2O1)TFSI. 2nd generation (GEN2) electrolyte was defined as a mixture of 1 volume of GEN1 electrolyte plus 1 volume of LiTFSI:N122(2O1)TFSI (1:9 mol. ratio).

The use of different in situ and ex situ analytical tools for analysis of Li-S batteries helped us to better understand the mechanism of Li-S battery and impact of different chemical environments on the electrochemical properties.

Within project other Li-S technologies has been assemble. The thin ceramic membranes and all solid-state Li-S batteries have been prepared and electrochemically tested. Alternative ionic conductive coating of particles by using ionic liquids polymers was performed and tested as an alternative approach. We envisaged alternative Si electrode as replacement for the negative lithium.

During the project, an aqueous binder system has been successfully developed for the laminated cathode. Three different generations of 18650 prototypes were assembled and tested. Estimated cell energy has been significantly improved from 45 Wh/kg (GEN1) to 130 Wh/kg (GEN3). However, electrolyte quantity is still high (even if it was reduced locally from 12 to 6 µL/ mgS) to reach project objectives and significant cell design improvements are required in order to increase specific energy and to improve overall electrical results (especially life duration).

Abuse tolerance test has shown a potential safety issue in overcharge in term of toxic gases and heat generation. The design of the cell will also need to take into account this issue.

It was highlighted that Li-S chemistry has a great potential for lowering the environmental impact of the battery in a full electric bus application.

Key research challenges
The EUROLIS project deals with fundamental questions of polysulphide solubility, diffusion and reactivity and also with technological questions of materials production, processibility and integration. The work within the project has a large scientific and technological impact on the potential LSB production in the future.

Scientific impact:
- development of new materials with properties required for LSB;
- use of a modelling approach with the aim to understand interactions between different components and polysulphides;
- application of different analytical tools which are described in published scientific papers;

Technological impact:
- developed know-how about cylindrical LSB production;
- implementation of laboratory results into prototype cells;
- protection of knowledge.

Project Results:
WP 1: Project administrative and financial management

All schedule reports, minutes and deliverables has been submitted. Project meetings and management committee meetings were carried out. The communication and everyday work kept the project smoothly running and supported effective and prosperous research and co-operation.

Task 1.1 Daily management
Management and coordination related activities has been carried out within this work package. The coordination of the project has been carried out via technical coordinator and a project coordinator. The project management has been carried out in accordance with the Project Road Map. Exception has been the meeting planned to be organised by SAFT that has been carried out by University of Montpellier 2. Risk and contingency planning of quality problems and substantial delays in deliverables and taking appropriate measures has been set. Problems have been solved through everyday communication and organisation of the exceptional meetings.

Task 1.2 Reporting
To monitor the project status 6-months internal partner reports after every meeting has been prepared. The WP leaders have also prepared 6-months internal WPs reports and the coordinator has assembled them in the internal project report. The report has shown to be useful for the preparation of the periodic report at M 18, M 36 and M 48 for EC.

WP 2: System definition

The system definition contained the coordination of the system selection and integration into prototype with final testing procedure according to the defined standards. It included unified materials characterisation using standardized analytical characterization tools which purpose had been a comparable overview on the behaviour of different materials.

Task 2.1 RTD coordination and management
RTD coordination and management was an on-going task for the whole duration of the project. It was carried out by email exchange, Skype and telephone conferences, and informal meetings during ALISTORE-ERI meetings, conferences and with short visits between partners (brainstorming meeting in Paris and informal meetings at the conference IMLB in Como, Italy). To the EC various RTD reports has been delivered.

Task 2.2 System definition
We were monitoring the development of the cathode composite (successfully improved porosity of carbon host matrices which contributed to increase ratio of sulphur in the carbon/sulphur composite), combined with a modelling and development of electrolyte and separator (we were able to minimize layers of separators and consequently overall quantity of electrolyte in the prototypes cells) and supported by the development of analytical tools for Li-S batteries. Selected materials (chemical environment) were tested and the components with the best electrochemical performance were characterized and integrated in the prototype cells. Quantities and timetable of materials transfer were adopted from the 1st prototype generation which turned out to be optimal for the preparation of 12 prototype cells for the 2nd generation of prototype cells.

Task 2.3 Test procedure
The goal of test procedure was to define and select the testing procedures for performance analysis, for efficiency assessment, calendar life and cycle life assessment and last made safety and abuse testing. We define overall sulphur content in cathode composite need to be 50 wt. % and the loading per cm2 had to be between 3-5 mg/cm2. Melting procedure at 155 ˚C was the most suitable method for the impregnation of host matrix with sulphur. To enable required conversion of sulphur the accessibility of electrolyte to the sulphur surface had to be assured. Based on the porosity of the cathode the amount of the electrolyte was recalculated (it should not exceed 2-5 µL/mg of sulphur in the hybrid electrolyte).

WP 3: Cathode composite

Direction for further improvements was based on the results from the studies of different physical and chemical parameters of the cathode composite with regard to energy density and charge efficiency (polysulphide shuttle mechanism). For that purpose, we studied:
- Influence of the materials and the composite morphology (BET surface area, type of porosity, pore size, pore volume and shape of pores).
- Impact of the type of additives (electron conductive additive, semi-conductor, ceramic adsorbent ...) and their surface properties (acidity, hydrophobicity, surface defects...).
- Ratio between sulphur and the additive(s) (normalized per one of the additive properties – like surface area, pore volume, pore size).
- Optimized cathode thickness and porosity.

Task 3.1 Literature Review of Li/S of previous studies
It contained an overview of different host materials used for the impregnation with sulphur and it discussed the importance of binder selection and the role of current collectors. Literature review on cathode composites showed importance of carbon based matrix and influence of oxide based mesoporous structures on the electrochemical properties of Li-S batteries. There was a strong message in the literature that PVdF as a binder might not be suitable and some concerns as well related to the use of classical current collectors.

Task 3.2 Mesoporous based carbonaceous host structures of sulphur impregnation
Three different families of carbonaceous host structures (HS- hollow carbon nano spheres; NF – mesoporous carbon obtained from ionic liquids using salt templating method and TC1x – mesoporous carbon obtained from sugar with salt templating method) has been prepared. The NF and TC1x types of carbon host materials had starting capacities that fit into the objective of this task around 1000-1200 mAh/g. Low cost processing and useful capacities (in the starting cycles) of TC samples were the major reason to use this sample as a host matrix for the 1st generation of the prototype cells. For a required larger batch cathode composite synthesis for the 1st generation prototype the salt templating method using ZnCl2 as a salt template and N - glucose as carbon precursor were used. All salt templating carbons had relatively high specific surface areas, but only the undoped carbon has shown a pore volume which reached and clearly exceeded the target of 1.26 cm3g-1. The maximum quantity of the sulphur that can be infiltrated into pores was checked. Conclusion was that the most pronounced pore filling was occurring in the micropore regime. In the run of increasing sulphur loading a successively reduced contribution of small pores was observed. Even at high sulphur loading (70 % S) residual porosity continuously distributed through the meso- and macropore range was observed. Different configurations of a binary salt template were tested. With 3:1 mixture of KCl/ZnCl2 to glucosamine the porosity compared to the 1st generation was increased and consequently the capacity retention in the formation cycles was improved. Carbon host matrix based on the KCl/ZnCl2 template synthesis method was used for the 2nd prototype generation. For the 3rd prototype generation we decided to use a commercial carbon from Imerys called ENSACO®. This carbon has an astonishing porosity for a commercial carbon, it can be obtained in large quantities and gave good results in the ongoing H2020 follow-up project of EUROLIS (HELIS).

Task 3.3 Mesoporous silicate type host structures for sulphur impregnation
The hydrophilic pores of mesoporous silicates reversibly adsorbs the hydrophilic polysulphides and releases them near the end of discharge so that they can be further reduced in the pores of the mesoporous carbon. The influence of pore size and specific surface of four different mesoporous silicates (SBA-15, KIL-2, MnKIL-2, FeKIL-2) on the discharge capacity was tested. The silicate with the smallest pores and the biggest specific surface (SBA-15 (818 m2 g-1)) had the highest and the most stable discharge capacity and by decreasing the specific porosity of mesoporous silicates additives the specific discharge capacity was consequently decreasing. Second group of tested additives were zeolites with high specific area and precise and uniform pore size. Incorporation of transition metals into their framework generated different types of catalytically active sites. With addition of 9 wt. % MnS-1(mw.) zeolite in the cathode composite the capacity fading was decreased, the cycle efficiency (95 %) was improved and the polarization was decreased and stabilized. The result was interesting due to two times lower specific surface and six times lower pore volume of zeolite MnS-1(mw.) than mesoporous SBA-15. We assumed that such a good result attributed to Mn2O3 sticks attached to zeolite surface, which have additional polysulphide absorption ability together with adsorptive zeolite structure.

Sputtering thin films of the vanadium oxide on the top of fibber separator in Li-S cells and by heating changing its oxidation state were tested. Vanadium with oxidation state V5+ gave the best electrochemical properties.

Task 3.4 Lab scale sulphur impregnation, electrode preparation and electrochemical tests
Three different methods of sulphur impregnation were tested initially (precipitation of sulphur, infiltration through the liquid phase and infiltration during thermal treatment in inert atmosphere at 155 °C). The thermal treatment method due to its simplicity and reliability was selected. The method was optimized on batches of 1-5 g, however for the need of 1st prototype generation this procedure was successfully upscaled to 50 g batch at the beginning and latter to 150 g batch, that was the maximum quantity that could be pre-treated in the laboratory furnace. All batches were carefully checked if sulphur was homogenously distributed and we found variations between 61.5-63.0 wt.%. Overall quantity of sulphur was estimated to be 62 wt.%. On the laboratory scale we studied the influence on the electrochemical behavior when we used various carbon based materials like carbon nanotubes, carbon black or graphene nanoribbons. Additional we performed various tests with different binders (PVdF, PVdF-Kynar and PTFE). A general conclusion was that for each composition special attention should be devoted to the optimized type of carbon additive and type of binder including their quantities. In laboratory scale the cells were tested in coffee bag cells. Electrochemical tests and porosity measurements has shown that 50-60 wt.% of sulphur in the carbon composite still preserves enough porous volume and electrodes are active, while 66 wt.% starts to close porosity.

Task 3.5 Preparation of 1 kg batches of sulphur – host structure composite
The synthesis of mesoporous carbon has been successfully scaled-up, the sulphur impregnation was performed in four batches. In the first batch we mixed 50 g of carbon and 100 g of sulphur (obtained from Aldrich), for the other three batches we mixed 450 g of carbon and 900 g of sulphur. All samples were ball milled at 300 rpm for 30 min before heat treatment at 155°C. Due to harmful, corrosive and dangerous properties of ZnCl2, a closed system for the preparation of larger batch of the carbon host matrix was built. The big batches were prepared by using ceramic bowls as crucibles. For each generation of carbon big batch samples, we perform a first run where a medium batch with a size of 25-50g were produced and characterized. Chemical composition and nitrogen sorption characteristics were tested and compared to the small batch reference sample. Comparing lab scale synthesis with a big batch synthesis we observed slight decrease in the nitrogen content BET surface and in the overall porosity. Received carbons were typically in the form of larger agglomerated particles that have to be reduced in size by the high-energy ball milling before sulphur impregnation. For all three generations, similar procedure was applied.

WP 4: Electrolytes and separators : Formulation and Modelling

- Impact of ionic strength of the electrolyte (salt concentration, additives etc).
- Impact of the donor numbers, polarity and the viscosity of solvents.
- Influence of the coordination properties on the diffusion of polysulphides.
- Importance of the composition of the separator.
- Importance of the electrolyte retention in the separators;
- Design of novel ionic liquids suitable for Li-S batteries.

Task 4.1 Review on electrolytes for Li/S batteries
This review was dedicated to both an overall assessment of the current state-of-the-art Li-S electrolytes, but in particular the challenge of the solubility of polysulphides. The connection between the electrolyte solvent and salt chemical and physical properties and the effects on solubility as well as performance and life-length was addressed. The objective of this review was also to provide the basis for the selection of chemical parameters and systems prioritized for the modelling.

Task 4.2 Synthesis of different ionic liquids for Li-S batteries
We started with ionic liquids based on three main cation molecular structures (1-alkyl-1- methylpyrolidinium (PYR1R+), 1-alkyl-1-methylpiperidinium (PIP1R+) and 1-alkyl-3-methylimidazolium (IM1R+) ) in combination with two main cations (Bis(trifluoromethanesulfonyl)imide (TFSI )and Bis(fluorosulfonyl)imide (FSI)). The length of the alkyl chain (R) chemically grafted on the cation was varied for every cation family. First set of 8 ionic liquids have been tested a mixture between ionic liquid and LiTFSI salt in molar ratio 9:1 . All tested batteries showed rapid capacity fading, although some of ionic liquids showed relatively high electrochemical activity in formation cycles. General conclusion from that measurement was that none of tested ionic liquid could be used in the 1st generation of prototype cell. Improved stability and excellent coulombic efficiency was achived with another IL family with TFSI with in the combination DEME cation - N122(2O1)TFSI (N,N-diethyl-N-methyl-N(2methoxyethyl)ammoniumbis(trifluoromethanesulfonyl) imide). The beneficial stability of Li-S batteries with N122(2O1)TFSI IL was attributed to the methoxyethyl chemical functionality grafted on the nitrogen atom of the IL cation. The synthesis of ILs has continued, knowing that methoxyethyl chemical functionality had to be grafted on the nitrogen atom of the IL cation. Several new ILs were then synthesized by Solvionic with the parameters mainly explored in the synthesis effort were the length of the nonmethoxy alkyl chains grafted on the ammonium cations and the anion type. For the purpose of the 2nd prototype generation N122(2O1)TFSI was selected out of a wide range of molecules, 2 L of the IL N122(2O1)TFSI were therefore synthesized at Solvionic, successfully showing the feasibility of up-scaling from a gram scale (50-100 g) to a kilogram scale (2-3 kg).

Task 4.3 Selection of different electrolyte formulations
As electrolyte for the 1st generation of the Li-S prototype cells we decided to use 1M LiTFSI in TEGDME/DIOL (1:1 v/v) (Tetraethylene glycol dimethyl ether / 1.3-Dioxolane). Form an industrial production point of view, this electrolyte was acceptable since TEGDME does not highlight any risk label, while 1.3-dioxolane is classified as highly flammable liquid and vapours.

Most of our efforts were devoted to the electrochemical characterization of the DEME-TFSI (N122(2O1)TFSI) IL. Cells with the DEME-TFSI IL based electrolytes showed very high stability and excellent Coulombic efficiency. One cell was cycled for 450 cycles. We observed continuous capacity fading, but no polysulphide shuttle effect and the Coulombic efficiency was higher than 99%. A higher temperature were leading to visibly higher capacities, but were accompanied by capacity fading. Due to the generally high viscosity of the ILs, detrimental for ion transport, mixtures between organic solvents and ILs were also tested as matrices. Based on both stability and capacity we concluded that the 1:1 mixture of GEN1 electrolyte and DEME-TFSI was optimal for Li-S batteries as it enabled almost the same capacity as the electrolyte from 1st generation and at the same time, it has shown a much better cycling stability. Due to electrochemical analyses of partners involved in the task, the 2nd generation (GEN2) electrolyte was defined as a mixture of 1 volume of GEN1 electrolyte plus 1 volume of LiTFSI:N122(2O1)TFSI (1:9 mol. ratio). The selection of electrolyte formulation for prototyping was also guided by toxicity issues of the organic solvents used in mixtures with ionic liquids. TEGDME, which is the glyme used in GEN1 and GEN2 electrolytes, has the lowest toxicity level of the following series of glymes.

Task 4.4 Evaluation of different separators
Work on separators has become very important in terms of stabilizing capacity fading and with this enabling to meet one of the general objectives in the project: cycling stability at least 1000 cycles. With the help of lithium conducting ceramic separator that conducts lithium and prevents polysulphide shuttle the effective separation was achieved. By use of fluorinated graphene oxide as self-standing films or layers on the top of the commercially available separators we managed to stabilize capacity and apart from the drop of capacity in the formation cycle. There was an incremental improvement compared to battery where only glassy fibber separator was used. Measurements suggested that non-modified reduced graphene oxide (rGO) influence the cycling stability. The fluorinated rGO were further used for preparation of self-standing membranes, which could be used as a separator in Li-S batteries. Finaly membranes containing chemically modified rGO and membranes without rGO were laminated together and tested in Li-S batteries. We found out that membranes without rGO could function as separators, while the laminated membrane did not work – the most probable reason being blocked porosity due to the lamination process.

Typical separators used in all other experiments are mainly glass fibber separators or separators available from Li-ion batteries. Additionally we tested two different types of separators provided by Fraunhofer ISI where the major difficulty in use of that separator was low storage volume for the electrolyte.

Task 4.5 Modelling of the chemical parameters of the solubility of polysulphides
The modelling of Li-S electrolytes have been focused on the interactions of several polysulphides (PS): S22-, S42, S62- and S82- with a) the ionic liquid (IL) cations PYR13+ and PiP13+, and b) the organic solvents DME, DIOX, and TMS. The objective was to address the solubility of PS in different electrolytes, which was expected to be intimately related to the strength of PS-solvent interactions. New ensembles were made in order to prepare additional molecular systems with low energies (stable) to serve as input for the construction of relevant artificial IR, Raman and UV-vis spectra for electrolyte components and dissolved species – including also super-structures.

The overall conclusions:
- Multiple configurations of interacting species were possible and increase in number with PS length.
- The trends in interaction energies: PYR13:PS ~ PP13:PS >> TMS:PS > DME:PS ~ DIOX:PS Li+ coordination to Solvent: PS complexes reveal reversed order – favoured for weaker Solvent: PS complexes.
- Experimental importance of the differences in PS:electrolyte interactions were largely unknown.
- Spectroscopic analysis was possible based on the predicted spectroscopic properties of low energy complexes.

WP 5: Analytical tools

Different in situ and ex situ analytical tools for analysis of Li-S batteries at different stages of charge and discharge with the aim to understand the mechanism of Li-S battery behaviour and to understand the impact of different chemical environments on the Li-S battery electrochemical properties were used:
- XPS (X-ray Photoelectron Spectroscopy) and SAM (Scanning Auger Microscopy) for the study of lithium surface after cycling in different environments, for the mapping of elements and their distribution and for the study of the surface reactivity.
- Different spectroscopic techniques (UV-Vis, Raman spectroscopy, in situ XAS) were coupled with galvanostatic charging/discharging of the battery for study of Li-S battery behaviour and polysulphides formation.
- Electrochemical techniques (iono-selective micro electrode for polysulphides, electrochemical reduction) were used in multi electrode cells for the study of Li-S battery behaviour and polysulphides formation.

Task 5.1 Studies of electrolyte electrode interface using XPS and SAM
Study of a modified zeolite additive in the positive electrode
Most of the beneficial mechanisms have occurred at the cathode side only. The decrease of polarization in the discharge/charge profiles, very well observed with the MnS-1 additive, was attributed to improvement of kinetics at the positive electrode side. This enhanced kinetics was due to better retention of polysulphides in the pores of carbon positive electrode with the additive. Their concentration within the electrode remained rather high (at least upon the first half of the discharge) and their reduction into shorter-chain polysulphides were not diffusion-limited, which resulted in the improvement of the kinetics, with a beneficial impact on Li–S cell performance.

Study of a fluorinated reduced graphene oxide (F-rGO) separator
A careful study of the first electrochemical cycle by XPS has shown that a kind of «activation» process of the F-rGO interlayer was necessary to have the modified separator fully operating. This process occurred at the first cycle and consisted in a reaction of lithium with the free fluorine remaining in the separator to produce LiF. This stage consumed lithium and that a rather weak proportion of S8 has been reduced during this first cycle. At the beginning of the second cycle (2nd discharge), the Li-S was fully ready for reversible lithium exchange. The reversibility of the electrochemical reaction appeared to be much better for the cathode using the F-rGO interlayer separator: more short-chain polysulphides have been re-oxidized into long-chain polysulphides and elemental sulphur upon charge. With F-rGO, no Li2S signal could be detected, and the amount of polysulphides was weaker. This was the proof that the F-rGO interlayer has been efficient in blocking the diffusion/migration of polysulphides, although some of them have reached the lithium electrode.

Effect of polymer ionic liquid (PIL) membranes
The XPS S 2p spectrum of the lithium electrode after the first discharge has detected only short-chain polysulphides at the surface of Li, as indicated by the bridging to terminal sulphur atoms ratio. No Li2S signal could be detected. In this way, the XPS signature of the Li electrode at the end of the discharge was rather similar to that observed when using the F-rGO separator. This proved the efficient behaviour of the PIL membrane as retardant for polysulphides diffusion/migration. However this PIL membrane couldn’t stop the redox shuttle for more than a few cycles and this explains the rapid capacity fading of the cells.

Task 5.2 Study of electronic state of sulphur with XAS
For the purpose of reliable measurements we have developed two types of cells, first was based on “coffee bag” technology and it had 3 µm thick Mylar window which was thin enough to enable measurements in the reflection fluorescence mode. The second type of the cell was based on the cell body with a replaced bottom part of the cell with a 13 µm thick beryllium window. The cell could be used for at least three different spectroscopy measurements. In addition, a chamber with a controlled overpressure of He has been developed which allowed us the use of the cell in the He compartment in the reflection mode.

We had two sets of measurements: sulphur K-edge XANES measurements in operando mode, which enabled us to measure several XANES spectra from the back side of the cathode electrode during discharge and charge process, and ex-situ measurements, with which we have obtained XANES spectra from the cathode side that was faced to the lithium anode. The measurements provided complementary information about the mechanism of polysulfide formation and diffusion in the selected system. The conclusion from that work was that cathode composite that had no sorption affinity for polysulphides was not suitable for Li-S batteries since most of sulphur during first reduction was solubilized as a form of long chain polysulphides who migrated in the direction of the separator and lithium metal.

In the second set of the measurements we confirmed that zeolites have affinity to polysulphides adsorption since during the discharge process much higher quantity of sulphur species remained on the back side of the electrode.

We found out that the quantity of sulphur in the electrolyte was much higher than the quantity of sulphur involved in the electrochemical reaction and for reliable analysis we required large number of variable parameters (three for each shell of neighbours for each S site in the mixture).

For the first time we have measured a battery containing a sulphur free electrolyte. In this way we were able to do a full EXAFS analysis to evaluate possible interactions between the sulphur and the host matrix and/or the electrolyte. The use of a silicalite as adsorption additive, reducing the diffusion towards the separator and anode of the polysulfide, allowed us to have a complete view of the cell, although the penetration depth of the x-rays was limited to less than 10 µm. The major outcome from this work was the fact that we have three equilibrium states within the cell during discharge.

In the context of continuous effort to improve the performances of Li-S batteries reducing or avoiding the polysulphides shuttle mechanism, we have measured in operando conditions in an alternative configuration with a focus to fully understand better the mechanism of the 1st oxidation of Li2S. The battery design allowed us to work at high overpotential in the 1st oxidation which showed a negligible amount of polysulphides formation in the electrolyte during 1st oxidation.

Task 5.3 In situ analytical work using Raman and UV-VIS spectroscopy
The UV-Visible spectroscopy is based on the interactions between polysulphide molecules and the UV-Visible electromagnetic radiation is depending on the length of polysulfide chain, on the alkali metal and on the solvent where polysulphides are dissolved. That in principle enables distinction between polysulphides having both different length of chains and different concentration.

The improved stability of battery with modified separator (using rGO-F) layer has been studied by in operando UV-Vis spectroscopy. Use of a separator with the rGO-F interlayer showed remarkable difference in the mechanism. In contrast to the fibrous separator, observed concentration of long-chain polysulphides remained constant also during the low voltage plateau. Recalculated concentrations of mid and short chain polysulphides were much lower in the case of separator with rGO-F interlayer. We presumed that due to hydrophobic nature of rGO-F the precipitation of end discharge products were more homogenous through all volume of the electrode.

Raman spectroscopy can also play a major role in gathering information on the Li-S reaction in the different components of the cell (electrodes, electrolyte, etc.). To characterize the formation of polysulfides in the electrolyte by Raman spectroscopy we used the newly built cell. The observed mechanism (two sets of peaks corresponding to two different types of polysulphides assigned to S32- and S62-) was in good agreement with the results obtained by UV-Vis and XAS spectroscopies.

Finally, a specific cell was developed for ATR-IR spectroscopy, it allowed us to follow the concentration of solvated Li+ at the electrode side during electrochemical cycling. The results have shown a rapid and strong decrease of solvated Li+ when Li2S precipitated at the end of discharge. On the other hand, the dissolution of Li2S at the beginning of the charge resulted in a rapid increase of the concentration of solvated Li+.

One set of UV-Vis measurements was a support to XAS measurements with non sulphur containing electrolyte. For that purpose we used TEGDME: Dioxolane solvents with LiTDI salt. Generaly there was no difference between polysulphide diffusion into separator between LiTFSI and LiTDI based electrolytes. We summarized that the diffusion of polysulphides follows the same sequence in the presence of LiTDI salt in the electrolyte as observed in the LiTFSI based electrolyte.

Task 5.4 Reliable cell for in situ electrochemical analytical work
The application of 4-electrode modified Swagelok cell used in the study of polysulphide shuttle mechanism with rGO-F modified separator was presented. These results were in the agreement with the obtained mechanism with UV-Vis spectroscopy what indicated that developed techniques were reliable. Namely, we were able to get the same information by using two independent analytical tools which principle of operation is completely different.

Additionally, we have worked on the impedance spectroscopy. The comparisement of the galvanostatic method (a non-linear excitation method) with the impedance method (a linear excitation method) presented a potential problem in the interpretation. Both methods were comparable only in cases where the degree of nonlinearity was small enough. In order to check the possible effect of non-linearity on impedance data, we have had to develop and perform so called dynamic impedance measurements (using a current excitation signal so we also call it “galvanic impedance”). To compare both methods we performed additional measurements by using the high-amplitude electrochemical noise and transforming the data into impedance. We made two important conclusions. The first was confirmation of the meaning of the high-frequency arc which includes the resistances of contacting including the resistance of passive layers (that is all contribution due to charge migration). Most importantly, by performing the dynamic impedance measurements across the whole charge-discharge curve, we could confirm that ca. 20-50 % of voltage was lost due to contacting, passive layers, electrolyte migration etc., while the rest was lost due to electrochemical reaction including diffusion of species to reactive spots.

Task 5.5 DFT modelling of spectroscopic measurements
Insight into the absorption characteristics of individual species were obtained via DFT modelling of UV-Vis and Raman spectra for isolated and complexed polysulphides in different environments. The simulated spectra aid the experimental identification of polysulphide species in Li-S battery electrolytes by predicting the contributions of pre-defined species. The simulated UV-Vis spectra were based on the thirty electronic transitions of lowest energy and the bands were plotted assuming a spectral half-width of 30 nm. With the number of transitions included in the simulated UV-Vis spectra, the dominant absorption is in the UV region. As the lengths of isolated polysulphides were increased, the absorption was shifted to longer wavelengths. The continuum approach best reproduced the experimental trends observed ex situ in the ~400 nm region and the S3*– radical absorption at longer wavelengths. An explicit organic solvent or large cation coordinating to S22-, on the other hand, had only a very small effect on the position and intensity of the absorption band. Therefore, weaker spectral changes with respect to an implicit only approach were expected for the longer polysulphides using a similar combined explicit/continuum approach.

Raman spectra were simulated with the vibrational energies having a Lorentzian broadening with a half-width of 5 cm-1. We have shown that UV-Vis and Raman spectra of polysulphides could be simulated using different models for the polysulphide surroundings.

WP 6: Benchmarking of other Li-S technologies

One of the main problems of Li/S technology comes from the solubility of polysulphides. This solubility induces a loss of positive active materials and modification of the SEI layer on the lithium negative electrode. This mechanism was detrimental to the capacity retention. In this Work Package, we envisaged to replace the negative lithium by alternative Si electrode. Another possibility that was evaluated consisted of the preparation of thin ceramic membrane and all solid-state Li/S batteries. Alternative ionic conductive coating of particles by using ionic liquids polymers was performed and tested as an alternative approach.

Task 6.1 Development of ceramic membranes
- Synthesis of LZCP
Li1.2Zr1.9Ca0.1(PO4)3 (LZCP) has been synthesized. The electrochemical stability of LZCP versus lithium has been tested by mixing LZCP with carbon as an electrode versus lithium. As a conclusion, we reported a new ionic conductor LZCP with a good conductivity and with a better electrochemical stability than LAG compounds.
- Tape Casting preparation of LAGP
We have developed our proper ceramic membranes that were based on LAG (Li1.3Al0.3Ge1.7(PO4)3) and have been prepared by the tape casting process. By optimizing the tape casting method, we succeed to obtain dense ceramic membranes with good mechanical properties. These ceramics could be easily designed to any shape and several ceramics membranes that could be introduced into coin cells were produced. The ceramic membrane was acting efficiently with very stable capacity and 100 % of coulombic efficiency.
- Use of ceramic separators in Li/S batteries
The function of ceramic separator has been demonstrated by using commercially available lithium conductive ceramic membrane obtained from OHARA. To enable better ionic contacts, we put a few drops of electrolyte on both sides of ceramic membrane what enabled us to have a cell with electrochemical characteristics of normal Li-S battery. Importance of efficient separation of sulphur cathode and lithium had been clearly demonstrated by long cycle life and very high coulombic efficiency. This was possible while ceramic membrane conduct only lithium ions and it prevents diffusion of polysulphides to the lithium side of battery. We achieved cycle life of 600 cycles however, after long term cycling other problems related to Li-S battery technology has appeared.

Work on Solid State Batteries:
- Cold press method: We have designed a home-made system for the evaluation of different active materials and electrolyte in solid state batteries. We have identified a new electrolyte Li6PS5Cl coming from the argyrodite family. We concluded that solid state batteries using C/S active materials are a very interesting alternative to classical Li/S batteries since we were able to get close to the full theoretical capacity of (1450 mAh/g) at current density C/20.

- SPS solid state batteries: Our goal was to assemble solid state batteries by SPS and to have self-standing batteries. We were able to assemble a full solid state batteries and by testing we had identified that there was a problem in the interface between Li2S and LAGP.

Task 6.2 Use of alternatives negative silicon negative electrodes
Alternative electrode for lithium can be a key point for this technology. We have optimized the electrochemical properties of Li2S compound which is well known after its very low electronic conductivity. We had found out that for good capacity retention we needed an optimized silicon electrode and C/S (or Li2S) electrode but at the same time both of them needed their own electrolyte.

Task 6.3 Use of polymeric ionic liquids
We prepared self-standing films and use them as a separator and secondly we prepared composite cathode with a self-standing film by infiltrating PIL based electrolyte into the porous sulphur cathode. Both Li-S batteries configurations were considered as an all solid state configuration which should in theory stop polysulphide solubility and consequently redox shuttle mechanism. Result at elevated temperature has shown the possibility of using PIL membrane as ion conductor in the Li-S batteries. Impregnation of the sulphur cathode with PIL resulted in improved/increased contact between sulphur and PIL where PIL membrane could effectively prevent polysulphide redox shuttle.

The best electrochemical performance was obtained by battery configuration where additional electrolyte has been added only on the lithium side with additional Celgard separator between PIL membrane and metallic lithium as a buffer for the electrolyte storage. The Coulombic efficiency close to 100 % was suggesting the complete reversibility of the charge transfer, but capacity fading pointed out some problems related with accessibility of active material. Additionally we observed very well preserved lithium surface which confirmed the potential of PIL membranes which not only stop polysulphide redox shuttle but also prevent degradation of the lithium surface.

The PIL based electrolyte used in this study was proven to be very promising for Li batteries. The study has already shown that our ionic liquid polymer based electrolyte prevents from polysulphide redox shuttle.

Task 6.4 Flow battery using catholyte and ceramic membrane
The idea of working on flow batteries was coming from for the sulphur electrode where the dissolved polysuflides in the electrolyte were used as active material. Our goal was to upscale these results to high quantity of polysulfides. Polysulfides were the positive electrode, our flowable negative electrode was silicon. Good electrochemical performances were obtained with silicon when it was tested as a slurry despite the huge volume expansion of silicon upon lithiation.

WP 7: Integration, scale up, testing, life cycle assessment and benchmarking

The key objectives for each generation of cathode and electrolyte design were:
- To assemble coin cells with matched electrodes with the suitable capacity per surface for testing and wound cylindrical cells in accordance with the target of energy density improvement for testing;
- Assess the performances of the different generations of Li-S full cells and compare them to available Li-ion
batteries and safety behaviour of the different generations of cylindrical Li-S;
- From the performance and safety results and perform failure mode analysis and identify possible axes of improvements.

Task 7.1 Laminated cathode
During the development and manufacturing of the carbon sulphur cathodes, three major challenges were addressed: a low drying temperature (< 90 °C) in order to avoid sulphur sublimation, a good adhesion on the current collector for the winding of the electrodes, a high porosity for optimal electrolyte diffusion and good conductivity for maximum performance.

The focus was on the development of using the aqueous binder systems in the electrodes. The major problem that had to be addressed was the high viscosity of the electrode slurry due to the high surface of the cathode and the addition of MWCNT. For the aqueous binder system a new formulation was developed: 85 % Active Material (C/S), 7 % Carbon black (C45), 1% CNT, 4% Binder (SBR) and 3% Binder (CMC). The prepared electrodes have shown a high adhesion on the current collector and no Al corrosion.

Task 7.2 Prototype assembly
Impregnation method for GEN 1 cell assembly (electrolyte introduction under vacuum) has led to very high excess of electrolyte relatively to the electrode and separator porosity. In order to reduce this electrolyte excess, a maximum of 50 % excess of electrolyte has been used, but the risk that all porosity has not been filled was possible especially since GEN 2 electrolyte contained ionic liquid making the electrolyte more viscous. First cells with GEN 2 electrolyte were assembled to check the capacity. Capacity at 1st cycle at C/20 has been only 550 mAhg-1. Therefore the 12 prototype cells for GEN 2 were assembled using GEN 1 electrolyte. 6 cells were sent to Renault in mid of June for testing by using EV profile and 6 cells have been tested at Saft for initial performance, high rate dischargeability, low and high temperature, self-discharge at RT and 45 °C and cycling at RT & 45 °C measurements. For GEN3 we have reduced the electrolyte quantity in the cells (7 to 13 µL/mgS for GEN3, in comparison with 16 µL/mgS for GEN2). Therefore, prototype energy density was significantly improved for GEN3 compared to GEN2 cells. The initial performances were promising, but the cell life for GEN3 was short.

Task 7.3 Life cycle testing
GEN1: Prototype cells with electrolyte excess shown high capacity but lower than laboratory cells whereas cycle life was comparable. More than 150 cycles were performed. Prototype energy density was low due to high electrolyte excess, low cathode loading and thick anode. This generation of samples has enabled a better understanding of prototype behaviour upon cycling, which led to adaption and improvement of our cycling protocols.

GEN2: Discharge capacities were close to 600 mAh/gS (C/3) at the beginnings of the cycling which was in good accordance with capacities measured initially at C/2 or C/5. Average capacity losses after 50 cycles were -12 %, after 100 cycles were -20 % and after 150 cycles were -28 % (based on C/3 discharged capacities). A good reproducibility was observed during RT cycling.

GEN3: Initial capacities obtained promising values were above 1000 mAh/gS. The best performances were obtained with the Design 2 (DME electrolyte): we have obtained around 700 mAh/gS at C-rate. Average capacity losses after 50 cycles were significant: between -40 to -56%. Charge efficiencies were above 85% on first cycles but they were decreasing very quickly during cycling. One Cell of Design 4 has undergone a sudden death at 70 cycles. The cell life for GEN3 was very short in comparison to GEN2.

Task 7.4 Safety tests
Initial performance and cycling under constant current: At the beginning, it was observed that measured capacity is very sensitive to the discharge rate, particularly between C/2 and C rate. The capacity fading was relatively high at the beginning of cycling and become quite acceptable after 20 cycles. At RT, GEN 2 cells with GEN 1 and GEN 2 electrolytes were cycled up to 500 cycles. GEN 2 electrolyte was not selected for the other tests since capacity was too low at C/5 discharge rate. The cells with GEN 1 electrolyte exhibit a regular and linear decrease of capacity up to 300 cycles. The 20 % capacity loss representing the end of life was reached after around 100 cycles. Cycling at 45 °C has been carried out and 175 cycles have been done. Capacity was decreasing regularly and end of life criterion was reached between 20 and 40 cycles. Observation of the disassembled cell with GEN 2 electrolyte has shown very dry separator and very fragile lithium electrode. In our opinion is one of the major causes for aging was the electrolyte consumption.

DST cycling at RT and 45°C: GEN 1 durability at 45 °C was unconfirmed by reproducibility experiment. GEN 2 prototype revealed higher initial capacity (266 to 289mAh, vs 184-202 expected). These results reflected both improvements in chemistry and cell design but also earlier test initiation. At 25 °C, 75 % of initial DST discharge capacity was reached within 30 cycles. At 45 °C a stronger decrease in capacity was observed than that for GEN 1. With DST cycling tests on GEN 2 cells, a continuous capacity fading has been observed.

Task 7.5 LCA
Conducting a Life cycle assessment (LCA) in our case includes mining, processing and transporting the raw materials as well as the production of the cells and its components. For this purpose, a model was developed in the Software GaBi. The model based on a given number of input parameters that need to be available in order to make the assessment. The model covered three different cases of Li-S batteries, representing the current best performance as well as two future performances meant to illustrate what environmental benefits that could be found if the development progresses.

Initial evaluation of the results from the assessment showed that the lower weight of the Li-S pack is a very critical factor for the environmental performance, in many of the interesting categories. As it was assumed that the battery runs on EU average electricity, there was still some CO2 and other emissions released in the use phase and as the use phase spans over such a long time, even a small change in weight could be very important for this life cycle stage. Additionally, a lower weight means less material used, which lowers the impact from the manufacturing and extraction phases. The first evaluation of the results also showed that the electronic components in the packs have a large impact on the overall environmental impact, and that the decreased amount needed for the Li-S cases imply an improvement.

Potential Impact:
Lithium sulphur batteries are considered as a next generation of high energy density post Li-ion cells that will have impact on increased driving range. Technology developed within EUROLIS project show that cells are appropriate for use in electric cars, however cycling stability and power should be improved before large scale commercialisation. During the project, we have established a chain of knowledge where each partner has an exact role in the production of prototype cells. For that we have used components which are sustainable and cheap and with that enable low carbon foot print. This is of high importance since not only capability of batteries to have high energy density, but also energy consumption during production has a huge impact on CO2 footprint. Furthermore, established chain of knowledge about the production of Li-S cells enables further development of Li-S cells, specially within direction of assessing safety issues and ageing of cells and battery packs. Bringing this technology closer to the market have impact on different segments of industry, primarily on automotive industry since OEM’s can get information’s about the performance from prototype cells build at research lab of battery producer. Another segment of industry that can have benefits from Li-S batteries are space and aeronautic industry and others where a huge demand for energy density cells is present. Commercialisation of Li-S cells will for sure have impact on some new technologies which are not feasible with Li-ion cell which are currently used and at the same time commercialisation of Li-S cell and their use in electromobility will enable much longer distances driven without charging.

Main dissemination activities

The dissemination activities have been focused on stakeholder in the European scientific and research communities. Two dissemination workshops have been held during the ALISTORE-ERI meetings as has been planned in DoW. First dissemination workshop was carried out in June 2016 at ALISTORE-ERI meeting in Delft, Netherland. Second dissemination workshop has been performed at two different events. This decision has been made within the consortium to target two different target groups: industrial partners, European research/ scientific target groups. The project results have been presented at EUCAR annual reception and conference by two project partners (NIC and SAFT) and ALISTORE-ERI meeting by three project partners with four presentations (NIC, CNRS/IPREM and CNRS/UM). All the activities have ensured the transfer knowledge among industrial, research and academy stakeholders.

Knowledge management
All partners have fully agreed on consortium agreement (CA) on 15 July 2013 and the CA have been signed on 10 September 2013. There have been no annexes or amendments to CA.

At the beginning of September 2013 NIC started with a process of filling in a patent application to protect the inventive approach of polysulphide blocking separator that can be used in Li-S batteries. The application was submitted Slovenian Intellectual Property Organization under no. SI20130000414 and on 27.02.2014 the application was filed at European Patent Register (no. WO2015088451). The inventors are: Alen Vižintin (NIC), B. Genorio (CO-NOT), M. Gaberšček (CO-NOT), R. Dominko (NIC), and the title of patent application is: Chemically converted graphene as a separator for lithium sulphur (Li-S) batteries.

The plan for the use and dissemination of foreground has been prepared. The plan contains beneficiary’s preliminary intention towards exploitation the project results to support their activities as laid down in the initial version of the project DoW, and in relation with the CA and the GA.

The exchange of the results and ideas between project EUROLIS and project STABLE has happened in Ljubljana on 14th of April 2014. The second exchange of results was carried out with on project STABLE clustering workshop held in Brussels on 28th of May 2015. The coordinator was also present on Battery Materials Cluster Brainstorming Workshop in Brussels on 29th May 2015 where the exchange of ideas happened between different project representatives.

List of Websites:
The project website [] was in fully functional for the external (public) visitors in February 2013. For partners, we have established also internal confidential part of the website. The website was technically supported, maintained and refresh by NIC. The domain name was register for the period of 5 years’ ant the website will be maintained at least 1 year after the project end.

To understand the usage of the website, we have monitoring the website visits with Google Analytics statistic programme and is shows:
- 4.621 sessions;
- average number of unique visitors per month is 70;
- 32,9 % of returning visitors;
- 4,84 number of pages per session;
- 67,1 % of new visitors;
- 2:34 is average visit duration;
- 69,4 % of visits in Europe;
- Number of internal web site users: 20

Logo, PPT, project letter, deliverable and reports templates has been prepared.