Advanced, High Performance, Polymer Lithium Batteries for Electrochemical Storage
SAPIENZA INNOVAZIONE CONSORZIO
Viale Regina Elena 295 Pal Uff Complesso Regina Elena
€ 205 932
Marina Busi (Ms.)
Sort by EU Contribution
CHALMERS TEKNISKA HOEGSKOLA AB
€ 646 318
€ 68 038
€ 99 360
ETC Battery and FuelCells Sweden AB
€ 340 696
UNIVERSITA DEGLI STUDI DI ROMA LA SAPIENZA
€ 847 654
SAES GETTERS S.P.A.
€ 164 785
Stena Metall AB
ZENTRUM FUR SONNENENERGIE- UND WASSERSTOFF-FORSCHUNG BADEN-WURTTEMBERG
€ 586 339
UNIVERSITA DEGLI STUDI GABRIELE D'ANNUNZIO DI CHIETI-PESCARA
€ 163 003
STENA RECYCLING INTERNATIONAL AB
€ 97 980
ROCKWOOD LITHIUM GMBH
€ 112 193
Grant agreement ID: 265644
1 June 2011
31 May 2014
€ 4 601 443
€ 3 332 298
SAPIENZA INNOVAZIONE CONSORZIO
Final Report Summary - APPLES (Advanced, High Performance, Polymer Lithium Batteries for Electrochemical Storage)
Trade-offs between choice of no toxic substances in production and performance in their use phase is the present challenge of substituting fossil fuels for car propulsion with batteries. The substitution success depends on battery performance, which is not yet sufficient. Accordingly, the main goal of APPLES was to fill this gap by designing and developing at industrial scale an advanced lithium ion polymer battery, with the following characteristics:
The nanostructured Sn-C electrode shows very stable performances in Li cells. The particular carbon matrix ensures an efficient buffer for large volume variations usually occurring during Li-Sn alloying/dealloying processes and guarantees that the electrode retains proper electrical contact in all the investigated conditions. Innovative techniques have been proposed for the preparation of nanostructured Sn-C powders. In particular, relevant success has been obtained by the fine optimization of the Sn-C composite synthesis according to a new strategy based on the electrospinning technique It was shown that the synthesis protocol used at laboratory benchmark allows producing batch sized of 5-10g of Sn-C electrode materials. The issues relating to the scaling up was afforded and solved by the consortium.
Target of this activity was the development of a high voltage spinel cathode (LiNixMn2‐xO4) for a 1 – 2 Ah cell in pouch design. The overall development process from the active material powder to the optimized cathode configuration followed several iteration cycles for fine-tuning. According to this line an optimized cathode has been developed. The resulting electrodes have been calendared to their final density. Electrochemical tests confirm the high reproducibility of the developed process. The optimized electrodes were implemented in the final cells.
The strategy of the APPLES project was to investigate the possibility to fully or partly replace LiPF6, to investigate the combination of other organic solvents, and to introduce ionic liquids as the solvent. For the Li-salt the full or partial replacement of LiPF6 with LiBOB (Lithium bis(oxalato)borate) was considered since it has the potential of having a fluorine-free salt and also the fact that LiBOB is known to form stable passivation films during the first cycle on the electrode, improving the safety of the battery. One of the objectives was the optimization of the electrolyte materials with respect to performance, safety and adaption of synthesis procedure for up scaling. A second step was the implementation of a gel polymer electrolyte membrane instead of the normal separator. The work builds on two concepts previously developed by the partners based on gel casting of membranes and on electro-spun membranes.
Materials scaling up
The APPLE project orientates on a pre‐industrial production level, which ensures wide availability of all materials. Thus a transfer of laboratory work to larger scale processes was the main scope. Spherical high density LMNS was available as a blend in 4 kg scale. A double side coated cathode has been developed with this material and with optimized loading. The process for large batch Sn/C preparation, properly developed, consisted of high temperature thermal treatment procedures with subsequent high-energy ball milling steps. The scaled Sn/C anode material was optimized with respect to its tin content and particle size. For the prelithiation of the Sn/C anode material fine lithium metal powder can be employed. A significant improvement can be reached in terms of processability and control of the degree of lithiation by adequately chosen amounts of Li powder. The polymer electrolyte scale-up activities were targeting on (i) reduction of the membrane thickness from 700 µm to 50 µm to obtain suitable membranes that match the requirements for cell assembling and (ii) enlargement of the membrane from several cm² up to 28 x 24 cm².
High performance getters with high absorption capacity
Safety is a key requirement for EV batteries. APPLES project has addressed this issue by developing getters to trap gases, which may evolve due to unexpected battery faults that in turn may induce serious risk of fires or even explosions. The gassing behavior of APPLES cells was studied, revealing that the internal pressure increases very rapidly during the cell formation, and it keeps on increasing (although with slower rate) during cycling. The gases formed by cycling APPLES-type cells are mainly CO2 and H2. It is believed that a getter able to sorb one or both of these gases could help to control the internal pressure, enhancing therefore the cell safety and performance. A getter for CO2 based on LiOH was designed, manufactured and tested, revealing high sorption capacity for CO2 and compatibility with battery electrolytes.
Target of the project was the development of high voltage prototype cells on a pre-industrial manufacturing line. The following activities have been carried out: Step I: definition of the lay out design of the cell. Step II: electrode development. Step III: evaluation of different electrolytes. Step IV: pretreatment of electrodes. Step V: prototyping; several iteration cycles between the individual steps had been carried out for fine-tuning. Different cell chemistries, electrolytes and cell concepts have been taken into account. For comparison cells with commercially available materials have been developed in parallel in comparable design as a reference. Two different design concepts have been taken into account. a) a multilayer-stacked design, and b) a modular bi-cell unit, which follows an alternative process of cell stacking. Based on these designs, overall three generations of high voltage cell prototypes have been developed successfully within the scope of the project. The results obtained by the prototype pouch cells clearly confirm the behaviors observed in the small cells and give a positive technical outlook on this chemistry. All three generations cells show reasonable performance and reach their intended capacity.
Recycling of production waste and end-of-life Li-ion (EOL) batteries covers the important aspects regarding battery recycling, from solutions for recycling of production waste to the recycling of individual end-of-life battery cells. APPLES dealt with the following aspects:
• Recycling plan for production waste: the waste from the production of battery cells consists of waste from the input materials (mainly different foils), waste from cutting the pouches, discarded cells and slurries from coating and electrolyte;
• Collection plan for EOL batteries: APPLES batteries are legally covered by two Extended Producer Responsibility (EPR) systems, the ELV-directive (End of Life Vehicle) and the WEEE-directive (Waste of Electrical and Electronic Equipment);
• Recycling plan for EOL Li-batteries: the proposed recycling process for Apples cells consists either on mechanical shredding or on a pyrolysis step.
Industrial manufacturing process
A production process for industrial manufacturing of the lithium-ion batteries has been defined. The development of the process has been based with the end use in automotive applications. The final Apples cell under optimal conditions may reach 57.2 Ah at a weight of 865 grams and a nominal voltage of 4.7 Volts. This is equivalent to an energy density of 310.7 Wh/kg. The material cost is estimated to 13.8€ per cell with high uncertainty in cost of the active electrode material. The total production cost including electrode production, assembly and waste management is estimated to 30.7€ per cell, excluding labor. An energy optimized battery pack has been designed using the APPLES industrial size cell. Battery dimensioning of power and energy needs has been performed by simulating the NEDC European drive cycles. The resulting battery to supply 200 km of driving distance was a 236 kg battery with 176 cells divided into two battery packs comprised of 88 cells each. The nominal voltage was designed to 413.6 V and the total capacity of the system 114.4 Ah of which 89.2 Ah were usable. The usable energy performance reached 156.6 Wh/kg where 64.6% of the system mass consist of cells. These favorable aspects resulted in an overall cost reduction at system level of 100 €/kWh. This is well below the costs indicated at the beginning of the project, which was 150 €/kWh.
Environmental and sustainability assessment
Within APPLES environmental sustainability has been assessed by Life Cycle Assessment methodology (LCA) following the ILCD handbook prepared by the EU JRC at Ispra. Three other techniques, namely mass flow analysis (MFA), technology change analysis and risk assessment were used. Mass flow analysis methodology were used to compare the total need for metals in APPLES batteries if they were used in large scale for all cars and the present global production and reserves of the metals.
Project Context and Objectives:
The present energy economy based on fossil fuels is at a serious risk due to a series of factors, including the continuous increase in the demand for oil, the depletion of non-renewable resources and the dependency on politically unstable oil producing countries. Another worrying aspect of the present fossil fuel energy economy is associated with CO2 emissions, which have increased at a constant rate, with a dramatic jump in the last 30 years.
The urgency for energy renewal requires the use of clean energy sources at a much higher level than that presently in force. The CO2 emissions, and the consequent air pollution issues in large urban areas, among other urgent actions, also requires the replacement of internal combustion engine (ICE) cars with ideally, zero emission vehicles, i.e. electric vehicles (EVs) or by controlled emission vehicles, i.e. full hybrid electric vehicles (HEVs) and/or plug-in electric vehicles (PHEVs).
Lithium ion batteries are seen as the power sources of choice for sustainable transport because they are considered the best options, which can effectively guarantee the progressive diffusion of HEVs, PHEVs, and EVs at high levels. In HEVs the synergic combination of ICE with an electrochemical battery provides high fuel utilization with proven benefits for fuel economy and therefore, for pollution emission control, as well as favoring driving performances which are similar if not superior to those of pure gasoline cars.
The main objectives of the project were to:
1. Optimize preparation procedures of the electrode material;
2. Improve the demonstrative battery behavior in terms of safety, energy density, cycling life and power capability;
3. Integrate the battery system for potential automobile applications;
4. Define the recycling process of battery once it has reached its spent condition.
The achievement of the above mentioned objectives lead to the realization of demonstrative 300 Wh/kg Lithium Ion cells based on nanostructured Sn-C anode, on optimized gel polymer membrane as the electrolyte and high voltage cobalt-free cathode.
Moreover, despite the cells scale up process resulted to be technically very complicated because of the membrane application, the results obtained by pouch cells clearly confirm the behaviors observed in the small cells and give a positive technical outlook on this chemistry. Compared to coin cells the pouch cells work more stable. Precious information and the direction to develop and improve this technology have been obtained.
1 - Optimized anode configuration
The objective of this activity was the synthesis refinement of the anode materials. In particular, starting from the synthesis procedures already established in the laboratories involved, innovative techniques have been applied. For the anode material a lithium-tin (Li-Sn) metal alloy has been proposed. This alloy is one of the most promising negative electrodes proposed to replace the common carbon based materials, showing a specific capacity of the order of 990mAhg-1. The main issue of the use of Sn-based anode is the large volume expansion-contraction, which occurs during the cycling process. Within the project, this problem has been solved by optimizing the Sn synthesis moving to a nanostructured configuration able to control the large volume change.
Different approaches have been attempted. The first one iconsists on the optimization of a synthesis procedure based on the infiltration of an organometallic tin precursor in an organic resorcinol-formaldehyde gel, followed by calcination under argon. The composite nanostructured anode material has been characterized by XRD diffraction, BET measurements, SEM and TEM analyses and by XPS tests. The mechanical properties of the Sn-C composite material have been also studied by anelastic spectroscopy (AS) technique. The morphological analyses confirmed the presence of sphere-like tin particles confined within the carbon matrix with proper particle distribution. XRD experiments revealed the presence of crystalline tin phases, and in particular the quantitative analysis reported the presence of about 88% of Sn (β phase), 7% of SnO (Romarchite) and 5% of SnO2 (Cassiterite). The mechanical properties of the Sn-C material, studied by anelastic spectroscopy tests, evidenced that in this sample there were not any parasitic dynamic processes. Preliminary electrochemical tests have been carried out in lithium-based cells by using conventional LP30 electrolyte. The results of prolonged galvanostatic cycling test at 100mA g-1 clearly evidenced the high irreversible capacity at first cycle, due to the formation of a solid electrolyte interphase (SEI). However, this irreversible process does not affect the electrode stability and cycling ability, so that the electrode is able to deliver more than 400 mAh g-1 after 100 cycles. The electrochemical characterization of the Sn-C electrode has been also investigated at low-temperature (limit temperature ≈ -30 °C) at different C rates and temperature ranges. It has been shown that the cell can be cycled at low temperature (T below 0°C) and at high C rate (up to 1C). In particular, at C/4 (100 mA g-1) the cell retains 65 % and 25 % of the initial capacity at T = 0 °C and T=-30 °C, respectively.
The second synthesis approach development comprised the use of the high-energy mechanical milling (HEMM) technique to produce a tin-carbon composite. The HEMM consists of different steps that should be carefully controlled. The composite morphology and structure have been studied by scanning electron microscopy, transmission electron microscopy and X-ray diffraction, and its electrochemical behavior has been characterized by cyclic voltammetry and galvanostatic cycling in lithium cell. The electrode evidenced highly nanostructured morphology and enhanced lithium–tin alloying–de-alloying process stability, with capacity ranging from 500-to 400 mAh g−1. Further important characteristic of the tin–carbon nanostructure reported was the very high rate capability, extending up to 2 A g−1, that finally allowed to its application in a high voltage, high rate lithium ion cell using the LiNi0.5Mn1.5O4 spinel cathode. The cell showed a working voltage of 4.3 V and a capacity of 120-mAh g−1 obtained at 1 C rate. The very promising features of the cell, its high energy and power density, and the low cost of the involved materials, suggest that the electrode reported here can be efficiently used as an anode in advanced configuration lithium ion battery.
Relevant success has been obtained by the fine optimization of the Sn-C composite anode synthesis according to a new strategy based on the electrospinning technique. An anode material composed by metallic nanosized tin particles embedded in electrically conducting porous multichannel carbon microtubes (Sn-PMCMT) has been synthesized by coelectrospinning followed by stabilization of the composite fibers in air and carbonization in Ar/H2 atmosphere. The physicochemical and electrochemical properties of this composite, nano-engineered material have been extensively studied. An Sn-PMCMT electrode with 5 wt. % of Super-P carbon and 10 wt. % of Na-carboxymethylcellulose binder exhibits a superior rate capability and exceptional cycle life in cell tests at room temperature. This electrode delivers a discharge capacity after 10 cycles of 632-mAh g-1 at 0.7 C with 99.3 % of efficiency and of 460-mAh g-1 at 1.4 C with 99.9 % of efficiency after 100 cycles. Even at the very high rate of 10 C the electrode still delivers 138-mAh g-1 after 200 cycles. These outstanding figures represent the highest performances reported so far for this type of electrode. Scanning electron microscopy and transmission electron microscopy confirmed the nanostructure of the electrode and that the nanosized Sn particles are well embedded in the carbon host matrix. An activation protocol aimed at maximizing anode mechanical stability and capacity retention upon cycling has been proposed. and the results obtained have been compared with those obtained with an anode using pristine PMCMT carbon as active material. The Li uptake and release processes by Sn and C have been evaluated by galvanostatic charge/discharge cycles, in order to differentiate the two contributions to the overall anode capacity. Electrochemical impedance spectroscopy (EIS) analysis has been utilized in order to evaluate possible improvements to charge-transfer kinetics due to the nanosize Sn particles dispersion. Finally, the performances of the composite Sn–PMCMT anode have been characterized at different charge/discharge currents and temperatures regimes. The Sn-C anode can deliver a capacity of 500 mAhg−1 for more than 300 cycles, most of them at 1C or higher charge/discharge rate, which confirms its very high, stable cycling performances. Moreover, the tailored nanostructured anode retains a relevant amount of capacity even in very demanding cycling conditions, as the case for very low temperatures. These results make the proposed Sn–PMCMT an ideal candidate anode for high-performance Li-ion batteries able to operate in a wide array of operating conditions.
2 - Optimized cathode configuration
Target of this task was the development of a high voltage spinel cathode for a 1 – 2 Ah cell in pouch design. The overall development process from the active material powder to the optimized cathode configuration can be divided into the following main steps. Step I: preliminary definition of the cathode specification; Step II: materials selection; Step III: optimization of the electrode; Step IV: process development for large-scale coating; Step V: electrode production. According to this line an optimized cathode has been developed. Several iteration cycles between the individual steps have been performed for fine-tuning.
Step I – specification: the development has been started with preliminary calculations of the required cathode loading. Calculations were based on the energy density targets and on the final cell design. Electrochemical properties of the cathode material LMNS, LiNi0.5Mn1.5O4 and of the anode materials graphite and tin-carbon, have been taken from first small scale samples. A reasonable range for the active material loading both for cathode and anode has been defined. In later states of the development the model has been adjusted to the experimentally obtained electrochemical data of the optimized materials.
Step II – material selection: the materials selection was executed via an iterative approach. It was directly attended to the materials development and optimization steps. The focus of selection has been laid both on electrochemical performance and processability of the powder materials. The influence of particle size and particle size distribution, tap density, crystallite size, crystal structure and stoichiometry has been investigated in order to select the most suited material. For this purpose a close interaction between material and electrode development took place. As a final material a stoichiometric LMNS with spherical morphology and a tap density between 2.3 and 2.4 g/cm3 was selected for electrode optimization and production. The selected material features a single high voltage plateau and high discharge rate capability.
Step III – electrode development: For electrode development various dispersions and coatings have been tested in small scale. Batch sizes were about a few grams for each batch. A small lab coating system was used. This scale was chosen as a reasonable compromise between material availability and obtainable information. In principle the electrode design can be optimized in this scale. Loading, homogeneity and porosity are key parameters. The loading has been controlled by the wet film thickness. During the development various loadings have been applied. Different ratios and types of active material, binder and conductive agent have been tested and best-suited components have been selected from the screened samples. The electrodes were compacted by different methods such as application of static pressure and calendaring and the influence of densification has been investigated. A standard electrode composition with a ratio of 90% active material, 6% carbon black and 4% binder finally was chosen. Several binder types have been tested and their influence on the sticking behavior and mechanical stability of the film has been evaluated. Best effect on the adhesion has been observed for a PVDF binder P5130 (SOLVAY SOLEXIS). As conductive additive, Super P (TIMCAL) showed the best performance. Graphitic additives have been avoided due to their electrochemical behavior at high charge potentials. Active material loadings have been varied between 6 mg/cm2 and 18 mg/cm2. Higher loadings are still working reasonably from an electrochemical point of view, however may increase rigidity of the electrode and reduce the power performance. The finally selected loading of about 15 mg/cm2 showed the best balance between energy density of the cell and rate capability of the electrode. The specification of the optimized electrode has been transferred into the next step as a target for the manufacturing process.
Step IV – manufacturing process: The electrode manufacturing process can be divided into three sub-tasks, which have to be optimized separately. These are a) slurry preparation, b) coating process and c) calendaring. All steps first have been optimized for a single side coating on an intermediate size coil coater in batch sizes of 200 g. After optimization parameters were transferred to double side coating. The sub-tasks of Step IV in detail: a) A slurry has to fit into several specifications to be suited for coatings in larger scale. Important for the processing behavior and electrode quality are the rheological properties and the homogeneity of the dispersion. Both are strongly dependent on mixing conditions, solvent, composition, and the binder type used. Furthermore, the type of the active materials, their chemical composition, surface properties and particle size distribution play an important role. Different conductive additives may be used and the mixing equipment and parameters are to be adapted to the disperse systems. All these parameters have been investigated in depths; correlations have been identified and used for further optimization. b) For the coating procedure wet film thickness, solid content, drying temperature, time and atmosphere may be critical factors and have been taken into account. Correction factors for the deviation between theoretical wet coating thickness and final real loading had been investigated and applied. Good adhesion between coating and metal foil are necessary as well as homogeneity of the materials distribution and the mass loading. All these parameters have been carefully optimized during development of the manufacturing process of the LMNS electrode. SEM and EDX investigations showed that the developed coating process leads to cathodes well suited for cell manufacture. The electrodes show a homogeneous distribution of the active material, binder and carbon black. The adhesion is high and the electrochemical performance of the large-scale coated electrode corresponds to the behavior of the small-scale electrodes. c) The resulting electrodes have been calendared to their final density. The influence of single and multiple calendaring steps has been investigated and electrode kinetics has been optimized.
Step V – electrode production: for higher cell capacity a stacked cell design is needed. This design requires double side-coated electrodes. For production of the double side coated electrodes a process transfer from the medium size coil coater to the large size coil coater had been carried out. The transfer of recipe and coating process could be successfully implemented. Due to the size of the large-scale coater, about 1 kg of active material is required for each coating batch. Iteration steps can be necessary and failure may occur. Thus a sufficient back-up amount of material was inevitable and 4 kg of homogeneous high-quality LMNS have provided from the scale-up process in WP5 for this task. The processing parameters obtained for single side coating have been directly transferred to the production of double side-coated electrodes. The final double-side coated electrodes from the large-scale coater had the intended loading of 15 mg/cm2 each side. The tolerance in loading lies below 3 %. The electrodes utilize more than 120 mAh/g of the active material for C rates between 0.2 and 2 C, which corresponds to an area related capacity of 1.8 mAh/cm2 each side. The discharge rate capability is excellent. Up to a rate of 2 C the capacity is nearly constant, and at 5 C still 90% of the initial capacity can be discharged. In order to provide sufficient electrode material for final development and production of multi-layered cells against graphite and tin-carbon composite electrodes two further double-sided cathodes of the target specification with a length of 15 m and 17 m have been produced. Electrochemical tests confirm the high reproducibility of the developed process. The optimized electrodes were implemented in the final cells.
3 - Optimized electrolytes
The objective of the activities was the optimization of the electrolyte materials with respect to performance, safety and adaption of synthesis procedure for up scaling, in terms of the optimization of composition and morphology of polymer membrane, type of electrolyte solution, Li-salts, additives, understanding molecular structure and interactions.
3.1 - Development of the electrolyte solution
The first step in the work was to develop and alternative to the most common commercial electrolyte solutions, such as LP30 (EC:DMC:LiPF6). Obviously these electrolyte solutions have many advantages since they are widely used in commercial batteries for portable electronics. However, they are also the cause of the safety problems with Li-ion technology with batteries catching fire or even exploding. The problem with the current technology is twofold. Firstly it is based on flammable organic solvents (low vapour pressure and low flame point). Secondly the widely used salt LiPF6 is inherently unstable and in the presence of even tracer amounts of water it can undergo hydrolysis with harmful and dangerous decomposition products, such as HF. Both effects compromise the safety of the battery.
The strategy of the APPLES project was to investigate the possibility to fully or partly replace LiPF6, to investigate the combination of other organic solvents, and to introduce ionic liquids as the solvent. For the Li-salt the full or partial replacement of LiPF6 with LiBOB (Lithium bis(oxalato)borate) was considered since it has the potential of having a fluorine-free salt and also the fact that LiBOB is known to form stable passivation films during the first cycle on the electrode improving the safety of the battery. An organic electrolyte solution can be further improved by adding other solvents and here we considered the addition of PC (propylene carbonate), possible since we are not considering a graphite anode. To fully go away from unsafe organic solvent we also considered ionic liquids, which have very low vapour pressure, are non-flammable and have high electrochemical and thermal stabilities.
To optimize the stability and low temperature behaviour of the organic electrolyte solution we suggest using a 1:1:3 mixture of EC:PC:DMC. This solution can also be combined with LiBOB. We showed that the 1:1:3 EC:PC:DMC solution doped with 0.7M of LiBOB is the best choice together with the APPLES electrode materials SnC (pre-activated) and LNMO (uncoated). When considering the ionic liquid electrolyte as the solvent we found that the ionic liquid Py14TFSI doped with 0.2m LiTFSI is a suitable composition. One can also consider mixtures of an ionic liquid and an organic solvent, where the ionic liquids can be viewed as a flame retardant. We found that up to 40 wt% of IL added to the organic solvent is favourable. The flammability is considerably decreased, but the Li-ion coordination is still favourable as in the case of the organic solvent, i.e. ensuring fast ion transport. Considering upscaling issues the organic solvent solution with LiBOB is the better choice. Ionic liquids are still quite expensive in particular at the purities required for battery applications. However, this situation can change in the future as the use of them becomes more widespread for various applications.
3.2 - Optimization and characterization of gel polymer electrolytes
A second step in the optimization of the electrolyte is the implementation of a gel polymer electrolyte membrane instead of the normal separator. The work builds on two previously developed concepts by the partners in the project based on gel casting of membranes and electro-spun membranes. In both preparation methods the polymer gel is based on PVDF (poly vinylidene fluoride) and is swollen with the desired electrolyte solution.
Gel casted membranes
In the project we adapted the previous preparation procedure for gel casted membranes to be compatible with either the optimized conventional electrolyte solution or with the ionic liquid electrolyte and high quality membranes were obtained in both cases. In the case of the ionic liquid a one-step synthesis procedure was developed where the ionic liquid was directly used in the preparation of the polymer membrane.
The gel casted membranes overall show good performance in combination with the APPLES anode and cathode materials, both with the new organic solvent based and the ionic liquid based electrolytes. Using the optimized conventional electrolyte solution we showed the desired functionality of LiBOB forming a protective passivating film during the first cycle, seen through the presence of a peak at around 1.7 V during the first cycle in the electrochemical tests. These membranes also showed a very high mechanically stability also after thermal cycling. In the case of the ionic liquid based electrolyte we showed that the amount of ionic liquid influence the crystallinity of the membrane and also that the thermal behavior of the electrolyte solution is influenced by the confinement in the membrane. By Raman spectroscopy we could determine that the Li-ions interacted weaker with the anion when confined in the membrane, a feature that can be beneficial for the ionic transport and thus the functionality of the membrane.
Electro spun membranes
In the project we optimized the preparation procedure for the electro-spun membranes by investigating the role of solvent, molecular weight of the polymer and process parameters (voltage, speed, etc.). The obtained membranes consisted of non-woven mats containing randomly oriented nano-fibres. A particular feature of these membranes is the very rapid uptake and good retention of the electrolyte solutions (also here one can use either the ionic liquid based electrolyte or the one based on organic solvents, both show fast uptake). The result of the optimization work showed that the best membranes were obtained using the ultra high molecular weight homopolymer (PVdF-UHMW). These membranes showed the best mechanical properties without compromising the electrochemical performance. However, the mechanical performance is still the Achilles heel of these membranes and thus hot-pressing has been used to improve the mechanical properties of the membranes. Our characterization of the electro spun and then hot pressed PVdF-UHMW membranes showed that the temperature during the hot pressing should not exceed 50°C in order to be able to apply a high pressure which is important for the mechanical properties.
In-situ gelation of membranes
For the gel casted membranes we developed a new technology the so-called in-situ gelification showing a potential route, on the lab-scale, for using these materials in a realistic battery process. We showed that the gel can be formed in-situ in a battery cell and that the electrolyte solution (here an organic solvent based electrolyte) could subsequently be infiltrated into the gel formed in the battery. This process was followed by both electrochemical and spectroscopic methods. The in-situ gelled membranes showed satisfactory compatibility with the APPLES electrode materials, the SnC anode and the LiNi0.5Mn1.5O4 cathode. Very low reactivity and good time-stability were found for all membranes with the LMNO spinel but the interfacial resistance increases upon time with the SnC anode, especially for the membranes containing LiBOB.
4 - Materials scaling up
The APPLE project orientates on a pre‐industrial production level. For the development of this cell all materials have to be available in sufficient amounts. Thus a transfer of laboratory work to larger scale processes was the scope of WP5.
4.1 - Cathode Material LMNS
The route of choice for the preparation of the LMNS cathode material is a two-step synthesis route with a continuous precipitation as step one and a subsequent thermal treatment as step two. Target was to adjust particle shape and size distribution to commercially available layered oxide materials as NMC or NCA. In a first attempt a model compound with spherical particle shape in reasonable size distribution for electrode coating has been obtained in 200 g scale. After verification of transferability of the oven procedures from small scale to larger scale thermal treatment have been checked and adapted. Intermediate analytical steps have been introduced to monitor and control crystallite related and compositional effects. The influences of process parameters have been proven. Material, cell and electrode properties are strongly interconnected. Finally the up scaling of the LMNS material has been finished in May 2013. The development turned out to before time and cost consuming than planned. To fulfill the requirements of electrode and cell development the synthesis activities have been significantly increased. Instead of two kg of high voltage spinel 6.5 kg had been synthesized in several batches. The individual batches have been combined to one single large-scale blend in order to fulfill the material needs for recipe development and several large scale coating batches. Spherical high density LMNS is available as a blend in 4 kg scale. A high current capable double side coated cathode has been developed with this material and with optimized loading, showing that cathode material scale‐up could be finished successfully before schedule.
4.2 - Lithium metal powder
For the prelithiation of the Sn/C anode material fine lithium metal powder can be employed as an alternative to lithium foil. A significant improvement can be reached by the lithium metal powder in terms of processability and control of the degree of lithiation by adequately chosen amounts of Li powder. For the scale-up of coated lithium metal powder production two processes have been evaluated with respect to their feasibility and reliability: a two-step process involving the atomizing of lithium metal in an inert atmosphere and a subsequent coating step or a one-step procedure based on a lithium metal dispersion, which has been established on a laboratory scale. After the evaluation based on a variety of parameter the dispersion process was chosen due to its better transferability to larger scales in a batch production. The obtained lithium metal powder is protected and passivated by a proprietary procedure resulting in a thin (1-5 µm thick) composite (inorganic or organic) coating. The produced grey to black lithium metal powders show low sodium grade and lithium content of > 98 % measured by gas volumetry. The mean size of the spherical particles is about 120 μm depending on synthesis conditions; influencing factors are the metal content in the dispersion as well as temperature of the process. Differential scanning calorimetry (DSC) tests proved that these materials are stable towards dry NMP (water content <200 ppm) up to ca. 100 °C for 15 h. There was no pressure evolution detectable. In dry carbonates like PC and DMC the lithium metal powder is stable up to 200 °C and a medium pressure increase is observed.
The scale-up of the lithium metal powder for pre-lithiation purposes within the project had been completed. Lithium powder was produced and supplied to the partners.
4.3 - Anode material tin/carbon composite
In the course of the project it became apparent that the proposed route to the Sn/C composite anode material via the resorcinol-formaldehyde process is not scalable in the sense of reproducibility. An alternative process is based on high temperature oven process with subsequent high-energy ball milling. Small batches (2 g scale) of the Sn/C composite were prepared via this new route to evaluate necessary parameters for a scale-up of this process. Several modifications had to be made to adopt this process based on the process description provided by the project partners.
With the changed parameters the Sn/C composite was synthesized in a 50 g scale. The 50 g batch showed a larger irreversible capacity loss and poor cycling stability. A larger tin particle size was pointed out as root cause for the different performance. Thus, an additional optimization cycle was necessary to adjust the material properties.
Finally the batch size could be increased by roughly two orders of magnitude (150x) to 300 g; a further scale up was not possible due to scarcity and lead times of one of the raw materials (nano-tin powder). The scaled Sn/C anode material was optimized with respect to its tin content and particle size and supplied to the partners for electrochemical characterization and cell manufacturing.
4.4 - Scaling-up of materials preparation/polymer electrolyte membranes
The scale-up activities were targeting on (i) reduction of the membrane thickness from 700 µm to 50 µm to obtain suitable membranes that match the requirements for cell assembling and (ii) enlargement of the membrane from several cm² up to 28 x 24 cm².
Membranes with the initial composition were cast in a glass mold with inner dimensions of 28 x 24 cm (~ 670 cm2). Finally, an average GPE thickness of 100 µm was achieved; it has been found in previous experiments that thicknesses below 100 µm unavoidably lead to an increased number of defects.
The thickness distribution was determined by spot measurements. It was found that the thickness varies from 50 – 200 µm caused by an error of the mold alignment of approx. 0.03° which could not be corrected any further even after multiple repeats.
The residual water content of the precursor membrane was determined by a Karl-Fischer method resulting in a water level of less than 20 ppm. The performance of the GPE was checked after activation of the pre-membrane by the described soaking method. For this purpose the precursor membrane was soaked with the standard electrolyte LP30 and subsequently an ionic conductivity of 10-3 S/cm was measured.
Precursor membranes were cast and sent to the partners for performance tests and cell manufacturing.
5 - High performance getters with high absorption capacity
The overall objective of APPLES project was to develop a lithium-ion battery based on new components with enhanced properties regarding energy, safety and cost. One way to address the safety issues was to study the possible use of getters inside the battery cell. The role of the getter would be to keep the internal pressure of the battery under a control value by removing the exceeding gas.
In order to detect which gases are generated in APPLES-type cells, the gassing behavior of APPLES cells was studied, revealing that the internal pressure increases very rapidly during the cell formation, and it keeps on increasing (although with slower rate) during cycling. GC/MS analyses were performed on LMNS/graphite bi-cells. The results showed that both CO2 and H2 gases are produced in large amounts during a series of cycling and ageing experiments. The proportion between H2 and CO2 depends on the type of experiment and on the electrolyte composition. During cycling, H2 and CO2 are produced in similar amounts (H2:CO2 = 49vol% : 46vol%). The CO2 production is slightly decreased by using vinylene carbonate as additive for the electrolyte (H2:CO2 = 49vol% : 33vol%). In flow charge experiments, the concentration of H2 increases compared to CO2, especially when using LP10 electrolyte (H2:CO2 = 63% : 22%). Also, minor amounts of CO, methane and ethane are formed in all experiments.
The experimental evidences collected during the project confirm what is indicated by literature about the possible mechanisms for gas evolution in lithium-ion batteries.
The salt commonly used in commercial electrolytes, that is, LiPF6 can decompose at the cathode. In presence of small traces of water, the salt can produce acidic products that cause rapid cell performance loss, particularly at elevated temperatures. Just as an example, one should note that 10 ppm of water (that is, 1 mg water/100 ml solution) are enough to produce about 1 ml hydrogen gas.
Also, the formation of Solid-Electrolyte Interface (SEI) on carbonaceous anodes occurs during the first cycle, but SEI conversion, stabilization and growth also proceeds during further cycling and storage. A catalytic reduction of the electrolyte solvents assisted by the carbon surface in correlation with its specific surface area was suggested to be at the basis of the decomposition mechanism. Solvent co-intercalation together with gas evolution can cause exfoliation and loss of Li, which in turn leads to the cell capacity fade. Therefore, gas evolution is not only an issue for the pressurization of the cell, but also for its performance.
Among the produced gases, CO2 is notable because it may be reduced at the anode to oxalate, formate (in presence of water), carbonate and CO. The oxalate is partially soluble in the electrolyte and it can migrate to the cathode to be re-oxidized to CO2, thus leading to a progressive cell self-discharge. These further considerations led to focus on CO2 as the target gas to remove from the battery internal atmosphere.
A range of materials able to capture CO2 are known, but the question addressed in the APPLES project is whether they are able to work in the battery environment. Among the candidate getter materials, zeolites, metal hydroxides, and ionic liquids (ILs) with amino groups have been considered.
Zeolite X is the most widely studied sorbent for CO2 capture via physical sorption. Zeolite 13X is known to be very selective to carbon dioxide, this indicating that this material can be used for the separation of CO2 from gas mixtures comprising H2, N2, CH4, CO. The selectivity is due to the CO2 quadruple moment. However, in presence of the electrolyte vapors, it was noticed that a strong and non-reversible interaction with the carbonate solvents caused the clogging of the zeolite pores. In fact, if heated up in presence of the solvents vapor, the zeolite turns its color in brown. This indicates a possible cracking of the carbonate adsorbed on the surface of the zeolite. Moreover, after the color change, the zeolite was no more able to capture any CO2, this suggesting the clogging of the pores. Due to the similarity between the CO2 and the organic carbonate molecules, no physical sorption mechanism can be effective in size-selecting the absorbate.
On the other hand, it was found that in order to capture CO2 via nucleophilic addition on ILs high partial pressures of CO2 are needed, around 7-8 bar. This is of course a condition not compatible with the application in batteries.
Finally, alkali metal hydroxides have been considered, among which a special attention was dedicated to LiOH, for its high theoretical capacity for CO2 (388 TorrL/g) and good compatibility with the solvents used in common electrolytes. Hydroxides have the drawback of reacting with CO2 by producing Li2CO3 and water. The retention of water can be addressed by adopting a specific configuration of the final getter solution, see below.
The study of the sorption properties of LiOH showed that the performance strongly depends on the experimental conditions.
LiOH/CO2 ratio: upon the same partial pressure of CO2, the sorption activity becomes significant only when a suitable amount of LiOH is used. In presence of 500 Torr of CO2, 30 mg of LiOH did not give any reaction (expected amount to be sorbed: 11,5 TorrL), whereas by using 300 mg of LiOH the reaction proceeded as expected. This suggests that there is a minimum LiOH/CO2 ratio below which the reaction would apparently not start.
CO2 pressure: even when using the proper LiOH/CO2 ratio, the CO2 partial pressure affects the reaction mechanism. Two possible routes have been found experimentally: below 1 bar CO2, the reaction of LiOH proceeds with slow kinetics, so that 2 days are needed to capture only 27% of the theoretical amount of CO2. When the CO2 pressure is raised up to 2 bar, the reaction is completed in few minutes, revealing a second possible mechanism characterized by fast kinetics.
Presence of the electrolyte: the presence of the organic solvent vapors did not affect the sorption properties of the LiOH powder, this confirming the principle of use LiOH as getter in batteries.
The configuration studied for the getter comprised the dispersion of LiOH in high-density polyethylene (HDPE) and the extrusion of the composite to form a film. To obtain suitable films, powders with reduced particle size (10 um) were used. HDPE was selected as the polymer matrix because it is compatible with LiB environment, highly permeable to CO2, and thermally stable. The extrudate was tested for its compatibility with commonly used electrolyte solvents, showing no degradation after immersion in a mixture EC/DEC. Also, the CO2 sorption performances have been measured in presence of DEC solvent vapors. Compared to the powder, the extruded getter has slower sorption kinetics. Most likely, in the case of the film the rate-determining step is the diffusion of CO2 into the polymer matrix. The sorption kinetics being slower, the water produced by the reaction can be progressively released during the sorption process. The saturation of the getter is around 55 wt% relative to the LiOH content, which is the stoichiometric weight gain expected by considering the complete release of the water produced by the reaction.
To avoid the drawbacks of the water release, the extruded getter sample can be enveloped into an external polymer layer (using for example HDPE or polypropylene) that might serve as a barrier for the water generated into the getter.
To conclude, the gases formed by cycling APPLES-type cells are mainly CO2 and H2. It is believed that a getter able to sorb one or both of these gases could help to control the internal pressure, enhancing therefore the cell safety and performance. A getter for CO2 based on LiOH was designed, manufactured and tested, revealing high sorption capacity for CO2 and compatibility with battery electrolytes. The issue of the processability of LiOH, especially when dealing with anhydrous, small particle-size powders, has been managed, and extruded prototypes have been obtained. The water released upon carbonatation of the LiOH can be partially restrained by using a polyolefin as the matrix for the extruded material. However, although this solution slows down the transport of water, coupling the CO2 getter with a dryer solution might be considered for a feasible application in real batteries.
6 - Cell prototype
Target of this task was the development of high voltage prototype cells on a pre-industrial manufacturing line at ZSW. During the process development several main steps had been carried out. Step I: definition of the lay out design of the cell. Step II: electrode development. Step III: evaluation of different electrolytes. Step IV: pretreatment of electrodes. Step V: prototyping. Several iteration cycles between the individual steps had been carried out for fine-tuning. Different cell chemistries, electrolytes and cell concepts have been taken into account. For comparison cells with commercially available materials have been developed in parallel in comparable design as a reference. During the project modifications of the original plan became necessary. Due to unforeseen technical obstacles, which showed up during scale-up of the active materials, the deadlines for the manufacture of large full cells could not be met completely for the defined cell chemistry. Thus a revised plan of cell manufacture has been developed. Two different design concepts, which both are known to be produced in industrial scale, have been taken into account. a) a multilayer-stacked design, and b) a modular bi-cell unit, which follows an alternative process of cell stacking. Based on these designs, overall three generations of high voltage cell prototypes have been developed successfully within the scope of the project. The development steps in detail are described below.
Step I – definition of the lay out design: The electrode dimensions are predefined by the available manufacturing line, while the overall thickness of the cell is determined by the dilatability of the pouch foil which had to be pre-shaped by a deep drawing process. Thus the cell dimensions limit to 40 x 60 x 14 mm. Based on the physical dimensions, on the target cell capacity and on the expected electrochemical data of the active materials, the preliminary requirements for electrode loading and electrode thickness were calculated as a zero point to start with. The influence of the electrode loading on the overall energy density of a cell stack was taken into account and a reasonable range has been defined. During the course of cell development several iteration steps were implemented in order to find the optimum cell configuration.
Step II – electrode development: For electrode development various dispersions and coatings have been carried out, starting in small scale and finally going to the pre-industrial scale production. The electrode design has been optimized with respect to loading, homogeneity, and porosity. Various loadings have been applied; different ratios and types of active material, binder and conductive agent have been tested. The best-suited components have been selected from the screened samples. Electrodes have been compacted by different methods such as application of static pressure or calendaring, and the influence of densification has been investigated. For the different cell generations two different sets of anodes have been developed, a graphite electrode, and a tin-carbon electrode. Balancing of the electrodes has been carried out using monolayer pouch cells with a metallic lithium reference electrode. Balancing strongly depends on the type of active material, the individual electrode kinetics, and the electrolyte. Thus several iterations cycles were made for each cell generation.
Step III – electrolyte: Two different classes of electrolytes have been defined for use, which are an optimized liquid electrolyte on the on hand and a gel-polymer type electrolyte, GPE, on the other hand. Both types have been developed within the project. They show different processing behavior and thus were implemented in different cell concepts.
Step IV – pretreatment of electrodes: Due to SEI formation on the anode, generally a certain amount of charge is irreversibly lost during the initial full cycle. A pronounced loss is observed especially for high-capacity lithium alloying materials like tin, which is the target anode material of the final cell configuration; tin is used as a tin-carbon composite, Sn/C. This effect strongly reduces cell capacity, if not compensated for by a well suited pre-treatment. Thus several concepts of prelithiation have been tested. Principle applicability could be proven in the intermediate scale for different routes from which finally the electrochemical prelithiation has been applied.
Step V – prototyping: after optimizing the individual steps, three generations of high-voltage prototype cells have been developed and delivered within the scope of the project. The first two generations are multilayered stacked cells. These cells use high-voltage lithium manganese spinel, LMNS, cathodes with an active mass loading of 15 mg/cm2, balanced graphite anodes and an optimized electrolyte. Cells were manufactured on an automatically working stacking winding machine; a polyolefin separator was used. To successfully take the step from bi-cell pouch design to the multilayer-stacked design with a new electrolyte, a sequential approach with several preliminary intermediate steps had been chosen. First, all optimization measures, such as balancing and compatibility tests with the optimized electrolyte have been investigated in small single layer pouch cells beforehand. This was done in order to verify compatibility of the cell chemistry. Secondly, the functionality of the 3-stack design has been checked with standard electrolyte LP30 (no VC). After verification of both presettings, cell manufacture was started. For GEN1 type cells, three stacks á 13 electrodes were combined to cells with an overall capacity of 1 Ah at C/3. Electrochemical formation of the assembled cells has been optimized for the new chemistry. It could be proven, that several full cycles were necessary to stabilize the cells for storage. Cells have been provided to the partners for testing. For the second cell generation, GEN2, several fundamental technical improvements have been developed in order to increase cell capacity. For this concept the pre-industrial stacking-winding machine has been adapted and newly programmed in order to increase the overall number of the electrodes in one stack. Instead of three individual cell stacks á 13 electrodes these cells now contain one single stack with overall 47 electrodes. The cell provides a practical capacity of 1.5 Ah, which corresponds to an increase of about 50% compared to Gen 1. In order to further increase the energy density, the overall electrolyte amount was reduced for the second generation compared to the first. Different cell sets have been delivered for electrochemical testing. Within the precycling cell capacity remains stable. The potential profile of the cell shows low over potential indicating high quality of electrodes and cells. During prolonged cycling a constant capacity loss has been observed. In depth post mortem analyses have been performed. The investigations clearly show, that the cells dried out during the cycling procedure due to electrochemical reaction of the electrolyte. Especially the inner part of the stack is affected. Re-testing of disassembled electrodes has been done in T-cells, showing pronounced ageing at the anode while the cathode is unaffected. It could be concluded, that the capacity loss of the Gen 2 cells is due to the not yet optimized amount of electrolyte. Higher stability can be obtained when this point will be addressed properly. The third generation of cells, GEN3, was made completely from new materials LMNS/GPE/SnC; these cells are designed as modular units that follow an alternative industrial process of cell stacking. The method of prelithiation had to be adapted and a new manufacture concept had to be developed. In order to understand the impact of each individual new component several intermediate stages of cells have been assembled. Pouch cells with capacities between 30 mAh and 50 mAh have been made. Overall five interim stage cell sets had to be developed, investigated and understood until the final cell could be assembled. Recipe and coating had to be optimized newly and several electrodes had to be processed and balanced. For the Sn/C anode several methods of prelithiation procedures have been screened. Finally electrochemical prelithiation of the anode has been proven to be the best controllable route for this specific application. The cells scale up process proved itself to be technically very complex in the step of the membrane application. The normal equipment and protocol used to produce pouch cells could not be applied to this kind of technology and the cell assembling had been done manually. Due to the prelithiation the final cell shows a reasonably low irreversible capacity loss during the first cycle and a coulombic efficiency >98% for the following cycles. Gen 3 clearly shows that this new cell chemistry can be transferred from coin to pouch design, with comparable electrochemical behavior. To summarize the prototyping activities: Overall three generations of prototype cells have been successfully developed during the scope of the project, Gen 1 to Gen 3, differing in cell design, materials and concept. Problems have been studied and questions experimentally addressed. Pros and cons of the new cell chemistries have been determined. The results obtained by the prototype pouch cells clearly confirm the behaviors observed in the small cells and give a positive technical outlook on this chemistry. Requirements for technical transfer above the given scale could be defined. It could be shown that all three generations show reasonable performance and reach their intended capacity. Immediate formation of the cells after completion has been proven to be inevitable to stabilize the cells. In small cells excellent cycling stabilities have been achieved even without using any additive in the electrolyte. In large cells major ageing effects can be attributed to electrolyte starvation. Based on the overall results of the project, measures to further improve cell chemistry for all three cell generations have been deduced. The electrochemical validity of the cells has been successfully proven for pouch format.
7 - Recycling process
Recycling of production waste and end-of-life Li-ion (EOL) batteries covers the important aspects regarding battery recycling, from solutions for recycling of production waste to the recycling of individual end-of-life battery cells.
The following summarizes the findings on three main areas:
• Recycling plan for production waste
• Collection plan for EOL batteries
• Recycling plan for EOL Li-batteries
7.1 - Recycling plan for production waste
The waste from the production of battery cells consists of waste from the input materials (mainly different foils), waste from cutting the pouches, discarded cells and slurries from coating and electrolyte. There will also be waste from assembling the battery. The different foils have a varying value and should be kept separated as waste. The handling of the waste slurries from production will be costly since they will have to be sent for destruction due to the PVDF content.
The waste from the input materials are the following:
• Pure copper foil: Sent to a primary refinery.
• Pure aluminum foil: Sent to a smelter.
• Coated copper foil: Copper with graphite, PVDF and possibly SnC. Sent to secondary refinery.
• Coated aluminum foil: Aluminum with PVDF, carbon black and LiMn1.5Ni0.5O4. Sent to energy recovery or to a recycler of battery cells.
• Nickel plated copper from tabs. Sent to primary smelters if not too much nickel.
• Slurries from coating: PVDF, graphite and organic solvent with SnC or LiMn1.5Ni0.5O4. Sent for destruction.
• Electrolyte slurry: At best sent for solvent recycling, otherwise sent for destruction.
• Waste from pouch cutting: Aluminum and polymer. Sent to energy recovery.
The focus in terms of waste minimization should be to optimize the slurries and the coating generating that both the amount of slurry and the amount of coated foil fraction are minimized.
7.2 - Collection plan for EOL batteries
Lithium ion batteries for vehicle applications are legally covered by two Extended Producer Responsibility (EPR) systems, the ELV-directive (End of Life Vehicle) and the WEEE-directive (Waste of Electrical and Electronic Equipment), which both will have a large impact on the collection system for the batteries. Also, safety issues related to the high energy content in the batteries are critical to create an efficient collection system. Important points for The APPLES batteries (and other Li-ion batteries for vehicle applications) regarding collection are:
• Li-ion EOL batteries for vehicle applications are generated as waste at workshops, car repair companies and car dismantlers. Thus, collection points for Li-ion batteries and lead acid batteries are similar; hence, collection should be coordinated.
• Li-ion batteries for vehicle applications may have a value for second use in other less demanding applications such as energy storage. In the future, it is possible that battery reparation will create business opportunities through e.g. the replacement of single cells in otherwise healthy battery packs.
• Safety issues have to be solved with regard to the batteries high voltage and energy content. Special authorization should be required to repair, dismantle and handle this type of batteries.
• The intelligence of the battery pack gives the opportunity to scan batteries for damaged cells and evaluate its status.
7.3 - Recycling plan for EOL Li-batteries
As electric vehicles and hybrid electric vehicles are relatively new on the market, experiences from recycling end-‐of-‐life Li-‐ion batteries for electric vehicles applications are very limited. The possibility that charge remains in an EOL battery, even after discharging, could be a safety hazard and will impact both collection systems and recycling processes. Also, solvents with low flash point and the risk of elemental lithium formation make recycling difficult from the safety point of view. In addition, handling of fluorinated salts and polymers is a challenge due to the formation of hydrogen fluoride when heated. To make recycling easier, description of type of Li-‐ion battery on cells and battery packs are recommended. Compared to Li-ion batteries with cobalt-chemistry, where cobalt has a high value, the material value is lower in APPLES batteries, which will impact the economy of recycling the batteries.
It is suggested in the project that the first recycling step for cells is either mechanical shredding or a pyrolysis step. Mechanical separation like sieving and possibly eddy current separation if the metal sheets are sufficiently thick, may be applied to remove elemental metals like aluminum and copper. For oxidized metals such as lithium, manganese and nickel, mainly hydrometallurgical methods are recommended.
The APPLES battery samples (LiMn1.5Ni0.5O4/C) received from University of Camerino were opened in a glove box with argon atmosphere. Various methods for substrate separation from black mass have been investigated to assess which would be best suited for the APPLES battery.
The black mass obtained from these separation processes has been separately leached with three different inorganic acids: hydrochloric acid, nitric acid and sulfuric acid of 0.1 1, and 6 M concentrations. Leaching experiments at room temperature showed that hydrochloric acid is the best leaching agent and that the leaching time is about 24h for complete metal dissolution with 6 M acid. Black mass that was heat-treated at 400°C for 2h, gave better metal dissolution and faster kinetics. Leaching at higher temperatures showed, as expected, that complete recovery was considerably faster and was achievable with 1 M acid concentrations. Hydrochloric acid, followed by sulfuric acid, showed the best leaching capability whereas the black mass that was recovered by ultra-sonication gave faster kinetics. Leaching at 90°C increased the kinetics of dissolution as compared to 60°C.
Solvent extraction processes were chosen to recover and separate metal ions from leaching liquor. Two kinds of extractants were studied to investigate the relationship between extractant type and the effect on separation of metal ions: acidic extractants (CYANEX 272, CYANEX 301, DEHPA) and solvating extractants (CYANEX 921, CYANEX 923, CYANEX 471X, TBP). The AKUFVE apparatus was used for solvent extraction studies. It was experimentally verified that acidic extractants give sufficient extraction of manganese and nickel, while lithium is not extracted. It was concluded that separation of metals from Li-ion batteries is possible using both CYANEX 272 and DEHPA from HCl, HNO3 and H2SO4 media and at the extractant concentrations tested. Lithium remains in the aqueous phase in both systems, whereas manganese and nickel are extracted at a lower pH in DEHPA than in CYANEX 272. In order to avoid hydroxide formation at higher pH, it might be preferable to use DEHPA.
Based on the leaching and AKUFVE experimental data, hydrochloric acid has been chosen as a leaching medium for black mass obtained after mechanical and thermal pre-treatment. DEHPA diluted in kerosene has been chosen as the extractant for the aforementioned metals. The extraction was tested as a counter-current process, by using mixer-settler equipment. A process flow sheet has been proposed for up scaling.
8 - Industrial manufacturing process
A production process for industrial manufacturing of the lithium-ion batteries has been defined. The development of the process has been based with the end use in automotive applications in mind. The developments and results have required some assumptions in order to proceed with the work. In particular there is no large-scale anode production process and there is no available prelithiation method at industrial level that may reduce the irreversible capacity of the anode. The following is based on the assumption that both processes are available.
The cell is constructed from thin sheets of cathode, separator and anode. The sheets are built up to layers and the gaps between the layers are filled with liquid or solidified electrolyte.
The cathode is a HV spinel produced as powder with well-defined particle size range. This is coated on aluminium foil about 10-20 microns thick. The anode is produced in the same way by substituting the active cathode powder with anode powder and (for electrochemical reasons) coating onto thin copper foil.
The coating is performed by using a slurry produced from mixing the active cathode material with conductive agent (Carbon black), binder material (PVDF) and binder solvent (NMP). The mixing of the slurry is essential in order to obtain a good end result. Alternative water based systems to PVDF and NMP using water soluble binders has been evaluated as they offer many environmental and cost advantages but were found to be insufficiently stable and not supported by machine producers and therefore very difficult to implement at industrial level at this point.
Aluminum foil comes in different thicknesses and various alloys. Usual choices are the 1000- series (close to pure aluminum) and 3000-series (alloyed with manganese). The 3000-series offer higher tensile strength why this one is preferred allowing for thinner foils adding less dead weight to the final cell. A corona treatment is used to degrease the foil before use to improve adhesion. For the anode foil there are more alloys to choose from. However, the best choice was found to be electrodeposited copper (ED foil) as it allows for the best weight performance and sufficiently high tensile strength and no degreasing required.
The active cathode material and anode material slurries are spray coated using a die nozzle on both sides onto the surface of the aluminum and copper foils, respectively. The wet coating thickness of the cathode is about 250-300 micron before the solvent is evaporated in a dryer, more than 100 meters long. After drying the foils are calendared to compress material and improve adhesion. The dry foil is then cut and slit to sheets that are stored in cassettes similar to a printer. Anode and cathode sheets are fed to a stacking machine from the cassettes where they are stacked in the pockets of a z-fold separator foil (olefin based at this stage of progress for the industrial process). This step produces a cell stack that is parallelized by a machine that folds and ultrasonically welds together all positive terminals to one positive terminal and all negative terminals to one negative terminal before the stack is sealed in a pouch and electrolyte is added.
The sealed cell is subjected to a formation process where electrolyte is reduced on the surface of the anode to form the SEI layer that stabilizes the anode. During this process gas is evolved why the final sealing is performed in two steps. The first time excess pouch foil is used forming a gas pocket that later is trimmed to remove the evolved gas. The total formation time is 17 days of which 24 hours is load time with charge/discharge and the remainder is resting time.
Under the assumptions stated at the first section the final cell under optimal conditions may reach 57.2 Ah at a weight of 865 grams and a nominal voltage of 4.7 Volts. This is equivalent to an energy density off 310.7 Wh/kg. The material cost is estimated to at least 13.8 € per cell with high uncertainty in material cost of the active electrode material. The total production cost including electrode production, assembly and waste management is estimated to 30.7 € per cell, excluding labor.
8.1 - Battery system integration
An energy optimized battery pack has been designed using the APPLES industrial size cell. Battery dimensioning of power and energy needs has been performed by simulating the NEDC European drive cycles. Two different air-cooling concepts has been thoroughly investigated showing that air flow between the cells gives sufficient cooling and the best weight performance. To obtain the cooling concept composite frames to hold the cells in place. For connection of cells screw terminals is the best option since it allows for cost saving remanufacturing and easier recycling. The battery container is made of aluminum to offer high strength and low weight.
The target for battery system integration is an electric vehicle (EV) battery. The definite size of a manufactured EV battery system will be dependent on the characteristics and design of the vehicle in which it will be installed. This report demonstrates a general modular system to be incorporated in the bottom plate or the fuel tank space of a traditional internal combustion engine vehicle. It may be that the needed battery volume would need more space and need either a secondary battery or an extension of the battery pack. However, this is outside the scope of the project and will not be addressed.
To dimension the battery typical EV designs were evaluated and compared with real traffic recordings in the databank at ETC. Further, simulations of drive cycles such as NEDC were performed to model the energy requirements. The model results included efficiency for driveline and regeneration for the continuous combined NEDC showed that the energy requirement for the battery was 184.5 Wh/km with an average battery power of 6.73 kW and maximum battery power of 62.6 kW.
Based on the energy and power data calculated from simulations described above a cooling system was designed. For weight reasons an air-cooled system was decided, this also allows for a less complex system that is easier to demount and remanufacture than a liquid cooling system. This is obtained by assembling the cells in composite frames allowing for airflow between the cells or by assembling thin aluminum plates between the cells with heat sinks at the end of the plates. Thermodynamically evaluating the two choices showed that the difference was very small why the composite design was chosen as it allows lower weight.
The resulting battery to supply 200 km of driving distance taking into account a usable SOC windows of 80 % and 1 % overcapacity to cover for cell imbalance was a 236 kg battery with 176 cells divided into two battery packs comprised of 88 cells each. The nominal voltage was designed to 413.6 V and the total capacity of the system 114.4 Ah of which 89.2 Ah were usable. The system level usable energy performance reached 156.6 Wh/kg where 64.6% of the system mass consist of cells.
9 - Environmental and sustainability assessment
Within APPLES environmental sustainability has been assessed by Life Cycle Assessment methodology (LCA) following the ILCD handbook prepared by the EU JRC at Ispra. We also used three other techniques: mass flow analysis (MFA), technology change analysis and risk assessment.
LCA was carried out in four phases: goal and scope definition, life cycle inventory, impact assessment and interpretation. In the goal and scope definition, system boundaries are determined as well as allocation rules and data requirements. In the inventory phase, unit processes, and flow data for the technical system is collected and calculated relative to a reference unit. In the impact assessment phase, the inventory data are grouped and transformed in impact category indicator to better show their potential contribution to various environmental impact types. They may also be aggregated by different weights to single values, which may be used to find which impacts that are most significant and to aid in improvement assessment. In the interpretation phase completeness checks and sensitivity analyses are made. The methodological framework is standardized by ISO in two standards, ISO 14040 and ISO 14044. In the inventory part, an LCA model was made in Excel format.
Simulations with this model were made adding different input data. The model also included impact assessment and weighting to single indicators for the total environmental impact. The impact assessment methods used were Ecoindicator 99, EPS 2000d, and Recipe. The difference between the methods is mainly that EPS has a longer time perspective than the others and therefore metal resource depletion gets a relative higher weight than the two others. The difference between Ecoindicator 99 and Recipe, which were developed by more or less the same people, indicate the stability over time when applying largely the same approach.
Mass flow analysis methodology were used to compare the total need for metals in APPLES batteries if they were used in large scale for all cars and the present global production and reserves of the metals.
9.1 - Environmental life cycle assessment of APPLES cells
1) Goal and scope definition in the LCA
The goal of the study was to assess the environmental sustainability of the APPLES battery, to identify options for improvements and to give eco-design recommendations.
A LMO battery was used as a benchmark. The functional unit was 1 kWh of electrical energy delivered. The reference flow was one batch of cells produced. System boundaries included a foreground system with manufacturing and EOL processes for the batteries. The energy losses during the use phase were also included and allocated to the battery. Background systems include material and electricity production
Impact categories was selected both at the midpoint and endpoint level. At the midpoint level GWP100, Acidification potential, Eutrophication Potential, Ozone creation potential, Ozone depletion potential, Ecotox, Human tox and resource depletion potential was determined. At the endpoint level the EPS system v2000d was used. Weighting was made using EI99, Recipe and EPS v.2000d.
2) LCA model
The first LCA model was made without recycling. Tin was then used in the anode, and the production process for the active material involved organic tin, which was mixed with carbon and calcinated. Later during the project, only carbon was used for the active material in the anode, but the same model could still be used, quantifying the flows and energy use in the tin precursors and calcination as zero.
The model was made in Excel, so it could be transparent and used by anyone in the project. The input data consisted of choices of materials from drop-down menus and amounts of materials and energy. Type of electricity-mix for the electrode and cell manufacturing and for the use phase could also be chosen.
3) Results from simulations
The first models and simulations were made for the production and use of the batteries, but without recycling. The weighted life cycle impacts from the APPLE cell do not differ much from the LMO cell if average European electricity is used in production. If electricity from waterpower is used the relative difference become more pronounced. The highest contribution to the overall impact from this early cell concept comes from electricity production, production of active material for the cathode, copper foil manufacturing and production of active material for the anode (Sn). If the recycling process developed by WP9 is included for the APPLES battery the impact becomes significantly lower. It was not possible to compare against the LMO battery when recycling was included since no specific recycling process was made for the LMO battery.
The conclusions that can be drawn from the LCA analysis, is that significant improvements can be made by using low carbon electricity and efficient recycling.
4) Mass flows analysis
If the APPLES battery is to be used for a substantial replacement of fossil fuel powered cars, large amounts of raw materials are needed. Especially some metals may not be available in sufficient amounts. The amounts of metals needed if no recycling occurs, and if all cars are run on APPLES batteries are a 10 % increase in Al production compared to present virgin global production, a 23 % increase for Cu, a 22-fold increase for Li, 60% increase for Mn, 160 % for Ni and a tenfold increase for Sn.
5) Risk assessment
Several of the chemicals used in the production of the APPLES battery are toxic or irritating. Some are flammable. The risks are considered manageable.
6) Eco-design recommendations
The results from LCA, MFA and RA shows that recycling is recommendable and that many chemicals shall kept out of contact with humans.
7) Setting recycling targets for metals in a Li battery
It seems easy to agree that recycling is recommendable for many elements used in the APPLES battery, but to what extent? There are several comparative LCA studies of batteries used for propulsion in literature, but to our knowledge, none of them has addressed the question of what is an acceptable recycling rate. To do so, one has to be able to handle trade-offs between different environmental issues and other issues. One way of making these trade-offs is to use monetary measures of environmental assets and optimize the total cost for the recycling.
Depending on the recycling cost, in % of the metal market prize, we find that the targets vary for different metals, but in most cases, the recycling of Cu, Ni, Sn and Mn should be more than 90%. For Al and Li, the targets vary more with recycling cost.
8) Sensitivity analysis
From the LCA results we can conclude that the carbon intensity in production of electricity is very important, as well as the degree of recycling. Another important factor is the losses while charging and discharging. If the efficiency changes from 95% to 96%, there is a 20% decrease in the amount of electricity allocated to the battery.
No direct hinders is found for the production use and waste management of APPLES cells. But several measures need to be considered for the work environment, to decrease environmental impacts and safeguard scarce resources.
The measures required from the environmental sustainability assessment point of view are manageable, although they may increase the production and recycling costs, which may be negative in the sense that the largest real improvement to the environment from the APPLES battery is dependent of how much it can replace internal combustion engines, which in turn, very much is a cost issue. There is thus a risk for sub optimization if the system boundaries of the environmental assessment are too narrow.
The research and innovation proposed by the APPLES project contributes to the strengthening of the competitiveness of the European industry of automotive battery and electrochemical capacitors in global markets through the scaling up to an industrial level of a lithium ion battery based on a chemistry totally new in respect to that exploited the common versions of these power sources.
This new chemistry involves the use of lithium-metal alloy (Sn-C) as anode (with a practical specific capacity, i.e. 500 mAh/g, higher than that of the common graphite), a lithium nickel manganese oxide spinel, LiNi0.5Mn1.5O4 as cathode (with an operational voltage, i.e. 4.5V vs. Li higher than that of the common lithium cobalt oxide) and a composite, gel-type membrane as polymer electrolyte (with expected reliability and processability higher than those of the common organic carbonate solutions).
The high capacity of the anode, combined with the high voltage of the cathode, allowed obtaining a consistent enhancement in the energy density of the battery.
The final Apples cell under optimal conditions may reach 57.2 Ah at a weight of 865 grams and a nominal voltage of 4.7 Volts. This is equivalent to an energy density of 310.7 Wh/kg. The material cost is estimated to 13.8€ per cell with uncertainty in cost of the active electrode material. The total production cost including electrode production, assembly and waste management is estimated to 30.7€ per cell, excluding labor. An energy optimized battery pack has been designed using the APPLES industrial size cell. Battery dimensioning in terms of power and energy needs, has been performed by simulating the NEDC European drive cycles. The resulting battery to supply 200 km of driving distance was a 236 kg battery with 176 cells divided into two battery packs comprised of 88 cells each. The nominal voltage was designed to 413.6 V and the total capacity of the system 114.4 Ah of which 89.2 Ah were usable. The usable energy performance reached 156.6 Wh/kg where 64.6% of the system mass consist of cells. These favorable aspects resulted in an overall cost reduction at system level of 100 €/kWh. This is well below the costs indicated at the beginning of the project, which was 150 €/kWh.
In terms of safety, the replacement of the low vapor pressure, flammable liquid electrolyte with a more reliable, polymer membrane, consistently reduced the safety hazard. Moreover, the replacement of the conventional lithium cobalt oxide with lithium nickel manganese spinel at the cathode side also contributed for the safety improvement with respect to common lithium ion battery system. The safety control was also pursued by the design of a getter specifically directed to trap exceeding gases arising during the battery operation and that otherwise would build up the internal pressure of the cell.
Finally, the replacement of the toxic cobalt-based cathode with an environmentally compatible nickel-manganese compound improved the sustainability of the new demonstrative battery.
The determination of the effective impact of this battery confirmed its capabilities when scaled-up to a demonstrative industrial size. In fact, it was demonstrated: i) the effective value of energy density under a high capacity configuration and the feasibility of 300 Wh/kg; ii) the definition of the safety by a USBABC/Freedom Car abuse test procedure protocol; iii) the relevant feasible recycling processes and iv) the environmental sustainability by cycle life assessment evaluated according to International Reference Life Cycle Data System (ILCD) Handbook.
Dissemination and Exploitation of results
The overall aim of the dissemination and exploitation activities was to ensure that the project results are disseminated and that the results arising from the APPLES project are exploited. Specifically, the objectives of the dissemination and exploitation activities were to:
• Promote the dissemination of project results
• Organize dissemination management for the consortium
• Support the partners in the management of intellectual property rights (IPR)
APPLES partners agreed that the exploitation of project results is of a common interest. Therefore, dissemination and exploitation activities were planned, as much as possible, at project level and not at individual level.
To raise the public awareness of the objectives and the research applied in SECOA, the Consortium followed several paths including conference organization, public project presentation and typical academic publication activities. To such purpose, the dissemination activities concentrated primarily on:
• Presentations at conferences and workshops both at national and international level
• Publications in scientific peer reviewed journals (both national and international) and magazines
• Setting up and maintenance of the APPLES web site
• Organization of the APPLES final event
All the above-mentioned activities were targeted primarily at the International Scientific Community and at potential business partners.
During the project duration the following activities were carried out:
• 10 publications in peer reviewed journals, plus 2 prepared and submitted
• 21 presentations at national and international conferences and workshops,
• 11 poster presentations
• APPLES final event within the International Conference ILED 2014 (International Meeting on Ionic Liquids for Electrochemical Devices), attended by about 50 people, during which 10 presentation on project results were given.
The exploitation of the project results is of interest for the industrial partners of APPLES. In order to streamline exploitation activities, partners identified the following six key exploitable results:
1) New anode material – Tin / Carbon composite (Sn-C)
2) New cathode material - Lithium manganese nickel spinel (LMNS)
3) New electrolyte and in situ preparation of membranes (in Situ gelification)
4) Improved getters
5) Recycling process
6) Industrial manufacturing process
For each of the above-mentioned key exploitable results, partners already identified possible future paths for exploitation such as further research activities, patenting and licensing activities and direct use by some of the industrial partners.
List of Websites:
From December 2012 on, the website (www.applesproject.eu) has been linked with Google Analytics, in order to monitor and to acquire detailed information on website traffic.During the project duration, the site received 2,081 visits (54% new visitors and 46% returning visitors).
Grant agreement ID: 265644
1 June 2011
31 May 2014
€ 4 601 443
€ 3 332 298
SAPIENZA INNOVAZIONE CONSORZIO
Deliverables not available
Grant agreement ID: 265644
1 June 2011
31 May 2014
€ 4 601 443
€ 3 332 298
SAPIENZA INNOVAZIONE CONSORZIO
Grant agreement ID: 265644
1 June 2011
31 May 2014
€ 4 601 443
€ 3 332 298
SAPIENZA INNOVAZIONE CONSORZIO