Skip to main content

Stable Interfaces for Rechargeable Batteries

Final Report Summary - SIRBATT (Stable Interfaces for Rechargeable Batteries)

Executive Summary:
SIRBATT project, “Stable Interfaces for Rechargeable Batteries” consisting of 11 full partners from the European Area (six Universities, one Research Institute and four industrial partners) has been a large scale multisite collaborative RTD project funded by the European Commission under the ENERGY.2013.7.3.3 – “Understanding interfaces in rechargeable batteries and super-capacitors through in situ methods” call for research.

SIRBATT has put together a multidisciplinary team of leading European Universities (University of Liverpool (ULIV), Karlsruher Institut fuer Technologie (KIT), Politechnika Warszawska (WUT), Università degli Studi di Camerino (UNICAM) Universidade de Aveiro (AVE), Centre national de la recherche scientifique (CRNS)), Research Institutes (CIC Energigune (CIC)) and Industries (Johnson Matthey (JM), Johnson Matthey Battery Engineering Systems (JMBS), SGL Carbon (SGL), Iberdrola (IBR)) with the overall goal of fundamental understanding of the structure and reactions occurring at lithium battery electrode/electrolyte interface which would lead to longer lifetime lithium batteries.

Batteries are one of the main contenders for stationary energy storage with lithium-ion batteries having applications for providing uninterruptible power supply, power quality, transmission, distribution and load shifting. Batteries now have to demonstrate the ability to store energy efficiently, within certain power, lifetime and safety specifications, at a price-point that is ultimately affordable by the energy industry, before they can be considered for bulk power management.

Within the project Stable Interfaces for Rechargeable Batteries (SIRBATT) we have investigated lithium-ion battery electrode chemistries to deliver the long-term cycle life batteries required by future societies. Within the programme SIRBATT has undertaken a wide program of experimental and theoretical research of Li-ion battery electrode interfaces and also includes the innovative step of the integration of Li-ion cells with microsensors to monitor internal and external temperature.

The list of the scientific publications, published within the project three year timeframe– 12 in leading peer-review journals such as Advanced Energy Materials, The Journal of Physical Chemistry C and Chemical Communications, Journal of the Electrochemical Society - as well as numerous videos explaining the activities and scientific achievements of the project are available in the project website It is noted that 2 patent applications have been submitted to National Patent Offices.

In addition to technical aspects, SIRBATT has organised two international workshops in Spain and USA in 2015 and 2016 to disseminate the work from the project and to engage with the Sister programmes (Hi-C, Influence and Baccara) funded within the same call, as well as complementary European and USA research activity.

Project Context and Objectives:
The need to reduce carbon dioxide (CO2) emissions from the burning of fossils fuels has led to Europe looking towards increasing the development and use of alternative ‘green’ energies. However, electricity supplies gained from, for example, wind, wave and solar power are intermittent as they are dependent upon external factors. As a result the storage of energy produced by these means is now crucial in the levelling out of supply and demand. If such energies are to be used on a larger scale there needs to be a mass market solution, therefore major improvements in stationary energy storage technology are now paramount.

Batteries are one of the main contenders for stationary energy storage with lithium-ion batteries having applications for providing uninterruptible power supply, power quality, transmission, distribution and load shifting. Whilst at the moment energy storage in bulk power management is served by pumped hydro power and compressed air storage, their geographical specific locations may not be the most practical energy storage solution. Batteries now have to demonstrate the ability to store energy efficiently, within certain power, lifetime and safety specifications, at a price-point that is ultimately affordable by the energy industry, before they can be considered for bulk power management.

SIRBATT (Stable Interfaces for Rechargeable Batteries) is a multisite collaborative project consisting of 11 full partners from the European Area (six Universities, one Research Institute and four industrial partners). The diversity of the research organisations in the partnership was been chosen to provide a wide range of complementary expertise in areas relating to the study of battery electrode interfaces, covering both experimental and theoretical aspects of this important contemporary area. SIRBATT has developed microsensors to monitor the internal temperature of lithium cells in order to maintain optimum operating conditions to allow long-life times that can be scaled for use in grid scale batteries. These cells will comprise of candidate electrode materials in which the complex interfacial region and surface layers were well characterised and understood via utilisation of a suit of advanced in situ and ex situ measurement techniques, complemented by the application of transformative modelling methods. The knowledge from these studies is presently being exploited to develop candidate electrode materials with an optimised cycle life and stability, for example by the use of novel stable lithium salts and the inclusion of stable film forming additives into the electrolyte.

Scientific Aims and Objectives

The scientific aim of SIRBATT is, therefore, a radical improvement in the fundamental understanding of the structure and reactions occurring at lithium battery electrode/electrolyte interface which will be achieved through an innovative programme of collaborative research and development. The three scientific objectives were to:

1. Fully characterise the chemistry of interface layers (solid electrolyte interphase and passivation films) with a suit of advanced in situ analytical techniques and state-of-the-art computational methods on a selection of lithium battery electrodes in conventional and non-conventional electrolytes;

2. Use the generated understanding of the interface layers to develop electrode/electrolyte interfaces that allow high cycle life (for example to move towards achieving 5000+ cycles, to 80% initial capacity) of lithium cells, thereby demonstrating that lithium batteries are sufficiently robust for grid scale batteries, due to their long-term stability;

3. Develop microsensors to monitor lithium cells in order to maintain optimum operating conditions to allow the operation of long-life time devices that can be scaled for use in grid scale batteries.

Project Results:
SIRBATT has undertaken a wide program of experimental and theoretical research of Li-ion battery electrode interfaces and integration of Li-ion cells with microsensors. Below is a summary of results achieved from the project that are grouped in accordance to addressing the three scientific objectives:

1. Fully characterise the chemistry of interface layers (solid electrolyte interphase and passivation films) with a suit of advanced in situ analytical techniques and state-of-the-art computational methods on a selection of lithium battery electrodes in conventional and non-conventional electrolytes;

Modelling and simulation of electrode interface

Characterisation of buried interfaces between epitaxial Li4Ti5O12 (Li7Ti5O12) and alpha-Li2TiO3 focused on Density Functional Theory (DFT) investigation the lowest formation energy structure for the buried interface [alpha-Li2TiO3//Li4(7)Ti5O12] formed after the first Li-intercalation in spinel Li4Ti5O12 (111)-oriented single crystal.

The energy favourable models were characterised in terms of electronic structure and elastic properties. Based on the computed phase-diagram for the considered phases, growth of an extended α-Li2TiO3 layer on Li4Ti5O12 substrates, acting as solid electrolyte, could be achieved by suitably tuning bias and O2 contamination in the electrolyte.

Structural and electronic characterisation of Li4Ti5O12 and LiMn2O4 interfaces focused on interfaces of Li4Ti5O12 and LiMn2O4. Specifically, DFT screening of several α-Li2TiO3 surfaces was carried out with the aim of identifying the energy favoured termination of the phase experimentally detected after the first Li insertion (extraction) in (from) single crystal Li4Ti5O12 (111). The simulations indicate the α-Li2TiO3 (100) surface to be energetically favoured, which explains the experimentally observed tendency to faceting of α-Li2TiO3 phases grown epitaxially to Li4(7)Ti5O12(111). The role of Li vacancies in promoting organic solvent (ethylene carbonate, EC) decomposition was also investigated and its non-negligible role in activating formation of solid electrolyte interphase (SEI) on alpha-Li2TiO3 demonstrated.

Given the current debate on the optimal simulation approach for LiMn2O4 bulk and surfaces, we started by considering the problem of bulk and evaluated the role of different simulation protocols for the agreement with available experimental results on the structure, cation and magnetic ordering of the material. The bulk-focused work was extended by the investigation of the relative energy of several LiMn2O4 surfaces and the dependence of the results on the adopted simulation protocol for different magnetic orderings. For the lowest-energy surface models, we then studied the adsorption of both EC and dimethyl carbonate (DMC) on LiMn2O4 surfaces, paying special attention to the effects that different approaches to the description of both 3d sites (Mn) and van der Waals corrections may have on the simulated results. An extended characterisation of the electronic structure for the considered vacuum-exposed surface and LiMn2O4/solvent (EC and DMC) interfaces was presented and discussed with respect to both experimental and, where available, theoretical data. The current unavailability of simultaneously atom- and time-resolved experimental characterisation of the considered interfaces makes the results presented novel and of potential high interest to the experimental community. Most notably, the simulations revealed that Mn2+ sites in LiMn2O4, for specific crystallographic cuts [i.e. (111) Mn and Li terminated and (111) Li-terminated] are a ground state property of the surfaces, whose formation does not require electrochemical cycling of the substrate.

Identification of structural and electronic factors governing Li-ion diffusion and ageing of Li4Ti5O12 and LiMn2O4 interfaces focused mainly on LiMn2O4 given the larger amount of experimental characterisation results from the literature. Given the lack of published result on the matter, we focused on the surface-dependence interplay between cathode state of charge, Li-ion intercalation, and Li-ion co-adsorption in electrolyte/LiMn2O4 cathode interfaces.

Based on experimental compelling evidence of a marked surface-dependant difference in the reactivity, SEI formation and evolution of LiMn2O4 cathodes exposing (111) and (110) surfaces, we carefully considered the structure, electronic properties and dynamics of model (111) and (110) interfaces with a standard organic solvent for the electrolyte (EC) as a function of the state of charge of the system, which was found to induce semiconductor to metal transitions in the substrate. This line of work was extended by considering the role of strain in tuning the reduction drive of LiMn2O4 surfaces. The modelled effect was substantial with changes up to over one 1 eV in the substrate work-function for strain values as small as 5%. The results indicate that strain-engineering of LiMn2O4 surfaces via metal-oxide core-shell nanoparticle synthesis, could be very meaningfully explored towards definition of zero-bias SEI-forming layers. See exploitable result 6.

Furthermore, we carried out also a detailed study of the effect of Li-ion intercalation in EC/LiMn2O4(111) and EC/ LiMn2O4 (110) interfaces on the overall electronic properties of the system as a function of the intercalation depth of the Li-ion. This work was complemented by explicit simulation of the role of hole-excess, as expected during the initial stages of cathode discharge (battery re-charge), for the electronic properties and dynamical evolution of lithiated EC/LiMn2O4(111) and EC/LiMn2O4(110) interfaces. This work revealed an important role of intercalating Li-ions at the surface in altering the redox properties of surrounding electrolyte molecules, making the latter substantially more reducible. Based on these results, atomistic simulation of solvent decomposition on LiMn2O4 surfaces should not overlook the presence of ion adsorption on the surface.

Structural and electronic characterisation of metal-oxide coated LiMn2O4 interfaces and composition of Graphitic carbon anodes SEI’s considered Al2O3 coating of LiMn2O4 surfaces and modelling of redox shuttle electrochemistry towards understanding of the SEI passivating role for solvent reduction.

Following validation of novel procedures for computationally efficient pre-optimisation of amorphous Al2O3 models based on carefully controlled approximations, we investigated the possibility of coating-driven change in the relative energy, thence stability, of different LiMn2O4 crystallographic faces. The role of the amorphous Al2O3 coating and its interface with LiMn2O4 for the electronic properties of the composite was also studied. Notably, for all the considered Al2O3–coated LiMn2O4 substrates, we found the coating to strongly interact with LiMn2O4 and invariably donate electronic charge to LiMn2O4. We also validated directly the experimentally suggested proposition that hetero metal-oxide coating of LiMn2O4 should be effective in preventing Mn disproportionation (2Mn+3→ Mn+2 + Mn+4) at the LiMn2O4 surface resulting in dissolution of Mn+2 into the electrolyte. We found this not to be the case, given the electron transfer from the Al2O3–coating to the LiMn2O4 and the ensuing reduction of Mn3+ sites at the interface. Accordingly, electron-accepting coatings, i.e. coatings with a larger electron affinity than LiMn2O4 or providing different interface charge redistribution, should be designed to suppress Mn disproportionation in LiMn2O4. These elements, currently unavailable in the specialised literature, should be useful for future, better informed, strategies toward stabilisation of LMO cathodes via hetero metal-oxide coatings.

With the aim of improving our understanding of battery ageing, we investigated the passivation degree of the SEI towards solvent reduction. The passivating character is characterised by two properties: the electron transfer at the passivated electrode and the solvent transport through the SEI film. To describe these properties, a characterisation method based on introduction of a redox shuttle in the electrolyte was implemented. The way the electrochemistry of the shuttle is modified by the presence of the film provides information about the passivating character. Mathematical models were developed to simulate voltammograms at either (rotating or stationary) flat or porous electrodes, before or after the SEI formation. The models account for the presence of a soluble redox couple that is used as an electrochemical sensor of the SEI properties. Ferrocene/ferrocenium was selected as the shuttle. From a model-based analysis of data collected after SEI formation with different formation charges at flat electrodes, SEI-related parameters such as its porosity and the effective kinetic parameters of the shuttle reaction (modified because of the presence of the SEI) could be determined. The larger the formation charge, the lower the SEI porosity. The kinetic rate constant of the shuttle reaction was also found to decrease with an increase in formation charge. These parameters directly reflect the SEI quality in terms of passivation. The approach was then extended to graphite porous electrodes. We found that SEIs formed in similar conditions as those at a glassy carbon electrode are way less passivating toward the shuttle reaction. Finally, voltammograms of the shuttle at pristine porous graphite electrodes were analyzed with the model in order to determine geometric parameters of the porous electrode, such as its electroactive surface area and its tortuosity.

Lithium-ion Electrode Interfaces in Conventional Liquid Electrolytes

Materials, electrodes and charge/discharge behaviour. Two generations of active anode and cathode materials have been synthesised and characterised (structure, morphology, and electrochemical properties). As first generation anodes, graphite SBG-1, ZnFe2O4-C and Li4Ti5O12 (LTO, prepared by flame-spray pyrolysis (FSP)) have been synthesised. As first generation cathodes, LiMn2O4 ((LMO, synthesised by flame-spray pyrolysis and comparative solid state literature routes) and LiFePO4 (LFP commercial sample supplied by ALEEES) have been used. As second generation anodes, improved graphite SBG-2, Zn0.9Fe0.1O-C C-coated LTO have been synthesised. As cathodes, several surface coated LMO and LiNi0.5Mn1.5O4 (LNMO) samples have also been synthesised.

It has been demonstrated that for all materials a carboxymethyl cellulose (CMC)/H2O formulation allows the preparation of stable electrodes, being at the same time cheaper and greener than the use of polyvinylidene fluoride (PVdF)/ N-methyl-2-pyrrolidone (NMP) binder/solvent system. See exploitable result 7.

Regarding the first-generation materials, LTO prepared by FSP shows good performances, suggesting that the FSP synthesis is a viable method for obtaining high-performance LTO. In particular electrodes based on this optimised LTO nanopowder, using CMC as binder, and copper as current collector revealed excellent rate performance, providing specific capacities of 133, 131, 129, 127, 124, and 115 mAh g-1 when applying C rates of 1C, 2C, 5C, 10C, 20C, and 50C, respectively. Targeting the commercial application of thus synthesised nanoparticles, we optimised also the electrode composition, comparing three different binding agents (CMC, PVdF, and poly(acrylic acid), PAA) and substituting the copper current collector by aluminium. The results of this comparative analysis show, that the combination of nanoparticulate LTO, CMC, and an aluminium current collector appears most suitable toward the realisation of environmentally friendly and cost-efficient lithium-ion anodes, presenting very stable cycling performance for more than 1000 cycles at 10C without substantial capacity decay.

In comparison, the LMO sample prepared by FSP shows poorer performances (capacities lower than 100 mAh g-1) than those prepared by comparative route. This has been ascribed to the nano-dimensional nature of the material, promoting additional degradation mechanisms.

Cycling data of the graphite (SBG-1), ZFO-C and LFP show capacity values in line with the theoretical ones (ca. 350, 1000 and 170 mAh g-1 respectively).

Progressing onto the Second-generation materials, SBG-2 graphite (with a modified surface) shows improved rate capability than SBG-1 at room temperature (25oC), but reduced stability in LP30 (ethylene carbonate, EC: dimethyl carbonate DMC 1:1 w/w 1M LiPF6) at high temperature (50oC) (probably because of high surface reactivity). C-coating LTO using cellulose results in an improved rate capability and interfacial stability. At 50oC the stabilities among uncoated and C-coated LTO are comparable. Li2TiO3 phases were also grown over TiO2 substrates and their bulk and surface compositions characterised. Coating of LMO by oxides, results in improved rate capability and interfacial stability, especially at high temperature (50oC), although C-coating is ineffective. LNMO prepared by a solid state route exhibits good performances, in contrast to one prepared by FSP. Oxide coating of LNMO resulted in decreased performances. Zn0.9Fe0.1-C is more stable at high temperature than 1st-gen ZnFe2O4-C.

Ex situ measurements

An innovative solution for monitoring the growth of the SEI was developed for X-ray methods. Arsenic (As) atoms present in the electrolyte solution were used as a local probe for monitoring the SEI growth on different electrodes.

The electrolyte salt commonly used in Li-ion cells is LiPF6, even with potential issues coming from possible HF formation because of PF6– decomposition in the presence of moisture. X-ray experiments have been performed using a mixture of LiPF6 and LiAsF6 electrolyte salts to exploit X-ray beam cross-section at the As K-edge within the composite electrodes. In fact, while LiAsF6 has been discarded for commercial applications because of its toxicity, however its lithiation reactions are similar to those of LiPF6 in which for both electrolytes the LiF and Li(As/P)Fx phases are expected to be the components of the inorganic, inner layer of the SEI.

High-quality As K-edge spectra were collected in fluorescence mode for a set of carbon-coated ZnFe2O4 and graphite (SBG-1 and SBG-2) electrodes allowing the characterisation of the inorganic components of SEI formed over graphite and ZFO-C electrodes. Detailed structural results were derived for different Li contents in different electrodes including the estimated thickness of the SEI layer, contribution of different As oxidation states, and As local structure.

In different experiments, involving “soft X-rays” interaction with a SEI formed by decomposition of LP30 electrolyte (LiPF6 1M in EC:DMC 1:1), the combination of X-ray absorption spectroscopy (XAS) in total electron yield (TEY) and total fluorescence yield (TFY) modes allowed to follow the dynamics of growth of SEI over ZFO-C (initial nucleation around carbon, then percolation through electrode and growth of a bulk SEI).

Correspondingly, XAS/X- ray photoelectron spectroscopy (XPS) combined spectroscopies followed SEI formation over ZFO-C and demonstrated the reversibility of an external “gel-type layer” of SEI. Its composition consisted in part of di-(Li alkyl carbonates) and it was found to contribute to the reversible capacity. Evidence of a partially reversed redox process, taking place within the SEI by charge–discharge cycling, was also obtained.

This was confirmed by XPS analysis of SEI surface over graphite (SBG-1 and SBG-2) binder-free electrodes, finding different distributions of organic (Li2CO3, ROCO2Li, polyethylene oxide (PEO), ROLi) and inorganic species (LiF and LixPFyOz).

Besides conventional XPS experiments, hard X-ray photoemission (HAXPES) was performed on carbon-coated ZFO. These HAXPES experiments allow to obtain compositional information in the same way XPS do but with increased probing depth which is very useful for SEI studies. The combination of XPS and HAXPES provided a picture of the outermost surface layer (SEI region) combined with subsurface information (electrode/SEI interface). In the case of ZFO, the formation of lithium oxide at the very end of the conversion reaction of C-coated ZFO has been identified. This could be associated to a gradual reaction from the core to the particle surface, even segregation of metallic species. The formation of metallic Fe and Zn is detected by means of XPS and HAXPES as well as the thickness of the SEI layer.

In situ measurements

EIS characterisation of LTO anodes prepared with PVDF, CMC, or PAA binders allowed to assess the stability of CMC-based electrodes. The EIS response of all 1st and 2nd generation materials was investigated at T = 25 °C and T = 50 °C, in order to monitor the ageing of the cells and, particularly, of the passivation layers formed at electrode/electrolyte interface. All the Nyquist plots revealed features related to passivation layer, charge-transfer and diffusion to blocking electrodes. Among the other results, it has been demonstrated that, in LP30 electrolyte, for the graphite, SBG-2 undergoes a more pronounced ageing than SBG-1 at high temperature. Coated LTO and LMO appear instead more stable, even at high temperature, than the respective uncoated first-gen LTO and LMO. Also Zn0.9Fe0.1-C showed higher resistance to ageing at high temperature than ZnFe2O4-C.

For in situ Raman, a three-electrode configuration cell has been used to collect spectra during the first discharge-charge cycle of different anode (SBG-1, SBG-2, ZFO, ZFO-C, Zn0.9Fe0.1O-C and LTO) and cathode (LFP, LMO, LNMO) materials. The spectra show the evolution of characteristic peaks of each material as the lithium is inserted or extracted. Among the most interesting results: (a) SBG-2 graphite presents a sharper G peak and the relative intensity of this peak with respect to the D band is also higher when compared with the SBG-1 spectrum, suggesting a more ordered structure in SBG-2 or the presence of fewer defects at the edges; (b) LTO shows the presence of an anatase form, which has a great contribution in both the electrochemical and spectroscopy results; (c) the carbon coated ZnFe2O4 shows the appearance and disappearance of peaks that can be related to SEI compounds at potentials between 0.80 and 0.69 V during the discharge process, with a temporary enhancement of the Raman signal (surface enhanced Raman spectroscopy, SERS) produced by Zn nanoparticles formed during the conversion reactions in the same potential window, while at lower potential these bands disappear due to an increase in the SEI thickness inhibiting the SERS or the formation of LiZn during the alloying reaction; (c) for Zn0.9Fe0.1O-C the SEI-related signals observed for first-gen ZFO-C are lacking due to the fact that the signal-enhancing Zn is formed at much lower potentials (below 0.3V) than those corresponding to SEI formation. See exploitable result 5.

In situ XAS measurements were carried out using a modified closed electrochemical cell from EL-CELL. The cell geometry was modified for the X-ray absorption measurement optimisation. The In-situ measurement was carried out on Fe K-edge during the first half cycle of the lithiation. Theoretical spectra of model compounds with several possible cations and O lattice arrangements have been predicted in order to fit experimental data. Preliminary results confirm the metallisation and amorphisation of the Fe metal in below 0.7 V and the inversion of the spinel structure within the lithiation process. Further analysis is in progress.

Lithium-ion Electrode Interfaces in Non-Conventional Electrolytes

SIRBATT synthesised new salts and electrolyte additives and then tested them within novel electrolyte formulations against selected generation one and two electrode materials as half cells versus lithium metal.

Most of the synthesis are one step synthesis and cut the material used in final electrolyte of almost by half. New salts LiTDI and LiPDI (lithium 4,5-dicyano-2-trifluoromethylimidazolate and lithium 4,5-dicyano-2-pentafluoroethylimidazolate) 1st generation, as a new non-conventional salt were synthesised and sent to partners with a procedure of making liquid electrolyte. As an additive the LiHDI (lithium, 5-dicyano-2-heptofluoropropylimidazolate) was proposed in 1st generation of additive material. Based on our experiments the 2nd generation of non-fluorine salt LiPCP (lithium pentacyano propenide) was proposed and LODI as an additive. New salts exhibit very high thermal stability up to 350oC and were stable within a wide electrochemical stability window.
In order to assess the stability of the new lithium salts, soluble redox organic molecules were used as reducing agent. The reduced form of these molecules were first prepared by lithiation in tetrahydrofuran (THF) and the new salts were then added to the as-obtained solution. Reduced forms of the molecules are radical anions and provide a strong colour to the solution. A change in colour upon addition of the salt is ascribed to the oxidation of the radical anion along with the chemical reduction of the lithium salt. Hence, it is a handy and visual method to assess the salt stability.
Results indicate that both FSI- and FTFSI- easily reduce (presumably because of the F-S bond cleavage). On the other hand, there is no colour change for PF6-, TFSI-, TDI- and HDI-, even in the reduced form of dibenzene, which indicates that they are stable at potentials of 0.5 V vs Li or higher.

Electrochemical stability of new salts

In parallel with the chemical stability experiments, the SEI formation on graphite electrodes in carbonate electrolytes containing the different salts was studied. SEI formation tests were conducted on two different graphite materials, namely the SBG1 graphitic carbon, as well as a synthetic graphite (KS-6, supplied by Imerys). The comparison of the differential capacity plots on both graphite materials shows consistently a large magnitude of the SEI formation peaks with KS-6, which is ascribed to its large BET surface area. Results also show that the onset of the SEI formation peak is also shifted to slightly higher potential for KS-6 (10 to 15 mV). This effect is minor with respect to the salt effect, but seems to be consistent amongst all samples (every experiment was repeated twice).
The type and concentration of the Li salt in the electrolyte is shown to have a greater effect on the potential shift of the SEI formation peak; the trend of the onset potential is LiHDI > LiTDI > LiPF6 (maximum of the peak at 0.93V > 0.85V > 0.78V respectively). This is clearly an effect of the anion.

Ionic liquids

Ionic liquids are salts in which the ions (cation and anion) are poorly coordinated, which results in these solvents being liquid at room temperature. Seven types of Ionic liquids have been synthesised as a possible candidate for SIRBATT baseline materials. Of the obtained ionic liquids, two ionic liquids were selected out of the preliminary screening by conductivity measurements and electrochemical stability tests: these were 1-Butyl-3-methylimidazolium 4,5-dicyano-2-(trifluoromethyl)imidazolide (BMImTDI) and 1-Butyl-3-methylimidazolium 4,5-dicyano-2-pentafluoroethylimidazolate (BMImPDI). All ILs were characterised thermally and electrochemically showing the potential in application as an alternative for conventional electrolytes.

Investigation of lithium metal – electrolyte interfaces

SIRBATT characterised the lithium metal interface with a variety of ex situ and in situ techniques to better understand how the SEI changes with cycling and under certain storage conditions. In order to best study the lithium deposition process a new in situ setup for light (or optical) microscopy was developed specifically for the research of Li metal electrodes. This method is ideal to investigate the dendrite growth in situ, one of the greatest safety concerns of Li metal anodes.

The initial literature survey led to a focus of our research on non-conventional electrolytes as carbonate solvent-based systems typically lead to low cycling efficiency and strong dendrite growth.

Work was dedicated to the synthesis of electrolytes suitable for the use with Li metal electrodes. As identified within the literature survey, novel electrolyte systems, mainly based on ionic liquids, were prepared. This work went in close collaboration with WP5, to make use of the novel anions developed within that part of the SIRBATT project.

Various experiments were performed to develop a deeper understanding of the SEI on Li metal and how this affects the electrochemical performance as well as safety issues (e.g. dendrite growth).

We could prove that the stable cycling of Li metal electrodes in ionic liquid based electrolytes is possible for several thousands of hours. During a long-term cycling experiment in Pyr14TFSI with LiTFSI (9:1 mol ratio), no dendrite growth that would represent a safety problem was detected. However, it was found that during long operation, the SEI grows thicker and thicker, which finally to a large overvoltage during cycling limiting the energy efficiency.

Key for the safe operation of lithium anodes is the choice of an appropriate electrolyte and the permanent control of the cell voltage (i.e. electrode potential).

A great variety of experiments was performed to clarify what testing conditions actually lead to a stable and safe operation.

If the electrode is overloaded, i.e. a too strong current or voltage pulse is applied, strong electrolyte decomposition as well as SEI crack formation is found, which directly goes along with dendrite growth and pitting dissolution.

Post-mortem SEM investigations of cycled electrodes revealed that dendrites form in nanofibres in bundles inside cracks of the SEI. It is likely this dendrite growth process actually causes the SEI cracking. To further clarify the process of morphological changes of Li metal electrodes, a novel setup for in situ light microscopy was developed. Experiments with this setup allowed the investigation of dendrite growth and unveiled that pitting dissolution, despite not being of direct safety concerns, is another deleterious process that happens at strong electrochemical biases. The so far mostly neglected presence of pitting dissolution on Li metal electrodes, proven with in situ light microscopy, might potentially be the first damage to the Li surface and the trigger for dendrite growth. To this end, three electrode measurement of symmetrical Li metal cells, enriched with impedance spectroscopy, clarified that the electrode potentials of the plated and stripped electrode strongly diverge within a half-cycle. In fact, it was observed that the stripped electrode is polarised more strongly, pointing to the option that Li stripping is more difficult than plating, which is rather unexpected following the omnipresent discussion of solely dendrite growth.

To get more insight into the chemical composition of the SEI on Li metal electrodes, in situ Raman and IR experiments were conducted. However it was found that the high energy of Raman laser irradiation inevitably leads to electrolyte decomposition on the electrode. Therefore, conventional Raman experiments (in situ or ex situ) were found to be unusable and the SHINERS technique was applied. As the process of Au nanoparticle integration to the metal surface typically involves aqueous solvents, an alternate route was followed. The SHINERS nanoparticles were deposited on the Raman window instead of on the electrode itself. First measurements proved that, in case of close proximity of the Raman window to the electrode (thin electrolyte film), this approach is a valid option, as the Au nanoparticles led to strong signal enhancement of species on the Li metal surface despite not being integrated in the surface film directly. This method, as soon as fully developed, has the potential to generally increase the usefulness of SHINERS as it allows a much broader range of samples to be investigated.

A great step towards the further understanding of Li metal electrodes is clearly the investigation of the primary SEI (the native SEI and artificial alternatives) conducted within this project. A large set of UHV experiments, in collaboration with DFT modelling of WP3, focused on the products of exposure of ultraclean Li surfaces with N2, CO2 and O2, gases that are typically present during the synthesis of Li foils for battery research. It was found that Li does not easily react with dinitrogen to form Li3N, but that this process usually involves surface contaminants like LiOH. The products formed by O2 and CO2 exposure lead to a lower work function than for pure Li, a very uncommon feature of a “passivation” layer.

Following this fundamental research, a new setup for the creation of pure Li surfaces and in situ treatment with liquids was developed that can be applied with standard lab equipment (no UHV necessary). This work was intended to replace the omnipresent native SEI film with an artificial primary SEI. First experiments with amines led to a stable Li surface and are a promising start in this new research area.

The properties of LiTDI and LiHDI in EC/DMC solvent mixtures were investigated as possible future electrolyte systems. It was shown with impedance spectroscopy that the SEI formed on lithium metal anodes in those novel electrolyte mixtures is more stable than standard LP30. Especially the addition of low amounts of LiHDI leads to a stable interfacial resistance. These imidazolium anions based new anions, developed mainly in WP5, were also implemented into polymer electrolyte mixtures on a PEO basis (PEO + Li salt + corresponding ionic liquid). However, those ternary systems were found to be hardly conductive and hence, their performance for lithium plating/stripping experiments was consequently inferior to other existing technologies.

SIRBATT has therefore set up and developed a wide range of ex situ and in situ measurement techniques to critically examine the chemistry of the bulk and the surface of electroactive materials, as well as to follow the development of passivation and SEI films on electrode surfaces. SIRBATT has synthesised a wide range of anode and cathode Li-ion battery electrode materials via a number of synthetic routes and tested them in conventional and additive containing electrolytes at different temperature and cycling regimes. Green electrode formulations through processing from water using a fluorine-free binder has been a further highlight.

2. Use the generated understanding of the interface layers to develop electrode/electrolyte interfaces that allow high cycle life (for example to move towards achieving 5000+ cycles, to 80% initial capacity) of lithium cells, thereby demonstrating that lithium batteries are sufficiently robust for grid scale batteries, due to their long-term stability.

From the detailed electrode/electrolyte interface studies it was decided to evaluate long cycle life of a lithium-ion coin cell, consisting of a graphitic anode and lithium iron phosphate (LiFePO4; provided by ALEEES) as the cathode. This cell was cycled for over 2500 times, showing 40% capacity fade. 1750 cycles were achieved with 20 % capacity fade. This exceeds the project target for 500 cycles and moves us a step closer to the long-term goal for Li-ion batteries to the achievement of 5000+ cycles. When the capacity values obtained by single side coin cell electrode are projected considering the assembling of the final cells, cell level specific energy and energy density values of 95 Wh/kg and 150 Wh/dm3, can be estimated that were lower than our original target due to the compromise of selection of a lower voltage Li-ion system to enable longer life cycling.

Furthermore from work on lithium metal it was shown that stable cycling of Li metal electrodes in ionic liquid based electrolytes is possible for several thousands of hours. During a long-term cycling experiment in the ionic liquid electrolyte 1-butyl-1-methylpyrrolidinium bis(trifluoromethanesulfonyl)imide (Pyr14TFSI) with lithium bis(trifluoromethylsulphonyl)imide (LiTFSI) (9:1 mol ratio), no dendrite growth that would represent a safety problem was detected. See exploitable results 3 and 4.

3. Develop microsensors to monitor lithium cells in order to maintain optimum operating conditions to allow the operation of long-life time devices that can be scaled for use in grid scale batteries.

Micro-Sensors for Real Time Cell Control and Performance and Management of Large-Scale Batteries

The main objectives of were (i) the definition of large scale application scenarios, (ii) the investigation of micro-sensors for lithium ion batteries, and (iii) cell performance analysis. A detail review of the last advances reported in the literature was carried out in order to identify the most interesting solutions to be applied to the cells. Fibre Bragg grating sensor networks were designed for real time monitoring of internal temperature and pressure, during charging and discharging under various conditions. The performance of such micro-sensors was evaluated in the lab for generic operating conditions of the cells.

Definition of large scale application scenarios

The requirements in backup electrical batteries systems for nuclear power plants were modelled, as substitute of actual fuel systems. So, realistic requirements for such batteries were developed, for usual operations and for emergency conditions, enough to guarantee a proper performance in all circumstances. We derived the design requirements (such as ambient conditions, operating conditions, discharge curves, and others) for batteries that could operate in nuclear power plants are defined. A specific scenario for lithium-ion batteries was proposed.
Also, specific requirements such as qualification procedures that shall be undertaken in order to qualify the batteries for nuclear plants, were described in this deliverable. So, type testing (test plan, test sequences, acceptance criteria) needed in order to fulfil this qualification process was described in this document. Based on information from the application analysis, large-scale analysis (including: operating temperature conditions duty cycles, energy throughput, etc.), cell level testing analysis was carried out through cell models.
These models were utilised to perform a full electrochemical and thermal characterisation, key parameters for the calibration and optimisation for the Battery Management System. The data obtained were used to build an electrothermal virtual model which will allow support the application analysis through simulation.

Investigation of micro-sensors for lithium ion batteries

For theses analysis, we have supplied characteristics data for one of the plant battery banks, in order to simulate a similar bank capacity with the batteries developed. Data supplied was related to system voltage, nominal capacity, internal resistance, short-circuit current, equalisation voltage, discharge current chart (A/V), and the demand curve for battery (A/t).

Fibre Bragg sensors were chosen for performing real time measurements of the battery performance. They have the advantages of being relatively small and flexible and they offer immunity to electromagnetic radiation and multiplexing capabilities. The sensors were recorded on commercially available photosensitive optical fibres. The thermal calibration of sensors was made inside the thermal chamber in a range of temperatures between 10 oC and 35 oC, in steps of 5 oC.

Bragg sensors assembly in pouch cells was successfully achieved. Internal sensors were located between the separators, one placed in the central part of the active area and other placed near of the tab-electrode connection. In addition, external sensors were placed in direct contact with the surface of the pouch cells and thermal paste was used to increase contact surface and conductivity. Five charge-discharge cycles at various C-rates (1C, 2C, 5C and 8C) (please note a C-rate is a measure of the rate at which a battery is discharged relative to its maximum capacity. A 1C rate means that the discharge current will discharge the entire battery in 1 hour) were carried out in order to analyze the response of the sensors. As expected, the response followed the galvanostatic cycling and the highest wavelength variation occurred at 5C and 8C, with Δλ values between 0.05 and 0.07 nm. In general, it was found that the battery performance did not change significantly upon the sensor network integration and it could be monitored successfully.

To summarise the sensor work within SIRBATT research, we were able to integrate optical fibre sensors into Li-ion pouch cells. Such sensors can be used to investigate different physical and chemical parameters. For the discrimination of pressure and temperature within the pouch bag, a hybrid sensor composed by two sensors connected in series (a fibre Bragg grating and a Fabry-Perot interferometer) has been developed. Both sensors exhibited different sensitivities to temperature and pressure. Using the appropriate signal processing, it was possible to infer the variations of these two parameters. This sensor revealed to be much more precise and reliable than any other existing technology (integration into battery gives much faster and more relevant results than external measurement). And our approach solves the separation of pressure and temperature influence to the actually measured signal shift, which has not been fully addressed by competing research groups. It should be highlighted that temperature and pressure control is very important for the Li-ion cell safety as well as lifetime.

Cell performance analysis

The initial part of the work was the investigation of the performance at cell level to support the application analysis for large-scale devices. Originally, it was planned to perform testing on three of the six enhanced cells with integrated sensing units. One cell was to be utilised to perform a full electrochemical and thermal characterisation, thus providing technical understanding of the performance of the cell as a function of temperature and state of charge and key parameters for optimisation for the Battery Management System. The second and third cells were to be used for ageing tests.

The results from the tests conducted do show that the embedded Fibre Bragg Grating micro-sensors inside the cell only show a small thermal lag to exterior environment, have faster response to pulse testing and display a higher recorded temperature compared to exterior thermocouples. This testing study has shown that internal sensors have the benefit of being more responsive and have more accurate temperature readings of the cell under a variety of conditions compared to that of the traditional measurement of thermocouples placed on the outside of the cells.

The second part of the work to be carried out was for three of the prototype cells with embedded micro-sensors to be built into integrated module system. However, this was no longer feasible due to unforeseen cell situation. As an alternative, JMBS proposed to model and simulate a 12 V module based on data from testing data generated within SIRBATT.

For development of the conceptual 12 V module, performance data; rate capability and thermal behaviour recorded during testing by both external and internal sensors was used. This data however was not extensive and robust enough to fully develop the virtual module, the model was therefore developed partially upon data obtained from previous work package characterisation and supplemented with commercial LFP based pouch cell data.

To account for various features in the module, two models were developed in two different environments: a detailed Electrochemical-Thermal model available for integration at system level in Matlab-Simulink; and a detailed cell thermal model for integration into a module/battery thermal model in ANSYS.

In summary, Li-ion for stationary storage has been considered via a paper study as an alternative to lead (Pb)-acid for back-up power for nuclear power plants. The requirements for such a system have been investigated and the key parameters that a Li-ion system must meet have been documented.

Microsensors that can sense both the change in temperature and internal pressure of the cell have been incorporated into pouch (or coffee-bag) cells containing Li-ion negative and positive electrodes and tested under real cell conditions. It was revealed that the microsensors were much more time responsive than conventional thermocouples. Furthermore microsensors have been used to as an HF sensor via within battery electrolytes. The early detection, and in real time, of HF is of utmost importance for the Li-ion cell safety as well as lifetime. See exploitable results 1 and 2.

By combining the data of the requirements for a backup power batter system and electrochemical land thermal data (from the microsensors) a cell –level model was developed.
Potential Impact:
As described in the previous sections and shown by the scientific publications published in leading peer-review journals (see the Annex I: “scientific publication list”) SIRBATT has undertaken leading research in the fields of organic and inorganic materials chemistry, electrochemistry, advanced electrode and electrode interface characterisation and modelling and micro sensor technology. The significant impact outcome from the research was the successful incorporation and testing of microsensors within a Li-ion pouch cell. The microsensors were shown to be much more time responsive than conventional thermocouples.

Specifically microsensors for batteries could have substantial impact in the design, maintenance and controlling safety of small and large stationary battery installations. Further development of these sensors would involve optimisation of delivery of both temperature and pressure within a single sensor.

A potential short term impact wold be the use of these sensors by cell manufacturers to test cell safety profiles during production and or battery module and pack designers to understand how to manage heat load and distribution within their large format batteries.

The SIRBATT project has had impacts through the development of microsensors in-line with the Cooperation Work Programme (2013 EC (2012) 4536 as of 09 July 2012) which is based on the EU strategy document: ‘Europe 2020: A strategy for smart, sustainable and inclusive growth' in that it will led to growth that is smarter, more sustainable and more inclusive. Specifically SIRBATT will boost research and innovative in a range of key EU priority areas relating to climate change and energy:

Smart Grids (and Energy Storage) through the European Technology Platform for Electricity Networks of the Future (SmartGrids ETP) that is encouraging the integration of decentralised renewable energy sources, more flexible and smarter electricity grids with appropriate storage options, better facilitate the connections and operation of generators of all sizes and technologies and significantly reduce the environmental impact of the whole electricity supply system.

Employment of stationary energy storage technology to support a more distributed and diverse power generation network through the development of microsensors that allow this interconnection and flow of information on the status of health of large battery installations.

Under the Energy work Programme (Work Programme 2013 – Cooperation theme 5 – ENERGY – EC C(2012) 4536 as of 09 July 2012.) it is stated that there is an urgent need to bring new, high performance technologies to the market and to ensure EU leadership on low carbon energy technologies and that this scale requires a European approach.

The achievements of the SIRBATT project was fully in-line with supporting the goals of The Energy Road Map 2050 – that targets a 20% reduction in energy usage by 2020 and an 80-95% reduction by 2050 through a shift in the EUs current energy system to largely decarbonise the energy – This can only be achieved if localise energy storage technologies can be developed. Lithium batteries remain one of the most promising technologies to enable us to achieve this. By demonstrating extremely long cycle life a lithium-ion coin cell, SIRBATT has highlighted the prospective use of Li-ion technology as a for long life time applications. The SIRBATT coin cell consisted of a graphitic anode and lithium iron phosphate (LiFePO4; supplied by ALEES) as the cathode, has been cycled for over 2500 times, showing 40% capacity fade. 1750 cycles were achieved with 20 % capacity fade. This result exceeded the project target for 500 cycles and moves us a step closer to the long-term goal for Li-ion batteries to the achievement of 5000+ cycles with less than 20% capacity fade. The challenge is to take this excellent result at the coin cell level and replicate it at scale in order to demonstrate the robustness of Li-ion technology for grid applications.

Exploitable Results. Significant scientific outcomes from SIRBATT project:

Seven key findings from this three-year collaborative research programme are as follows, what partners will do with the exploitable foreground can be found in OVERVIEW TABLE WITH EXPLOITABLE FOREGROUND.

1. Integrated Microsensors for Temperature and Pressure Control

Fibre-Bragg Grating sensors have been successfully integrated in Li-ion pouch cells and were shown to be more responsive and more accurate compared to traditional thermocouples placed on the outside of the cells under a variety of testing conditions.

SIRBATT partners hope to take forward the knowledge of the microsensor integration gained from this work to develop sensing technology for safety monitoring next generation lithium-ion cells and other alternative battery chemistries.

2. Integrated Microsensors for HF Detection

Hydrogen fluoride, HF results from side reaction of the lithium salt lithium hexafluoro phosphate, LiPF6. A microsensor was developed that could be used to detect HF concentration within a Li-ion cell electrolyte

3. Artificial SEI generation by treatment of fresh lithium surfaces with liquid

It was experimentally observed that the history of Li foils (storage in e.g. dehydrated air vs. argon) has a large impact on the electrochemical performance of such pristine Li anodes. Alternative equilibrium based chemical strategies were explored to create a working SEI or SEI-precursor layer on Li metal anodes, before contact with the liquid or polymeric electrolyte. The gas-solid approaches are described in another key result due its largely differing application method. First of all, an experimental setup was developed that allows reproducible and easy preparation of liquid pre-treated Li foils. SIRBATT results with an amine as precursor and a non-reactive chemical (n-pentane) as reference indicate that the surface cleaning and pre-treatment with suitable precursor compounds leads to a Li metal anode with very reproducible electrochemical properties. This is very important for reliability of research data. Furthermore, and even more important, once the right compound is found, a strong improvement in electrochemical performance of Li metal anodes is to be expected, which would most certainly be commercially exploitable in next-generation batteries.

4. Safe operation limits of lithium metal anodes

The safety of Li metal anodes, mainly related with uncontrolled dendrite growth, still poses a large hindrance to the application of this material in rechargeable batteries. Via a combination of in situ light microscopy and electrochemistry the limits of the safe operation were determined, i.e. dendrite-free plating during charge. As long as the limiting current density is not exceeded, no dendrite growth or pitting dissolution was observed. In line with the determination of the limiting current density, the Li metal electrode potential leaves its characteristic shape during galvanostatic testing, which signalizes detrimental processes. Hence, controlling the current density or the electrode potential both allow the safe operation of Li metal electrodes. Our results give a guideline to how the necessary values can be obtained for different electrolyte systems.

5. SEI detection using temporary Raman enhancement

An unexpected detection of SEI growth products during the lithiation process of carbon-coated ZnFe2O4 was made. SEI components on the carbon coating were assigned from the Raman data as the polyethylene oxide type and organic alkyl carbonates (RCO2Li). This result is important as it demonstrates the suitability of in situ Raman spectroscopy to detect SEI products on an electrode interface under potential control. It enables a convenient method to follow the progress of SEI formation in real time. It also demonstrates that in situ generated metal nanoparticles could act as a temporary generator of the SERS effect.

6. Zero Bias SEI formation on transition metal oxide cathodes

Computational studies that show that strain-engineering of lithium manganese oxide spinel (LiMn2O4) surfaces (a lithium-ion battery cathode material) via metal-oxide core-shell nanoparticle synthesis, could allow the generation of zero-bias solid electrolyte interphase (SEI) forming layers. This result provides a new rational route for material design whereby the surface is fabricated to form a stable interface just by putting the material in contact with the battery electrolyte (removing the requirement for an electrochemical formed layer), thus conceivably removing the requirement for formation cycles during battery manufacturing.

7. Green Electrode Formulations

Green electrode formulations have been demonstrated to provide similar electrochemical performance to more conventional and widely industrially used electrode formulation. In particular, it has been demonstrated that for a wide range of Li-ion anode and cathode electrode formulations, a carboxymethyl cellulose/water formulation allows the preparation of stable electrodes, being at the same time cheaper and greener than the more conventional use of the fluorine-containing binder polyvinylidene fluoride and the harmful solvent N-methyl-2-pyrrolidone binder/solvent system.

Main dissemination activities

Partners were profoundly engaged in dissemination activities and attended and presented talks and posters on SIRBATTs research at numerous national and international conferences.

SIRBATT organised two international workshops in Spain and USA in 2015 and 2016 to disseminate the work from the project and to engage with the Sister programmes (Hi-C, Influence and Baccara) funded within the same call, as well as complementary European and USA research activity.

The first workshop in Spain, in September 2015 entitled “Understanding Lithium Battery Interfaces” had over 60 attendees. Representatives from Hi-C, Influence and Baccara gave the keynote and invited talks. Further external leading academics in battery interfaces from Europe gave invited presentations. SIRBATT work was well disseminated with talks from partners Universidade de Aveiro (Portugal), Universita Degli Studi di Camerino (Italy), Karlsruher Institut fuer Technologie (Germany) and Centre National de la Recherche Scientifique (France). Industry partners from the SIRBATT network presented their experience of energy storage with talks Johnson Matthey Battery Systems, UK) and Iberdrola, Spain.

Overall the Workshop provided a wide overview of the work currently being undertaken within battery science across Europe. It also provided an opportunity for networking and collaboration within the community.

The Project held its second Workshop ‘Controlling Lithium Battery Interfaces’ in the USA on in May 2016 which was attended by some 50 battery scientists. External leading academics in battery interfaces from the United State of America gave keynote and invited talks. Furthermore, two representatives from BACCARA, one of the other three successful projects with SIRBATT’s EC call, attended with talks, outlining results from their project’s work. From the University of Liverpool talks and posters were presented on battery interface research.

Talks and posters provided a wide overview of research currently being undertaken within the subject; and the Workshop ended with a summary and discussion of the current state of battery science across Europe and the USA. Future areas of research and possible exploitation were outlined and overall the event offered a unique opportunity for networking and collaboration within the wider battery community.

A Networking Workshop in May 2014 between China MOST and EU was attended by representatives from SIRBATT (ULIV and CIC) which gave an opportunity to disseminate European battery research activity within SIRBATT to the Chinese. The meeting gave an excellent opportunity to network with Chinese battery researchers.

Knowledge of the project goals was disseminated through an initial project leaflet, followed up with an updated leaflet with early results and achievements highlighted. Both leaflets where distributed to all the partners and they dispersed them to co-workers in their institution, visitors and to external collaborators, as well as circulating them at nation and international meetings and conferences.

An end-of-project brochure was produced that highlighted the main scientific outcomes from the project. This was very useful to disseminate the SIRBATT project in different events such as scientific/technology and industrial seminars, workshops, conferences and exhibitions.

Exploitation of Results

The application of microsensors within Li-ion cells that can provide both temperature and pressure readings has been identified by the SIRBATT consortium to be of great interest, as it fits into the long-term technological goal of advanced sensors to monitor and control the battery chemistry in order to prolong life. Additionally, the sensors can thereby be used to provide data, so that an accurate second life and economic value can be ascertained at the end of a particular application when it could be reused within another. As such a patent will be filed, potentially bringing in license money to fund further activities in the field of integrated microsensors. For the purpose of the patent application, SIRBATT partners are presently continuing our joint research work despite having no project funding. We will also contact other potentially interested companies for R&D projects. In addition, joint research proposals will be written if there appears a suitable call in H2020 or national funding agencies.

SIRBATT took advantage of the Support Services for Exploitation of Research Results (SSERR) and a meeting was held in Aveiro on the 30th November 2016 to assist Aveiro in how best approach the exploitation of key results 1. Integrated Microsensors for Temperature and Pressure Control and 2. Integrated Microsensors for HF Detection.

List of Websites: