Basic fundamentals of lithium metal battery system
In a general way, a Li-ion battery is constituted by two electrodes (positive and negative) separated by a porous separator soaked by a polar electrolyte. The current trend for the next generation of batteries goes with drastic safety requirements, especially when it concerns LMB technology. In this frame, the battery design intends to either modify lithium surface through passivation (ex and in situ) to reduce or control the lithium dendrite growth or by the replacement of volatile organic electrolyte by non-flammable electrolyte. Another approach is emerging and consists on building an all-solid-state battery by replacing the separator-electrolyte system by a solid electrolyte such as a polymer electrolyte (SPE), gel polymer electrolyte (GPE) or ceramic conducting electrolyte.7–10 The operating principle of this battery is depicted in Figure 1.3. The negative electrode is lithium metal. The positive electrode is composed by a host material ([H]) that allows the insertion and disinsertion of the Li+ ions. In electrochemistry, we distinguished the anode as the electrode at which the oxidation reaction occurs, and the cathode as the electrode at which the reduction reaction occurs.
Figure 1.3 Scheme of the redox processes happening in a battery during charge and discharge. During the discharge step, a lithium oxidation takes place at the negative electrode (anode). The Li+ ions cross the electrolyte to be inserted into the host material, whereas the electrons travel through the external circuit and participate to the reduction at the positive electrode (cathode). During the charge process, the reverse reactions occur: the oxidation takes place at the positive electrode (anode) and the reduction at the negative electrode (cathode). These processes are resumed in Table 1.1.
From practical perspectives, the lithium batteries must provide high energy density, long cycle life, high safety, and being low cost. The cycle life of a battery is strongly affected by numerous factors: the solid electrolyte interphase (SEI) formation, dendrite growth, irreversible redox process, etc., that are driven by temperature, cycling rate.
The choice of electrode active materials must meet several requirements such as high potential with regards to Li potential, good reversibility of the intercalation/de-intercalation processes, high specific capacity together with high ionic/electronic conductivity.
Since the first TiS2 electrode materials,11,12 research was focused on layered oxide transition material,13 such as lithium cobalt oxide (LiCoO2, LCO), lithium nickel cobalt manganese oxide (LiNi1-x-yMnxCoyO2, NMC), lithium nickel cobalt aluminum oxide (LiNi1-x-yCoxAlyO2, NCA) and tends to reach ultimate high potential metal oxide electrode LiNiO214 The commercial LCO-based LIB can achieve a high voltage up to 3.7 V vs. Li+/Li and delivers a quite stable capacity of ∼150 mAh.g-1.15 The major limitations are high cost, low thermal stability, and capacity fade at high current rates or during deep cycling.16
In the same family of lamellar compounds, LiNi0.33Co0.33Mn0.33O2 (NMC 111) has already been commercialized for automobile applications due to its high capacity and outstanding safety characteristics. However, its capacity of 155 mAh.g-1 is quite low for use in next-generation EVs applications.17 Recently, a large part of the efforts have been focused on the variation of the stoichiometry between Ni, Co and Mn, knowing that i) Ni provides a high capacity but poor thermal stability, ii) Mn maintains the structural stability insuring the cycle life and safety, iii)Co offers increased electronic conductivity resulting in an excellent rate capability. For instance, LiNi0.8Co0.1Mn0.1O2 (NMC 811) can achieve capacities up to 230 mAh.g-1 within a potential range of 3 to 4.5 V vs. Li+/Li.18
NCA has high usable discharge capacity (200 mAh.g-1) and long storage calendar life compared to conventional Co-based oxide cathode. NCA cathode has found relatively widespread commercial use, for example, in Panasonic batteries for Tesla EVs.19 However, it was reported that capacity fade may be severe at elevated temperature (40– 70°C) due to solid electrolyte interphase (SEI) growth together with micro-cracks growth at grain boundaries.16,20,21
In addition to the layered structures, two main other families have been investigated: the spinels,22 such as LiMn2O4 and the olivines,23,24 such as LiFePO4 (LFP).
Spinel materials are a very attractive choice as cathode material for lithium-ion rechargeable batteries due to their economic and environmental advantages over LCO, however it suffers from a strong Mn dissolution which reduces the cycle life. Besides, a significant improvement in energy density (vs. Li) and cycle life was reported when adding a second transition element, i.e. LiMxMn2−xO4 with M= Ni, Co, and Cu.25 Unfortunately, high potential (4.7 V vs. Li+/Li) LiNi0.5Mn1.5O4 always suffers from severe capacity deterioration and poor thermal stability because of the instability of carbonates based electrolytes,26 and decomposition of LiPF6.27
LFP is the representative material for the olivine structure with a capacity of 170 mAh.g-1, a low cost compared to LCO and a high intrinsic safety. However, it owns a very low electrical conductivity. To improve this, carbon-coated LFP is an attracting cathode material for application in batteries designed for high power applications. Nonetheless, LFP has a low potential 3.43 V vs. Li+/Li which reduces its energy density (vs. Li). Attention is also focused on alternate olivine material , such as LiMnPO4 and LiCoPO4.28
The vanadium pentoxide (V2O5) is a cathode material with a layered crystalline structure. This typical intercalation compound has large theoretical capacities of 294 mAh.g−1 for two Li+ or 437 mAh.g−1 for three Li+ intercalation/deintercalation,29 which are much larger than those of traditional cathode materials, such as LCO and LFP.
However, V2O5 has disadvantages of low electrical conductivity and slow lithium ion diffusion, resulting in poor cycling stability and rate capability.30
The vanadium oxides, LiV3O8 present interesting performance thanks to their high specific capacity with 3 Li intercalations/deintercalations about 280 mAh.g-1. Nonetheless, this cathode suffers also from an inherently low electronic conductivity (10−6 S.cm-1) and reduced Li+ diffusion coefficients (10−13 cm2.s-1).31
Because of its properties, lithium seems to be the best choice as a material for negative electrode. It owns the lowest reduction potential of all known elements (-3.04 V vs. the standard hydrogen electrode (SHE)) and a theoretical specific capacity of 3860 mAh.g-1 allowing the realization of cells with an extremely high specific energy depending on the cathode material (Li-lithium transition oxide 440 Wh.kg-1, Li-S 650 Wh.kg-1 and Li-air 950 Wh.kg-1).3 Nonetheless, lithium metal based batteries suffer from two main problems: the growth of dendrites and the continuous consumption of lithium due to the continuous healing of SEI along cycling formation (i. e. low Coulombic efficiency).32
To circumvent this issue, graphite has been proposed giving rise in the 90th to the well-known Li-ion technology. It shows a rather flat potential profile at potentials below 0.5
V vs. Li+/Li, offering a specific capacity of 370 mAh.g-1. Although graphite is the most used material in commercial LIB, it is still weak in terms of safety, as solvents can forcibly co-intercalate and exfoliate the graphite with large release of heat.33,34 Several metalloid lithium alloys like Si, Sb, Sn, have been proposed as anode materials with capacities of 660, 994 and 3579 mAh.g-1, respectively.35 However, such materials suffer from volume expansion (until 400 % for the Si).36 To overcome this issues several strategies including size control through nano-structuration or using their oxide form such as SnO, SnSiOx.37
Very promising alternative for graphite especially in power application is the spinel-structured Li4Ti5O12 (LTO). However, LTO operates at a potential of about 1.55 V vs. Li+/Li with a capacity of 170 mAh.g-1. In spite of this inconvenient, LTO potential is far from the lithium voltage plating and no dendritic formation can occur, which ensures good safety.38
The role of the electrolyte is to allow the ionic transfer between both, positive and negative, electrodes within broad electrochemical stability window, between 0 and 5 V vs. Li+/Li, in order to avoid their degradation upon lithium intercalation/de-intercalation.
Electrolytes must have high ionic conductivity, above 1 mS.cm-1 at room temperature and low electronic conductivity. We can distinguish two families of electrolyte, solid ionic conductor based on ceramic or polymer, and liquid electrolytes.
Liquid electrolytes are mainly lithium salts (e.g., lithium hexafluorophosphate (LiPF6), lithium bis(trifluoromethanesulfonyl)imide (LITFSI)), dissolved in aprotic organic solvents (ethylene carbonate (EC), diethyl carbonate (DEC), dimethyl carbonate (DMC) or ethyl methyl carbonate (EMC)) in order to assure the conduction of Li+ ions in between the two electrodes. For instance, the combination LiPF6/EC/DMC is the prototype of electrolyte for LIB.
Among the alternative lithium salts, LiTFSI has the inconvenient to de-passivate the aluminum current collector of the positive electrode inducing its corrosion. This obstacle can be overcome by adding proper additives such as fluoroethylene carbonate (FEC) or vinylene carbonate (VC).39 Recently, lithium bis(fluorosulfonyl)imide (LiFSI) has been of great interest because of their ability to form a relatively homogeneous (SEI) rich in LiF on the negative electrode without corroding of the aluminum current collector.40 However, this salt has a poor thermal stability (decomposition temperature: 343 K) in comparison with the very stable LiTFSI (657 K).41
Regarding the electrolyte solvents, alkyl carbonates used in LIB are not the best choice for LMB because of their poor cathodic stability against lithium metal.34 In contrast, the ether solvents favor the formation of compact and flat Li deposits with relatively higher cycling efficiency whereas carbonate solvents lead to porous and dendritic Li deposits with lower cycling efficiency.40 However, carbonate solvents have better oxidative stability, which is essential for the high-voltage Li batteries.
Ether solvents include 1,2-dimethoxyethane (DME, glyme), tetraethylene glycol dimethyl ether (TEGDME, tetraglyme), 1,3-dioxolane (DOL), 1, 4-dioxane (DX) and tetrahydrofuran (THF).42–44 In Table 1.2, a comparison of electrolyte ionic conductivities is given. Note that the conductivities are in the range of 1-10 mS.cm-2.
However, the high vapor pressure and flammability of all these solvents lead to strong safety issues which imped the mass development for large li-ion battery. One possible alternative is the use of ionic liquid.
Ionic liquids based electrolytes
ILs are defined as molten salts having melting point at temperatures below 373 K. As ILs are constituted only by ions. The cation is commonly organic whereas the anion could be organic or inorganic. These ions are typically large and asymmetric in order to reduce the interaction between them. This explain their low melting points. They own some remarkable properties including retarded flammability, low vapor pressure (~10-10 Pa at room temperature), high temperature of decomposition (up to 573 K), and the ability to dissolve inorganic, organic and some polymer materials. They have been signaled as potential candidates for electrolytes solvents, electrodeposition, double layer supercapacitors, fuel cells, dye-sensitized solar cells, carbon dioxide reduction, water splitting and bioscience.33,41,46–49 ILs are of great interest because of their safety properties. During the process of a battery charge, the heat developed, when it is not effectively dissipated, produces a fast increase of temperature that could trigger exothermic reactions with frequently destructive results of the main components of the battery and the risk of explosion and fire. This mechanism of battery failure is known as thermal runaway.4,50–52 However, the main inconvenient with ILs based electrolytes is their high viscosity which limits their transport properties. The low mobility of Li+ results in low limiting current density in the Li-ion cells.53
Positive electrode materials for the next generation of batteries need electrolytes with a high stability at high potentials (> 4.5 V vs. Li+/Li). Most of the electrolytes are not stable against lithium metal. contrary to the combination of conducting polymer and ILs, named as gel polymer electrolytes, which are at the spot light of lithium metal based battery.54
In battery applications, the most popular choices of anion are by far the bis(trifluoromethanesulfonyl)imide (TFSI- or NTf2-) and the bis(fluorosulfonyl)imide (FSI-).41 The main cations and anions for batteries are given in Figure 1.4.
Figure 1.4 The list of most used cations and anions of ILs based electrolytes for lithium battery technology.55
Regarding the cations, the 1-alkyl-3-methylimidazolium cation (CnC1Im+) is constituted by an alkyl chain that can be tuned (n = 1, 2, 3, etc.) in order to get different electrochemical properties. Tokuda et al.56 have studied the effect of the variation of the alkyl chain length on melting points, density, viscosity, conductivity, and self-diffusion coefficients. For instance, C1C2ImTFSI exhibits a higher density and conductivity than C1C4ImTFSI and C1C6ImTFSI.
However, several reports indicate that imidazolium-based ILs are not stable at low potentials because of the acidity of the proton at the position C2 of the imidazolium ring. 41,57,58 Thus, deprotonation of the imidazolium ring can occur with the formation of N-heterocyclic carbenes (NHCs) and other decomposition products.59 Watanabe et al.41 indicated that cathodic decomposition at 303 K of C1C2Im+ and pyridinium cations with TFSI anion occurs at 0.6 V vs. Li+/Li and 1.6 V vs. Li+/Li, respectively. These potentials are more positive than those for aliphatic quaternary ammonium (AQA, including pyrrolidinium and piperidinuim) and aliphatic quaternary phosphonium (AQP) cations.60,61
Improvement of the stability of the alkyl imidazolium cation by grafting a methyl group (i.e., C1C1CnIm+) has been reported by Schmitz et al..62 Despite of this, the performance of C1C1C4ImTFSI/LiTFSI (0.3 mol.L-1) in a Li/LFP system decrease significantly after 90 cycles at 313 K. However, same authors have previously reported that imidazolium-based ILs could be used in LMB with the addition of additives such as the fluoroethylene carbonate (FEC).57
Regarding, the use of imidazolium-based ILs with additives, Srour et al.63 have studied the combination C1C6ImTFSI/LiTFSI (1 mol.L-1), with 5 % of vinylene carbonate (VC), in graphite/LFP cells, obtaining a capacity of 120 mAh.g−1 at 0.1 C beyond 30 cycles at 333 K.
Finally, the replacement of TFSI by FSI anion could be a solution to work with low potential anodes. In this context, Matsui et al.64 have reported the performance of C1C2ImFSI/LiTFSI (0.45 mol.L-1) in a Li/NMC system giving a capacity of 163 mAh.g-1 during 50 cycles at a regime of 1 C. It has also been reported that when FSI is selected as anion of the neat IL, e.g., C1C2ImFSI/LiTFSI can provide a stable, reversible capacity for a graphitized negative electrode without any additives or solvents at ambient temperature.65
A study of reference of the properties of imidazolium-based ILs that takes into account different anions (TFSI, FSI), the role of the methylation of the cation and the role of the additives (FEC, VC) from a fundamental approach is necessary to remove any controversy. For this reason, we made such a study (described in chapter 2) that will serve as point of departure for analyzing the reactivity of these ILs based electrolytes against lithium metal.
In contrast with the imidazolium, the pyrolidinium family of ILs has been described by several authors as more stable at low potentials and, as a consequence, more suitable for working with lithium metal electrode.57,66–68 The most popular cations of this family are N-methyl-N-propylpyrrolidinium (Pyr13+) and N-butyl-N-methylpyrrolidinium (Pyr14+).
Schmitz et al.57 have demonstrated an enhanced cycling performance in a Li/LFP system in the presence of Pyr14TFSI + 0.3 mol.L-1 LiTFSI, giving a capacity of 130 mAh.g-1 at 0.2 C at 313 K over 150 cycles, in comparison with C1C2ImTFSI and C1C4ImTFSI. Nevertheless, the addition of 5 % (FEC) to these imidazolium ILs results in a similar performance than in the case of Pyr14TFSI.
New types of ILs based electrolytes are continuously being testing in different LIB systems. They include for example, N-methoxy-ethyl-N-methylpyrrolidinium TFSI (Pyr12O1TFSI), N-N-diethyl-N-methyl-(2-methoxyethyl)ammonium TFSI (DEMETFSI)69 or the phosphonium family (P111i4FSI + 3.8 mol.kg-1 LiFSI).70
Recently, Chen et al.71 reported a novel imidazolium-based ionic liquid 1-trimethylsilylmethyl -3-butylimidazole bis-(trifluoromethylsulfonyl)imide or [SiM-BIM]TFSI, which achieves a capacity of 151 mAh.g-1 and a Coulombic efficiency of 99.7% at a 0.1C rate when is cycled in a Li/LFP system.
In Table 1.3 the main electrochemical properties for selected ILs have been collected. Note that in general ILs based on FSI anion exhibit a higher ionic conductivity than their analogous with TFSI anion. The electrochemical windows of ILs based on pyrrolidinium and piperidinium are reported wider than the imidazolium based ILs.
Table of contents :
CHAPTER 1: General aspects and state of the art
1.2. Basic fundamentals of lithium metal battery system
1.3. Electrode materials
1.3.1. Positive electrodes
1.3.2. Negative electrodes
1.5. Ionic liquids based electrolytes
1.6. The solid electrolyte interphase
1.6.1. Historical background
1.6.2. Characterization of the SEI
1.7. SEI in IL-based electrolytes accumulators
1.8. Challenges of the SEI face to lithium dendrite growth
1.9. Characterization of the dendrite nucleation and growth
1.9.1. Morphology of the dendrites and the Chazalviel’s model
1.9.2. Dendrite observations with IL based electrolytes
1.10. Strategies to mitigate dendrite growth
1.10.1. Stabilizing the SEI with additives
1.10.2. Building an artificial SEI layer
1.10.3. Enhancing of the shear modulus of the electrolyte
1.10.4. Effect of surface roughness
1.11. The operando protocols of characterization
1.12. Conclusions and objectives of the thesis
CHAPTER 2: Methodologies and reference studies
2.3. Experimental part
2.3.1. Imidazolium ionic liquids based electrolyte synthesis
2.3.2. Electrolyte preparation
2.3.3. Assembly of the electrochemical systems
2.3.4. Ionic conductivity measurements
2.4. Characterization protocols
2.4.1. Electrochemical impedance spectroscopy
2.4.2. Cyclic voltammetry
2.4.3. X-ray photoelectron spectroscopy
2.4.4. Scanning electron microscopy and Auger spectroscopy
2.5. Reference study of the ILs electrolytes
2.5.1. Ionic conductivity of the ILs
2.5.2. Characterization by XPS of the neat ILs, the associated electrolytes and organic doped ILs
2.5.3. Characterization of the pristine lithium electrode by XPS
CHAPTER 3: Stability of ionic liquids based electrolytes in contact with lithium metal in open-circuit condition
3.2. Study of the stability of IL based electrolytes at OCV
3.2.1. Evolution of the electrolyte and interphase resistances
3.2.2. Chemical environment evolution of the bulk electrolyte and the interface upon aging at OCV conditions
3.2.3. Intermediate conclusions about the degradation mechanisms
3.3. Effect of cation methylation and ILs based electrolyte solvent doping
3.3.1. The effect of the methylation
3.3.2. The effect of the additives
CHAPTER 4: Behavior of the ionic liquid based electrolytes in contact with lithium metal under polarization
4.2. Electrochemical window of the IL-based electrolytes
4.2.1. Literature review
4.2.2. IL-based electrolytes behavior under oxidation
4.2.3. Effect of the additives and the cations
4.2.4. Behavior of the IL-based electrolytes under reduction
4.2.5. Performances of the IL-based electrolytes below 0 V vs. Li+/Li
4.5. Study by XPS of the reduction of the system lithium/copper
4.6. The Li/Li system under the application of an electrical current
4.6.1. Experimental part
4.6.2. The Chazalviel’s model
4.6.3. Experimental verification of the inverse-square law τ ∝ 1/J2
4.6.4. Transport properties of the electrolytes
4.7. Post-mortem XPS analysis of the cells after chronopotentiometry
4.8. SEM/AES analyses of the electrodes after chronopotentiometry
4.8.1. The regime of divergence of the potential
4.8.2. The regime of short-circuits
CHAPTER 5: Operando XPS a novel approach for the study of lithium/electrolyte interphase
5.2. Operando characterization for battery technology
5.2.1. Optical techniques
5.2.2. Electron beam based techniques
5.2.3. X-ray techniques
5.2.4. Scanning Probe Microscopy
5.2.5. Neutron based techniques
5.2.6. X-ray photoelectrons spectroscopy
5.3. OXPS Experimental set-up
5.3.1. XPS spectrometer and sample holder up-grade
5.3.2. Cell design
5.3.3. OXPS sampling and protocol
5.3.4. OXPS sample optimization and connectivity to the potentiostat
5.3.5. X-ray beam and acquisition parameter optimization
5.4. OXPS study of the interface lithium/imidazolium based IL electrolyte, result and discussion
5.4.1. OXPS study of lithium/IL based electrolyte/lithium under chronoamperometry 180
5.4.2. OXPS study of lithium/IL based electrolyte/lithium at the OCV
5.4.3. OXPS study of the interface Lithium/IL electrolyte under polarization
APPENDIX 1: Electrochemical impedance spectroscopy
APPENDIX 2: X-ray photoelectron spectroscopy
APPENDIX 3: Binding energies, relative atomic concentrations and FWHM of the XPS core levels of IL based electrolytes