ALL-SOLID LITHIUM-ION BATTERIES FOR HIGH SAFETY AND STABILITY

Operational safety and cycle stability are important attributes for all rechargeable batteries. To meet these stringent demands specifically for biomedical applications, an all-solid lithium-ion battery (ASLIB) consisting of a polyethylene oxide (PEO)-based polymer electrolyte with a lithium bis(trifluoromethanesulfonyl) imide (LiTFSI) salt, lithium titanate (LTO) anode and lithium iron phosphate (LFP) cathode is proposed. This work implements fabrication methods, composition optimizations and an assembly procedure, all tailored to the unique cell chemistry and ending in the all-solid LTO-PEO/LiTFSI-LFP cells. Subsequently, these ASLIBs are tested close to body temperature at 40 °C. This assures solid-state, but augments bulk electrolyte and interfacial resistance compared to frequent investigations of polymer electrolyte cells at even more elevated temperatures. In spite of these drawbacks, LTO-PEO/LiTFSI-LFP cells are successfully charged/discharged with a C-rate of C/20. In order to understand observed capacity fading, the cycling behavior of these cells is related to several electrochemical phenomena through impedance measurements and investigations of respective halfand symmetric cells. In the end, a unique electrode composition and assembly procedure is proposed to minimize interfacial resistance.


Motivation and Challenges
Conventional lithium-ion batteries are favored for small and large-scale applications, from portable electronic devices (cell phones and laptops) to electric vehicles.
Furthermore, they represent a promising battery chemistry to fulfill the increasing demands of stationary energy storage, and thus could play an important role in the transition from fossil fuels and nuclear energies to renewable alternatives. All in all, no other battery system threatens its dominance today. This is illustrated in (a) Worldwide battery market [2] (b) Ragone-diagram [3]  The triumph of LIBs is due to a combination of several advantages: First, they show a favorable specific energy and power density, as shown in the Ragone-diagram in Figure 1.1(b). The lithium-ion chemistry covers a great range in this diagram with great combinations of specific power and energy density. This enables a huge number of applications across numerous different energy or power requirements. In addition to that, the lithium-ion system exhibits an excellent cycle and calendar life, great efficiency, low maintenance and a small self-discharge rate [4]. Most striking reason for the superiority of LIBs is the alkali metal lithium. Under standard conditions (0 ℃, 1 bar), it is not only the lightest metal but also the lightest solid element on the periodic table, and consequently exhibits a high gravimetric capacity of 3862 Ah/kg. Moreover, it shows a low ionization energy (520.22 kJ/mol) and the highest standard potential in the galvanic series of metals, -3.02 V vs. SHE (Standard Hydrogen Electrode), resulting in the high power density of the battery [5].
In spite of these great characteristics, the classical structure of LIBs still exhibits disadvantages. Conventional LIBs are composed of two electrodes with a liquid electrolyte in between (see Section 2.1.1 for more detail). It is this liquid electrolyte which causes several limitations. Decomposition due to its low electrochemical stability window against the active materials induces the deposition of a solid electrolyte interphase (SEI) on the anode. The initial formation of this SEI, continuous decomposition of the electrolyte and an overall loss of lithium-ions within this layer lead to an ongoing capacity loss of the cell, limiting further cycling stability improvements [6]. Furthermore, the possible leakage and flammability of the liquid electrolyte cause additional safety risks [7]. Environmental incompatibility and human toxicity aggravate the safety concerns even further [1].
To overcome the drawbacks of liquid electrolytes, ASLIBs with polymer electrolytes were proposed. The underlying motivation can be summarized as follows: preventing dendrite propagation successfully [8,9].
(ii) Stability: ASLIBs are not only mechanically more robust, but also show a wider electrochemical stability window compared to liquid electrolytes. This leads to less reactivity towards the electrodes and consequently minimized decomposition of the polymer electrolyte. As a result, ASLIBs are less limited in their cycling stability as is the case for conventional LIBs [8].
(iii) Manufacturing: The all-solid construction results in a less expensive production cost of the battery since all layers can be fabricated with well developed coating techniques. Additionally, replacement of the thick metal containers with vacuum sealed plastic bags saves on volume. Furthermore, the use of solid electrolytes enables thin film batteries, leading to flexible and highly adaptable battery structures, facilitating novel device geometries [8,10].
Many different applications could benefit from these safe and stable ASLIBs. For example, this technology seems to be tailor-made for biomedical applications: New innovations in medical technologies include internet-or communication-enabled devices, such as fully internal cochlear implants. These novel technologies require miniature and safe batteries with particular high cycling stability to avoid additional surgery for only replacing the power source [11]. Furthermore, ASLIBs could be a potential solution for space energy storage systems. The Japanese start-up company ispace just started to test an all-solid battery on the moon [12].
So far, the extreme lunar temperatures, which can vary from -173 ℃ to up to 127 ℃, prohibited a use of conventional LIBs. Although the cold temperatures will lead to a much slower charge/discharge rate, ASLIBs could at least survive the extreme temperatures in space since the polymer electrolyte does not degrade at these high temperatures. Besides this great thermal stability, a good environmental compatibility and the spill-proof and nonvolatile nature of this technology also makes it attractive for unmanned underwater vehicles [13].
Despite the great advantages, two major drawbacks still have to be surmounted: First, the polymer electrolyte's resistance is orders of magnitude higher when compared to liquid electrolytes due to its lower bulk ionic conductivity.
This hinders high power applications. Secondly, a high interfacial resistance is present due to insufficient contact of the solid electrolyte with the electrode's active material particles. In contrast to liquid electrolytes, the solid alternative cannot soak into the electrodes and penetrate their pores, increasing the interfacial resistance. Also, since the interface is between two solids, volumetric changes of the electrodes lead to an increased amount of damage in this area [14].

Structure of this Work
This master's thesis is structured into a theoretical background in the beginning, followed by three parts of experimental investigations and a conclusion in the end.
The theoretical background in Chapter 2 reviews the present state on ASLIBs with a specific focus on the LTO-PEO/LiTFSI-LFP cell chemistry. Basics including the operating principle and cell structure are explained. This is followed by reviewing the single cell components: first, background on the polymer electrolyte as the key component differentiating all-solid from conventional LIBs is given; secondly, an overview of the electrode materials is presented, ending with the motivation of using LTO and LFP in this work. Attention is then devoted to the electrode/electrolyte interface due to its significance on the performance of ASLIBs. In the end of this chapter, electrochemical impedance spectroscopy (EIS) is introduced as the most frequently used measuring method in this work.
The first part of experimental investigations is presented in Chapter 3.
These include ex-situ experiments performed on the polymer electrolyte component with the goal of manufacturing the best-performing PEO/LiTFSI films for ASLIBs: EIS for measuring the electrolyte's conductivity is explained in the beginning, followed by the implementation of a fabrication method, optimization of the salt concentration and testing of BaTiO 3 filler particles as a possible electrolyte additive.
Chapter 4 presents the cycling behavior of the LTO-PEO/LiTFSI-LFP cells at a temperature of 40 ℃. In order to understand capacity fading, impedance investigations are performed and respective half-and symmetric cells are tested.
Based on this, electrochemical phenomena can be formulated to explain the obtained cycling results for the LTO-PEO/LiTFSI-LFP cells.
Chapter 5 proposes two approaches to reduce interfacial resistance. First, the electrode composition is investigated; secondly, a unique assembly procedure is tested in order to improve electrode/electrolyte contact.
Finally, this work ends by summarizing the most important results and with a personal recommendation on how to further proceed in this project is given.

CHAPTER 2
Theoretical Background

Basics about All-Solid Lithium-Ion Batteries
This chapter gives a background on all-solid lithium-ion batteries (ASLIB). The basic structure and operating principle of ASLIBs are explained. Next, some electrochemical fundamentals are described as a basis for understanding the investigations in this thesis. Frequently used battery terminologies in this work are introduced afterwards.

Operating Principle of Lithium-Ion Batteries
The operating principle of a conventional and all-solid LIB is illustrated in  In both cases, the cell includes two electrodes, which consist of an electrochemically reactive (active) material mixed with various additives and coated on a metal foil, which is referred to as the current collector. In a conventional LIB (compare  [15]. It is worthwhile noting that it is convention in the field of batteries to always refer to the negative electrode as the anode and to the positive electrode as the cathode. Usually, the anode is defined to be the electrode at which oxidation occurs while reduction takes place at the cathode. However, in the field of batteries, this matches the naming convention only for the discharging of the cell. The misleading naming has its origin in primary (non-rechargeable) batteries [15,16].
Driven by a potential difference of the electrodes, spatial separated oxidation and reduction of the active material particles takes place at the anode and cathode, respectively. For example, in the case of discharging, the anode active material is getting oxidized, emitting lithium-ions and electrons. Since the electrolyte is conducting only for cations but insulating for electrons, the charge carriers get separated. Lithium-ions get solvated by the liquid electrolyte or complexed by its solid alternative, and migrate to the cathode under the influence of the applied field as well as diffusive transport. A more detailed explanation of the charge transport mechanism in polymer electrolytes can be found in Section 2.2.4.
In order to satisfy charge neutrality, the electrons follow the lithium-ions via an external conducting path, resulting in an electrical current flowing through the energy consumer. At the cathode side, reduction of lithium-ions occurs via reaction with the electrons and the cathode active material. The separator membrane or solid electrolyte prevents an internal short circuit of the cell, since physical contact between the electrodes would induce a spontaneous reaction of the active materials without spatial separation of lithium-ions and electrons. During charging, these processes reverse [4].

Electrochemical Background
The driving force for the electrochemical reactions inside a battery is the cell voltage, defined as the potential difference between th cathode and anode: The voltage difference causes the migration of lithium-ions from one electrode to the other and changes the lithium concentration in the anode and cathode. This results in an increase or decrease in the concentration-dependent electrode potentials. If the battery is at rest, meaning that no current is flowing, the equilibrium electrode potentials Φ eq c/a can be modeled with the Nernst-equation [17]: where the concentration dependency is described with the activity a i , and the standard redox potential Φ eq,• c/a describes the potential difference to a reference electrode, tyically the standard hydrogen electrode or metallic lithium, at standard conditions. The difference of these resting potentials Φ eq c/a is called the open-circuit voltage (OCV) while the influence of lithium concentration on the OCV is shown in respective OCV/state of charge (SOC)-curves. However, the cell voltage also depends on the flowing current, which induces overpotentials η c/a . In order to calculate the electrode potentials under current, and subsequently the cell voltage, these have to be accounted for [18]: The overpotentials can be negative or positive, depending on the direction of the current in the respective electrodes. In the field of batteries, three main overpotentials have to be considered; (1) ohmic (2) charge transfer or activation and (3) diffusion overpotentials [18,5]: The ohmic overpotential describes the voltage drop produced by the ohmic resistances of individual battery components which current has to pass. At the interface between active material particles and the electrolyte, electrons and lithium-ions meet. As a consequence, an electrical double layer (EDL) forms here. The EDL has to be passed for the redox reaction to occur, requiring an activation energy. The accompanying potential difference for a passing current is called charge transfer or activation overpotential. A diffusion overpotential results if the transport of the reactants to the reaction sites is a limiting process. This is the case especially for high currents, when a large concentration gradient forms in the diffusive layer of the EDL.

Terminology
In this work, several battery specific terminology is used. Without claiming completeness, most frequent ones are explained in the following: C-rate: In order to compare the charge/discharge current for batteries of different sizes, a charge/discharge rate is specified relative to the cell's maximum capacity. The C-rate is the charge/discharge current normalized with the cell's maximum capacity: Thus, the unit of the C-rate is 1/h. For example, a 1 C-rate will charge or discharge the battery in 1 hour [19].
Glass transition: For polymers, the glass transition describes the change from a non-ordered, hard, "glassy" state with little mobility to a more "rubbery" state with more flexibility of the polymer chains. The glass transition temperature T g describes the temperature at which this phase transition occurs [20].

Impedance:
The electrical or complex impedance Z describes the relation between the complex representations of a sinusoidal currentÎ(ω, t) and volt-ageÛ (ω, t) [21]: Given its similarity to the ohmic law, the impedance is often referred to as AC resistance. It holds information about the magnitude of the impedance |Z|, which is related to the drop of the voltage amplitude U 0 when the current amplitude I 0 is applied or vice versa. Additionally, the impedance describes the phase shift ϕ between the voltage and current. Due to the complex nature of the impedance, a real and imaginary part can be defined: The respective plot of Im(Z) vs. Re(Z) is called the Nyquist-plot [22].

State of charge (SOC):
The SOC is a measure of how much the battery is charged. It is calculated from the relationship between currently available discharge capacity to the battery's total capacity in the completely charged state in the beginning of the cycle [23].
Transference number: The transference number t i describes the fraction of an electrical current carried by an ion species i: Consequently, the sum of the transference numbers for all of the ions in the electrolyte equals unity [24].

Polymer Electrolytes
In this work, a polymer electrolyte based on poly(ethylene oxide) (PEO) and lithium bis(trifluoromethane)sulfonimide (LiTFSI) is investigated for its use in ASLIBs. This is motivated by good ionic conductivity, transference number, thermal stability and non-toxicity compared to other combinations [25]. For the purposes of completeness it should be noted that alternatives including inorganic solid electrolytes exist [26]. This section focuses on giving a background on the general class of polymer electrolytes and specifically on the PEO/LiTFSI chemistry. After a short historical introduction, general required properties for polymer electrolytes are defined. Next, the most frequently used materials for the two component polymer/salt system are presented, and the specific PEO/LiTFSI chemistry is characterized. Subsequently, the major conduction models and influencing factors are explained. The most investigated modifications to optimize the properties of the polymer electrolyte are described afterwards. Finally, an overview of the most popular fabrication methods to obtain the thin film polymer electrolytes is given.

Historical Background
Interactions of inorganic salts with PEO were first investigated by Lundberg et al. [27] in 1966. They stated that an incorporation of 10-30 % of salt in the polymer host leads to a decrease in crystallinity while retaining compatibility with the PEO. Moreover, they stated a polyelectrolyte-like behavior of these mixtures at low salt concentration: the introduction of the salt changed the polymer from insoluble to soluble in methanol. Further investigation of the ionic nature of these systems was done by Fenton, Wright and coworkers [28] in 1973, resulting in the discovery of the first ion conducting polymer material: They reported moderate ionic conductivity for PEO complexes with dissolved sodium and potassium thiocyanates and even performed measurements of σ at different temperatures. However, the relevance of this publication was not considered until 1978, when Armand et al. [29,30,31] finally pointed out the potential of these materials as polymer electrolytes. It was this work which heightened research interests of polymer electrolytes, including experimental studies of the charge transport, modeling of conduction mechanisms, investigation of chemical structure and evaluation of physical/chemical properties [8].

Required Properties
The polymer electrolyte needs to satisfy several requirements to develop a functional ASLIB. From a practical point of view, the most important ones are the following: (i) Ionic conductivity: It is crucial for the basic operating principle of every battery that the electrolyte is both, a good ionic conductor and an electrical insulator. Commercial liquid electrolytes typically consist of a solution of the salt lithium hexafluorophosphate (LiPF 6 ) in mixtures of organic solvents, such as ethylene carbonate, dimethyl carbonate or propylene carbonate. These show ionic conductivities in the range of 10 −3 to 10 −2 S/cm [32]. The polymer electrolyte should exhibit a conductivity of at least 10 −4 S/cm at ambient temperatures for practical battery applications [8].
(iii) Mechanical properties: On the one hand, polymer electrolytes need to exhibit a certain degree of mechanical robustness: They must stop hazardous dendritic growth of metallic lithium from the anode surface, which would lead to an internal short circuit of the cell when getting in contact with the cathode. In addition to that, the thin electrolyte sheet has to remain structurally stable during its processing and the cell assembly. Especially for large scale manufacturing a high mechanical stability is desirable. On the other hand, polymer electrolytes need to be soft to promote good contact with the electrodes' active material particles. High stiffness of the polymer electrolyte sheet would lead to gaps at the electrode/electrolyte interface and therefore greatly increase the interfacial resistance [8].
(iv) Compatibleness with the electrodes: The solid electrolyte has to be inert against both electrode materials used in the ASLIB. No chemical reactions at the electrode/electrolyte interface should take place. The electrochemical stability window needs to be high enough to avoid electrolyte decomposition [33].

Materials for a Two Component Polymer/Salt Complex
The traditional polymer electrolyte consists only of two components: an ion coordinating and high molecular weight polymer host with a dissolved lithium salt.
In this section, typical representatives for both components are introduced and the specific PEO/LiTFSI chemistry used in this work is characterized.
Several different host materials have been investigated in the past. The most frequently ones are shown in Table 2.1, together with their respective chemical formulas and phase transition temperatures.  LiTFSI) and lithium tris(trifluoromethanesulfonyl)methide (LiC(SO 2 CF 3 ) 3 ), as they offer great thermal stability and exhibit the highest known ionic conductivity for organic anion-based lithium salts [35].
In this work, the polymer electrolyte is based on a PEO/LiTFSI chemistry. A higher thermal stability and an increased ionic conductivity has been observed compared to other combinations [25]. Structural formulas of PEO and LiTFSI are shown in Figure 2.2:    [37]. Glass transition of the amorphous parts of pure PEO with a molecular weight of 10 5 g/mol occur at T g = −67 ℃, the melting temperature of the crystal regions is T m = 65 ℃ and the flash point is stated to be T f = 229 ℃. PEO is susceptible to oxidative degradation when in contact with air. Therefore, contact with oxygen or processing at higher temperatures should be avoided if this degradation is undesirable [36,38]. At room temperature, PEO has a high degree of crystallinity of about 70-84 %. The remaining amorphous elastomeric phase is trapped by these crystals [39].
The hydrophilic salt, LiTFSI, with a molecular weight of 287.09 g/mol has a melting temperature of about 236 ℃ [40]. It consists out of a Li cation and a bistriflimide anion. Kalhoff et al. showed that the use of LiTFSI instead of the commercially common, but toxic and environmentally unfriendly lithium hexafluorophosphate (LiPF 6 ) leads to substantially safer electrolytes [41]. Since the bistriflimide anion is large relative to other investigated lithium salts, anion migration in the polymer electrolyte is minimized, resulting in a higher lithium ion transference number. Anion migration is undesirable in LIBs as it not only reduces the transference number, but might also induce a self-discharge of the cell and lead to electrode surface degradation. LiTFSI also has a plasticizing effect on the PEO. This is not only attributed to the anions size, but also its high dislocation of the electrons. These characteristics decrease the interactions between the polymer chains and therefore hinde crystallization, leading to a more flexible and amorphous polymer/salt complex [25]. A superior conductivity, salt diffusion coefficient and transference number for LiTFSI compared to other salts was also shown by Buriez et al. [42].

Conduction Mechanism
Understanding the charge transport mechanism in polymer electrolytes is of tremendous importance to optimize their conductivity. However, because of their complex structure, the conduction mechanism still remains not fully understood. In general, conduction can be described on a macroscopic and microscopic level.
In this equation, A 0 is a pre-exponential factor related to the number of available charge carriers and therefore depends on the salt concentration. E A is a pseudo-activation energy related to segmental motion of the polymer chains. T 0 is the ideal glass transition temperature and k b the Boltzman's constant. The equation has an empirical origin, but can be derived from quasi thermodynamic models such as the free volume theory and the configurational entropy [33,43].
(ii) Microscopic approach: In the original publications by Wright [28] and Armand [29,30], a microscopic approach for explaining conductivity in polymer electrolytes was introduced. They proposed that it was the "hopping" of lithium cations inside of the polymer helix which induces charge transport. However, this model was not supported by structural studies of the polymer salt complex: Hibma et al. [44] found that the alkali ions are mostly located outside of the polymer helix. Based on conductivity studies on mixed anion complexes of PEO with BH 4 and BF 4 -, Dupon et al. [45] suggested that cation transport in PEO is not limited by motion through the helical channel, concluding that conduction between the polymer chains would be the dominant process. Subsequent investigations showed that the alkali ion in the polymer/salt complex is tightly coordinated to the ether groups in the polymer chains. In fact, based on molecular dynamics simulations, it was suggested that lithium ions are complexed to polymer chains via approximately five ether oxygens [46]. These interactions between the polymer chains and the lithium ions would result in lower mobility of the cations. Taking this into account, polymer dynamics play an important role in the conduction mechanism.
At present it is believed that lithium transport in the polymer host predominantly arises from a combination of two effects, which are illustrated in Figure 2.3:  The first major conduction mechanism is believed to be the hopping of lithium ions between two complexation sites, shown in For both conduction mechanisms, the polymer is required to be in an amorphous state as mobile polymer chains are necessary for the transport of the lithium ions. This explains the observations, that conductivity in the amorphous phase is several orders of magnitude higher than in the crystalline ones [53].

Influencing Factors
Electrical where n Li + is the number of lithium ion moles and n EO the number of ethylene oxide moles.   to decrease. This is mainly caused by the formation of ion pairs, triplets or even larger cluster of ions, which decrease mobility in the polymer host and result in a reduced ionic conductivity [25,56]. Ion association at high salt concentrations was verified in NMR studies by Bruce et al. [57]. The optimal salt concentration is dependent on the specific chemistry of the polymer/salt complex, which can be influenced by the presence of other components and experimental conditions (e.g. temperature).
(ii) Temperature: The behavior of polymer electrolytes is strongly dependent on temperature. As such, the phase and thus the stiffness of the system varies with it. This influence is shown in respective temperature/composition phase diagrams for specific polymer/salt systems (see [58,59] for example).
In addition to these mechanical properties, temperature also influences both the conductivity and transference number of the polymer electrolytes, as shown in Figure 2.5. In general, an increase in temperature leads to a rise of (a) Conductivity [60] (b) Lithium ion transference number [61]   Anions coordinating to OH-end groups would be more stable than the lithium cations coordinating to the oxygen atoms in the polymer chains. Therefore, an increased temperature would primarily enhance the mobility of the lithium ions due to the weak oxygen-cation bond, whereas the OH-anion bond is only marginally affected. This results in an increasing contribution of lithium ions in the overall charge transfer [62].  This behavior can be explained by the influence of the polymer's chain length on the conduction mechanism: Intersegmental cation hopping and vehicular diffusion were presented as the predominant mechanisms. The former takes place independent of the polymer's molecular weight, while the latter is mainly present for only short chain lengths. This is due to a higher diffusion coefficient of the lower molecular weight polymer [50]. In spite of the enhanced ionic conductivity of polymer electrolytes with low molecular weight hosts, they are typically not used for ASLIB applications. This is due to the well-known direct correlation between polymer stiffness and chain length [63].

Plasticized Polymer Electrolytes (PPE)
Considerable enhancement of the ionic conductivity is observed after introducing a polar or ionic liquid into the polymer/salt complex. As a consequence, adding a plasticizer to traditional polymer electrolytes is one of the most common approaches to enable room temperature functionality. The enhanced conductivity is attributed to the increase of amorphous regions in the polymer electrolyte.
Improved segmental motion of the polymer chains promotes cation hopping between different complexation states. Also, the liquid facilitates salt dissociation, increasing the number of available charge carriers. However, the introduction of liquids in the electrolyte composition is accompanied by a deterioration of the mechanical robustness of the electrolyte film. In addition to that, the electrolytes corrosive reactivity towards the metal electrodes typically increases. Consequently, a compromise between the solid and highly plasticized liquid state of the complex has to be found [25].
When large amounts of plasticizers are added, they are typically referred to as gel polymer electrolytes (G-PE). A gel is defined as a dilute cross-linked polymer network swollen with a solvent. It does not flow and is obtained by physically or chemically crosslinking the polymer chains. In order to enhance this process and therefore obtain a greater mechanical stability of the gel, additives which can easily be crosslinked or thermally cured are often introduced [8].
Many different plasticizers have been investigated. They can be classified as organic solvents, low molecular weight polyethers and ionic liquids [64]:  Furthermore, they compared PEGDME with EC and PC as plasticizers to determine their effect on transport and electrochemical properties. In this study, PEGDME led to a superior conductivity and diffusion coefficient [73,74]. Compared to traditional dry polymer electrolytes they not only observed enhanced conductivity, but also a wide electrochemical stability window and good lithium stripping/plating performance.

Rubbery Polymer Electrolytes (RPE)
In contrast to the traditional 'salt-in-polymer' complex, this group of polymer electrolytes is often referred to as a 'polymer-in-salt' system: a small amount of high molecular weight polymer is added to a large amount of salt. Typically, the glass transition temperature of these mixtures is low to maintain a rubbery state of the electrolyte at room temperature. Although providing improved ionic conductivity and good electrochemical stability when compared to traditional polymer electrolytes, the disadvantage of this group is poor mechanical properties. As a result, applications are limited because the salt tends to crystallize at lower temperatures, leading to brittle electrolyte films [25,64].
The R-PE group was first studied by Angell et al. [78], who mixed different lithium salts with small quantities of PEO and PPO. They stated a conductivity of up to 10 −2 S/cm at ambient temperature. However, the mechanical properties of this electrolyte prohibited its use in LIBs with metallic lithium anodes. rubbery electrolytes with PAN and its copolymers as the polymer. This is because interactions between the nitrile groups and lithium ions have been attributed to stabilize ionic clusters in the polymer-in-salt system. In contrast to traditional polymer electrolytes, it is well accepted that ion aggregation is desirable for the ion transport in polymer-in-salt systems [64]. To name only one of many studies on R-PE with PAN polymers, Zalewska et al. [80] studied a LiAlCl 4 system. At room temperature they measured a conductivity of 10 −4 S/cm.

Composite Polymer Electrolytes (CPE)
Composite polymer electrolytes have developed into one of the most active research areas in the polymer electrolyte field. In this group, small quantities of filler particles are dispersed in the polymer electrolyte. Contrary to expectations of classical theories, where adding small amounts of an insulator to an electrical conductor is believed to decrease conductivity, an increasing effect has been observed in the case of micro-or nanosized particles. In addition to that, an enhanced mechanical stability and reduced electrode/electrolyte interfacial resistance has been stated for C-PEs [25]. Different approaches to explain these effects were proposed: It has been suggested, that various Lewis acid-base interactions at the surface of the particles would induce pathways for lithium ion transport and result in the enhanced transport properties. Furthermore, the small particles might act as crosslinking centers for PEO segments and salt anions, hindering reorganization of the polymer chains [64]. Regardless, their introduction in the polymer/salt system would lead  [64]. The effect of these on polymer electrolytes has first been studied by Weston et al. [81] in 1982. In their pioneering work, they added α-alumina to a PEO/LiClO 4 polymer electrolyte and found an enhanced mechanical stability of the material, improving the ability of this polymer electrolyte for use in LIBs. However, they stated only negligible changes in the ionic conductivity and lithium transference number. In contrast to that, many following works showed the positive effect of the fillers on the charge transport properties of these systems. For example, Croce et al. [82,83] studied the influence of 13 nm sized TiO 2 and 5.8 nm sized Al 2 O 3 particles on PEO/LiClO 4 systems by comparing the electrical and mechanical properties of these composite systems with the respective ceramic particle-free electrolytes.
They not only observed excellent mechanical stability in a temperature range of 30 to 80 ℃ and great electrochemical stability, but also an increase of ionic conductivity. In fact, for the same temperature range, systems without ceramic particles showed a conductivity of around 10 −  (iii) Other fillers: Different fillers have been studied for composite polymer electrolytes. For example, the use of cellulose was suggested as a reinforcing agent to enhance mechanical strength of the electrolyte. Samir et al. [92] investigated cellulosic whiskers in a PEO/LiTFSI system and indeed found a reinforcing effect while ionic conductivity was retained. Furthermore, carbon based materials as fillers were investigated to increase conductivity and interfacial stability. In fact, Appetecchi et al. [93] were the first to propose the use of carbon powders as filler materials in polymer electrolytes. Ibrahim et al. [94,95] studied carbon nanotubes in PEO/LiPF6 system plasticized with EC and stated a rise in ionic conductivity.

Fabrication
The literature suggests different fabrication methods for the polymer electrolyte.
Although exact conditions and procedures are varying, the approaches can generally be separated to a solvent-based and solvent-free method. In both approaches, all solid components used in the fabrication process of the electrolyte are dried for a sufficient amount of time at an elevated temperature beforehand. Also, the steps are primarily carried out in a dry argon or nitrogen environment to avoid moisture contamination.
If processing in open air is required, the materials are placed in a sealed envelope before bringing them out of the dry box.

Solvent-based method
The solvent-based approach is the traditional method for fabricating the polymer electrolyte. In this process, the LiTFSI and PEO are dissolved in a suitable solvent under magnetic stirring. Most common solvents are acetone, acetonitrile, tetrahydrofuran (THF) or methanol. Typically, the dissolution of the two components is done separately and the solutions are mixed together. To ensure salt complexation in the polymer host, the stirring process is done for up to 24 hours. During this step, possible additives are inserted [25,33,96,97,98]. The downstream processing of the resulting clear and viscous mixture differs in the literature. In most cases it is poured into a high surface area cavity, usually a Petri dish or PTFE-mold. Here, the film formation takes place by slow evaporation of the solvent. This step is often accelerated by pulling vacuum or raising the temperature to up to 120 ℃. To ensure complete solvent evaporation, the drying is done for up to one week [25,33,96,97].
In other procedures, the mixture is heated to evaporate the solvent and then hot pressed at 90 ℃, leaving behind the polymer electrolyte film [98].

Solvent-free method
This method presents an alternative to fabricate the polymer electrolyte without the use of any solvent. In its original proposal by Gray et al. [99], dry powders of the polymer, complexing salt and possible additives are physically mixed by ball milling. The homogeneous powder is then heated slightly over the melting temperature of the polymer under continuous mixing to ensure salt complexation.
The soft and sticky mixture is then placed between Mylar sheets and hot pressed at elevated temperatures and pressures: First, 5 mm pellets are obtained by hot pressing at a temperature between 70 ℃ and 110 ℃. After the cooling of the pellets, the second step is performed. Here, a pressure of 19 MPa is applied at room temperature, followed by hot pressing at 7.6 MPa and 70 ℃ to 110 ℃.
This hot pressing method was adopted by many research groups with slight modifications in the procedure and conditions of the individual fabrication steps: After mixing the powders in a mortar instead of a ball mill for two minutes,

Electrode Materials
This chapter reviews the most frequently used electrode materials in LIBs. After the basic structure of the composite electrodes is introduced in the beginning, an overview of the conventional anode and cathode active materials is given. Here, a specific focus is attributed to lithium titanate (LTO) and lithium iron phosphate (LFP), which represent the electrode materials in this work.

Basic Structure
A schematic structure of the composite electrodes in LIBs is shown in Figure 2.8: Figure 2.8: Schematic structure of the composite electrodes in LIBs adapted from [15] In general, the electrode's coating is based on a mixture of active material, binder and conductive fillers. With mass fractions around 90 wt%, the active material forms the major part in this composition. They determine the capacity and potential of anode and cathode. The binder is needed for a good cohesion between the electrode particles and sufficient adhesion to the current collector (metal foil).
Classical representatives are polymers such as polyvinylidene fluoride (PVDF) or styrene butadiene rubber (SBR). Since electrons have to be transported through the electrode structures for the redox reactions to occur, conductive fillers are inserted to enhance the electrical properties of anode and cathode. Carbon black is the most popular representative of this group. The amount of these inactive additives in the electrodes depends on the energy-and power requirements of the battery.
Generally, a high power design utilizes a high proportion of binder and carbon black, whereas for high energy applications the active material proportion is maximized.
The electrode coating is placed on thin metal foils, the current collectors. Typically, aluminum is used for the cathode side due to good electrical conductivity, low cost and weight, as well as sufficient electrochemical stability in the potential range of the cathode. For the lower potential anode, aluminum would form lithium alloys, which is why copper is generally used on this side as an alternative [15].

Anode Active Materials
Pure lithium metal was originally investigated as an anode material for LIBs. It exhibits the highest specific capacity (3861 mAh/g) and best standard potential of all proposed alternatives so far. However, metallic lithium proved to be hazardous for rechargeable batteries as it leads to dendrite formation on the anode's surface during charging, increasing the risk of internal short circuiting, subsequent thermal run-away and explosion. Major efforts have been undertaken to enable the use of lithium metal in rechargeable batteries. For example, solid electrolytes might provide a possible solution to prevent the safety risks since their great mechanical robustness can suppress the formation of lithium dendrites. In spite of that, the use of lithium metal as the anode's active material is still restricted to primary (non-rechargeable) LIBs at the present state. In the following section, an overview of the most popular alternatives is given first. Special focus is then attributed to lithium titanate, which represents the anode active material in this work [104].

Overview
An overview of the most studied anode active materials including their characteristic capacities and potentials is shown in Figure 2.9: Figure 2.9: Potential as a function of gravimetric capacity for popular anode active materials [105] In general, they can be categorized based on their reaction mechanism into intercalation, alloy or conversion materials.
(i) Intercalation materials: Intercalation describes the reversible insertion of molecules or ions into interstitial vacancies of the parent material. In this process, no covalent bonds are formed or broken [106]. The most popular intercalation hosts are carbon based. Commercially, graphite is the most frequently deployed anode material due to a great combination of properties: Graphite provides a high standard potential that is close to lithium metal (0.125 V vs Li/Li + ). It shows good mechanical, thermal, chemical and electrochemical stability, resulting in better safety when compared to lithium metal.
A favorable reversibility of the intercalation reactions results in moderate cycling stability. Also, it shows high lithium diffusivity and electrical conductivity, as well as acceptable volume changes during lithiation/delithiation.
Finally, the easy availability implies low-cost and consequently makes it very attractive from a commercial standpoint [104,105]. The most striking disadvantage of graphite is its limited capacity of 372 mAh/g. Other carbon based materials are currently being investigated to improve the specific energy.
These include "hard" carbons with a random alignment of small crystallites such as carbon nanotubes, carbon fibers, porous carbon, or graphene [107].
Another popular group of intercalation materials are titanium based oxides, which show favorable properties for high safety applications. A typical representative of this group is lithium titanate, which is extensively studied in Section 2.3.2.2 [106]. representative of the alloy anode materials is silicon, which provides a high theoretical specific capacity of 4200 mAh/g, but a volume expansion of over 400 %. Other alloy materials such as germanium or tin oxide exhibit a similar behavior. Different strategies have been proposed to circumvent the described difficulties. One of the most popular approaches is alloying composites with active or inactive materials. In this case, the inactive/active materials serve as a buffer that provides free volume for the expansion of the active material particles. Highly investigated representatives of this group are carbon-silicon composites. Second, it is proposed to reduce the active material particle size from micro-to nanoscale. This would induce more homogeneous lithiation/delithiation and minimize differential expansion, thus reduce particle pulverization. Other strategies attempt to stabilize the SEI by encapsulating the alloying material particles using electrolyte additives or inserting binders into the electrodes, which increase the mechanical stability [104,105]. : Classical representatives include iron oxides, cobalt oxides and metal phosphides, sulphides and nitrides. The advantages of such materials can be summarized with high theoretical capacities, environmental abundance, and low material costs. However, similary to the alloy anode materials, they suffer from large volumetric expansions, resulting in poor capactiy retention. The additional large potential hysteresis keeps this group from penetrating the commercial market at the present state [105].

Lithium Titanate
Lithium titanate (Li 4 Ti 5 O 12 ), also referred to as lithium titanium oxide and abbreviated as LTO, has emerged as a promising anode active material. This spinel structured material belongs to the group of intercalation anode materials. Based on the underlying redox couple Ti 3+ /Ti 4+ , LTO exhibits a working potential of 1.55 V vs. Li/Li + . When charged/discharged in the potential range of 1 to 3 V, the electrochemical reaction can be expressed as which results in a theoretical specific capacity of 175 mAh/g [108].
The advantages of this material can be summarized as the following: The non-toxicity and high operating potential of LTO results in favorable safety characteristics. Also, the anode's operating potential is above the SEI formation potential caused by electrolyte reduction of the anode's surface, minimizing continued electrolyte decomposition. Yet it should be noted that a complete absence of an SEI is not experimentally verified at the present state. This is due to the possibility that oxidation products from the cathode can migrate to and deposit on the anode surface. Nevertheless, the irreversible capacity loss due to electrolyte decomposition in the first cycles is tremendously reduced. Furthermore, LTO exhibits a negligible volume change upon lithiation/delithiation, resulting in an excellent cycling stability of this material. The constant particle volume minimizes mechanical strain (LTO is also called "zero-strain material") of the active material and the SEI, leading to less particle fraction and SEI self-pulverization. In addition to these favorable electrochemical properties, the elements of LTO show a relative environmental abundance, resulting in low material costs and enabling commercialization [106,109].
The favorable safety and stability of this material due to the high operating potential is contrasted by low power-and energy densities. Further difficulties that have to be surmounted are a low electrical conductivity (∼ 10 −8 to 10 −13 S/cm) and poor lithium-ion diffusivity (∼ 10 −8 to 10 −13 cm 2 /s), limiting the charge/discharge rate performance. Different strategies have been proposed to overcome these drawbacks, including anion or cation doping, surface modification with conductive coatings and nanostructuring [109,110].
One attempt to illustrate the unique cubic spinel structure of LTO in the form Li 4 Ti 5 O 12 is shown in Figure 2.10(a). The oxygen atoms arrange approximately as a cubic closed packing (ccp), located at the 32e Wyckoff-position.
Lithium-ions sit in the tetrahedral 8a sites of the lattice, whereas the first half of the octahedral 16d sites are randomly occupied by 1/6 Li + and 5/6 Ti 4 + . The second half of the octahedral 16c sites remain empty though play a fundamental role during lithium intercalation [110].

Cathode Active Materials
For conventional LIBs, the cathode represents the most expensive and highest weight component, justifying extensive research on next-generation cathode materials in the past years. A broad overview of these is given in the first part of this section, followed by a more detailed characterization of lithium iron phosphate as the cathode material utilized in this work.

Overview
An overview of the currently most investigated cathode materials is shown in    blocking of lithium-ion diffusion paths by the nickel atoms, limiting rate capability [111,112]. Extensive research has led to considerable improvements of their cycling performance. Currently, the main disadvantages of this group include limited specific capacity as well as electrolyte decomposition in the first cycles due to the high operating potential of this material [111,113]. Nevertheless, a poor energy density requires more research for this derivative of the olivine group [104,113].
(v) Conversion compounds: Conversion or alloying cathode materials are proposed as an high energy density alternative to the intercalation compounds.
However, the arising products of the underlying conversion reactions can change the structure and chemistry of the electrode with the reaction mechanisms often remaining unknown [114]. Classical representatives of this group include fluorine and chlorine compounds, sulfur and lithium sulfide, selenium and tellurium, as well as iodine [104].

Lithium Iron Phosphate
Lithium iron phosphate, with the chemical formula LiFePO 4 and abbreviated as LFP, attracted attention as a possible next-generation cathode material for power tools, electric vehicles and stationary energy storage. It belongs to the group of olivine structured cathode materials and provides an OCV of 3.4 V vs. Li/Li + .
Based on the redox couple Fe 2+ /Fe 3+ , the electrochemical insertion/extraction reaction is expressed as which results in a specific capacity of 170 mAh/g.
Research in LFP is motivated by several advantages: It provides a characteristic flat discharge curve and great cycling performance due to good reversibility of the insertion/extraction reaction. The olivine structure benefits from higher safety compared to LCO and thermal stability up to 400 ℃. Relative abundance of its elements results in low material costs which make it attractive for commercialization. Furthermore, the non-toxicity and environmental compatibility of LFP results in a 'green' alternative to other cathode chemistries [115].
Major limitations include the low electrical conductivity (∼ 10 −9 to 10 −11 S/cm) and lithium-ion diffusivity coefficient (∼ 10 −11 to 10 −13 cm 2 /s), resulting in poor rate performance. Different strategies were proposed to surmount these drawbacks, such as conductive material coatings, decreasing the LFP particle size or doping LFP with cations to improve intrinsic conductivity [116]. Another disadvantage is the low tap density of LFP: nanosized LiFePO 4 shows a tap density of 0.6−1 g/cm 3 , which is much less compared to 2.6 g/cm 3 for commercial LCO. In spite of the reasonable gravimetric capacity, this reduces the volumetric energy density, increasing cell size. In order to overcome this disadvantage, new sythesis methods are proposed to control morphology and improve homogeneity of the LFP particles [115].
The complex olivine structure of LFP is illustrated in

Electrode/Electrolyte Interface
It is generally accepted that the performance of lithium-ion batteries is not only dependent on the electrode and electrolyte, but also on the properties of the electrode/electrolyte interface. Besides the limited ionic conductivity of the polymer electrolyte, high interfacial resistance is known to be one of the two major drawbacks in ASLIBs. This is due to various reasons: Most important, the active surface area at the interface is much less compared to conventional cells, where the liquid electrolyte soaks into the pores of the electrodes and provides enhanced active material contact. This is not the case for a solid electrolyte, which greatly increases the interfacial resistance. Furthermore, since the interface is between two solids, electrode volumetric changes during lithiation and delithiation cycles lead to increased damage to this area [14]. In spite of the fact that polymer electrolytes exhibit a much higher electrochemical stability window compared to liquid electrolytes, a surface layer formation was also observed for ASLIBs. This surface layer can substantially increase the interfacial resistance. In the following section, the understanding of the surface layer formation in polymer electrolyte based ASLIBs is described. Based on this, different attempts for decreasing resistance at the electrode/electrolyte interface are presented.

Surface Layer Formation
In liquid electrolyte cells, a so-called solid electrolyte interphase (SEI) forms during the first charge/discharge cycles on the surface of the negative electrode.
This is due to the mismatch of the anode's electrochemical potential and the electrochemical stability window of the liquid electrolyte. As a consequence, the electrolyte is reduced when in contact with the active material of the anode, inducing its decomposition. Finally, this results in the deposition of the reduction products on the electrode's surface and formation of the SEI, which schematic structure is shown in Figure 2.13: Figure 2.13: SEI formation on the anode's surface in liquid electrolyte cells [117] The evolution of this thin and solid layer can be observed by an irreversible capacity loss in the first cycles of the battery. It is well known today that a good and controlled formation of the SEI is of major importance for the cell's performance.
The film is insulating to electrons, but permeable to lithium-ions. Therefore, the layer shows properties of a solid electrolyte, hence the name. Furthermore, these properties allow for spatial separation of the electrons and lithium ions in the anode, which is crucial for the operating principle of the battery. In addition to that, the SEI prohibits physical contact between the electrode and electrolyte, preventing ongoing electrolyte decomposition and protecting the electrode from exfoliation. Due to the importance of this layer in the battery's performance, a tremendous amount of research on the SEI has been carried out. Overall, this led to a decent understanding of the layer's composition and formation mechanism [118,6].
Compared to liquid electrolyte cells, the interfacial chemistry in ASLIBs has been explored much less. This might be due to a theorized higher chemical and electrochemical stability of the polymer electrolyte. In addition, the film is ideally well attached to the electrode, making investigations of the interface between the two solid layers more difficult. Nevertheless, the presence of an interfacial resistance has been observed after cell assembly based on electrochemical impedance spectroscopy (EIS). Several studies attributed this to the formation of a passivation film on the surface of the electrodes: In 1984, Fautex [119,120] described the formation of an ionically conducting passivation film between a lithium electrode and a PEO/LiCF 3 SO 3 electrolyte, stating that this film would act as a solid electrolyte interphase similar to liquid electrolyte systems. They supported their results with EIS studies, in which they attributed the higher frequency semicircle in the impedance data to the electrolyte resistance R e and the lower frequency circle to the interface resistance R i , as shown in Figure 2.14 (a).
(a) assignment of R e and R i to semicircles (b) time dependency of R i Figure 2.14: Impedance diagrams for an Li-PEO/LiCF 3 SO 3 -Li system with electrolyte resistance (R e ) and interface resistance (R i ) [119] As expected, both arcs decreased with temperature. However, the time dependency differed, as shown in Figure 2.14 (b): a slight reduction of the high frequency arc was observed after cell assembly, which was attributed to a small creepage of the polymer electrolyte film leading to a decrease of its bulk resistance. In contrast, the low frequency semicircle greatly increased with time. This evolution of interfacial resistance would indicate the growth of an ionically conducting surface film at open-circuit conditions. Subsequent investigations [121,122] similarly described the development of a resistive layer on the anode.
Attemps at characterizing the film formation mechanism and composition followed. However, it is worth noting that both are still not well understood yet. Besides, the surface layer is highly influenced by the processing conditions and specific chemistry of the investigated system. Most of the research has concentrated on the interface between a lithium anode and PEO-based electrolyte, while literature on the formation of a passivation layer on other electrode materials is rare [86]. Thus, the following explanations have to be taken within this in mind.
Peled et al. [123] suggested that the layer builds due to reaction of lithium with water, salt anions and other impurities. They proposed a schematic of the Li/PE interface, which is shown in Figure 2.15(a).   For the graphite/PE interface, a high formation of LiOH was observed, which was assigned to reactions of water with lithium ions. Besides that, they found LiTFSI decomposition on the graphite surface. The newly developed species at the Li/PE interface were mainly based on lithium fluorides and lithium alkoxides. In contrast, the SEI of respective liquid electrolyte system was based on decomposition products of the electrolyte such as carbonate species or PEO-type polymer.

Reduction of the Interfacial Resistance
Different attempts for decreasing the interfacial resistance have been proposed.
One of the most popular is the introduction of filler particles into the polymer electrolyte. The resulting composite polymer electrolyte was introduced in  This study illustrates that an introduction of only 1.5 wt% of 0.1 µm sized BaTiO 3 particles can reduce the interfacial resistance to around 15 % compared to the filler-free system. As stated previously, surface film formation is believed to be induced by reactions of lithium with water and other impurities. It is assumed that filler particles might absorb the water, therefore suppressing the reactions between lithium and the water. This would result in a stabilizing effect of the interface [126].
Many other studies show the positive effect of various fillers on the interfacial resistance for various electrode/polymer electrolyte chemistries [83,90,127,128].
A further attempt to inhibit surface layer formation was proposed by Masona et al. They placed a self-assembling monolayer (SAM) of copolymer molecules with the formula H( -CH 2 ) 32 ( -CH 2 CH 2 O -) 10 H onto the surface of the polymer electrolyte films. The PEO like head of the molecules would absorb into the polymer system, leaving the nonpolar tail for self-assembling. This SAM surface structure, which is illustrated in Figure 2.18(a), was confirmed with FTIR spectroscopy. Investigating the interfacial resistance development with EIS, they found a much slower formation of the passivation layer compared to untreated polymer electrolytes as shown in Figure 2.18(b). It is suggested that the hydrocarbon tails "hide" the lithium anode from water or other impurities in the electrolyte and therefore hinder surface layer formation [129].
(a) SAM structure (b) effect on interfacial resistance Figure 2.18: Self-assembled monolayer (SAM) on the polymer electrolyte to inhibit passivation at the Li/electrolyte interface [129] Besides inhibiting a surface layer formation, increasing the contact between the electrode's active material particles and the polymer electrolyte would reduce the interfacial resistance. In order to do this, temperature and pressure can be tuned.
However, this is limited by the mechanical properties of the polymer electrolyte and the application environment. To increase the specific surface area between the solid electrolyte and the electrodes, thin films of the respective electrodes are used.
Nevertheless, charge transport to deep active material particles remains one of the major challenges in the field of ASLIBs [130].

Electrochemical Impedance Spectroscopy
Electrochemical Impedance Spectroscopy, also called AC Impedance Spectroscopy and abbreviated EIS, is an analytical characterization tool for electrochemical systems. It represents a non-destructive method to analyze electrochemical processes in batteries and is the most popular instrument for measuring ionic conductivity in polymer electrolytes [6].  which is typically presented in the form of a Nyquist-plot [132,131].
This Nyquist-plot represents the basis for the subsequent data analysis.
In general, an equivalent-circuit model (ECM) can be built to fit the obtained impedance spectra. In doing so, the investigated system is modeled with a combination of electrical elements, such as resistors, capacitors and inductors.
These are then assigned to electrochemical processes in the system and used to interpret the experimental data [133].   .1) [135,136].

CHAPTER 3
Manufacturing the Polymer Electrolyte

Introduction
The polymer electrolyte is the major component differentiating ASLIBs from

Electrochemical Impedance Spectroscopy
Electrochemical Impedance Spectroscopy (EIS) is the standard for measuring ionic conductivity in solid electrolytes. To obtain reproducible results, this section introduces a consistent way of measuring ionic conductivity by testing different setups and excitation amplitudes at several measurement temperatures.

Experimental
The bulk electrolyte resistance R b is represented by the lower frequency minimum in the Nyquist plot of the cell's impedance curve and is therefore determined by EIS. For the two shown setups, the thickness of the polymer electrolyte is evaluated by measuring the thickness of the single stainless steel parts (electrodes and/or covering), and subtracting them from the setup's total thickness, which is taken before each EIS measurement. The thickness of the stainless steel parts is measured only once at room temperature before each setup is put together, and are assumed to be constant for all subsequent measurements.    (d)). This is due to a bad signal-to-noise ratio: a low excitation amplitude only induces a small current, which can easily be impacted by external noise.

Results
(2) Compared to the measurements with small voltage amplitudes, an excitation of 20 mV or 30 mV leads to a drop of the impedance at lower frequencies. This results in the observed variation of impedances from the characteristic 45 • straight line at low frequencies for these measurements. The effect is caused by the nonlinearity of the system: EIS assumes linearity by only taking into account the first harmonic frequency response. However, especially at lower frequencies, the system reacts with higher frequency harmonic signals, which fails to satisfy the assumption of pseudo-linearity.
(3) A frequency shift in the measurements occurs when changing the temperature.
This becomes visible with the change of the characteristic lower frequency minimum (indicated with ) that is taken as R b the ionic conductivity calculation. As a consequence, the lower frequency semicircle attributed to the resistance and capacitance of the bulk electrolyte is not visible anymore.
This cannot be adjusted by changing the frequency range of the measurements due to the upper limit of the potentiostat of 1 MHz. However, the frequency shift does not impact ionic conductivity measurements, since only the lower frequency minimum of the semicircle is needed for the calculation, and it can still be approximated form these curves.
Ionic conductivity is evaluated for all measurements and summarized in Figure 3.3.
For both setups and at all temperatures, the calculated ionic conductivity is negligibly affected by the excitation amplitude. At 20 and 40 ℃, differences in the values for the two different setups are insignificant, considering also that two different polymer electrolyte coins were investigated. In contrast to that, a difference in the ionic conductivity with these two setups is observed at 60 ℃. This is due to

Conclusion
To establish a consistent way for measuring the ionic conductivity of the polymer electrolyte, two different setups were tested with several excitation amplitudes and temperatures. A bad signal-to-noise ratio at low excitation amplitudes and nonlinearity of the system at high amplitudes leads to the choice of 10 mV for the subsequent EIS measurements. No significant differences in the results for the two different setups are observed. In both cases, reproducible measurements can only be performed below the melting of the polymer/salt complex, which occurs around 60 ℃. Due to simplicity, setup A is chosen for following ionic conductivity measurements.   The objective of this section is to implement a fabrication method that leads to reproducible results and optimal mechanical and electrical properties of the obtained polymer electrolyte films. As stated in Section 2.2.7, two general approaches have been suggested in the literature, and both are investigated in the following: a classical solvent-based procedure and a solvent-free method. For the latter, an annealing step at a temperature above the melting point of the polymer is often performed to help salt complexation to the PEO chains. However, this step is time consuming as the polymer electrolyte is typically annealed for no less than 24 hours. Therefore, the actual effect of the annealing step on the electrical properties of the polymer electrolyte is determined by also testing the solvent-free method without the annealing step.

Experimental
The classical solvent-based approach (method A), the completely solvent-free procedure with an annealing step (method B) and without one (method C) are summarized in Figure 3.5.    The ionic conductivity increases with temperature due to the rise in mobility (2) The scattering of the measured ionic conductivity values increases with temperature, which can be observed in the larger error bars at higher temperatures.

Results
As already stated in Section 3.2, this effect is due to the measuring setup: A higher temperature causes a softer polymer/salt system. Since the electrolyte film is sandwiched between two stainless steel electrodes under slight pressure (compare Figure 3.1(a)), this leads to some electrolyte squeezing out of the sides. Consequently, this induces a deterioration of the electrode/electrolyte contact and a non-uniform thickness. As a result, the variability of the data increases.
(3) The dry method does not lead to a decrease in ionic conductivity compared to the solvent-based procedure. In fact, the solvent-free method produces a slightly higher ionic conductivity in the polymer electrolyte based on this data.
However, this statement has to be taken with some uncertainty due to the high scattering of the conductivity values at high temperatures. Nevertheless, taking into account that the solvent-free method is less complex and more time efficient, it can be concluded that the dry method shows advantages over the solvent-based one.
(4) The polymer electrolytes that were fabricated with Method B and C only exhibit a negligible difference in the ionic conductivity. Consequently, it can be stated that the annealing step in the dry fabrication method has no substantial effect on the ionic conductivity of the polymer electrolyte. This is an important result, as an additional annealing step would increase processing time and complexity.

Conclusion
Three different fabrication methods were investigated: a solvent-based approach as well as a solvent-free method with and without an annealing step. The ionic conductivity for polymer electrolytes from each fabrication procedure was measured in the temperature range from 20 to 50 ℃. The temperature dependence on ionic conductivity matches the expectations from the VTM-model. Furthermore, the less complex and more time efficient solvent-free method does not lead to a decreased ionic conductivity compared to the solvent-based procedure. In addition to that, results show no major differences when the polymer electrolyte is annealed. As a consequence, the dry method without the annealing step is chosen for the subsequent processing of the polymer electrolyte.

Optimization of the Salt Concentration
This section seeks to find the best salt concentration for the two component PEO/LiTFSI mixture. This is not only important to maximize the ionic conductivity of the system, but also to optimize the mechanical properties of the thin electrolyte films: they have to be mechanical robust enough for a good processability, but soft enough for a good contact at the electrode/electrolyte interface.

Experimental
Polymer electrolyte coins with a PEO molecular weight of 10  g/mol, respectively, the conversion from molecular ratio r to weight ratio x can be calculated.  Again, three polymer electrolytes are investigated per salt concentration, and the obtained maximum, minimum and calculated average is evaluated.  The observed behavior is expected due to the fact that a higher salt concentration deceases the melting temperature of the polymer/salt system. In order to find the best mechanical properties for the polymer electrolytes, a compromise has to be found: From the processing perspective, a high mechanical robustness would be favorable to easily peal off the polymer electrolyte film from the Mylar sheet, and furthermore to punch out the coins and assemble the cell without damaging the polymer electrolyte. In contrast, a soft sheet will increase contact at the electrode/electrolyte interface, resulting in a decreased interfacial resistance.

Conclusion
Polymer electrolytes were fabricated with different molecular ratios of salt-to-

Testing BaTiO 3 Filler Particles
The striking disadvantages of ASLIBs with polymer electrolytes are the poor ionic conductivity and high interfacial resistance especially at moderate temperatures.
Considerable research efforts were undertaken to bring down the operating range of the polymer electrolyte to room temperature. To improve the electrolyte's ionic conductivity and suppress a surface layer formation, the introduction of nano-or microsized filler particles into the polymer/salt system has proven to be successful. Furthermore, an increase in the amorphous regions of the polymer electrolyte due to the introduction of the BaTiO 3 particles has been suggested. Acting as nucleation centers, they would lead to a higher nucleation rate, resulting in an accelerated solidification [138,139].
In this section, the effect of barium titanate (BaTiO 3 ) nanoparticles on the ionic conductivity of a PEO/LiTFSI system is investigated. CPEs with different BaTiO 3 weight factions are fabricated with both, the solvent-based and solvent-free procedure. is used for these measurements, and the bulk electrolyte resistance is taken from the lower frequency minimum in the impedance curves and used to calculate the ionic conductivity of the several CPE coins.  The diagram gives rise to the following results:

Results
(1) For the solvent-free method, a decrease in the ionic conductivity is observed when the BaTiO 3 powder is introduced. Supported by the visual results from do not match the enhancement of ionic conductivity stated in the literature [86,138,139,140,141]: Most likely, the carried out solvent-based fabrication procedure could not fully prevent agglomeration. Zhang et al. [139] investigated the influence of the BaTiO 3 particles size on the ionic conductivity.
Comparing sizes between 5 nm and 500 nm, they stated an increasing ionic conductivity with a decreasing particle size. As a result, the effect of 500 nm filler particles was small, matching the obtained results from

Electrolyte Preparation
PEO/LiTFSI based polymer electrolytes are manufactured in a nitrogen-filled glovebox with a solvent-free fabrication method (see Section 3.3) and salt concentration of r = 0.055 (compare Section 3.4). The procedure results in ∼ 300 µm thick films, which are subsequently dried in a vacuum oven overnight at room temperature and transferred to an argon-filled glovebox. Here, these films are hand-punched into polymer electrolyte coins of 19 mm diameter and stored until cell assembly.

Electrode Preparation
LTO anode and LFP cathode films are prepared with a conventional solution casting method which is summarized in  Carbon black is needed for increasing the electrical conductivity of the electrodes.
Finally, PEO is inserted to improve the charge transport of lithium-ions from the polymer electrolyte to the active material particles in the bulk electrode.

Cell Assembly
A schematic of the produced CR2032 coin cells is shown in   This combination (C) is added, polymer-side-up, to the positive casing (D). Next, a gasket is put in the coin cell (E), which not only ensures good contact of the pressed cathode-electrolyte sheets to the casing, but also avoids physical contact between the stainless steel and conducting coin cell parts, thus preventing short-circuiting.
Subsequently, the anode is pressed on top of the cathode-electrolyte system (F), followed by a stainless steel spacer and spring (G). Finally, the negative case of a CR2032 coin cell is pressed on top of the setup (H) and crimped closed with 1 ton of pressure. In the conclusion of these steps, an all-solid lithium-ion battery is obtained, which is brought out of the argon environment for cell testing.

EIS Measurements and Cell Cycling
The

Experiment Motivation and Description
This section presents the first cycling data obtained for the LTO-PEO/LiTFSI-LFP cells at 40 ℃. EIS is a powerful instrument to assign impedance changes to electrochemical processes. Thus, the cell's impedance behavior is investigated during cycling in order to relate capacity changes to electrochemical phenomena.   Impedance curves are shown for one of the cycled (Figure 4.4(a)) and one of the stored cells (Figure 4.4(b)). As an indicator of the cells' resistance, R cell evaluates the real impedance part of the lower frequency minimum for the cycled cells.

Results
For the stored cells, the lower frequency minimum diminishes with time. Here,   For the latter, it is important to note that the film formation would have to be cycling induced, since no arcs are observed for the stored cells. This does not match the proposed surface layer formation mechanism in the literature which is based on the reaction of lithium with impurities such as water, independent of cycling.
(iv) Generally low capacity: In the best state of the cell after cycle five, still only 27 % of the cathode active material is lithiated. The obtained cell resistance R cell for the LTO-PEO/LiTFSI-LFP ASLIBs at that state is around 2000 Ω. This is much higher compared to the liquid electrolyte equivalents. The generally high resistance and low capacity is due to the two major drawbacks in the field of ASLIBs: first, the polymer electrolyte's lower bulk ionic conductivity; secondly, a high interfacial resistance is present due to insufficient contact between the solid electrolyte and the electrodes' active material particles.

Conclusion
LTO-PEO/LiTFSI-LFP cells were investigated over cycling and storage at 40 ℃.
For a constant rate charging/discharging at C/20, results show an increase of the average specific capacity to a maximum value of around 47 mAh/g LFP after

LTO-PEO/LiTFSI-Li and LFP-PEO/LiTFSI-Li cells (referred to as LTO-or LFP-
half-cells respectively) are investigated in order to separate LTO-or LFP-specific effects from the cycling behavior described in the LTO-PEO/LiTFSI-LFP full-cells.
The well-understood lithium metal electrodes do not contain any carbon black, binder or other additives. Consequently, their use simplifies the cell setup as well as the occurring electrochemical processes, and thus facilitates interpretation of the cycling results: Electrical conductivity and lithium-ion diffusion issues can be eliminated when using these lithium electrodes since lithium is highly electrically conducting and no lithium-ion diffusion into the bulk electrode takes place during cycling. Furthermore, no current collector is needed for lithium-metal electrodes.
Three LTO-half-cells and three LFP-half-cells are fabricated as explained previously.
Their specifications can be found in Appendix C. The cells are cycled with a C-rate of C/40 in the first 5 cycles and C/20 in the following 20 cycles at a temperature of 40 ℃.  The qualitative capacity behavior of the LTO-half-cells matches the cycling results obtained for the LTO-PEO/LiTFSI-LFP full-cells: A capacity increase is followed by a drastic drop and then a slower decrease of the specific capacity. Quantitatively, a much higher maximum capacity is obtained (∼80 mAh/g LTO after cycle three).

Results
However, this difference can be explained not only by the "quasi-infinite" amount of available charge carriers due to the use of metallic lithium as one of the electrodes, but also ascribed to a difference in the cycling procedure: in the first five cycles, a C-rate of C/40 is used instead of C/20, resulting in a higher capacity at the beginning of the cycling procedure. Similarly to the full-cells, the coulombic efficiency stabilizes at around 90 % when cycling at C/20 after the first 5 cycles.
However, no major capacity drops take place here. This is due to the fact that EIS is not performed and the cells are cycled non-stop.

Experiment Motivation and Description
In order to investigate the bulk polymer electrolyte during cycling, Li-PEO/LiTFSI-Li symmetric cells are tested. Again, the use of lithium metal electrodes simplifies the occurring electrochemical processes due to the absence of any electrode additives.
Thus, interfacial and bulk electrolyte phenomena dominate the cycling behavior, facilitating data interpretation.
The lithium stripping/plating procedure for the testing of symmetric cells differs from the conventional cycling of LIBs with at least one composite electrode. Due to the quasi-infinite amount of available charge carriers in both lithium metal electrodes, a completely lithiated/deliathiated state is not reached during cycling.
Thus, a natural voltage range for cell cycling does not exist. Consequently, instead of defining cut-off voltages, cell testing is performed with a constant charge/discharge capacity: cycling occurs with a specified current and polarity is switched in constant time intervals. In doing so, each electrode alternately acts as a source (with lithium stripping from the surface) and as a sink (with lithium plating on the surface). The voltage at the end of each cycling step is then characteristic for the cell resistance.
In the following, symmetric cells are tested with two different cycling procedures.
First, stripping/plating occurs with a current of 25 µA (equivalent to a current density of about 12.5 µA/cm 2 ) for 10 h, which approximately mimics conditions in the half-and full-cells when tested at a C/40-rate. Secondly, the same current of 25 µA is applied but for a significantly shorter time of only 2 h. EIS is performed before, during and after the cell cycling to evaluate the cells' impedance behavior, specifically in terms of the bulk electrolyte resistance.

CHAPTER 5
Investigating the Electrode Composition and Cell Assembly Method

Introduction
As shown in the previous chapter, insufficient contact at the electrode/electrolyte interface is a major drawback for the LTO-PEO/LiTFSI-LFP cells at the investigated temperature of 40 ℃. In this chapter, two approaches are presented for improving the electrode/electrolyte interface. First, it is proposed to put small quantities of polymer electrolyte into the electrodes. Secondly, a different cell assembly procedure is tested by not stacking the several layers together, but melting the polymer electrolyte on the electrode surface in order to improve contact.

Electrode Composition
In conventional LIBs, electrode pores are soaked with liquid electrolyte, facilitating lithium-ion migration into the bulk electrode. In contrast to that, this is not the case for ASLIBs, since the rigid structure of the solid electrolyte prohibits its penetration into the small electrode pores. In order to circumvent this and end up with a comparable electrode structure, it is proposed to put small quantities of polymer electrolyte into anode and cathode during electrode processing. This would provide continuity at the electrode/electrolyte interface and diffusion pathways for lithium-ions into the electrode, thus facilitating accessibly to a larger portion of active material particles and consequently increasing cell capacity. Table 5.1 shows the three electrode compositions investigated in the following.

Experimental
Composition 1 does not contain any lithium salt but does consist of 2.5 wt.% PEO.
Electrode composition 2 exhibits the same amount of polymer but also consists of 0.9 wt.% LiTFSI. This is equivalent to a molecular ratio of salt to ethylene oxide repeating units of r = 0.055, representing the found to be optimal salt concentration for the polymer electrolyte at the given conditions (see Section 3.4).
In composition 3, the amount of polymer electrolyte in the electrode is doubled.
Carbon black and binder weight percentages are held constant, such that the rise of PEO or lithium salt content is compensated by a decrease of the active material content.  Afterwards, cell testing is performed by cycling with C/5. All LECs are kept at a constant temperature of 20 ℃ during cycling.

Conclusion
For optimizing lithium-ion transport into the bulk electrode, it is proposed to put LiTFSI into the electrodes.

Assembly Procedure
For optimizing contact at the electrode/electrolyte interface, a melting assembly procedure is proposed. So far, LTO-PEO/LiTFSI-LFP cells were assembled by stacking and pressing the anode/electrolyte/cathode layers together, see Figure 4.2(b) on page 88. However, this procedure allows small gaps to remain at the interface due to the rigid and uneven polymer electrolyte. By melting the polymer/salt mixture on the electrode sheets at an elevated temperature under vacuum conditions, gaps could be minimized and the electrolyte might be able to penetrate the pores, thus improving contact with the active material particles.

Experimental
Polymer electrolyte and electrodes are fabricated as explained in Section 4.2 but without punching them in the respective coins. Thus, electrode and electrolyte sheets result from these procedures. Subsequently, the melting assembly procedure is as follows, illustrated in Figure 5.3. This assembly procedure is tested with electrode composition 1 and 2 from   The behavior indicates failure due to a high cell resistance. There are different effects which could have induced this behavior. First of all, when melting the polymer electrolyte on the electrode sheets, the viscous and sticky melted polymer/salt mixture moves due to a change in density. This might have damaged the electrode/electrolyte interface, since the sticky electrolyte could have separated active material particles from the bulk electrode, inducing particle isolation. Next, it is well-known that gassing of LTO is aggravated by elevated temperatures [142], which might have worsened the electrode/electrolyte contact. Furthermore, the melting of a polymer electrolyte sheet on both electrodes also results in an additional polymer eletrolyte/polymer electrolyte interface. This might increase the total resistance of the cell and hinder lithium ion hopping from one electrode to the other. In addition, it is also worth noting that the total polymer electrolyte thickness in this setup is doubled compared to the stacking procedure, since two polymer electrolyte sheets are present in this setup.

Conclusion
For optimizing contact at the electrode/electrolyte interface, it is proposed to melt  This chemistry seems to be tailor-made for biomedical applications, which is why all investigations in this work were performed close to body temperature at 40 ℃.
The manufacturing of the polymer electrolyte was investigated in the first part of this work. In order to evaluate the ionic conductivity of the polymer/salt system as one of the critical electrolyte properties, a reproducible way to measure EIS was determined in the beginning. Afterwards, two different approaches for fabricating the polymer electrolyte were tested, a solvent-based procedure and a solvent-free one. Negligible differences in ionic conductivity and mechanical properties between both fabrication methods were obtained. Thus, the less complex and more time efficient solvent-free procedure was taken for the subsequent processing of the polymer electrolyte. Next, the salt concentration in the two component polymer/salt blend was optimized. Polymer electrolyte coins with several molecular ratios of ethylene oxide repeat units to lithium-ions (r) were fabricated. They were investigated regarding their ionic conductivity, processability and assumed wettability at the electrode/electrolyte interface. Based on that, the optimal salt concentration at a temperature of 40 ℃ was found to be r = 0.055.
As a possible additive for improving ionic conductivity and mechanical robustness of the polymer electrolyte, nanosized BaTiO 3 filler particles were investigated.
However, the results do not match the enhancing effects as described in the literature, which is likely due to obtained agglomeration of filler particles in this work.
In the second part of this master's thesis, the cycling behavior of the All in all, LTO-PEO/LiTFSI-LFP cells were manufactured by implementing fabrication methods, optimizing compositions for the components and refining an assembly procedure. The ASLIBs successfully cycle at a temperature of 40 ℃, but show need for further optimization due to a low and inconsistent capacity. The cycling behavior of these cells was related to several electrochemical phenomena based on impedance measurements and investigations on respective half-and symmetric-cells. In the end, a unique electrode composition and assembly procedure was proposed to optimize interfacial resistance.

Outlook
In the following, I present my personal recommendations on how to proceed in fabricating the polymer electrolyte, the electrodes and assembling the components together.
For the polymer electrolyte, the solvent-free fabrication procedure with a salt concentration of r = 0.055 results in a promising combination of ionic conductivity and mechanical properties for applications at 40 ℃. From the results in this work, it seems that the major drawback is interfacial resistance due to poor contact at the electrode/electrolyte interface and not the bulk electrolyte resistance.
Although not a priority, further minimizing the electrolyte's resistance would still

A.1 Problem Specification
The general layered structure of an ASLIB is shown in Figure A.1(a). For the subsequent modeling, this structure is simplified to consist only of a PE film, as shown in Figure A.1(b). Obviously, this simplification is not close to reality, as to the other layers of the battery, the PE exhibits lowest electrical as well as thermal conductivity and is the thickest layer in the described set up. This justifies the given simplification to obtain a first estimation of the resistive heating of the polymer electrolyte, as most heat will be generated in this part of the battery. To further specify this heat transfer problem, the polymer electrolyte is described to be a circular disc of thickness 2b and diameter d. It is surrounded by air with a constant temperature T ∞ , which is also the initial temperature throughout the polymer electrolyte.

A.2 Calculation
The general energy equation is given by where ρ is the material's density, c v the isochoric heat capacity, T the temperature, v the velocity,r the internal heat generation, q the specific heat, P the pressure and τ the sheer stress tensor. In the following, it is assumed that the diameter of the PE is large compared to it's thickness (b << d), and that the heat flux over where k is the thermal conductivity, the energy equation simplifies to: The internal heat generationr caused by the resistive heating can be described bŷ where the material's electrical resistance R is expressed with the ionic conductivity σ to R = 4b σπd 2 , and the PE's volume is V = π 4 d 2 b. To solve the partial, inhomogeneous differential Equation A.3, the temperature profile T (x, t) is separated into a steady state solution T ss (x) and unsteady state part θ(x, t): T (x, t) = T ss (x) + θ(x, t) (A.5) The specified problem in Figure  The PE layer is defined to be at the constant environment temperature T ∞ in the beginning, and the initial condition therefore becomes: where both sides can be plotted over C * and the intercepts will give different solutions C * n to this equation: Because the constants C * n will show up in the exponential term of the total temperature solution (equation A.33), the smallest values for C * n will impact the total solution the most. In the following, the smallest five C * n are calculated numerically and taken into account for the following calculations. The location dependent part of the unsteady state solutionX(x) is therefore approximated bŷ It is important to note, that this location dependent term still depends on the constants B n , and the initial condition needs to be evaluated after setting the total solution together to eliminate it.
Second, the time dependent partT (t) of the unsteady state solution needs to be calculated with the differential equation By multiplying both sides of this equation with cos(C * n x) and integrating from 0 to b, the summation term can be simplified to the integral of E n cos 2 (C * n x) from 0 to b. Dividing through this integral then gives the formulas for E n :

A.2.3 Total Solution
The total solution is obtained by summing up the steady state solution from Equation A.11 with the unsteady state solution form Equation A.29. Only accounting for the first five terms, this then gives where E n can be found from Equation A.32, C n from numerically solving for the intercepts in Figure A.2 and A = − 16I 2 π 2 d 2 σk .

A.3 Simulation
The calculated equations are implemented and then simulated with Matlab ® .  The steady state solution gives a quadratic temperature profile with a maximum in the middle of the polymer electrolyte, as shown in Figure A.3(a). The total temperature profile at different time points is plotted in Figure A.3 Table A.2 Qualitatively, the quadratic profile of the steady state solution is expected, as heat is generated throughout the polymer electrolyte, but cooling only takes place at the edges of the film. Quantitatively, only a negligible temperature increase is observed compared to the constant environment temperature of T ∞ =313.15 K for the given parameters. This is due to the very low chosen current of I=0.05 mA, which is why the heat production in the extremely thin polymer electrolyte sheet is small. The total solution starts at the initial temperature T ∞ = 313.15 K, and then converges against the steady state solution with increasing time.
Although the temperature profiles seam to be straight lines on the first view in this figure, it is important to note that the curves are of a quadratic nature for t > 0.

A.4 Conclusion
The temperature profile in the polymer electrolyte of a cycling ASLIB was modeled.
In order to reduce complexity of the system, the battery's structure was simplified to only consist of the polymer electrolyte film. The general approach of the calculation is based on the unsteady state energy equation with an internal heat generation term due to resistive heating. This gives an inhomogeneous partial differential equation, which is solved by separating the total temperature profile in a steady state and unsteady state part. Simulations were performed in Matlab ® with estimated parameters for a PEO/LiTFSI polymer electrolyte at 40 ℃. The steady state solution is of quadratic nature with a maximum in the middle of the polymer electrolyte, which fits with expectations. For the experimental conditions used in this master's thesis, no significant temperature increase is obtained with this model.
However, when the current is raised, the polymer electrolyte is warming up quickly due to increased resistive heating. This effect can be counteracted efficiently by reducing the thickness or increasing the conductivity of the polymer electrolyte sheet.

APPENDIX B Supplementary Material for the Polymer Electrolyte's Ionic Conductivity Measurements in Chapter 3
The following sections present the measured bulk electrolyte resistances R b (in form of the lower frequency minimums in the impedance curves obtained from EIS) and the measured film thicknesses. Both are necessary for the calculation of the in Chapter 3 presented polymer electrolyte ionic conductivity curves.   The following Tables C.1 to C.5 present the cell specifications for Chapter 4 and 5.
Consistently, anode diameter is 15 mm and cathode diameter 14 mm. Based on the active material percentage of the electrodes and an average weight of the respective current collector discs, the active material weight in each composite electrode can be back calculated. This is then taken to calculate the cell's theoretical capacity based on the limiting electrode, and the anode-to-cathode capacity ratio. Specific theoretical capacity for LFP is 170 mAh/g and for LTO 175 mAh/g.