Characterization and Morphological Analysis of Porous Electrodes for Lithium-Ion Batteries
Climate change is one of the greatest challenges of the century. Compared to the pre-industrial era, the average global temperature has already risen by 1 °C (1.5 °C in Germany). The temperature increase is caused by the emission of greenhouse gases, which convert sunlight reflected into the atmosph...
|PDF Full Text
No Tags, Be the first to tag this record!
|Climate change is one of the greatest challenges of the century. Compared to the pre-industrial era, the average global temperature has already risen by 1 °C (1.5 °C in Germany). The temperature increase is caused by the emission of greenhouse gases, which convert sunlight reflected into the atmosphere into heat. Around 20% of the CO2 emissions in Germany are attributed to the transport sector. Electro mobility represents a possible solution to this problem. Powerful lithium-ion batteries (LIBs) are needed for mobile storage of electricity from renewable energy. Three main approaches are being pursued to improve the performance of LIBs: the search for new materials, the development of new battery concepts, and the optimization of existing systems. This work takes the latter approach by imaging and studying the electrode morphology using tomography. Detailed morphological analysis and simulations are used to identify microstructural kinetic limitations. The results are compared with electrochemical characterization methods. In the following, the results of the five chapters of this cumulative dissertation are summarized. Chapters 1–3 are related to the study of transport limitations in batteries using a liquid electrolyte, Chapter 4 deals with all-solid-state batteries, and Chapter 5 applies the reconstruction approach to a hierarchical porous material.
In Chapter 1, transport limitations of an electrode are detected by both reconstruction-simulation (RS) and electrochemical measurements, and the results of the two approaches are compared to each other. The aim of the study is to determine the ionic tortuosity in both ways to quantify transport limitations in the pore space, filled by a liquid electrolyte. Graphite, which is a common anode material, is chosen as the active material. First, graphite electrodes with different thicknesses are investigated by electrochemical impedance spectroscopy (EIS) in a symmetrical cell setup. The resulting spectra are fitted using the transmission line model (TLM), which describes the impedance of porous electrodes. The analysis reveals an ionic tortuosity of τ_EIS=7.3. Second, one graphite electrode is physically reconstructed over the entire cross-section using FIB-SEM tomography. For this purpose, the pore space of the electrode is infiltrated by an osmium-based contrast agent. The space which is filled by the liquid electrolyte in normal battery operation is thus directly imaged and the contrast of the resulting image stack is enhanced, facilitating an accurate reconstruction. A comprehensive morphological analysis is conducted featuring porosity profiles, the geometric tortuosity, and a chord length distribution (CLD) of the solid phase and void space. All analyses are performed with regards to the spatial direction, showing that the flaky graphite particles form a distinct anisotropic microstructure. This leads to strong transport hindrances in the direction perpendicular to the current collector. The reconstruction volume is verified to be representative by a finite-size analysis, which is indispensable in order to obtain reliable results. Diffusion simulations based on a random-walk approach yield a similar tortuosity value of τ_RS=6.55, which is within the experimental error of τ_EIS. Consequently, this study shows that long-range transport simulations (without considering double-layer formation) and EIS combined with TLM (ion transport in the pores and double-layer formation) give comparable results even for a highly anisotropic microstructure. Compared to FIB-SEM tomography along with numerical simulations, EIS is significantly faster, cheaper, and easier to apply, and it is available in almost every electrochemical laboratory. However, the underlying microstructural features causing steric transport hindrances can only be analyzed by appropriate tomography methods. EIS screenings can be used to detect transport limitations of newly designed electrodes. Thus, the results of this study may contribute to the future development of more powerful electrodes.
In Chapter 2, the impedance of electrodes with variable thickness is examined for different liquid electrolyte systems. For this purpose, batteries are first cycled using a tetraglyme-based solvate ionic liquid (IL), a conventional carbonate-based electrolyte, and a LiFSI in IL electrolyte system. The area-specific resistances are estimated based on the overvoltages at 50% state of charge, which increase in the order; carbonate-based electrolyte < IL < solvate IL. The different electrolyte systems are characterized based on the electrode thickness by means of EIS. Special attention is paid to the impedance at 10 4 Hz, since this frequency corresponds approximately to the time scale of typical cyclization rates of 1–2 C. The impedances of the electrolyte systems increase in the same order as it was observed in the cycling experiments. Next, the analytical model of Huang and Zhang is used to shed light on the individual contributions to the overall electrode impedance for the carbonate-based electrolyte and the solvate IL. This model calculates the electrode impedance, taking into account salt concentration polarization in the electrolyte-filled pore space. It is applicable to electrolyte systems consisting of one type each of cation and anion in a solvent. Therefore, the LiFSI in IL electrolyte cannot be analyzed by this model. At 10-4 Hz, only a weak dependence on the electrode thickness is observed for the real part and the modulus of the complex impedance in the range of 50–100 µm. Using a generalized TLM, the impedance contributions of ion transport and thickness-dependent charge transfer as well as solid phase diffusion are analyzed separately. The impedance of ion transport for both the solvate IL and the carbonate-based electrolyte is higher than the contribution from charge transfer and solid phase diffusion at 10 4 Hz and for thicknesses between 50–150 µm. This explains the low dependence on thickness of the impedance spectra and leads to the conclusion that greater electrode thicknesses than the conventional 80 µm would be possible, given that the morphological properties can be kept constant over the entire electrode.
Chapter 3 examines the influence of the carbon-binder domain (CBD) on Li+ charge transport in the electrolyte phase by using a RS approach and compares the results with EIS experiments. The morphology of the electrolyte-filled pore space in LIBs is influenced by the microstructure of the solid components: active material (AM) particles, binder, and conductive carbon. The binder and conductive carbon form an interpenetrating nanoporous phase, the CBD. While the µm-scaled AM particles can be easily reconstructed by 3D tomography, the CBD is often not taken into account due to its small feature size. In this chapter, a LiCoO2 (LCO) composite cathode is physically reconstructed by means of FIB-SEM tomography to determine the Li+ transport tortuosity and to morphologically characterize the CBD. EIS experiments in the framework of the TLM are conducted to determine the ionic tortuosity experimentally and are compared with the RS approach. The three-phase reconstruction provides both the hitherto highest reported resolution down to a voxel size of (13.9 × 13.9 × 20.0) nm3, and an unprecedented large volume with a minimum edge length of 20 µm. This enables a representative description of the interstitial pore space. A detailed morphological analysis is presented to characterize the morphology of the void space featuring CLD, specific surface area determination, connectivity analysis, and calculation of the geometric tortuosity. The results show that the microstructural properties of the cathode are affected by the presence of the CBD spanning the void space as a convoluted network and leading to more tortuous and constricted Li+ transport pathways. Pore-scale numerical diffusion simulations reveal a significantly higher ionic tortuosity of 1.9 when the CBD is taken into account compared to 1.5 without CBD, which cannot be solely attributed to the lower porosity. The RS analysis underscores that only pore-scale simulations in physical reconstructions including the CBD can reproduce experimental tortuosity values derived from EIS.
In Chapter 4, the morphology of two sheet-type all-solid-state battery (ST-ASSB) cathodes with different solid electrolytes (SEs) is investigated to identify kinetic limiting features. The slurry-based manufacturing process of ST-ASSBs is comparable to that of conventional lithium-ion batteries and is thus relevant for eventual mass production. The sulfur-based SEs are β-LPS (β-Li3PS4) and LPSI (2 Li3PS4∙LiI) with conductivities of 0.2 mS cm 1 and 0.8 mS cm 1, respectively. While β-LPS is composed of mesoporous nanoparticles, the LPSI particles exhibit sizes up to the µm range and no intrinsic porosity. Small state-of-the-art NMC 85|05|10 particles coated with LiNbO3 are used as cathode active material (CAM). Three-phase FIB-SEM based reconstructions of large cathode volumes in high resolution reveal structurally representative and realistic models of the SE, CAM particles, and void space. The binder, which is distributed as a thin layer over all surfaces, cannot be resolved due to its small feature size and poor contrast. The volume fractions found in the reconstructions suggest that, for β-LPS, the binder accumulates predominantly within the nanoparticulate SE phase due to the high intrinsic surface area. For LPSI, it is distributed over all interfaces. Void space is dead space in ASSBs as it makes transport paths in SE more tortuous, prevents charge transfer at the SE–CAM interface, and reduces the volumetric energy density of the battery. For the β-LPS-based cathode, a small void fraction of 1 vol% can be found, while LPSI exhibits a much higher fraction of 11 vol%. The voids in the LPSI-based cathode are larger compared to β-LPS and mainly found at the SE or SE–CAM interface. For β-LPS, the voids are predominantly surrounded by CAM. This explains the larger active surface area of 87% for β-LPS, while 62% of the CAM surface is in direct contact with the SE for LPSI. An analysis of CAM connectivity shows that >99% of the CAM volume is directly connected in each case, making electron transport within the cathode uncritical. Numerical transport simulations show that the ionic tortuosity of the electrolyte phase of the LPSI sample is twice that of the β-LPS sample. In contrast, cycling experiments reveal that the LPSI sample has a higher discharge capacity (178 mAh/g vs. 150 mAh/g) and lower overvoltage. Using a general TLM, the individual contributions to the battery impedance were estimated to draw conclusions about kinetic limitations in the two samples. The charge transfer at the SE–CAM interface accounts for by far the largest impedance, while Li chemical diffusion in the CAM and ionic transport in the SE phase account for only a comparably small fraction. Due to their similar chemical composition, both electrolytes exhibit a similar charge transfer resistance. However, due to the higher effective SE–CAM interface of the LPSI sample, a lower effective charge transfer resistance is obtained, which explains the lower overvoltage. Consequently, especially in lowering the interfacial impedance, there is still great potential to further improve the performance of ST-ASSB.
In the fifth Chapter, the physical reconstruction technique is applied to hierarchical porous materials (HPMs), which have a high potential for use in the field of energy storage and conversion. HPMs are a class of functional materials characterized by a large specific surface area and an interconnected pore space with high accessibility. The study presents a universal laser-based procedure for generating metal oxide HPMs with cauliflower-like morphology. Based on a facile nanosecond pulsed laser-treatment, the manufacturing process is easy to implement, solvent-free, and scalable. The resulting hybrid micro-/nanostructures can be generated on a variety of metallic substrates over a wide range of melting points. The morphology of the superstructures can be directly controlled by varying the laser parameters. The formation process is investigated in detail by means of FIB-SEM tomography. For this purpose, the cauliflower-like structures are generated on copper metal, embedded in epoxy resin, and physically reconstructed. Cross-sections of the superstructures show a ring-like pattern similar to tree rings. These rings can be fitted by ellipses with a constant center point and linearly rising elliptical axes, which makes the structures resemble an ideal ellipsoid. The distance between the single rings is constant and depends on the laser scan line distance. The porosity increases towards the outer surface, resulting in a large external surface area. A hierarchical network of pores with diameters from nanometer to micrometer is created. During generation, the laser scans the metal surface in a linear pattern, causing material to melt and partially evaporate whereby the metal partially oxidizes. In a self-organization process, microstructures are created which grow layer by layer through stepwise recondensation due to the meandering laser path. A complex and over several orders of magnitude self-similar morphology is formed. Interestingly, the determined fractal dimension corresponds to that of natural cauliflower. The concept can be applied to a variety of materials, especially transition metals, such as those used as cathode material in LIBs.
In conclusion, this work provides new insights into the microstructure of battery electrodes. For this purpose, a protocol for two- and three-phase reconstructions is developed and applied to a variety of different samples. It is shown that only direct imaging of the morphology provides reliable conclusions about the reason for transport limitations and morphological heterogeneity. FIB-SEM tomography is the method of choice for physical reconstructions of electrodes as it achieves a sufficiently high resolution and provides a high sensitivity towards light elements, such as those found in the CBD. Optimization of the electrode morphology to reduce transport limitations will help to make LIBs even more efficient in the future.