Chemical solution deposition of oxides for energy storage applications: the breakthrough in coating high aspect ratio 3D structures

Abstract

Since the need for clean and efficient energy storage devices is expected to become immense, as the field of applications is constantly growing, the thesis did firstly focus on the synthesis of materials that can be used in those systems. Particularly, it was the intention to deposit functional multi-metal oxide films on planar substrates, interesting for thin-film capacitors or Li-ion thin-film batteries, and this via the use of chemical solution deposition. Aqueous “sol-gel” precursor chemistry was used, which is known to be cheaper, easier and more environmentally friendly compared to traditional sol-gel routes using organic solvents. Moreover, it allows the fabrication of complex multi-metal oxides due to the intimate mixing of metal ions and the resulting good homogeneity of the precursor.[1–3] The first group of materials which is very attractive for energy storage devices and in particular for thin-film capacitors are ultra high-k materials. Due to their extremely high dielectric constant (k>10000), they can possibly be used for the miniaturization of the energy storage devices and the increase of their energy density. For orthorhombic LuFeO3 ceramics, a dielectric constant of 10000 (frequency ≤ 1kHz) had been reported at room temperature. [4] Nevertheless, research on the deposition of LuFeO3 thin films, indispensable for the thin film applications envisaged, was still very scarce, especially when it did come to wet chemical deposition. A stable aqueous Lu/Fe multi-metal ion precursor was prepared by mixing citrate complex-based Fe and Lu solutions in the desired ratio. Prior to thin film deposition, the thermal decomposition of the precursor and the phase formation process towards bulk, crystalline LuFeO3 were studied. It was shown that phasepure orthorhombic LuFeO3 could be formed at 900°C, indicating a high temperature requirement. These insights, combined with a profound study on the key deposition and process parameters, were successfully employed for the formation of orthorhombic LuFeO3 thin films on Si3N4 via spin coating. Impedance spectroscopy analyses did confirm their associated ultra high dielectric constant (> 10000) at room temperature for frequencies lower than or equal to 1kHz, making them promising for the use in thin film capacitors. The downside of the material is that the dielectric loss is quite high at low frequencies. Future work could focus on the deposition of orthorhombic LuFeO3 on a metallic substrate in order to create a thin-film capacitor. The characteristics of this device, and consequently the usefulness of the dielectric layer, could then be determined in a more direct way. The further development of Li-ion batteries and specifically all-solid state Li-ion batteries would significantly improve battery safety and could bring a higher energy density at system level. A new process was developed for the deposition of Li4Ti5O12 films. This material is known as a very stable high-performance electrode material.[5] While compensating for the lithium loss during the deposition process in the stable Li/Ti precursor, a phase-pure film could be achieved via spin coating after a post deposition anneal at 700°C. The phase-pure film was characterized by a very high capacity, particularly ca. 800 mAh/cm3 for the low C-rates (0.5C-1C), while a mixed Li4Ti5O12/anatase film’s capacity was 500 mAh/cm3 at 0.5C and 1C. Since 800 mAh/cm3 is higher than what is theoretical possible, probably the amount of active material was underestimated. However, it seems that this aqueous solution-gel process is able to deliver spinel Li4Ti5O12 films which can be incorporated in thin-film Li-ion batteries. A more precise determination of the amount of active material and a more thorough electrochemical study should be performed to evaluate all electrochemical characteristics including the exact capacity and cycle life. (LixMg1-2xAlx)Al2O4 had been suggested as a novel Li-ion electrolyte with spinel structure and thus structurally compatible with the Li4Ti5O12 and LiMn2O4 spinel electrodes.[6] A stable Li/Mg/Al multi-metal ion precursor was developed by mixing citrate-based Li, Mg and Al ion solutions in different ratios so that the theoretical x-value in (LixMg1-2xAlx)Al2O4 was varied (0≤x≤0.5). Prior to film deposition, a crystallization study of the Li/Mg/Al powders showed that for x ≤0.3 phase-pure spinel (LixMg1-2xAlx)Al2O4 could be formed at 600°C in air. Crystalline spinel films could be achieved after a post deposition anneal at 800°C. A too high Li:Mg ratio for X=0.2 lead to the formation of LiAlO2 as secondary phase, deteriorating the intrinsic film quality and morphology. Electrochemical impedance spectroscopy showed that none of the deposited films possess a decent ionic conductivity. This is probably due to the formation of an amorphous Al2O3 interface at the grain boundaries, caused by an excess of Al available in the films. Despite the potential good lattice matching between these spinel electrolyte films with spinel Li4Ti5O12, the low ionic conductivity (much lower than 10-6-10-8 S cm-1) tackles the potential use of these solid electrolyte films in thin-film Li-ion batteries. Future work should focus on the study of the grain boundaries, e.g. by HRTEM/EDS characterization, to demonstrate or disprove the proposed hypothesis. It is believed that if the formation of Al2O3 could be avoided, e.g. by a more elaborate tuning of the Li:Mg:Al ratio, this could result in a higher ionic conductivity. Then, the material could be stacked with Li4Ti5O12 as anode material to obtain Li-ion battery half-cell as proof-of-principle. It is obvious that a similar process for deposition on a 3D substrate would take these materials even a step higher. Various known wet chemical methods were investigated. Those were mostly based on reactions with particular substrates and/or could not deliver a complete coverage on microstructured substrates. The knowledge from these experiments was further used to develop a novel chemical solution deposition (CSD) method for the deposition of (multi-)metal oxide coatings on 3D substrates, i.e. silicon micropillars coated with 20 nm TiN. The CSD method was shown first for TiO2, a candidate anode material. As starting point, the aqueous citrate-peroxo Ti precursor was optimized, in particular by adding ethanol, to allow coating on high aspect ratio features. The whole was spray coated at elevated temperature (180°C) and by controlling all levels of the process, complete coverage of the silicon pillars with the Ti precursor was obtained. Since the deposition was performed at elevated temperature, the solution was instantly transformed into a gel, consisting of coordinated complexes in which Ti ions were immobilized. Further heat treatment of the pillars lead to the complete combustion of the gel and thus to titanium oxide formation, it also triggered the crystallization of the deposited layer. A thorough crystallization study did indicate that phase-pure anatase TiO2 films could be obtained by a welldefined post deposition anneal at 500°C. Lithium-ion insertion and extraction behavior was observed for the anatase film deposited on TiN-coated silicon pillars, further illustrating its application potential for 3D Li-ion batteries. Further electrochemical experiments are necessary to determine its exact potential towards the proposed application. The presented technique was also successfully applied to deposit Li4Ti5O12 on the microstructured substrate. Furthermore, the method was used for the deposition of Lu/Fe oxide films, as described earlier, promising as dielectric material for thin film capacitors. More experiments should be defined to investigate their characteristics. The determination of exact thickness and conformality of the resulting coatings is not straightforward. Focused ion beam technology (FIB), combined with SEM, could be the solution to counter this issue. On-going research, using the “embedding technology”, shows that the reported novel method, which implies operational simplicity, low cost and relatively short deposition times, can deliver coatings with a conformality of ca. 13 % (standard deviation). Furthermore, the process could be successfully transferred to the more interesting 2 µm X 2 µm pillars, enabling the highest possible capacity, as discussed before. The technique is also being applied to achieve a series of other multi-metal oxide coatings (e.g. LLTO, LiWO3) on 3D substrates, bringing various novel applications such as commercial thin-film 3D Li-ion batteries one step closer to reality. Since various other multi-metal oxide coatings on microstructured surfaces seem to be feasible (as long as ethanol can be added to the aqueous precursor solution), obviously, there are a lot of opportunities to exploit the novel CSD method even more. The ultimate goal, at the level of applications, is thus to develop a 3D allsolid state battery, consisting of an electrolyte film stacked between two electrode films, obtained via the described method. Furthermore more fundamental research could be performed to unravel the exact deposition mechanism of the spray coating technique. This requires an extra experimental study (e.g. by the use of a high-resolution camera) together with a profound modeling study.