| dc.description.abstract | Since the discovery of piezoelectric ceramic materials in the mid-20th century, piezoelectric materials have become an important research field and is currently used in a whole range of applications such as sensors and actuators. The most important piezoelectric materials today are those based on the lead containing perovskites such as Pb(Zr,Ti)O3 (PZT). However, lead is known to be toxic, and safety and environmental concerns have led the EU and other governments to encourage the replacement of lead containing materials in electronic devices. The development of lead-free piezoelectric ceramics has therefore emerged as an important research field in the last decade. The Norwegian company poLight AS has developed a tunable autofocus lens (TLens) based on a PZT actuator. The present thesis was part of a project which focused on materials and technology for the next generation TLens, and the objective of this work was to develop a synthesis route to a lead-free piezoelectric material to replace the PZT actuator. Bi0.5Na0.5TiO3 (BNT) was chosen as the most suitable lead-free alternative due to the superior actuator properties reported for bulk BNT-based materials.
In the first part of the study, an aqueous synthesis route to BNT thin films was developed for the first time. Chemical solution deposition (CSD) based on water as the solvent was selected as the synthesis method due to the overall goal of producing non-toxic and environmentally friendly piezoelectric devices. In the sol, Ti isopropoxide was stabilized by citric acid (CA), Bi citrate was stabilized by ethanolamine, Na was added as hydroxide, and the pH of the sols was adjusted to neutral by ammonia solution. The synthesis route provided sols with no sign of aging still after two years. An alternative system based on Bi and Na nitrates with low pH to avoid Bi precipitation did not provide sols with sufficient stability, and was therefore not pursued further. BNT thin films were deposited on SrTiO3 (ST) and platinized silicon (SiPt) substrates by spin-coating, and the films were pyrolized and annealed by rapid thermal processing in O2. The BNT perovskite phase was observed to form after calcination at 450 °C by X-ray diffraction, while thermal decomposition of the gel was shown to be completed at 550 °C by thermogravimetric analysis. The microstructures of the thin films were shown by electron microscopy to be homogeneous and dense at 550 °C. Decomposition of the gel was thoroughly investigated, and the conditions to obtain phase pure materials were identified. The main challenge of the CSD route was related to transient reduction of Bi3+ into metallic Bi during pyrolysis, which resulted in formation of a Bi-rich pyrochlore secondary phase. Re-oxidation of Bi occurred above 450 °C, and phase pure BNT thin films could be obtained by pyrolysis at 550 °C for 5 min, while pyrolysis at 500 °C required additional thermal annealing.
The second part of the work was devoted to incorporation of Ba doping in the BNT synthesis route. Large strain in bulk BNT is only obtained upon compositional engineering, and Ba is the most common dopant. Successful stabilization of the sols with Ba was obtained by using Ba nitrate stabilized by EDTA and CA. It was demonstrated that the presence of Ba promoted the formation of the pyrochlore phase, and that Ba-rich BNT compositions required heat treatment at 800 °C to obtain phase purity. Alternatively, phase pure as-pyrolized thin films were obtained by adding 10 % Na excess to the sols. Thin films of both BNT and BNT doped with 6 % Ba deposited on Nb-doped ST were demonstrated to be ferroelectric, with leakage current characteristics comparable to state of the art BNT thin films. Further efforts to optimize the microstructures of BNT thin films were carried out. Films deposited on SiPt were homogeneous, dense, and defect free using a heating rate of 40 °C/s in the pyrolysis step, while thin films deposited on ST required lower heating rate of 1.67 °C/s to 450 °C and 40 °C/s in the interval 450-550 °C. Slow heating rate during the entire pyrolysis event caused porous films, while rapid heating resulted in thin films with defects formed due to gas evolution during thermal decomposition of the gels.
The synthesis route was further developed to incorporate more complex BNT compositions, including the cations K, Li, Sr, and Nb in addition to Ba. These dopants are also known to cause large strain response in bulk BNT. Hydroxides were used as alkali cation precursors, Sr nitrate was stabilized by EDTA and CA, while Nb was added as Nb ammonium oxalate (NH4NbO(C2O4)2). It was demonstrated that the complex compositions of the sols promoted formation of the pyrochlore secondary phase in the thin films. The first layers were phase pure, but an increasing pyrochlore content appeared with additional layers. Also in this case, Na excess was demonstrated to promote phase purity and more efficient perovskite formation, while equal amounts of Bi and Na excess did not sufficiently suppress the formation of the pyrochlore. Thin films of approximately 800 nm thickness (15 layers) were predominantly phase pure after annealing at 700 °C. Ferroelectric properties of the thin films were demonstrated, and leakage current and tolerance to high fields were comparable to state of the art BNT thin films. Compositional dependence of the ferroelectric performance could not be determined due to clamping of the substrate and leakage contributions. These dopants comprise a representative selection of the most common modifications to BNT, demonstrating that the synthesis framework developed in this study is robust and able to accommodate the most significant BNT based compositions.
The fabrication of BNT thin films on SiPt substrates is contained in the final part of the study. SiPt may be considered the industry standard for deposition of piezoelectric thin films. It consists of a single crystal Si wafer coated with SiO2, TiOx and a Pt electrode. It was found that a high reactivity of Pt and Bi from the BNT film caused a detrimental weakening of the TiOx-Pt interface. This led to delamination of the films as they reached approximately 300 nm thickness. A BaTiO3 (BT) protective layer was introduced to spatially separate the BNT layer from Pt, leading to thicker achievable BNT films before delamination occurred. The microstructure and morphology of the BT protective layer was improved by adjusting the number of depositions and heat treatment procedure, but a sufficiently dense BT film was not achieved, and reaction between Bi and Pt was not completely eliminated. The films delaminated at a thinner critical thickness than what can be predicted from evaluating the stress due to thermal expansion mismatch of Si and BNT. Nevertheless, the feasibility of applying a protective BT layer for deposition of BNT on SiPt was demonstrated, and its potential was established.
In summary, the content of this work was the successful development of a novel aqueous BNT CSD method with implementation of the relevant dopants. Challenges related to transient Bi reduction, formation of the pyrochlore secondary phase, and Pt-Bi reaction were the main aspects requiring in-depth investigation. | |
