Preparation and properties of KNbO3-based piezoelectric ceramics

During the last several years, new lead-free piezoelectric materials have been developed to replace the lead-based materials, such as PZT. Presently, the family of lead-free ceramics showing the most promising piezoelectric properties is based on potassium sodium niobate (KNN). The KNN ceramics have been investigated since 1960s, but many problems arise, especially during the synthesis. Potassium and sodium based powders are moisture sensitive leading often to disintegration of samples after sintering and to moisture dependent properties. Another drawback is the poor densification during sintering. Elaborated procedures (hot pressing, special handling of powders) are needed to produce high quality KNN ceramics in a reproducible way, and their properties are inferior to those of PZT. Li, Ta and Sb modified KNN, however, were reported in 2004 to exhibit properties comparable to those of PZT. These modified compositions promise to be a new generation of environmentally safer piezoelectric materials. The goals of this thesis are to prepare selected compositions within this family and examine their properties relevant for applications in medical transducer. The emphasis of the work is on piezoelectric properties and their stability with respect to the temperature, humidity, and preparation conditions. The unmodified, and lithium (K, Na, Li)NbO3 and the lithium with tantalum modified KNN ceramics, (K, Na, Li)(Nb, Ta)O3 have been produced by the conventional solid state synthesis. The conventional processing steps have been adapted with a goal to obtain reproducible high quality samples without using complex techniques such as hot pressing or special powder handling. In particular, the particle grain size and particle size distribution have been controlled for all the steps; this control starts with the initial powders. To reduce the particle sizes the most efficient milling method has been found to be attrition milling. Another important point is the compositional homogeneity. To improve this homogeneity, a second calcination step has been added to the process. Finally, the sintering step is sensitive, the sintering temperature range in these compositions is as narrow as 5 °C and in some cases, the dwell time is reduced to minimum (several minutes) to avoid grain growth. The densities of the so-obtained ceramics are higher than 95%, but some compositional inhomogeneities have been observed in electron microscopy. The electromechanical properties at room temperature are promising for example d33 = 240 [pm/V], kt = 51%, kp = 45% with e = 919 [-] and tanδ = 2.6% for ceramics modified with 7 at% lithium, and d33 = 310 [pm/V], kt = 46%, kp = 46% with e = 829 [-] and tanδ = 2.4% for ceramics modified with 3 at% lithium and 20 at% tantalum. These properties are especially enhanced for compositions with the orthorhombic to tetragonal phase transition close to room temperature. Dielectric and piezoelectric (resonance, converse and direct) measurements as a function of temperature have demonstrated that the orthorhombic to tetragonal phase transition is strongly dependent on temperature unlike in PZT, where it is controlled primarily by the composition. The thermal behaviour of ceramics is thus influenced by the presence of the phase transition in the analysed temperature range. The thermal stability has been analysed in the perspective of medical transducers, which are subjected to sterilisation cycles up to 140 °C. A depolarisation after the first cycle has been observed for all the compositions, the smallest being for the lithium and tantalum modified KNN ceramics. After the second cycle, the properties stabilise, being the best for the compositions (K0.465Na0.465Li0.07)NbO3 and (K0.485Na0.485Li0.03)(Nb0.80Ta0.20)O3. As the enhanced piezoelectric properties of these materials are related to the presence of a phase transition, the determination of the contributions of the domain walls in either phase is then of practical importance. The intrinsic and extrinsic contributions of the piezoelectric response can be separated by analysing the piezoelectric response as a function of the applied stress (direct measurements). It has then been shown that the intrinsic contribution is always higher than the irreversible domain wall motion contribution. This latter contribution is nevertheless higher in the orthorhombic phase than in the tetragonal phase. Finally, as the potassium niobate based ceramics have shown in the past their moisture sensitivity, aging measurements have been done in different atmospheres. Surprisingly, the properties of the modified KNN ceramics studied are not found to be dependent on the moisture and only small aging has been noticed. After 100 days, the (K0.485Na0.485Li0.03)(Nb0.80Ta0.20)O3 ceramics show a decrease of the electromechanical properties below 10% and the coupling coefficient are almost stable with time. With these electromechanical properties and their time/thermal stabilities the modified KNN ceramics are promising substitutes to lead-based materials, in particular the lithium and tantalum modified ceramics.