Hardening effect and temperature dependence of lead-free ferroelectric and relaxor materials for high-power applications

Hard piezoceramics are essential for a wide range of high-power applications, such as ultrasonic cleaners, ultrasonic welders, and high-power transducers. For these applications, piezoelectrics are operated at high frequencies, typically in kilohertz (kHz) to megahertz (MHz) range, to generate high output power. The performance of these piezo-resonators is heavily influenced by material losses, as those can lead to their thermal breakdown. Therefore, minimizing losses is crucial when developing lead-free alternatives to state-of-the-art lead zirconate titanate (PZT). Piezoelectric materials used in high-power applications are typically assessed by the quality factor (Q), which is a parameter inversely proportional to the losses of the system, and the maximum reachable vibration velocity (v). An additional crucial requirement that has to be satisfied is the stability of their electromechanical properties across a broad operating temperature range. Despite this, there has not been a systematic study exploring the behavior of these materials in high-power mode under varying temperatures. This work aims to address this research gap. Solid solutions of sodium bismuth titanate and barium titanate (NBT-BT), as well as potassium sodium niobate (KNN), are discussed as the two most promising candidates for PZT alternatives in the literature. As acceptor doping is a well-established technique to ferroelectrically harden a material and consequently reduce losses, Mg-doped NBT-BT and Cu-doped KNN are investigated regarding their behavior under high-power conditions. First, temperature-dependent synchrotron X-ray measurements have been conducted to investigate the effects of temperature on the poling-induced domain texture of relaxor ferroelectric undoped and 0.5 mol% Mg-doped NBT-6BT. The aim was to assess whether the high degree of domain texture and, consequently, stability of Q is retained at different temperatures. Structural findings were correlated with the piezoelectric properties of Mg-doped and the undoped NBT-6BT quantified with small and large signals and under high-power conditions. The results on structure highlight a high degree of domain texture close to the theoretical maximum in both systems at room temperature, but temperature-dependent structural changes revealed a phase transition range from 70 °C to 90 °C. At this temperature range, piezoelectric properties are deteriorated, suggesting a strong influence of the structural phase transition on their stability. Subsequently, the effect of different amounts of Cu doping on the hardening behavior of KNN samples driven at high-vibration velocities was investigated. KNN-xCu (x = 0, 0.5, 1.0 mol%) ceramics were examined using the pulse drive method with a burst excitation on a customized setup at temperatures ranging from -40 °C to 140 °C, which is a typical operating temperature range for most high-power applications. Similar to hard-PZT, the quality factors Q of KNN-Cu doped samples decline with increasing vibration velocity. However, the Q of KNN-Cu doped ceramics decreases at a lower rate than PZT. Therefore, KNN-Cu doped ceramics were able to vibrate at velocities above 2 m/s, while PZT stopped vibrating at around 1 m/s. The 1.0Cu-doped composition exhibits the highest Q at room temperature (~1200), while the 0.5Cu-doped demonstrated superior thermal stability of Q at 1 m/s. Pulse drive measurements combined with time-resolved in-situ high-energy X-ray diffraction acquisitions were conducted to investigate the microscopic origins of the electromechanical response observed at the macroscopic level. This allowed us to decouple the strain contributions of ferroelectric domain wall motion and the lattice. The quantification of these contributions has only been conducted on PZT and NBT-BT systems, while it has never been reported for KNN. The result revealed that the transverse strain of both the 1.0Cu-doped composition and the undoped version originating from non-180° domain wall motion is lower than the one originating from lattice distortion. Driving the material to higher vibration velocities leads to intensification of both contributions, as also reported for hard and soft PZT. However, Cu doping reduces the ratio of non-180° domain wall motion over the total strain of KNN to 28%, which is a higher percentage compared to hard PZT (10%–15%) and NBT-BT(3%–5%).