JOURNAL OF MATERIALS SCIENCE LETTERS 10 (1991) 668-670 Large hydrostatic piezoelectric coefficient in lead magnesium niobate:lead titanate ceramics D. J. TAYLOR, D. DAMJANOVIC, A. S. BHALLA, L. E. CROSS Materials Research Laboratory, The Pennsylvania State University, University Park, PA 16802, USA Lead magnesium niobate (PbMgi/3Nb2/303, PMN) and its solid solution with lead titanate (PbTiO3, PT), which exhibit relaxor ferroelectric characterist- ics, have recently been studied extensively for possible use in actuators and transducers [1, 2]. The temperature range in which the dielectric permittiv- ity is a maximum is of special interest for these applications, even though in this relatively large temperature region the remnant polarization is absent. However, it has been demonstrated [3, 4] that a very strong piezoelectric effect may be induced in the material by the application of an external electric bias field. The possibility of control- ling the amplitude and the phase of the piezoelectric response of a material by an external field is of considerable interest in applications for transducers operating at high frequencies (such as in ultrasonic tomography) and at lower frequencies, for example, under conditions of hydrostatic pressure. Clearly in the latter, materials with high hydrostatic piezo- electric coefficients d h and gh are essential. In this letter we evaluate the hydrostatic piezoelectric coefficient, dh, in 0.9PMN:0.1PT ceramics, using the values of the transverse, d31, and longitudinal, d33, piezoelectric coefficients which were deter- mined experimentally using an ultradilatometer [5]. This composition with its maximum in dielectric permittivity around 40 °C at 100 Hz is particularly interesting for switchable transducers that operate near room temperature. The 0.9PMN:0.1PT composition was prepared using the Columbite precursor method [6]. Sintering was carried out at 1250 °C for 4 h. After sintering, the samples were annealed in oxygen at 900 °C for 4 h. The annealing step was necessary to make the samples free of any ageing effects [7]. Fig. 1 shows, for a typical sample, the dielectric constant as a function of temperature for selected frequencies. These measurements were performed during a heating cycle after the sample was held at room temperature for several days. No dielectric ageing behaviour was observed. In previous works the standard resonance tech- nique [3] as well as a modified resonance technique [8] were successfully employed in studies concerning the field-induced piezoelectric effect in relaxor ferroelectric ceramics. The advantages of the reson- ance methods include, first, the possibility of finding values of not only the piezoelectric coefficients but also the elastic and dielectric coefficients of the material and, secondly, in a more sophisticated technique [8], the ability to obtain phase information 668 for each of the above listed coefficients. Recently Pan et al. [4] have shown that the piezoelectric coefficients in relaxor ferroelectric ceramics depend strongly on frequency, particularly in the temperature range where the dielectric per- mittivity is also dispersive. This requires that all the piezoelectric coefficients be measured at the same frequency. For the resonance techniques this means that the lengths of the transverse (d31) and the longitudinal (d33) resonators should be approxi- mately the same. The resonance techniques are then unsuitable for the longitudinal mode of the relaxor ferroelectric resonators are d.c. electric bias field necessary to induce piezoelectricity may become impractically large. In addition, our experience shows that resonance of the longitudinal mode tends to be distorted under large electric bias fields, possibly because along the length of the resonator the field is no longer homogeneous. For the above reasons we have measured d31 and d33 as a function of the electric bias field using a laser interferometer [5]. The interferometer allows for measurements of ultra-low displacements (below 0.1 nm) over a wide frequency range and the measurements of both coefficients may be made on the same sample. The details of the experimental approach to measure d33 are described in [5], and details for measuring d3i are described below and represented schematically in Fig. 2. First, a thin sample was prepared in the shape of a rectangle. Using conducting epoxy, one of its sides was then mounted across the width to the sample mount and a mirror, made from glass that was gold sputtered, was attached to the other side (parallel to the sample mount) with 5 min epoxy. Care was taken in mounting the mirror parallel to the mount. Silver 30000 IOOHz-~ . oooo XX I0000 ale IM I I I I I I I I el I I I I I I I I I |OO Temperature(°C) Figure1 Dielectric constant of 0.9PMN:0.1PT as a function of temperature for selected frequencies. 0261-8028/91 $03.00 +. 12 © 1991 Chapman and Hall Ltd.