Abstract
Thick PZT films (1 - 20 m) have been prepared using a composite sol gel technique whereby PZT powder and a PZT producing sol are formed into a slurry and spin coated onto silicon wafers. The maximum relative permittivity obtained was approximately 80% of that exhibited by bulk PZT of comparable composition. However, the d33, f and e31, f[1] piezoelectric coefficients were shown to be significantly lower than that of bulk PZT. It has been proposed that the measured value of d33, f is affected appreciably by particle-particle rotation and substrate clamping leading to reduced poling efficiency which may also greatly reduce the value of e 31, f observed. Samples with high levels of porosity have been shown to exhibit a reduced value of d33. This was attributed to 31 and 51 mode piezoelectrically generated charges caused by the bending and shearing of particle-particle bridges. The effect of substrate clamping, on d 33, f and poling, has been studied by monitoring the changes in position and intensity of the (200)/(002) X-ray diffraction (XRD) peaks of composite films. The presence of the substrate was found to introduce tensile stresses parallel to the film plane which distorted the unit cell. Subsequent permanent polarisation following poling was found to be reduced due to the presence of these stresses. The discrepancies between the values of d 33 measured on thick films and bulk ceramics were highlighted as being of particular importance if thick film materials are to be modelled for device applications. Thick film piezoelectric coefficients (i.e. those of the combined film-substrate structure) should not be used in place of material piezoelectric coefficients when attempting to model the behaviour of devices. Such actions would inevitably lead to erroneous results. © 2002 Taylor & Francis.