Determination of Cross Sectional Variation of Ferroelectric Properties for Thin Film (Ca. 500 nm) PZT (30/70) via PFM

Abstract

There is much speculation about the origin of fatigue in ferroelectric thin films that have been grown by sol-gel on various substrates. One of the most important substrates for growing thin film ferroelectric materials is native SiO2. After the deposition of suitable electrode, most common are Ti-Pt, a thin film of a ferroelectric material can be grown on the substrate. Procedures to grow thin film lead zirconate titanate (PZT 30-70) have been well publicized and it is now routine that almost 100% [111] orientated PZT with a columnar structure of ca. 100 nm is grown on SiO2 substrates. Recent studies using PFM have shown that it is possible to determine a variation in hysteresis loops with a spatial accuracy of 50 nm on the surface of a ferroelectric material. This study was to determine the variation in d33 for a thin film sample of ferroelectric in terms of a depth profile. It was not suitable to just turn the sample on it's side and analyze the d33 in terms of depth as the continuous back electrode would mean that the resulting vibration of the ferroelectric surface would be a combination of d33 and d31. The de-convolution of this signal was outside the scope of the current investigation. In order to determine the d33 of the material in terms of depth the sample was machined (with an accuracy of 10 nm) to release the ferroelectric from the underlying electrode and electroded in the correct orientation for d33 to be investigated in cross section.Samples of perovskite PZT(30/70) have been produced via sol-gel using a spin coating technique that are ca. 500 nm in cross sectional thickness on Pt/Ti/SiO2 back electrodes. The PZT is shown to be highly [111] orientated and has in the past been well characterized both at the macro and nanoscale using a variety of techniques. Using focused ion beam (FIB) milling the PZT layer has been released from the SiO2/Ti/Pt substrate and electroded in such a way as to determine the δ33 properties of the film in terms of thickness of the as-deposited film. Figure 1 indicates the machining process used to generate the free standing PZT block. In Fig. 2, the results of the FIB machining process are shown. The ‘bridge’ structure has the dimensions of 500 nm in cross section and is long enough to ensure that electrode cross talk between the d33 and d31 modes is reduced. Hysteresis loops have been generated using the technique of piezo force microscopy (PFM) at discrete locations across the surface of the film corresponding to a depth profile of the as-deposited PZT. Variations in the shape of the hysteresis loops, calculated δ33 and coercive field for the film have been related to the variation in oxygen defect density in the film. This study shows that the PZT thin film has a variety of fundamental piezoelectric constants that are associated with the variation in local crystal chemistry.