Electrode stability, interdiffusion, phase purity and deviation from stoichiometry at the PZT-electrode interface are key issues in PZT thin film integration. This article highlights the use of transmission electron imaging combined with energy dispersive spectroscopy (EDS) for the investigation of these phenomena. The accuracy of the EDS analysis is discussed. It will be shown that using a standard PZT sample and controlled conditions, reliable analysis can be performed. Diffusion mechanisms have been studied for Pt based electrode systems and RuO2-based electrode systems developed for direct integration onto silicon. The materials studied were composed of stacks of silicon-silicon oxide, an adhesion layer (Ti, Ta or TiOx), an electrode (Pt or RuO2) and PZT (45/55). The PZT was deposited by sol gel using the same parameters to allow for comparison of the different electrodes. Four different electrode / adhesion layer materials were compared (Pt/Ti, Pt/TiOx, Pt/Ta and RuO2 / TiO2). In the case of Pt, lead and oxygen diffusion through the electrode is observed. While the initial Ta layer transforms into a homogeneous pyrochlore phase, the Ti adhesion layer is heavily deformed. In the case of TiOx the lead is mainly incorporated at the interfaces with Pt and with SiO2. No lead diffusion to the adhesion layer is observed for the RuO2 electrode. In-situ sputtering and sol-gel deposition of PZT are also compared. The sol-gel films are close to the right stoichiometry for the perovskite while the sputtered films contained an excess of lead. No sign of second phase is found by X-ray diffraction (XRD), by EDS and by high resolution transmission electron microscopy (HRTEM) which suggests that the excess lead is accommodated in the perovskite lattice.