Scavenging mechanical ambient energy is a promising solution to ensure the autonomy of wearable transducers. Nevertheless, the development of soft electrostatic generator (DEG) is up to now slow down due to the use of high bias voltage, high strain and innovative mischaracterized materials. The aim of this Ph-D thesis is to propose innovative solutions to these technological barriers. Firstly, a complete characterization of the electrical and mechanical properties of two commonly used dielectric polymer (acrylate VHB 4910 and silicone Polypower) has revealed the true physical properties of these polymers and especially the influence of the pre-stress and the nature of the electrode used. Thanks to these data, reliable analytic laws have been proposed and inserted into our thermodynamic model in order to predict the output power and scavenged energy density for the DEG. Moreover, our model allow us to propose improvements for the materials used in these applications. The second challenge is to propose an alternative to the high bias voltage needed for these soft generators. We have proposed an innovative solution combining an electret and a dielectric elastomer. Various configurations of hybrid generators in 2D or 3D geometry have been modelled and evaluated. Finally, a prototype has been designed allowing scavenging 35µW. With an appropriate optimization of our prototype, hundreds of µW can be scavenged.
