Résumé
This work focuses on the different technological characteristics and physical phenomena that can produce a dispersion of performance on surface acoustic wave (SAW) sensors. Using of a sophisticated statistical approach for the study of the devices variability requires a large amount of data and can be difficult to produce or compare. Modeling may improve the understanding of the physical and technological effects, thus it can help to estimate and reduce the dispersion produced by refining the design and / or the manufacturing process. A list of design parameters assumed to be at the origin of the variations of the SAW performance is proposed. In this work, measurements of approximately one thousand samples were grouped according to their characteristics and compared by using a Finite Element Model (FEM) analysis. The obtained results support the interest of the model and of the method developed to predict the influence of geometric and physical parameters on the acoustic wave devices behaviour, paving the way for a more exhaustive stochastic approach.