Porous Silicon (PoSi) is becoming a widely used material for applications including microelectronics, as substrate for RF devices or in gas sensors for example. The manufacturing processes are nowadays well known. Here, the sample etching is performed using a galvanostatic anodization in a HydroFluoric solution. After the etching process, essential parameters to control thickness or porosity are often determined empirically. Classical measurement methods to evaluate the properties of the porous layer after etching are mostly destructive. In order to monitor the material modifications during the etching process, ultrasonic methods seem to be a promising way since they could allow insitu contactless non-destructive evaluation of samples to be performed. The aim of this study is to investigate the feasibility of such a monitoring method which implies the development of an inverse problem resolution method to obtain the thickness, density, longitudinal wave velocity and attenuation coefficient of the porous layer.
First, a one-dimensional model of the immersed wafer, i.e. a bulk silicon layer and a porous layer between two water layers, is implemented to calculate theoretical transmission/reflection coefficients. Several samples have been produced: they are divided in two sets with different porosities, each set containing samples with different etched thicknesses in order to determine how ultrasonic parameters are affected. The thickness of the porous layer can be measured by optical microscopy (this being partly destructive) and the density can then be deduced from the change in weight before and after etching.
Longitudinal wave velocity and attenuation coefficients can finally be deduced from water immersion ultrasonic measurements using wide bandwidth transducers with a centre frequency around 15 MHz. These results show that the direct problem can be considered as solved.
To solve the inverse problem, two measurements are performed: one without the sample, which becomes the reference of the transmitted signal, and one with the sample. First, the effective longitudinal ultrasonic wave velocity in the whole wafer is deduced directly from these measurements, then the inverse problem resolution is performed using a genetic algorithm in order to extract the four PoSi layer parameters (i.e. thickness, porosity, longitudinal wave velocity, attenuation coefficient) from ultrasonic measurements only. It is shown that the values of the parameters are close to expected ones and discrepancies as well as uncertainties are discussed.
Future work, which is currently being initiated, will aim at taking into account the shape and size of pores, which has an influence on longitudinal wave velocity and attenuation coefficient. This will require the investigation of specific homogenization models for the porous layer. Then, a specific ultrasonic instrument for in-situ monitoring of the etching process will be developed.