Image processing for the characterization of porous silicon nanostructure
Résumé
This paper presents a new method based on image processing (IP) for the characterization of porous silicon (PS) nanostructures. It was developed using porous silicon layers (PS1 and PS2) having different nanostructures. According to gravimetric measurements, these layers present the same porosity of 45 %. First we characterized the PS layers by Barrett-Joyner-Halenda (BJH) theory applied to sorption data. Then applying BJH theory, we found that the mean pore diameter of PS1 and PS2 are 4.3 and 5.5 nm respectively. With Brunauer-Emmet-Teller (BET) theory, we found that PS1 has a specific area of 330 m²/g, and PS2 has a specific area of 223 m²/g. We obtained images of the PS layer surface by high resolution scanning electronic microscopy (HRSEM). The processing of these images leads to the estimation of mean pore diameter: 5.6 nm and 7.5 nm for PS1 and PS2 respectively, and to the representation of pore size distribution. According to a geometrical model, we estimated the porosity and found 55% and 48% for PS1 and PS2 respectively. The calculation of specific area according to the same model leads to the values of 343 m²/g and 217 m²/g for PS1 and PS2 respectively. The close agreement of results obtained by IP to those obtained by sorption theories shows the validity of our method. Advantages of images processing are discussed.