Reflectance anisotropy spectroscopy (RAS) for in-situ identification of roughness morphologies evolving during reactive ion etching (RIE)

When using reactive ion etching (RIE) as a semiconductor processing tool, techniques to recognize upcoming roughness and to identify the roughness morphologies at the etch front in-situ and in real-time can significantly increase yield and thus productivity of the technological processes. Here, reflectance anisotropy spectroscopy (RAS) is applied in combination with sophisticated statistical data analyses to recognize or even define the possible roughness morphologies on RIE-etched monocrystalline substrates in real time. We exemplarily use GaSb, but our approach will also work for other III/V semiconductors and even silicon, as long as the material is monocrystalline. In the first step of the evaluation of RAS spectra, principal component analysis (PCA) and linear discriminant analysis (LDA) are performed. The algorithms are capable of calculating eigenvectors that produce the maximum variance from the spectroscopic data, thus allowing for a reduction of dimensionality of the dataset and for a distinction of different roughness morphologies. In a second step, consulting a base of earlier RAS data from former own samples, such results can be employed to identify the current surface morphology based on its spectra in-situ and in real-time.