| AVS 61st International Symposium & Exhibition | |
| Spectroscopic Ellipsometry Focus Topic | Friday Sessions |
| Session EL+AS+BI+EM+SS-FrM |
| Session: | Application of SE for the Characterization of Organic and Biological Materials |
| Presenter: | Jian Li, University of Toledo |
| Authors: | J. Li, University of Toledo L. Mansfield, National Renewable Energy Laboratory P. Pradhan, University of Toledo H. Du, National Renewable Energy Laboratory S. Glenn, National Renewable Energy Laboratory J. Mann, National Renewable Energy Laboratory A. Norman, National Renewable Energy Laboratory K. Ramanathan, National Renewable Energy Laboratory R.W. Collins, University of Toledo G. Teeter, National Renewable Energy Laboratory D. Levi, National Renewable Energy Laboratory |
| Correspondent: | Click to Email |
Spectroscopic ellipsometry (SE) is a powerful tool for studying thin films, including the thickness and dielectric function, the latter being closely related to important properties such as composition, phase, grain size, porosity, and stress. The sub-nanometer sensitivity of SE is best exploited if all interfaces between layers, at substrate/layer and layer/ambient are abrupt and smooth. Even for the simple structure of substrate/film/ambient, however, whereby the film is fabricated in a uniform process, surface non-idealities including roughness, oxides, compositional variations, or a combination of these, are inevitable. If an accurate film dielectric function is of interest, then the widely-used effective medium approximation (EMA) treatment of the surface roughness can distort the result, especially in photon energy range of strong absorption.
In this work, an improved SE methodology has been developed, tested, and applied to study thin films with significant surface non-idealities. The investigated materials include Cu(InGa)Se2, Zn(O,S), Cu2ZnSnS4, and Cu2SnS3 deposited on transparent substrates by co-evaporation, sputtering, or chemical bath deposition. The film thicknesses in this study range from ~ 20 to 4000 nm, with potential applicability of the methodology over an even wider range. The key component of the SE methodology is integration of through-the-substrate (TS) SE with standard film-side (FS) SE. The following successes have been demonstrated.
(1) When the surface non-ideality is predominantly roughness within the EMA applicability, two-side (FS+TS) SE can minimize dielectric function distortion caused by the EMA assumptions.
(2) When the surface non-ideality is outside the EMA applicability and traditional SE methodology becomes unreliable, accurate results can be obtained using the FS+TS SE methodology, in which the dielectric functions of the surface and bulk layers can be allowed to vary wavelength by wavelength independently. Most thin films of this study fall into this category.
(3) When the surface is macroscopically rough and scatters light, films can be grown intentionally thick and hence rough enough to suppress specular reflection from the surface. In this case, through-the-substrate SE alone can be used to extract the bulk film dielectric function.
An important criterion for evaluating SE analysis on semiconductor films is that the ε2 spectrum should be flat and essentially zero below the band gap. It is demonstrated that the dielectric functions obtained through the above SE methodology either satisfy or better satisfy this criterion compared to previous studies. The limitations of the SE methodology will also be discussed.