Large-Area Compositional Mapping of Cu(In$_{1-x}$Ga $_{x}$)Se $_{2}$ Materials and Devices with Spectroscopic Ellipsometry

Puruswottam Aryal, Dinesh Attygalle, Puja Pradhan, Nikolas J. Podraza, Sylvain Marsillac, R. W. Collins

IEEE Journal of Photovoltaics · 2012 · 35 citations · 18 references

Concepts

Abstract

A spectroscopic ellipsometry (SE) capability having the potential to scan production-scale areas at high speed has been developed and successfully applied to map the alloy composition of copper-indium-gallium-diselenide (CuIn <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1-x</sub> Ga <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</sub> Se <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> : CIGS) thin films. This technique not only generates a compositional map but simultaneously provides maps of the more typical SE-determined properties as well, including bulk layer and surface roughness layer thicknesses. As a result, the methodology is suitable for characterization in online production-scale applications. In order to develop the mapping capability, CIGS films having different molar Ga contents x and fixed copper stoichiometry were deposited and measured in situ by SE in order to extract the complex dielectric functions (ε = ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> +iε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ) of these films. For mathematical interpolation between the available alloy contents, the (ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> , ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ) spectra were parameterized using an oscillator sum. Best-fitting equations were obtained that relate each oscillator parameter to the Ga content x, as determined by energy dispersive X-ray analysis. This approach reduces the number of fitting parameters for (ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> , ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ) from several to just one: the Ga content x. Because (ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> , ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ) is now represented by this single parameter, the chances of parameter correlations during fitting are reduced, enabling production-scale compositional mapping of chalcopyrite films by SE.

References

18