Resonant Raman scattering based approaches for the quantitative assessment of nanometric ZnMgO layers in high efficiency chalcogenide solar cells

Abstract

This work reports a detailed resonant Raman scattering analysis of ZnMgO solid solution nanometric layers that are being developed for high efficiency chalcogenide solar cells. This includes layers with thicknesses below 100 nm and compositions corresponding to Zn/(Zn + Mg) content rations in the range between 0% and 30%. The vibrational characterization of the layers grown with different compositions and thicknesses has allowed deepening in the knowledge of the sensitivity of the different Raman spectral features on the characteristics of the layers, corroborating the viability of resonant Raman scattering based techniques for their non-destructive quantitative assessment. This has included a deeper analysis of different experimental approaches for the quantitative assessment of the layer thickness, based on (a) the analysis of the intensity of the ZnMgO main Raman peak; (b) the evaluation of the changes of the intensity of the main Raman peak from the subjacent layer located below the ZnMgO one; and (c) the study of the changes in the relative intensity of the first to second/third order ZnMgO peaks. In all these cases, the implications related to the presence of quantum confinement effects in the nanocrystalline layers grown with different thicknesses have been discussed and evaluated.

Figure 1

Schematic representation of a high efficiency Cd-free chalcogenide solar cell, including a back Mo contact deposited on the substrate (typically SLG), the CIGS absorber layer, a nanometric Zn(O,S) buffer layer, a ZnO based window layer and a nanometric ZnMgO interficial layer between the buffer and window ones. For high efficiency pure sulfide CIGS devices, the ZnMgO layer replaces the Zn(O,S) buffer layer.

Figure 2

Raman scattering spectra (normalized to the intensity of the main LO peak) of ZnMgO layers grown with a relative Mg/(Mg + Zn) content of 10% and a thickness of 15 nm on the different kinds of substrates.

Figure 3

Raman scattering spectra (normalized to the intensity of the main LO peak) measured from layers series (i) that were grown with different Mg concentration.

Figure 4

Mg/(Mg + Zn) relative content versus frequency of the main LO Raman peak. Starts: experimental data from layers series (i). Circles: data reported from references^, . Full line shows the fitting of the data according to Equation 1.

Figure 5

Top: Raman spectra measured from series (ii) layers with Mg/(Mg + Zn) relative content 10% and different layer thicknesses. Bottom: Integral intensity of the main LO peak (highlighted in red in the Raman spectra plotted in Fig. 4 top) versus layer thickness for layers series (ii) (Mg/(Mg + Zn) relative content 10%) and layers series (iii) (Mg/(Mg + Zn) relative content 26%). Dotted lines are added as guide for the eye.

Figure 6

Raman spectra (normalised to the intensity of the main LO peak) measured from series (ii–a) and (iii–b) layers. (c) Relative intensity of the ZnS-like peak from the Zn(O,S) subjacent layer in relation to the main LO ZnMgO Raman peak (corresponding to the peaks highlighted in red in the Raman spectra) versus the layer thickness for layers series (ii) (Mg/(Mg + Zn) relative content 10%) and layers series (iii) (Mg/(Mg + Zn) relative content 26%). Dotted lines correspond to the fitting of the data with linear relationships in the log-log scale.

Figure 7

Top: Raman spectra (normalized to the intensity of the LO peak) from layers series (ii) with Mg/(Mg + Zn) = 0.1 and different thickness. Bottom: Relative intensity of the first order LO peak in relation to that of the second order and third order peaks versus layer thickness.