Study of polycrystalline Cu₂ZnSnS₄ films by Raman scattering

Abstract

Cu₂ZnSnS₄ (CZTS) is a p-type semiconductor that has been seen as a possible low-cost replacement for Cu(In,Ga)Se₂ in thin film solar cells. So far compound has presented difficulties in its growth, mainly, because of the formation of secondary phases like ZnS, Cu_xSnS_x+1, Sn_xS_y, Cu_2−xS and MoS₂. X-ray diffraction analysis (XRD), which is mostly used for phase identification cannot resolve some of these phases from the kesterite/stannite CZTS and thus the use of a complementary technique is needed. Raman scattering analysis can help distinguishing these phases not only laterally but also in depth. Knowing the absorption coefficient and using different excitation wavelengths in Raman scattering analysis, one is capable of profiling the different phases present in multi-phase CZTS thin films.

This work describes in a concise form the methods used to grow chalcogenide compounds, such as, CZTS, Cu_xSnS_x+1, Sn_xS_y and cubic ZnS based on the sulphurization of stacked metallic precursors. The results of the films’ characterization by XRD, electron backscatter diffraction and scanning electron microscopy/energy dispersive spectroscopy techniques are presented for the CZTS phase. The limitation of XRD to identify some of the possible phases that can remain after the sulphurization process are investigated. The results of the Raman analysis of the phases formed in this growth method and the advantage of using this technique in identifying them are presented. Using different excitation wavelengths it is also analysed the CZTS film in depth showing that this technique can be used as non destructive methods to detect secondary phases.

Introduction

Cu₂ZnSnS₄ (CZTS) with the kesterite/stannite structure is a p-type semiconductor with an absorption coefficient higher than 10⁻⁴ cm⁻¹ and a band gap energy close to 1.5 eV [1], [2]. Because of these characteristics and the fact that it uses no expensive or toxic elements like In, Ga or Se it is seen as a future replacement for Cu(In,Ga)Se₂ (CIGS) in thin film solar cells. However, CIGS laboratory solar cells achieved efficiencies up to 20.3 % [3], while for CZTS solar cells the maximum reported efficiency is 6.8% [4]. Before exploring the reasons for this difference, first it is necessary to understand the problem of phase identification in CZTS. X-ray diffraction (XRD) has been widely used as the main tool to analyse the structure of CZTS thin films. In this paper, we are going to show that XRD alone is not sufficient to resolve the different phases that may be present and that Raman scattering is a useful complementary technique not only to identify secondary phases but also to localize them spatially.

The Raman scattering technique was already used for the study of chalcopyrite compounds such as CuInS₂[5]. The in depth study of the absorber layer of solar cells using Raman scattering has been tested using three different approaches:

• Takei et al. [6] performed a cross sectional scanning of the absorber layer and studied the spectra;

• Calvo-Barrio et al. [7] changed the depth of analysis focusing the measurement point deeper in the absorber layer;

• the most common approach applies successively sputtering runs to expose the surface of the deeper layers and analyses each exposed surfaces. Álvarez-García et al. [8] used this methods combined with Auger electron spectroscopy to study the quality of CuInS₂ polycrystalline films.

The need of cutting the samples to look at the cross section and the use of sputtering to erode the surface make both techniques destructive. The main disadvantage of the second method refers to the time consumption needed to acquire a workable Raman signal intensity. Note that thin film absorber layers are characterized by a large absorption coefficient. The in depth method that we propose in this work combines several wavelengths with different focus depths. The use of longer wavelengths allows deeper focus on the sample without loss of Raman signal intensity. This is due to an inverse relationship between wavelength and absorption coefficient. Another feature refers to the possibility of using excitation wavelength with an energy close to the band gap of the material. This means that the measurement conditions are in a quasi-resonant mode which significantly increases the intensity of the Raman signal [9].

The first part of this work, Section 2, compiles the information of CZTS and related chalcogenide phases XRD analysis. A brief description of the structural properties of this compounds is presented. The main part of this section is devoted to explaining the limitations of this technique in terms of phase identification.

A brief description of the details used to grow the various phases are presented in Section 3. The experimental procedure described in the section is based in previously published work [2], [10], [11], [12], [13], except for the growth of the cubic-ZnS phase.

The Section 4.1 presents preliminary results of the characterization of a CZTS film based on structural and phase identification using the XRD and electron Backscatter diffraction (EBSD) techniques. The composition of the CZTS film is also analysed using energy dispersive spectroscopy (EDS) and inductively coupled plasma mass spectrometry (ICP-MS). In Section 4.2 it is presented the Raman scattering analysis’ results for the various binary and ternary phases that may be present in a CZTS film. These results are very important for the ensuing discussion. In the last part, Section 4.3, a detailed study of the CZTS depth Raman scattering analysis employing several excitation wavelengths is shown. Additional information can be obtained when compared with the results obtained by XRD and EBSD in Section 4.1.