Discrimination and detection limits of secondary phases in Cu2ZnSnS4 using X-ray diffraction and Raman spectroscopy

• XRD cannot be used to identify < 10% ZnS or < 50% Cu2SnS3 in a mixed Cu2ZnSnS4 sample.
• Green light Raman spectroscopy can't discern < 30% Cu2SnS3 in a Cu2ZnSnS4 sample.
• Green light Raman spectroscopy can't discern ZnS in a mixed Cu2ZnSnS4 sample at all.
• Multiple-wavelength Raman spectroscopy is useful to discern ZnS from Cu2ZnSnS4.
• Cu2ZnSn3S8 could not be confirmed in a thin film made by a typical growth process.

The formation of single-phase Cu2ZnSnS4 thin films is known to be challenging, mainly due to the difficulties to detect secondary phases in the Cu2ZnSnS4 system. Here, the ability to quantitatively discriminate the most likely secondary phases ZnS and Cu2SnS3 from Cu2ZnSnS4 using common approaches but also using more complex and time-consuming Rietveld refinement analysis techniques to analyse X-ray diffractograms is investigated in a comparative study to the peak analysis of Raman spectra measured with standard conditions. In studying not only individual samples of the respective phases but also a phase-gradient sample containing various amounts of Cu2SnS3 and ZnS alongside Cu2ZnSnS4, we found that refinement analyses can only discriminate more than 10% ZnS and 50% Cu2SnS3 from Cu2ZnSnS4, respectively. In comparison, Raman measurements performed with green-wavelength excitation can discern more than 30% Cu2SnS3 from Cu2ZnSnS4 while ZnS is indiscernible. The results show that the identification of secondary phases in the Cu2ZnSnS4 system is more difficult than currently assumed in literature. Furthermore, the potential of multiple-wavelength Raman spectroscopy as a tool to identify ZnS secondary phases is shown. Characterization of a Sn-rich sample (composition nearly Cu2ZnSn3S8) shows no sign of a Sn-rich quaternary phase, questioning its existence under typical annealing conditions.