Development of in situ methods for process monitoring and control and characterization of Cu-Zn-Sn-S based thin films

In recent years, kesterite Cu2ZnSnS4 (CZTS) has become an interesting alternative to copper indium gallium (di)selenide (CIGS) due to its non-toxic and earth abundant constituents. A variety of methods is being used to fabricate kesterite thin films, such as coevaporation, sputtering, electrodeposition, spray pyrolysis and others. Most of them include an annealing step to stimulate elemental mixing and interdiffusion. Although conversion efficiencies of kesterite solar cells have increased among different research groups, the record value of 12.6% set by IBM in 2014 has not been broken yet. Therefore, experimental and theoretical studies are needed to predict the effect of the secondary phases and detrimental defects on the electronical properties of the CZTS based solar devices.
The work presented here studies non-destructive techniques for in situ process control and monitoring. With the aim to detect phases and phase transitions to optimize crucial processing steps such as pre-annealing of metal precursors, high temperature annealing and vacuum deposition of Cu-Sn-Zn-S based thin films. The research consists of three parts in which Raman spectroscopy, X-ray diffraction (XRD) and reflectometry are used to explore this objective.
In the first part Raman spectroscopy is investigated as an in situ monitoring technique during high temperature annealing of thin films. It investigates whether the occurrence of CZTS can be monitored when it is created from annealing a Mo/CTS/ZnS layered thin film. CuS, SnS, ZnS and CTS (Cu-Sn-S) films are prepared by physical vapor deposition. The Raman scattering intensity was compared to investigate whether their specific vibrational modes can be distinguished from each other at room temperature. Then, the CTS film is annealed between 50 and 550 °C in order to investigate whether CTS vibrational modes can be identified at elevated temperatures and to see which transitions take place within the thin film. Also, a CZTS reference film is annealed between 50 and 550 °C for reference purposes. The temperature dependence of the main CZTS modes is examined to investigate whether it can be used for in situ temperature control. Finally, a ZnS layer is deposited on the unannealed CTS film to obtain a Mo/CTS/ZnS layered film. This film is used to study the conversion of CTS/ZnS into CZTS at elevated temperatures. It was found that Raman spectroscopy can successfully be used to monitor formation of CZTS by identifying its main vibrational mode during the annealing process. The intensity of the CTS modes reduces at elevated temperatures. At 450 °C, the main CZTS mode at 338 cm-1 can be clearly identified.
The second part also focuses on high temperature annealing. However, in this part the focus lies on annealing of the metal precursor films. It is explored whether specific alloys benefit or hinder the formation of secondary phases during formation of the CZTS absorber films. Also, to what extent this influences solar cell performance. In situ XRD was investigated for in situ monitoring of the pre-annealing process. Cu-poor metal precursor films are prepared by sputtering deposition. The precursors are annealed at 150 °C, 200 °C, 300 °C and 450 °C in a three zone tube furnace. The effect on the structural properties is analysed by XRD to study the formation mechanism of alloys. The precursor films are then sulfurized in a three zone tube furnace. The structural properties of the absorber are analysed and correlated with structures in the precursor. It is found that formation of SnS2 in the absorber is proportional to the remaining Sn in the pre-annealed precursor. Also, electron micrographs showed that pre-annealing temperature influences grain growth and surface precipitation of Sn-S and Zn-S. Pre-annealed absorbers at 450 °C did not exhibit these phases on the surface. Solar devices are fabricated from the absorber films and best performing devices were obtained from pre-annealed absorbers at 450 °C. They showed absence of Sn and SnS2 in, respectively, the precursor and absorber. It could be concluded that SnS2 phases are detrimental to device efficiency and that SnS2 XRD peak intensity follows an inverse proportionality with device efficiency.
The third part explores reflectometry as a method to monitor a growing film during thermal evaporation in a physical vapor deposition (PVD) system. A set of six CZTS absorbers is examined by ex situ Raman spectroscopy and reflectometry to study the influence of secondary phases CuS and ZnS on reflection spectra. Composition strongly influences reflection spectra and CuS leaves a characteristic dip in the reflection spectrum at about 600 nm. An integration method was used to analyze this phenomenon quantitatively. Subsequently, a reflectometry setup is designed, developed and integrated in the PVD system. Four different CZTS co-evaporated and multi-layered films are deposited. Structural, morphological and vibrational properties are investigated. The reflection spectra are monitored during deposition and time-dependent reflection spectra are analyzed for characteristic aspects related to properties such as thickness, band gap and phase formation. CuS could not be detected in the films by the integration method due to the superposition of the CuS dip with developing interference fringes during film growth. However, in multilayered CTS/ZnS film it is found that the onset of ZnS deposition can be detected by increased reflection intensity due to reduced surface roughness. Additionally, the shifting onset of the interference fringes to lower photon energies can be used as a characteristic fingerprint during the deposition process.
In conclusion, this work showed that Raman spectroscopy, XRD and reflectometry could be successfully implemented for in situ process control and monitoring of high temperature annealing and vacuum deposition of Cu-Sn-Zn-S based precursors and absorbers. The application of these in situ techniques can lead to the optimization of thin film material properties and solar cells. As such, this study has paved the way for further improvement of Cu-Sn-Zn-S based precursors and thin film absorbers.