Solution-processed kesterite solar cells: chemistry of dissolution, thin film deposition and device fabrication

Kesterite (Cu2ZnSn(S,Se)4) solar cells are an alternative thin film technology to CuInxGa(1-x)Se2 (CIGS) or CdTe, which possess the ability to reduce costs due to the constituent elements all being earth abundant. Kesterite solar cells possess an ideal band gap for photovoltaics and are therefore theoretically able to reach efficiencies to rival those of the established commercial technologies. However, the best kesterite efficiencies recorded to date are ~12.6%, a large reduction on the >20% efficiencies observed for CIGS and CdTe photovoltaics. This reduction in efficiency has largely been attributed to the presence of an open circuit voltage (VOC) deficit in kesterite solar cells, as well as difficulties in controlling the stoichiometry of the kesterite crystal structure. On top of these issues, the current best kesterite solar cells are produced with expensive/hazardous chemicals/methods, negating the advantageous aspects of kesterite solar cells, i.e. their relative inexpensiveness and low hazards.

Therefore, in this work the initial aim was to discover a method by which to produce the kesterite absorber layer using a low cost/risk method. Solution processing can be performed with low cost/risk, so was identified as the method to produce kesterite solar cells in this work. The current best solution processed kesterite solar cells used hydrazine,[15] which is undesirable due to the significant hazards it possesses. Therefore, an alternative solvent was required. Amine-thiol solution systems have been shown to dissolve the constituent elements of kesterite solar cells, but also possess significant hazards.[131] Using this knowledge, other, less hazardous, amine-thiol solution systems were identified which could dissolve the necessary elements. The most successful of which was a 10:1 mixture of ethanolamine and cysteamine, which dissolved Cu, Zn, and Sn in their elemental, oxide, and chalcogenide forms. Aqueous solutions of cysteamine and thiourea/urea were also shown to be successful in dissolving the constituent elements.

The solvent systems were analysed, predominantly by mass spectrometry (MS) and Infrared Multiphoton Dissociation (IRMPD), resulting in the identification of the complexes formed in solution by each metal. These results suggested that in most cases cysteamine acted as a chelating ligand, producing tetrahedral and octahedral complexes, as well as polymeric structures in the case of copper. IRMPD was used to further analyse these structures, confirming the exact shape of these structures; including even complex structures such as Zn(HOC2H4NH-C2H2NHC2H4S). Thermogravimetric analysis (TGA) was performed on these solutions to identify the thermal stability of these complexes and to help identify a suitable deposition technique for kesterite fabrication. TGA indicated that the ethanolamine/cysteamine (ETA/CA) solvent system was ideal for dissolving these metals as the majority of it evaporated below 200 °C, which was in contrast to aqueous solutions which struggled to fully evaporate even at around 350 °C. On top of this the analysis indicated that cysteamine was an ideal ligand for this work, as it was able to dissociate from its complexes at or below ~200 °C, whereas complexes involving thiourea took significantly longer to dissociate.

Using the insight from the TGA and the fact that ethanolamine solvent systems are too viscous for spray pyrolysis, spin coating was selected as the deposition method. Fabrication of 1-1.5 µm thick kesterite films was achieved via this method, which then underwent high temperature selenisations to induce kesterite crystallization. The substrate used in this work was optimised, as well as the processes which took place to grow kesterite films upon this substrate. Optimisation of the Mo back contact proved particularly beneficial, as the insertion of an MoNx layer allowed for the reduction in MoSe2 formation during selenisation. Using this optimised method, highly crystalline kesterite solar cells could be consistently produced, with efficiencies of up to 8.1%.

Finally, in order to improve the reproducibility of the kesterite photovoltaic properties, alkali metal doping was trialed. By adding Li, Na, and K in their chloride forms to the solution, the kesterite absorber layer was successfully doped with each element. The successful doping of the kesterite layer was indicated by scanning electron microscopy (SEM), X-Ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), external quantum efficiency (EQE), and capacitance-voltage (CV) analysis. The doping of the kesterite absorber layer individually by each alkali metal led to improvements in crystallinity, carrier concentration, and efficiency. The ideal mol% for each dopant was ~3%, which led to the average efficiency of the kesterite solar cells improving from 4.9% to 6.5% in the case of Na doping.