Perovskite solar cells based on methylammonium lead Iodide (MALI) have achieved efficiencies as high as 26%. This value is higher than the previous form of dye-sensitised solar cells, or cadmium telluride cells, and similar to silicon-based cells. In the last decade a large amount of research has been conducted into these materials. Mostly due to their ease of preparation not only of the photovoltaic material but the solar cell itself. The economic incentives to commercialise these devices are therefore high. This thesis investigated different compositions of materials (variable methylammonium iodide/lead iodide (MAI/PbI2) ratios), synthesis procedures and the influence of environmental factors such as moisture, temperature and oxygen on the electrical properties. Herein three different synthesis routes were investigated. (i) A solvent free solid state route using mortar and pestle; (ii) a mechanosynthesis synthesis route using ethanol, and, (iii) using organic solvents like dimethylformamide (DMF) and dimethylsulfoxide (DMSO). A combination of different starting stoichiometry’s and the different synthesis methods is used to elucidate which samples might be of high quality for fabrication and which synthesis route might be best for large scale applications. The quality of the samples was assessed by a combination of X-ray diffraction (XRD), low voltage scanning electron microscope (SEM), permittivity and dielectric loss (LCR) measurements, and impedance spectroscopy (IS).
This work was sponsored by Dyesol Ltd, a company that was based in Manchester. The plan was for solar cell construction and testing to be conducted at the company laboratories and for more fundamental, scientific based work, which involved working with bulk powders and characterising their structural and electrical properties in the Materials department at Sheffield University. This would allow insights into factors which affect the materials stability and behaviour to emerge, which would be of benefit in the fabrication of MALI-based cells.
The results showed MALI could be synthesised via the three different routes. The solid state method showed MALI to have a low formation energy and that it could be synthesised by mixing the precursors in a mortar and pestle at room temperature. Three different phase assemblage diagrams were constructed for each synthesis method, by varying the precursor amounts of PbI2 and MAI and the temperature. The solid state method shows incomplete reaction at room temperature and samples need to be annealed at 120 ֯C to drive reactions to completion/equilibrium. Heating above 140 ֯C initiates decomposition of the methylammonium organic cation, especially for samples annealed for prolonged periods (24hrs). Due to a combination of low annealing temperature and the ease of loss of powders during mortar and pestle grinding, samples of nominally stoichiometric MALI (x = 0.00) via solid state were not phase pure, and contained PbI2. However, going off-stoichiometry and increasing the MAI content by 5 or 10 mol% (x = 0.05, and 0.10) produced phase pure powders when annealed at 120 ֯C based on XRD results.
Due to the low sensitivity of XRD, low voltage SEM was conducted to verify phase purity. What emerged was regions of grey contrast (associated with MALI perovskite) and dark regions, associated with an organic MAI-rich phase. Therefore, these samples weren't phase pure. Thin film work from literature also indicates that the tetragonal MALI phase which appears grey in the SEM might be a mixture, which contains the stacked perovskite sheets, or α' phase. These can't be distinguished by XRD or SEM but may be responsible for some of the electrical heterogeneity observed by IS.
This trend (clean XRD results but more phases observed via SEM) continued for other samples made by the other two synthesis techniques. Synthesis of ball milled powders was conducted at 80 ֯C. x = 0.00 still had PbI2 present and so did x = 0.05, suggesting a slight shift of the PbI2/MAI ratio in these samples compared to the solid state reaction samples. x = 0.10 and 0.20 were clean by XRD. This was attributed to the easy loss of MAI during the reaction method. The organic solvent synthesis using DMF, showed x = 0.00-0.10 to be XRD clean when heated to 120 ֯C. The x = 0.00 sample via this route produced the cleanest SEM surface image, suggesting that it was the best quality sample prepared via the three routes.
The electrical results from both LCR and IS showed samples always had some hysteresis present between the first heating and cooling cycle. This is attributed to the easy intake of water by MALI. IS results always showed higher conductivity on the heating cycle than on the cooling cycle. The cooling cycle results are therefore best used to establish the intrinsic properties of the materials. IS data showed that samples with a small amount of excess MAI (generally x = 0.05 and 0.10) gave higher conductivity, even on the cooling cycle. This was not dependent on any specific synthesis route. These two phenomena were also observed from the LCR data. The permittivity and dielectric losses were higher on the heating cycle than on the cooling cycle with the more conductive samples having higher losses. The presence of moisture allowed the tetragonal to cubic MALI transition temperature to be elucidated at ~60 ֯C. This was because LCR gathered more data points and experiments were conducted in less time. LCR proved to be a quicker and easier method than IS to examine the electrical stability and behaviour of the samples.
