The work presented in this thesis focuses on both standard and advanced characterisation techniques applied to solution-processed CIGS and vacuum-processed CdTe thin film solar cells. Characterisation techniques such as capacitance spectroscopy, tempeture dependent J-V measurements, PDL Hall effect and along with other fundamental measurement techniques are used to extract the key parameters and material properties of these solar cells. Capacitance spectroscopy offers valuable insight to material properties such as defect density, energy level of defects and carrier concentration, which is especially important to understand in the operation of thin film solar cells and their limiting processes.
Antimony (Sb) was introduced into the solution-processing CIGS absorbers which led to an increase in EQE, hence an increase in J_sc. A bilayer structure of large-grain top layer and a fine-grain bottom layer was observed by SEM. The devices with Sb showed a low net carrier concentration and defect density compared to the devices with no Sb doping, indicating that adding Sb might be passivating the defect in the bulk, rather than doping the absorber. A shift in the long wavelength decay of EQE spectra for the device with Sb is observed, indicating a decrease in bandgap and only the change in In/Ga ratio able to explain this behaviour. From the V_oc (T) analysis, the main recombination mechanism for the device with Sb is found to be in the bulk of the absorber, whereas the device without Sb is dominated with the interface recombination. Adding Sb into the absorber layer has reduced the rollover seen at low temperatures in the J-V characteristics, and has removed the barrier at the CdS/CIGS interface. This may be due to the incorporation of Sb into the CIGS absorber, resulting an improved band alignment of CdS/CIGS. The admittance spectroscopy measurements for the device without Sb revealed an admittance step with an activation energy of E_A = 330 meV. This step is considered as a deep level defect, which has often been referred to as the N2 defect. The N2 defect has been removed with Sb doping and left with a shallow level defect, N1 (E_A = 42 meV). After 1 hour of light soaking under 100 mWcm-2 illumination at room temperature, the devices showed a significant increase in the cell performance especially in the junction capacitance and net carrier concentration.
A detailed study of the defects for N1 and N2 steps were performed using capacitance spectroscopy for devices with and without tellurium (Te) at the back-contact, and devices with and without CdSeTe (CST) layer of CdTe devices. The admittance measurements revealed different activation energies, E_A between 20 meV and 279 meV at different temperatures for each device. Adding Te to the CST/CdTe devices removed the N1 step. The N1 step corresponds to a shallow defect with low a E_A, suggesting that the defect might be located at the absorber/back-contact, since Te is added at the back of the device. For the CdTe devices without CST layer, Te has removed the N2 defect and left only the N1 defects. The CST&CdTe with Te devices showed high efficiencies around ~14-15% with high FF. The `CST with Te` device has shown an `S` shaped J-V curve, which can relate to a conduction band offset (CBO) or a presence of a significant current barrier. Non-diode behaviour J-V characteristics have been observed with the `CST/CdTe without Te` and this is due to having a low FF, a large R_s and a low R_sh. The dominant recombination mechanism by the activation energy 〖(E〗_a) is extracted from the intercept of V_oc (T) at T=0 K and this found to be at or close to the interface for all the devices. Adding Te to the CdTe devices has slightly reduced the interface recombination, leading to a smaller barrier height and improved J-V characteristics. Both C-V and DLCP measurements showed typical `U` shaped depth profiles with similar net carrier concentrations, except with the “CST/CdTe without Te”. The difference in depth profiles is due to high level of deep level and shallow defects. Thus, measurements are affected by the N1 and N2 steps.
The metastable behaviour found in CdTe devices with MZO buffer layers resulted in variations from the J-V characteristics depending on prior exposure history of light and environmental stresses, such as temperature and climate. Different preconditioning procedures have been studied that are used to recover the performance of the devices. J-V characteristics before preconditioning have shown an `S` shaped behaviour, with significant current loss in forward bias which is removed after preconditioning. Also, the depth profiles before preconditioning showed an unusual double minima, and the second minima towards the back of the device became less pronounced after preconditioning. The “Atmospheric” preconditioning procedure resulted in a significant recovery of the device performance compared to “Vacuum” preconditioning. Temperature dependent J-V and capacitance measurements before and after preconditioning revealed the presence of recombination centres and defect levels at the MZO/absorber interface. Light and voltage bias have improved the degree of these metastable behaviours by reducing the formation of a blocking layer at the interface. Despite all the preconditioning attempts, the recovery of PV parameters remained only for 3 days while the devices were maintained under vacuum in the dark. Since the temperature dependent capacitance and J-V measurements showed defects and recombination centres at the MZO/absorber interface, the Hall effect measurements have been studied in an attempt to extract carrier concentration, mobility and the conductivity of the MZO films. Conductivity measurements on MZO films showed no significant changes before and after preconditioning. However, CdCl2 treated MZO films showed slight improvement in the linearity of the I-V response. Although the response was improved, no linear sheet resistance or proper Hall signal were detected in order to extract reliable and accurate carrier concentration, or mobility of the MZO films. Alternatively, Ga as a dopant is used in MZO layer (GMZO), which should enhance the film conductivity and the carrier concentration. The annealed GMZO films showed an improved I-V response and a large improvement in conductivity after light soaking. However, a gradual decrease in Hall signal over time was observed and no reliable results could be extracted from the Hall effect measurements.
History
School
Mechanical, Electrical and Manufacturing Engineering