Search

Perovskite Solar Cells (PSCs) being the most advanced photovoltaic technology today have attracted global attention due to its solution processability and cost effectiveness. In this paper, we reported the performance of perovskite solar cells by integrating Reduced Graphene Oxide (rGO) as a buffer layer. The combined effect of Ultraviolet visible (UV-vis) spectroscopy, X-ray Diffractometer (XRD), Raman spectroscopy, Scanning Electron Microscopy (SEM), Four Point Probe and Solar Simulator were used to explore the absorbance, crystallinity, morphological, resistivity and current voltage behavior of the photoanodes and devices. The rGO was deposited on top the mesoporous Titanium dioxide (m-TiO2) (Device B), on methylammonium lead triiodide (CH3NH3PbI3) perovskite absorber (Device C) and on m-TiO2 & CH3NH3PbI3 perovskite absorber (Device D). The reference device (Device A) was fabricated without rGO as a bench mark for comparison. From the results obtained, the devices with rGO show significant improvement over the device lacking rGO. The reference device gave a Short Circuit Current Density (Jsc) of 8.151 mAcm-2, Open Circuit Voltage (Voc) of 0.575 V, Fill Factor (FF) of 74.4 % and Power Conversion Efficiency (PCE) of 3.49 %. The best performing device was the one with rGO incorporated on both m-TiO2 and CH3NH3PbI3. The device gave a Jsc of 8.737 mAcm-2, Voc of 0.836 V, FF of 68.1% and PCE of 4.98 %. The device experienced an enhancement of ~7.20 % and 42.69 % in Jsc and PCE over the pristine device. This investigation aids the fundamental understanding of the effect of Graphene Materials (GRMs) on the optoelectronic properties of PSCs, and it further confirms the promising prospects of achieving low-cost, efficient and stable PSCs for commercialization with graphene materials. [ABSTRACT FROM AUTHOR]

The application of machine learning algorithms to the detection of fraudulent credit card transactions is a challenging problem domain due to the high imbalance in the datasets and confidentiality of financial data. This implies that legitimate transactions make up a high majority of the datasets such that a weak model with 99% accuracy and faulty predictions may still be assessed as high-performing. To build optimal models, four techniques were used in this research to sample the datasets including the baseline train test split method, the class weighted hyperparameter approach, and the undersampling and oversampling techniques. Three machine learning algorithms were implemented for the development of the models including the Random Forest, XGBoost and TensorFlow Deep Neural Network (DNN). Our observation is that the DNN is more efficient than the other 2 algorithms in modelling the under-sampled dataset while overall, the three algorithms had a better performance in the oversampling technique than in the undersampling technique. However, the Random Forest performed better than the other algorithms in the baseline approach. After comparing our results with some existing state-of-the-art works, we achieved an improved performance using real-world datasets.