White light emission from thin-film samples of ZnO nanocrystal, Eu 3+ and Tb3+ ions embedded in an SiO2 matrix

• Main Author: Tay, Wayne
• Other Authors: Kantisara Pita
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering
• Online Access: https://hdl.handle.net/10356/158693
• Institution: Nanyang Technological University

This work is an extension of the previous study done in 2019 [8] and it presents a new method to obtain any colour light from thin film samples of Zinc Oxide Nano-crystals (ZnO-nc) co-dope with two different rare-earth(RE) ions, Europium (III) (Eu3+ ) and Terbium (III) (Tb3+) embedded in silicon dioxide (SiO2) matrix. In that 2019 study, a new method to obtain white light emission from Zinc Oxide Nano-crystals (ZnO-nc) co-dope with two different rare-earth(RE) ions, Europium (III) (Eu3+ ) and Terbium (III) (Tb3+) in a silicon dioxide (SiO2) substrate was developed. An empirical 4th-degree polynomial equation was used to determine the concentration of Eu3+ and Tb3+ that produces white light or any desired colour in the International commission of illumination (CIE) colour space. However, the method is tedious and in order to predict the concentration of the RE ions for a particular colour desired, trial and error is required. Hence, this work presents a more efficient method to predict the concentration of the RE ions required to be co-doped to produce a sample with light emission of desired colour without the need for trial and error. This method had two parts, the first was to use the 24 samples and their corresponding International commission on illumination(CIE) x, y and z chromaticity values to develop two different 4th-degree polynomial to create more simulated samples and subsequently neural networks was used to predict the concentration of Eu3+ and Tb3+ based on the colour desired, namely white light emitting sample and orange light emitting sample for this work. Due to insufficient samples/data, the simulated white light emitting and orange light emitting samples generated by the polynomial equations shows some variation to the actual samples. However, the neural networks model was able to learn and predict the concentration of Eu3+ and Tb3+ of the simulated white light emitting and orange light emitting samples fairly accurately. Therefore, with more actual fabricated samples in future works, neural networks would be an efficient and effective tool for the fabrication of sample of any desired colour.