IEEE PVSC 49
Search
SPLTRAK Abstract Submission
Prediction of Novel Phosphors using Machine Learning for Efficiency Enhancement of Silicon Solar Cells
Tae-Gwan Kim1, Eun-gyeong Kim1, M.Shaheer Akhtar1,2, O-Bong Yang1,3
1Graduate School of Integrated Energy-AI, Jeonbuk National University, Jeonju, South Korea
/2School of Semiconductor and Chemical Engineering, Jeonbuk National University, Jeonju, South Korea
/3New and Renewable Energy Materials Development Center (NewREC), Jeonbuk National University, Jeonbuk, South Korea

In order to achieve the theoretical efficiency over 30%, the mismatch between
solar spectrum and spectral response of Si solar cells is needed to improve. In general, the
thermalization of charge carriers, that usually generated by the absorption of photons having
higher energy greater than a bandgap of semiconductor is the most common loss mechanism
in Si solar cells. To control the thermalization of charge carriers, the introduction of
sensitizer ions (such as Eu2+) into phosphor fluorescence materials can enhance the
absorption of photons with lower energy. In this work, BaMgAl10O17:Eu+2 ,
SrMgAl10O17:Eu+2nano-phosphors were synthesized by solution combustion method. The
synthesized phosphors were extensively analyzed in terms of morphology, crystal structure,
compositional, photoluminescence, and electrical properties by various analytical tools like
X-ray Diffraction (XRD), Field Emission Scanning Electron Microscope-Energy Dispersive
X-ray Spectrometer (FESEM-EDX), PL Spectroscopy, and quantum efficiency analyzer, etc.
The database were extracted from raw data to enable the application of algorithms for
machine learning training. From optimized regression model like Random Forest Regressor,
Gradient Boosting, we can predict appropriate crystal structure and desirable wavelength of
phosphors. The establishment of predicted models is efficient to work, so researchers can
consider to find the potential materials such as ZnMgAl10O17 and CaMgAl10O17 in advance
and save the huge demand for energy resources. It is hope that this work can facilitate the
development of desirable synthesis methods under various conditions for phosphors with
minimal light scattering, and good luminous efficiency which may helpful for improving the
performance of Si solar cells.