Cation Substitution-Induced Partial Inversion to Pervade Short-Wave Infrared Light for Improving the Accuracy of Artificial Intelligence Image Recognition System
Yi-Ting Tsai,# Yu-Kai Huang,# Zhen-Feng Jiang, Yuan Yao, Pei-Hsuan Lo, Yu-Chiang Chao, Bi-Hsuan Lin, and Chun Che Lin*
2023/05/25
Short-wave infrared (SWIR) light is suitable for image recognition and biomedical applications due to the ability to perform unique absorption of material components. In this study, the partial inversion of a spinel structure was modified through cation substitution to induce an inverse behavior and charge variation. For MgGa2O4, the substitution of Ga3+ with Sn4+ expanded lattice parameters (a, b, c, and V), and Mg2+ was used to achieve charge balance. When the concentration of Sn increased, the T2g vibrational mode exhibited a significant decline at 638 cm–1, which was ascribed to GaO4, and another retentive T2g vibrational mode at 542 cm–1 was ascribed to MgO4. This demonstrated that MgGa2O4 was a local structure with partial inversion. The clusters of (Ga/MgO4–MgO4) from y = 0.5 to 0.7 revealed a disordered broadband signal in the same Raman shift, and octahedral sites retained their ordering structure. The local structure induced Ni2+/3+ to coexist and transform into pure Ni2+ at octahedral sites through increasing Sn concentration. The time-resolved photoluminescence and emission spectrum revealed high energy transfer efficiency (>90%). The long SWIR emission achieved through Sn substitution enabled the fabrication of an SWIR light-emitting diode device; this device, along with a two-dimensional convolutional neural network, increased the accuracy of an artificial intelligence-based image recognition system from 72.2% to 94.4%. This study promotes research on sequences for inverse tetrahedral sites such as Mg, Al, Ga, In, and other transition metals in spinel structures. In addition, the applicability of artificial intelligence to complete everyday tasks was demonstrated.
Research Highlights
