Convolutional Neural Networks Driven Synergistical Temperature Sensing and Chemical Concentration Recognition by Rationally Designed NIR‐II‐Emitting Composite Films
Qian Zhu, Huilin Chen, Ruichan Lv, Yuzhen Wang, Zhiguo XiaABSTRACT
Near‐infrared II (NIR‐II) emissions are unique and selective for non‐destructive detection owing to their specific response to temperature changes and the ability to cover the characteristic overtone absorption bands of molecular functional groups. However, cross‐interference in the spectral responses occurs for simultaneous temperature sensing and chemical reagent identification, therefore it is necessary to realize reliable decoding. Herein, the Nd 3+ / Er 3+ co‐doped Y 2 MgTiO 6 phosphors with NIR‐II emissions under 808 nm excitation have been designed and prepared, and the temperature sensing properties have been evaluated, in which the distinctive thermally enhanced luminescence is attributed to the efficient energy transfer process from Nd 3+ to Er 3+ ions. Besides, NIR‐II‐emitting composite film based on the rationally designed phosphors has been fabricated as the sensing platform, and the concentration recognition of water and methanol is enabled by the overtone absorption of C─H and O─H bonds. Finally, a Large Scale Convolutional Neural Network model is constructed to synchronously decouple, which can accurately extract temperature and chemical concentration from complex spectral data. This work paves the way for constructing a feasible and compact platform for dual‐function intelligent sensors.