Machine learning-assisted nitrite detection on smartphone-integrated μPAD using lychee-like perovskite nanocomposites.

A reliable and convenient nitrite detection method is crucial for environmental monitoring and food safety. This study presents a novel dual-response nitrite detection platform utilizing monodisperse, water-stable, and lychee-like CsPbCl: Mn@MSN@MnO (ClMnM@MnO) nanocomposites. The in-situ growth of perovskite nanocrystals within mesoporous silica nano-templates (MSN) significantly enhances the dispersion and stability of the nanocomposites in water. Mn doping introduces a second emission center, enabling potential post-synthetic designability. The MnO shell, synthesized through NaClO-mediated oxidation, exhibits excellent oxidase (OXD)-like activity. By integrating the nanocomposites with a microfluidic paper-based device (μPAD), we developed a portable system for quantitative nitrite detection by analyzing smartphone-captured images under both room and UV light. With the aid of machine learning (ML) through the ANN algorithm, this platform achieves nitrite concentration prediction and antioxidant additive discrimination, thereby expanding perovskite applications in biosensing and offering innovative solutions for nitrite detection in complex matrices.