Machine vision and deep learning dual-driven intelligent sensing platform for Zn ion and EDTA monitoring in soil filtrate, plant sap and tap water using ratiometric-fluorescence carbon dots-based paper chip

Highly-intelligent user-friendly sensing strategies are becoming desperately demanded and indispensable in under-developed nations and resources-insufficient areas. Herein, an intelligent sensing platform has been successfully established incorporating smartphone with machine vision and deep learning function, 3D-printed handheld device, and paper microfluidic chip stained with tri-emission ratiometric-fluorescence (FL) carbon dots (TCDs). The intelligent system can quantify zinc ion (Zn 2+ ) and ethylenediaminetetraacetic acid (EDTA) only through a single-step operation on smartphone. TCDs, engendering bright blue and cyan FL as well as weak red one at 370 nm excitation, were prepared hydrothermally using tartaric acid and alizarin carmine. As-prepared TCDs exhibit continuous FL color transformation from blue to purple to blue with the addition of Zn 2+ and EDTA. Subsequently, TCDs-based paper chip is explored to enable field-portable sequential discrimination of Zn 2+ and EDTA. Intriguingly, the proposed sensor is utilized for fast and intelligent quantification of Zn 2+ and EDTA in soil filtrate, plant sap and tap water with a low detection limit of 3.8 μM and 7.9 μM, as well as wide linear correlations of 0–280 µM and 0–560 µM, respectively. It is eagerly anticipated that our methodologies will become the strong impetus for new round of technological transformation for analyte assay.