Machine learning-assisted N-doped carbon dots for sensitive Pd2+ detection and high-performance LED applications in plant cultivation
摘要
A facile one-pot hydrothermal synthesis of strong yellow-emitting nitrogen-doped carbon dots (N-CDs) from o-phenylenediamine and 4-aminophenol for rapid detection of palladium ions (Pd2+) is reported. By leveraging the unique response of N-CDs to Pd2+, a ‘‘turn off’’ mode is developed to detect Pd2+ via fluorescence quenching of N-CDs, achieving a notably low detection limit (LOD) of 0.058 μM. Furthermore, a smartphone-incorporated sensing platform integrated with the deep learning algorithm allows for accurate and straightforward field-deployable quantification of Pd2+. Convolutional neural network (CNN) is trained on smartphone-captured images and used to predict unknown Pd2+ concentrations through color feature analysis. Integration with CNN enables an application to identify Pd2+ from archived sample images, achieving recoveries of 95.3–103.5% and RSD < 1.51%. Besides, due to the superior optoelectronic properties of N-CDs, warm white light-emitting diode (W-LED) boasting Commission International d’Eclairage (CIE) coordinates of (0.47, 0.43) is manufactured with the color temperature of 3353 K. The emission spectra of our LEDs, primarily within the blue and yellow regions, overlap closely with the light spectra primarily required by plants, thereby contributing to enhanced plant growth. Overall, our work will broaden the versatile applications of N-CDs in optoelectronic devices and intelligent environment monitoring.
Graphical abstract