Deep learning-assisted smartphone-based molecularly imprinted electrochemiluminescence detection sensing platform: Protable device and visual monitoring furosemide

Abstract

A novel, portable, and smartphone-based molecularly imprinted polymer electrochemiluminescence (MIP-ECL) sensing platform was constructed for sensitive and selective determination of furosemide (FSM). In this platform, MoSe2 nanoparticles/starch-derived biomass carbon (MoSe2/BC) nanocomposites as imprinted material, lucigenin (Luc) as the energy donor, CdS quantum dots (CdS QDs) were used as the luminophore (energy acceptor), and molecularly imprinted polymer (MIP) as the specificity recognition element to construct a MIP-ECL sensing system based on electroluminescence resonance energy transfer (ECL-RET) mechanism, which enhanced the sensitivity and the specificity of this system. Imprinted materials were characterized by SEM, TEM, XRD, FT-IR, etc. and the recognition performance of MIP was characterized using CV, EIS, and ECL methods. The elution and re-sorption of template molecules can be used as a switch to control ECL based on the signal that can be quenched by FSM. Interestingly, deep learning based on convolutional neural networks realizes batch processing of ECL signals. Additionally, this developed MIP-ECL method was established by using the traditional ECL analyzer detector for the assay of FSM with a detection limit of 4 nM in the range of 0.010 μM–100 μM. Besides, the consumer smartphone sensing platform based on deep learning showed an outstanding linear response between the R-value of the picture and the concentration of furosemide in the range of 1–70 μM with a detection limit of 0.25 μΜ, which is much lower than that the reported for other detection methods. More importantly, due to the transferability of deep learning, the smartphone-based MIP-ECL systems can facilitate the real-time monitoring of biochemical analytes in multiple fields.