Neural networks based fluorescence and electrochemistry dual-modal sensor for sensitive and precise detection of cadmium and lead simultaneously

Heavy metals are harmful and it’s meaningful to achieve co-detection. In this work, fluorescence (FL) and electrochemistry (EC) dual-modal sensors combined with neural networks are proposed to detect cadmium (Cd2+) and lead (Pb2+) without pretreatment for the first time. Dual-modal sensing eliminates individual limitations of FL and EC and combines their superiority. Quantum dots and sea urchin-like FeOOH are used as sensitive materials, among which FeOOH is used for the first time to detect Pb2+ with high repeatability and sensitivity. Combining with the proposed neural networks, the mean absolute error of Cd2+ and Pb2+ predicted are 0.2176 μg/L and 0.6002 μg/L, respectively, which are far better than traditional analysis methods. The R-Squared between the predicted value and the true value is 0.974 (Cd2+) and 0.999 (Pb2+), respectively, which verifies the feasibility of the designed sensor. This model eliminates the mutual interference between Cd2+ and Pb2+ based on the synergistic effect and can be used for low-level detection in water samples with complex background. In addition, the designed model could combine with other types of sensors to accurately monitor global-local waters. It also provides new ideas for data fusion, which expands the flexibility in environmental protection and health care.