Deep learning-assisted double-pulse laser-induced breakdown spectroscopy and AuAg nanocluster-modified nanofiber membranes for detection of Cs, Sr, and Co in soil

AuAgNCs-NFM is used for the efficient enrichment of HMs in soil. • The detection sensitivity and accuracy for Cs, Sr, and Co are enhanced. • The LODs for Cs, Sr, and Co reach 0.078, 0.085, and 0.16 mg/kg, respectively. • The 1D-CNN model is developed to improve predictive performance. Rapid, sensitive, and accurate detection of heavy metals (HMs) in soil is of importance for ensuring food security and advancing smart agriculture. However, conventional detection methods face numerous limitations, such as complicated operations and time-consuming procedures. This study proposed a novel method using AuAg nanoclusters (AuAgNCs)-modified nanofibrous membranes combined with double-pulse LIBS (DP-LIBS) for the sensitive detection of available cesium (Cs), strontium (Sr), and cobalt (Co) in soil. The AuAgNCs-modified nanofiber membranes were utilized to efficiently adsorb Cs + , Sr 2+ , and Co 2+ from soil solution, with simultaneous DP-LIBS detection. The results revealed that the limits of detection (LODs) for Cs, Sr, and Co in soil reached 0.078 mg/kg, 0.085 mg/kg, and 0.16 mg/kg, respectively. Furthermore, a one-dimensional convolutional neural network (1D-CNN) model was developed for accurate quantitative analysis and exhibited excellent predictive performance. The residual predictive deviations (RPDs) for Cs, Sr, and Co in soil were 6.0121, 5.1844, and 6.7082, respectively. This method holds great application potential in HMs monitoring, soil risk assessment, and human health protection. Download: Download high-res image (322KB) Download: Download full-size image