Detection of pesticide residues using flower-like silver SERS substrates based on flexible sponge

In response to the existing issues of cumbersome and time-consuming detection processes and limited application scope in current pesticide residue detection, this paper designed a novel flexible substrate for surface-enhanced Raman spectroscopy (SERS) by combining flower-like silver nanoparticles prepared by chemical reduction technology with a flexible sponge. The flexible substrate exhibits excellent SERS enhancement effects, with a minimum detection limit of 10-12 mol/L for the probe molecule rhodamine 6G (R6G) and an average enhancement factor of 6.63 x 105. For the commonly used pesticide thiram, the minimum detection limit is 0.1 mg/L, which is significantly lower than the maximum residue limits set by China and the USA for thiram. Further experiments confirmed the substrate's excellent uniformity and stability, and the use of finite difference time domain (FDTD) software revealed that the model combining flower-like silver nanoparticles with a sponge exhibited higher electromagnetic field intensity compared to the model without the sponge, resulting in abundant "hot spots". Additionally, the sparrow search algorithm (SSA) was used to optimize the backpropagation (BP) neural network for predicting the concentration of thiram pesticide. The experimental results indicated that the SSA-BP algorithm achieved a determination coefficient (R2) of 0.99974 and root mean square error (RMSE) of 300.321, demonstrating good network performance and meeting the requirements of actual detection needs.