Two-Dimensional LF-NMR Spectroscopy for Graphene Oxide-Based Magnetic Relaxation Switching (GO-MRS) Biosensing: Feature Analysis and Predictive Modeling for Acetamiprid Detection

The feasibility of using two-dimensional low-field nuclear magnetic resonance (2D LF-NMR) for detecting acetamiprid (ACE) residues was investigated based on a graphene oxide-based magnetic relaxation switching (GO-MRS) biosensor. Feature extraction and original spectral data analysis strategies were proposed for <i>T</i>₁-<i>T</i>₂ spectral signal analysis. Then Pearson correlation, principal component regression, and partial least-squares regression were employed to develop the prediction model for ACE residues. The model based on the feature extraction strategy exhibited greater efficacy, and the PLSR detection model demonstrated superior performance, with an <i>R</i>² of 0.9878, an RMSE of 1.8010, and a recovery range of 97.73% to 101.39%, with a relative standard deviation of less than 2.63%. The method also exhibited satisfactory specificity and sensitivity, with the LOD and LOQ values of 0.15 and 0.46 nM, respectively. This study also provides a reference for the application of 2D LF-NMR in the detection of insecticides by using MRS biosensors.