鉴定(生物学)
分析物
计算机科学
判别式
工作流程
质谱法
人工智能
化学
拉曼光谱
模式识别(心理学)
管道(软件)
分析器
色谱法
显微拉曼光谱
复矩阵
样品(材料)
样品制备
自编码
水准点(测量)
分析化学(期刊)
作者
Jingxiong Lin,Shuo Lei,Chengfeng Li,Zhangwu Chen,Can Tan,Chenyu Liao,Tianyu Wu,Shaokang Qi,Xuan Yu,Yue Zhang,Xunliang Su,Xinming Huo,Jianhua Zhou
出处
期刊:ACS Sensors
[American Chemical Society]
日期:2026-11-18
标识
DOI:10.1021/acssensors.6c04331
摘要
Rapid and accurate on-site identification of target analytes remains challenging, especially for structurally similar compounds in complex matrices. This challenge is particularly pronounced for illegally added drugs and their structural analogues covertly adulterated in food, where single-modality analytical techniques often provide insufficient discriminative information. To address this bottleneck, we propose an intelligent portable surface-enhanced Raman spectroscopy-mass spectrometry system (ipSERS-MS) for on-site identification of illegally added drugs in food through complementary dual-modal molecular sensing. The system integrates a miniature Raman-MS instrument, a surface microchannel membrane-based dual-modal substrate, and a convolutional autoencoder recognition algorithm based on dual-modal outer-product fusion. With a single sample injection, ipSERS-MS simultaneously acquires mass-to-charge ratio profiles and vibrational fingerprints, providing complementary evidence for distinguishing isobaric compounds and closely related analogues, while enabling automated identification without manual spectral interpretation. Both modalities showed high sensitivity, and the fusion-spectrum recognition algorithm achieved 99.65% classification accuracy for eight high-risk illegally added drugs, compared with approximately 70% for single-modality models. Analysis of 30 samples seized during regulatory enforcement showed complete agreement with benchmark LC-MS/MS, with the entire workflow completed within 2 min. ipSERS-MS combines portability, rapidity, and high accuracy, offering a reliable novel analytical approach with broad potential for on-site applications.
科研通智能强力驱动
Strongly Powered by AbleSci AI