Key Laboratory of Crop Physiology, Ecology and Genetic Breeding, Ministry of Education, Jiangxi Agricultural University, Nanchang, 330045, People's Republic of China.
Institute of Functional Materials and Agricultural Applied Chemistry, Jiangxi Agricultural University, Nanchang, 330045, People's Republic of China.
Mikrochim Acta. 2019 Jul 17;186(8):543. doi: 10.1007/s00604-019-3652-x.
A method for intelligent data analysis was designed by combining electrochemical sensing with machine learning (ML). Specifically, a voltammetric sensor is described for determination of the phytoinhibitor maleic hydrazide in crop samples. Carboxyl-functionalized poly(3,4-ethylenedioxythiophene) (PEDOT-C4-COOH) was electro-synthesized in aqueous micellar solution by direct anodic oxidation of its monomer. A nanosensor was then prepared by placing copper nanoparticles (CuNPs) on the PEDOT-C4-COOH film via electro-deposition of Cu (II) from aqueous micellar solutions. An artificial neural network (ANN) served as a powerful ML model to realize intelligent data analysis and smart transformation for digital output. Different established regression methods were selected for evaluating the ANN-based method that was found to be superior to known methods. The sensor has a wide working range (from 0.06-1000 μM), a low limit of detection (10 nM), good stability, selectivity and practicality. The method was applied to the determination of maleic hydrazide in (spiked) samples of onion, rice, potato and cotton leaf. Satisfactory results demonstrate that the feature of simultaneous data acquisition and analysis is highly attractive. Graphical abstract Schematic representation of an electrochemical sensor based on carboxyl-functionalized poly(3,4-ethylenedioxythiophene) (PEDOT-C4-COOH) and copper nanoparticles (CuNPs) by differential pulse voltammetry (DPV) to detect maleic hydrazide (MH). PEDOT-C4-COOH was electro-synthesized in 0.1 M LiClO aqueous micellar solution with 0.1 M sodium dodecyl benzene sulfonate (SDBS) by amperometry (CA). CuNPs was prepared by cyclic voltammetry (CV).
设计了一种将电化学传感与机器学习 (ML) 相结合的智能数据分析方法。具体来说，描述了一种用于测定作物样品中植物抑制剂马来酰肼的伏安传感器。在含有 0.1 M 十二烷基苯磺酸钠 (SDBS) 的 0.1 M LiClO 水胶束溶液中，通过恒电流电化学 (CA) 法电合成羧酸功能化聚(3,4-亚乙基二氧噻吩) (PEDOT-C4-COOH)。然后通过从水胶束溶液中电沉积 Cu(II) 将铜纳米颗粒 (CuNPs) 放置在 PEDOT-C4-COOH 薄膜上，制备纳米传感器。人工神经网络 (ANN) 作为一种强大的 ML 模型，用于实现智能数据分析和数字输出的智能转换。选择了不同的建立回归方法来评估基于 ANN 的方法，发现该方法优于已知方法。该传感器具有较宽的工作范围（从 0.06-1000 μM）、较低的检测限（10 nM）、良好的稳定性、选择性和实用性。该方法应用于洋葱、大米、土豆和棉花叶片中（加标）样品中马来酰肼的测定。满意的结果表明，同时进行数据采集和分析的特点极具吸引力。
