Wei Dali, Li Mengfan, Yang Yudi, Deng Chunmeng, Zhu Fang, Li Ming, Deng Yibin, Zhang Zhen
School of the Environment and Safety Engineering, School of the Emergency Management, Jiangsu University, Zhenjiang 212013, China.
Center for Medical Laboratory Science, The Affiliated Hospital of Youjiang Medical University for Nationalities, Baise 533000, China.
ACS Sens. 2025 Jul 25;10(7):5245-5255. doi: 10.1021/acssensors.5c01499. Epub 2025 Jul 15.
Phenolic pollutants pose a great threat to human health due to high toxicity, whereas existing methods are difficult to achieve the rapid recognition of multiple phenolic pollutants. In this study, we developed a novel machine learning-assisted sensor array based on ligand microenvironment-regulated Pt nanozymes for the simultaneous differentiation of five phenolic pollutants (phenol, 2,4-DCP, -chlorophenol, -chlorophenol, and -chlorophenol), wherein four cellulose ligands (carboxymethylcellulose, CMC; methylcellulose, MC; hydroxyethyl cellulose, HC; and hydroxypropyl methyl cellulose, HPMC)-regulated Pt nanozymes (Pt@CMC, Pt@MC, Pt@HC, and Pt@HPMC) with considerable laccase-mimicking activity were designed, and the Pt@CMC nanozyme exhibited the highest catalytic activity, which was about 7.5-folds than that of natural laccase. The calculation of density functional theory revealed that Pt@CMC had a stronger ability for capturing 2,4-DCP molecules, showing higher laccase-like activity. More importantly, the different cellulose ligands endowed four Pt nanozymes with laccase-like activity diverse recognition capability to phenolic compounds; thus, a nanozyme sensor array was developed for the differentiation of five phenolic pollutants. Moreover, the integration of a machine learning algorithm and the nanozyme sensor array successfully achieved accurate identification and prediction of the five phenolic pollutants in real water samples. Therefore, this study provided an emerging sensing strategy for the simultaneous identification of phenolic pollutants, carving a promising path for the application of sensor arrays and machine learning algorithms in environmental monitoring.
酚类污染物因其高毒性对人类健康构成巨大威胁，而现有方法难以实现对多种酚类污染物的快速识别。在本研究中，我们基于配体微环境调控的铂纳米酶开发了一种新型机器学习辅助传感器阵列，用于同时区分五种酚类污染物（苯酚、2,4-二氯苯酚、对氯苯酚、间氯苯酚和邻氯苯酚），其中设计了四种纤维素配体（羧甲基纤维素，CMC；甲基纤维素，MC；羟乙基纤维素，HC；羟丙基甲基纤维素，HPMC）调控的具有相当漆酶模拟活性的铂纳米酶（Pt@CMC、Pt@MC、Pt@HC和Pt@HPMC），且Pt@CMC纳米酶表现出最高的催化活性，约为天然漆酶的7.5倍。密度泛函理论计算表明，Pt@CMC对2,4-二氯苯酚分子具有更强的捕获能力，表现出更高的类漆酶活性。更重要的是，不同的纤维素配体赋予四种铂纳米酶对酚类化合物具有不同类漆酶活性的识别能力；因此，开发了一种纳米酶传感器阵列用于区分五种酚类污染物。此外，机器学习算法与纳米酶传感器阵列的整合成功实现了对实际水样中五种酚类污染物的准确识别和预测。因此，本研究为酚类污染物的同时识别提供了一种新兴的传感策略，为传感器阵列和机器学习算法在环境监测中的应用开辟了一条有前景的道路。
