大学化学 >> 2026, Vol. 41 >> Issue (1): 64-75.doi: 10.12461/PKU.DXHX202505088

所属专题： 化学实验数字化设计竞赛获奖作品

专题 上一篇    下一篇

基于机器学习与微流控技术的银纳米粒子合成实验创新研究

李作勇, 屠昊翔, 丁明炜, 刘美君, 杨婷   

  1. 东北大学理学院, 沈阳 110819
  • 收稿日期:2025-05-27 录用日期:2025-10-13 发布日期:2025-12-30
  • 通讯作者: 杨婷 E-mail:yangting@mail.neu.edu.cn Ting Yang
  • 基金资助:
    国家自然科学基金资助(22174011, 22374014)

Innovative Teaching Reform Study on the Synthesis of Silver Nanoparticles Based on Machine Learning and Microfluidic Technology

Zuoyong Li, Haoxiang Tu, Mingwei Ding, Meijun Liu, Ting Yang   

  1. College of Science, Northeastern University, Shenyang 110819, China
  • Received:2025-05-27 Accepted:2025-10-13 Published:2025-12-30
  • Contact: Ting Yang E-mail:yangting@mail.neu.edu.cn

PDF

229

赞

摘要： 银纳米粒子的合成与表征近年来逐渐被引入化学专业本科教学实验中，具有一定的综合性和前沿性。传统教学中普遍采用化学还原法，使用温和还原剂将银离子还原为银纳米颗粒，常在玻璃容器中配合水浴加热进行反应。但该方法存在试剂利用率低，产物粒径不均、难以精确调控粒子大小等问题，且实验条件可控性差，严重影响实验结果与学生学习体验。为了提升教学效果，本项目设计引入了微流控芯片合成与机器学习指导的综合方案对这个实验进行改进。通过在微流控芯片中控制流速、反应浓度和温度等参数，可提高反应可控性与纳米粒子合成的重复性，显著提高粒径一致性和合成效率；同时，借助机器学习方法对实验数据进行建模与预测，能够辅助学生理解不同反应条件对产物结构的影响，并实现目标产物粒径的可控合成。本实验的教学设计将微流控技术与数据驱动的人工智能方法引入到传统纳米粒子合成教学中，不仅丰富了教学内容，也对学生的操作能力、数据处理能力和问题解决能力提出了更高要求。通过项目实施，能够激发学生对现代化学研究方法的兴趣，帮助其构建对纳米材料化学反应机制更为深入的理解，提升其科学素养与综合创新能力。

关键词: 机器学习, 微流控芯片, 银纳米粒子, 数字化教学

Abstract: The synthesis and characterization of silver nanoparticles have been progressively incorporated into undergraduate chemistry laboratory curricula in recent years, demonstrating both comprehensive and cutting-edge characteristics. Conventional teaching methods predominantly employ chemical reduction approaches, where silver ions are reduced to nanoparticles using mild reducing agents in glass vessels with water bath heating. However, this traditional methodology presents several limitations, including low reagent efficiency, inconsistent particle size distribution, and poor controllability over particle dimensions, all of which adversely affect experimental outcomes and student learning experiences. To enhance pedagogical effectiveness, this project introduces an integrated approach combining microfluidic chip-based synthesis with the guidance of machine learning. Precise control of flow rate, reactant concentration, and temperature parameters within microfluidic channels improves reaction reproducibility and controllability, significantly enhancing particle size uniformity and synthesis efficiency. Furthermore, machine learning algorithms are employed to model and predict experimental data, enabling students to comprehend how different reaction conditions influence product characteristics and achieve targeted nanoparticle synthesis. This innovative teaching design incorporates both microfluidic technology and data-driven artificial intelligence into traditional nanoparticle synthesis instruction, not only enriching course content but also elevating requirements for students’ operational skills, data processing capabilities, and problem-solving competencies. Through this implementation, the project aims to stimulate student interest in modern chemical research methodologies, facilitate deeper understanding of nanomaterial reaction mechanisms, and enhance scientific literacy and comprehensive innovation abilities.

Key words: Machine learning, Microfluidic chip, Silver nanoparticles, Digital teaching