大学化学 >> 2026, Vol. 41 >> Issue (10): 370-378.doi: 10.12461/PKU.DXHX202510100

化学实验 上一篇    

科教融合导向下的本科化学实验设计——基于机器学习预测分子偶极矩的探索

刘国魁1, 李丽2, 杨一凡1, 韦瑶瑶1, 夏其英1   

  1. 1 临沂大学化学化工学院, 山东 临沂 276000;
    2 东北林业大学化学化工与资源利用学院, 黑龙江 哈尔滨 150040
  • 收稿日期:2025-10-28 录用日期:2025-11-26 发布日期:2026-10-13
  • 通讯作者: 韦瑶瑶, 夏其英 E-mail:weiyaoyao@lyu.edu.cn;xiaqiying@lyu.edu.cn Yaoyao Wei, Qiying Xia
  • 基金资助:
    临沂大学2023年实验教学改革与实验技术研究项目(SYJG2023Z06, SYJG2023M21)；临沂大学2024年校级本科教学改革研究项目(JG2024M24)；山东省研究生优质课程《量子化学》(SDYKC2024181)；临沂大学思政示范课程《量子化学》(SZKC23004)；2022年山东省本科教改项目(Z2022132)

Undergraduate chemistry experiment design guided by the integration of science and education: machine learning-based exploration for predicting molecular dipole moments

Guokui Liu1, Li Li2, Yifan Yang1, Yaoyao Wei1, Qiying Xia1   

  1. 1 School of Chemistry and Chemical Engineering, Linyi University, Linyi 276000, Shandong Province, China;
    2 College of Chemistry, Chemical Engineering and Resource Utilization, Northeast Forestry University, Harbin 150040, Heilongjiang Province, China
  • Received:2025-10-28 Accepted:2025-11-26 Published:2026-10-13
  • Contact: Yaoyao Wei, Qiying Xia E-mail:weiyaoyao@lyu.edu.cn;xiaqiying@lyu.edu.cn

PDF

144

赞

补充材料

1

摘要： 针对机器学习的飞速发展及其在各学科教学科研领域的重要应用，本文将机器学习融入物理化学中一个经典的实验“偶极矩的测定”。实验以正丁醇为模型分子，学生不仅要完成量子化学计算中的结构优化、势能面扫描和偶极矩计算，而且要进一步利用生成的量子化学数据集训练机器学习模型，实现对正丁醇偶极矩的快速准确预测。通过本实验，可以有效衔接本科教学与科研前沿，为学生普及机器学习这一强大工具，培养学生面向未来、跨学科的研究能力。

关键词: 机器学习, 偶极矩, 计算化学, XGBoost

Abstract: In light of the rapid development of machine learning and its widespread applications across various teaching and research disciplines, this study incorporates machine learning into the classic physical chemistry experiment “Determination of Dipole Moment”. Using n-butanol as a model molecule, students are required to perform quantum chemical calculations involving structural optimization, potential energy surface scanning, and dipole moment computation. Furthermore, they utilize the generated quantum chemical dataset to train machine learning models, enabling rapid and accurate prediction of n-butanol’s dipole moment. This experiment effectively bridges undergraduate education with cutting-edge research, introducing students to machine learning as a powerful tool while fostering their interdisciplinary research capabilities for future scientific endeavors.

Key words: Machine learning, Dipole moment, Computational chemistry, XGBoost