Dielectric properties of polymer nanocomposite interphases from electrostatic force microscopy using machine learning
Dielectric properties of polymer nanocomposite interphases from electrostatic force microscopy using machine learning
复制标题
DOI：
10.1016/j.matchar.2021.110909
复制
发表时间：
2021-01-30
影响因子：
4.7
通讯作者：
Sundararaman, Ravishankar
中科院分区：
文献类型：
--
作者：
Gupta, Praveen;Schadler, Linda S.;Sundararaman, Ravishankar
聚合物纳米复合材料中的界面区由于其纳米尺度而难以表征。静电力显微镜（EFM）提供了一种途径，局部介电性能的测量，但提取局部介电常数在复杂的相间几何形状从EFM测量仍然是一个挑战。我们证明了机器学习（ML）模型提取相间介电常数使用有限元模拟生成的合成EFM力梯度扫描的数据集的有效性。我们表明，支持向量回归（SVR）和随机森林（RF）算法能够？倒置？力梯度扫描以高精度预测介电常数。通过主成分分析（PCA）的特征约简改进了模型？的性能，并揭示力梯度对比度是最重要的功能，在介电常数检测。我们发现这些ML模型通过捕获EFM测量的显著几何复杂性而比分析方法表现得更好。
Interphase regions in polymer nanocomposite materials are difficult to characterize due to their nano-scale di-mensions. Electrostatic force microscopy (EFM) provides a pathway to local dielectric property measurements, but extracting local dielectric permittivity in complex interphase geometries from EFM measurements remains a challenge. We demonstrate the efficacy of machine learning (ML) models to extract interphase permittivity using a data set of synthetic EFM force gradient scans generated by finite element simulations. We show that both support vector regression (SVR) and random forest (RF) algorithms are able to ?invert? the force gradient scan to predict the permittivity with high accuracy. Feature reduction by principal component analysis (PCA) improves the model?s performance and reveals force gradient contrast to be the most important feature in permittivity detection. We find that these ML models perform better than analytical approaches by capturing significant geometric complexity of EFM measurements.