Isik Nimet
Department of Science Education,Mehmet Akif Ersoy University,15200 Burdur,Turkey.
Microsc Microanal. 2016 Apr;22(2):458-62. doi: 10.1017/S1431927616000118. Epub 2016 Feb 16.
Multi-element electrostatic aperture lens systems are widely used to control electron or charged particle beams in many scientific instruments. By means of applied voltages, these lens systems can be operated for different purposes. In this context, numerous methods have been performed to calculate focal properties of these lenses. In this study, an artificial neural network (ANN) classification method is utilized to determine the focused/unfocused charged particle beam in the image point as a function of lens voltages for multi-element electrostatic aperture lenses. A data set for training and testing of ANN is taken from the SIMION 8.1 simulation program, which is a well known and proven accuracy program in charged particle optics. Mean squared error results of this study indicate that the ANN classification method provides notable performance characteristics for electrostatic aperture zoom lenses.
多元素静电孔径透镜系统广泛应用于许多科学仪器中，用于控制电子或带电粒子束。通过施加电压，这些透镜系统可用于不同目的。在这种情况下，已经采用了许多方法来计算这些透镜的聚焦特性。在本研究中，利用人工神经网络（ANN）分类方法来确定多元素静电孔径透镜在像点处聚焦/未聚焦的带电粒子束与透镜电压的函数关系。用于人工神经网络训练和测试的数据集取自SIMION 8.1模拟程序，该程序在带电粒子光学领域是一个知名且经证实具有准确性的程序。本研究的均方误差结果表明，人工神经网络分类方法为静电孔径变焦透镜提供了显著的性能特征。
