Offline Secondary Electron Counting and Conditional Re-illumination in SEM
Offline Secondary Electron Counting and Conditional Re-illumination in SEM
复制标题
SEM 中的离线二次电子计数和条件重照明
DOI：
10.1017/s1431927620017249
复制
发表时间：
2020
影响因子：
2.8
通讯作者：
Berggren, Karl
中科院分区：
文献类型：
--
作者：
Agarwal, Akshay;Simonaitis, John;Goyal, Vivek;Berggren, Karl
最近在扫描透射电子显微镜（STEM）和扫描电子显微镜（SEM）中实施了自适应照明方案，以减轻成像期间的样品损伤[1，2]。在这样的方案中，通过仅照射样本像素的一部分来对样本进行成像。在STEM中，这些方案的主要实现因素是电子计数成像的实现[3]。我们最近提出了一种称为条件再照射（CR）的自适应方案，该方案可以通过在STEM上使用散射电子计数或在SEM中使用二次电子（SE）计数来显着减少样品损坏[4]。然而，SE计数是不常见的SEM，限制了该scheme的实施。在这项工作中，我们已经实现了SE计数成像和CR离线SEM（蔡司利奥1525）。为了实现SE计数，我们在2 GHz示波器上收集SE探测器信号，并通过研究其统计数据验证每个探测到的SE对应于一个信号脉冲[5，6]。我们收集了32个图像帧，像素分辨率为262 × 188像素，射束电流为2 pA，射束能量为10 kV，像素停留时间为440 ns。每个图像帧具有SEM扫描波形以及具有相同时间轴的透镜内和室内SE检测器信号。我们编写代码将探测器信号分割成像素，计算每个像素中的SE数量，并使用来自两个探测器和所有帧的计数创建参考图像。为了实现CR，我们的代码在每帧之后查看SE的数量（下文称为NSE限制的CR）或照明的数量（下文称为M限制的CR）。对于M限制方案，如果照明的数量超过最大阈值，则我们不使用来自后续帧的计数。对于NSE限制方案，如果像素的SE数量超过最大阈值，则我们在随后的帧中不使用该像素的SE计数。因此，每个像素的入射电子剂量均匀分布在所有像素的M有限的计划和非均匀分布的NSE有限的计划。为了将CR图像与参考SE计数图像进行比较，我们的代码通过最大帧数（32）与CR图像中该像素处使用的帧数之比缩放CR图像中每个像素的SE计数。我们使用两个指标来量化CR图像之间的比较-Thong的信噪比（SNR）[7]和平均绝对相对误差（MARE）[8]，平均绝对相对误差（MARE）是CR图像和参考图像之间差异绝对值的平均值，表示为参考图像平均值的分数。
Adaptive illumination schemes have recently been implemented in scanning transmission electron microscopy (STEM) and scanning electron microscopy (SEM) to mitigate sample damage during imaging [1, 2]. In such schemes, the sample is imaged by illuminating only a fraction of the sample pixels. In STEM, a major enabling factor for these schemes has been the implementation of electron count imaging [3]. We had recently proposed an adaptive scheme called conditional re-illumination (CR) that can significantly reduce sample damage by using scattered electron counting on a STEM or secondary electron (SE) counting in SEM [4]. However, SE counting is not commonly available on an SEM which limits the implementation of this scheme.In this work, we have implemented SE count imaging and CR offline on an SEM (Zeiss Leo 1525). To implement SE counting, we collected the SE detector signal on a 2 GHz oscilloscope and verified through a study of its statistics that each detected SE corresponded to one signal pulse [5, 6]. We collected 32 image frames at a pixel resolution of 262 by 188 pixels, beam current of 2 pA, beam energy of 10 kV, and a pixel dwell time of 440 ns. Each image frame had the SEM scan waveform and the in-lens and in-chamber SE detector signal with the same time axis. We wrote code to segment the detector signal into pixels, count the number of SEs in each pixel, and create a reference image using counts from both detectors and all frames. To implement CR our code looked at the number of SEs (hereafter referred to as NSE-limited CR) or the number of illuminations (hereafter referred to as M-limited CR) after every frame. For the M-limited scheme, if the number of illuminations crossed a maximum threshold, we did not use the counts from the following frames. For the NSE-limited scheme, if the number of SEs for a pixel crossed a maximum threshold, we did not use SE counts for that pixel in the following frames. Therefore, the incident electron dose per pixel was uniformly distributed over all pixels for the M-limited scheme and non-uniformly distributed for the NSE-limited scheme. To compare the CR images to the reference SE count image, our code scaled the SE counts for each pixel in the CR images by the ratio of the maximum number of frames (32) to the number of frames used at that pixel in the CR image. We used two metrics to quantify the comparison between CR images–Thong’s signal-to-noise-ratio (SNR)[7] and mean absolute relative error (MARE)[8] which is the mean of the absolute value of the difference between the CR image and the reference image, expressed as a fraction of the mean of the reference image.
通过使用无相互作用测量和条件再照明减少电子显微镜损伤
影响因子：
2.9
作者：
Agarwal, Akshay;Berggren, Karl K.;van Staaden, Yuri J.;Goyal, Vivek K.
通讯作者：
Goyal, Vivek K.