Influence of combinatory effects of STEM setups on the sensitivity of differential phase contrast imaging

Highlights

• Independent and combinatory STEM setups on the DPC sensitivity was investigated.

• Larger camera length and smaller overlap improve the DPC sensitivity.

• Camera length and overlap cannot be controlled independently in normal STEM systems.

• Multiple scattering followed by intensity redistribution decreases the sensitivity.

• For Nd₂Fe₁₄B, the best DPC signal is obtained with 65 nm thickness.

Abstract

Differential phase-contrast (DPC) imaging in the scanning transmission electron microscopy (STEM) mode has been suggested as a new method to visualize the nanoscale electromagnetic features of materials. However, the quality of the DPC image is very sensitive to the electron-beam alignment, microscope setup, and specimen conditions. Unlike normal STEM imaging, the microscope setup variables in the DPC mode are not independent; rather, they are correlated factors decisive for field sensitivity. Here, we systematically investigated the independent and combinatory effects of microscope setups on the sensitivity of the DPC image in a hard magnet, Nd₂Fe₁₄B alloy. To improve sensitivity, a smaller overlap of the electron beam with annular detectors and a greater camera length were required. However, these factors cannot be controlled independently in the two-condenser-lens system. In this linked system, the effect of the camera length on the DPC sensitivity was slightly more predominant than the overlap. Furthermore, the DPC signal was noisy and scattered at a small overlap of less than 11%. The electron-beam current does not evidently affect the sensitivity. In addition, the DPC sensitivity was examined with respect to the sample thickness, and the optimum thickness for high sensitivity was approximately 65 nm for the hard magnetic material Nd₂Fe₁₄B. This practical approach to the STEM setup and sample thickness may provide experimental guidelines for further application of the DPC analysis method.

Introduction

With the miniaturization and integration of devices, characterization to avoid unexpected physical phenomena in nanostructured materials is rapidly gaining in significance. In particular, transmission electron microscopy (TEM) is one of the most powerful instruments to observe the nanoscale structural and electromagnetic features of materials simultaneously. Thus far, Lorentz TEM (LTEM) and electron holography have been widely used to visualize the magnetic and electronic structure (Fuller and Hale, 1960; Hale et al., 1959; Kovacs et al., 2016; Mankos et al., 1996; Marshall et al., 1999; Pollard et al., 2017; Tonomura, 1983; Tonomura et al., 1980). LTEM is a convenient method to observe the magnetic structure of materials. Recently, its use has increased with the investigation of skyrmions (Seki et al., 2012; Yu et al., 2011, 2010). However, its moderate spatial resolution and large beam defocus limit the observation of nanoscale magnetic structures (Darlington and Valdré, 1975). The holography technique provides quantitative field information of materials, but it is limited in terms of the field of view and ability to provide intuitive information (Chapman, 1984; Gabor, 1948). Recently, differential phase-contrast (DPC) imaging in the scanning transmission electron microscopy (STEM) mode has been suggested as a new method to visualize electric and magnetic fields in materials (Lohr et al., 2012; Zweck, 2016).

The fundamentals of phase-contrast imaging in STEM were laid by Rose (1974b). An actual implementation was first reported with a split detector by Chapman et al. (1978) and Rose (1974a). The STEM DPC can visualize the nanoscale electric or magnetic fields by detecting the electron beam deflected by the field in the specimen (Lohr et al., 2012; Matsumoto et al., 2016; Sandweg et al., 2008; Shibata et al., 2017, 2015). However, successful acquisition of DPC data is challenging because the image contrast is very sensitive to the beam alignment, microscope setup, and sample conditions such as thickness, crystal orientation, and field strength. Therefore, some previous papers reported possible effective parameters based on calculations and simulations. F. Schwarzhuber examined the influence of camera length as well as overlap size on the calibration factor (κ_β) and suggested that the field sensitivity can be improved with a greater camera length and smaller overlap (Schwarzhuber et al., 2017). D. J. Taplin explored the effect of several factors on the center of mass shifts (Taplin et al., 2016). He also suggested the optimum specimen thickness, crystal orientation, detector type, and overlap size. However, these theoretical approaches have several limitations when applied to practical situations. For example, the camera length and overlap size between the detector and diffraction disk cannot be controlled separately in a normal STEM system. Thus, a more practical and systematic investigation of experimental setups and materials is necessary.

In this study, we investigated independent and combinatory effects of the microscope setup, such as the camera length, spot size, and aperture of the second condenser lens (C2) on the sensitivity of DPC imaging. Additionally, we demonstrate how DPC images are affected by multiple electron scattering and recommend the appropriate specimen thickness for DPC imaging.