Using STEM with quasi-parallel illumination and an automated peak-finding routine for strain analysis at the nanometre scale

Abstract

Strain engineering has become an important tool to allow the semiconductor industry to meet roadmap requirements for device performance in the face of limits to device scaling. In addition, strain and/or lattice distortion through chemistry or mechanical stress, can have significant effect on mechanical, electrical and magnetic properties in a wide range of materials. Therefore, determination of strain in a processed, failed or natural sample will have ramifications in almost all fields of material science and solid state physics. Currently, only transmission electron microscopy (TEM) has proven capable of measuring such buried strains at the required spatial resolutions. This contribution presents an automated methodology to measure strain with high accuracy, high reproducibility, and high spatial resolution yet without the need for elaborate experimental setup or highly trained operator. The methodology is first put in perspective and compared to other available methodologies. Important aspects of the experimental setup are then detailed as well as the specificity of the methodology and used algorithm. Three different cases are described: SiGe multilayer, strained transistor, and indented sapphire and the strain measured discussed.

Four-Dimensional Scanning Transmission Electron Microscopy (4D-STEM): From Scanning Nanodiffraction to Ptychography and Beyond

“…Many researchers have applied NBED to materials science studies, including Liu et al (2008), Sourty et al (2009), Favia et al (2010), Uesugi et al (2011), and Haas et al (2017). Very recent studies, including Han et al (2018), have even extended strain measurements to heterostructures in very weakly scattering 2D materials.…”

py4DSTEM: A Software Package for Four-Dimensional Scanning Transmission Electron Microscopy Data Analysis

“…Therefore, the spacing of the Bragg disks is inversely proportional to the real space atomic spacing. Precise measurements of the reciprocal lattice vectors can, therefore, be used to map the local strain present in a crystalline sample, given by the deviations of the lattice from the ideal spacing and angles (Usuda et al, 2004;Liu et al, 2008;Béché et al, 2009;Sourty et al, 2009;Favia et al, 2010;Uesugi et al, 2011).…”

Lattice strain analysis of silicon fin field-effect transistor structures wrapped by Ge2Sb2Te5 liner stressor

“…A larger convergence angle can be used to obtain a smaller probe size using the current setting, but a larger convergence angle gives larger diffraction spots which will complicate the peak localization. 23 Fig. 2(a) shows a TEM image of Si fin structures wrapped around by the c-GST stressor.…”

“…However, a thick specimen leads to blurred diffraction patterns which result from a combination of inelastic scattering and chromatic aberration of the TEM optics. 23,24 Therefore, in order for the peaks to be localized accurately, energy filtering was used to eliminate the blurring effect. Diffraction patterns were recorded by a 2048 Â 2048 pixels CCD camera at a series of ten reference points in the Si substrate far from the strained fins.…”

“…By measuring the shifts of the peaks in the diffraction pattern as compared to the diffraction pattern acquired in a reference unstrained region, one can obtain the lattice strain in the fin in the two directions perpendicular to the electron beam. [22][23][24][25] …”