Principle component analysis applied to high resolution cross sectional STEM imaging: Quantitative analysis of 2D heterostructures

Abstract

Monolayers of 2D transition metal dichalcogenides (TMDCs) provide excellent semiconducting counterparts to insulating hexagonal boron nitride (hBN) and conductive graphene.[1] Combining all three materials in a Van der Waals vertical heterostructure allows the electronic, photovoltaic and electroluminescent properties of the TMDCs to be studied.[2] Whilst transport and optoelectronic measurements can probe the properties of exotic charge carriers, and ARPES can map the bandstructure such systems, direct high resolution imaging of the buried interfaces is only possible via high resolution cross sectional (S)TEM imaging.[3] The nature of these van der Waals interfaces not only determines the carrier injection between components[4], but also affects the bandstructure of the device and its ultimate functionality[5]. Here we present a novel strategy to denoise high resolution HAADF STEM images of cross sections using principal component analysis (PCA).[6] Cross sections are fabricated in a dual‐beam FIB‐SEM instrument using the in situ lift‐out method and polished with low energy ions to achieve electron transparency.[7] Cross sections were imaged in high resolution HAADF STEM using a probe‐side aberration corrected FEI Titan G2 80–200 kV with an X‐FEG electron source. By treating each line profile perpendicular to the fringes as a signal, components associated with noise and scattering can be separated by their variance. Removing the noise components allows us to accurately determine the separation between dissimilar crystals at these unique interfaces. More widely, the approach developed here also has application in a variety of layered material systems.