Cryogenic Lorentz Transmission Electron Microscopy of van der Waals Ferromagnetic Heterostructures

Magnetic skyrmions are topologically protected magnetic spin textures that can exist both as individuals and in dense quasiparticle lattices. Manipulation of skyrmions then requires a detailed understanding of their stability. While individual skyrmions have been studied extensively, their collective behavior in skyrmion lattices is poorly understood, both on the micron-scale where lattice ordering is observed and the nanoscale where inter-skyrmion interactions occur. In this work, we will present the observation of Néel skyrmion lattice ordering in the van der Waals (vdW) ferromagnet Fe3GeTe2 (FGT) using cryo-Lorentz transmission electron microscopy (LTEM). We will discuss the magnetic domains formed in FGT under various conditions of applied field and as a function of temperature. We will discuss the stability of skyrmions under applied fields and formation of domains with higher topological charge. In order to understand the collective behavior of these Néel skyrmions, when imaged with LTEM, we performed imaging of skyrmion lattices over a large field of view and employed a machine learning algorithm which applies a convolutional neural network (CNN) to identify skyrmion centers. We will discuss the local and global lattice order, and the observed thermal hysteresis in the order. Based on an analytical model, we can elucidate the collective behavior of skyrmion lattices to be dependent on the competition between the Zeeman energy and the domain energy.