Abstract:The functionality of functional materials originates from field and local symmetry. The local symmetry in matter science is described by four degrees of freedom: Lattice, charge, orbit and spin. To understand and use the rich physical properties brought by the structural distortion of atomic nearest neighbors and second nearest neighbors under symmetry breaking, we need to grasp the fine structure of matter at the picometer scale. This paper reviews the exploration of the origin of functional materials by picometer scale electron microscopy in the three cases in the lattice degree of freedom, the expansion and contraction of the polyhedron, the tilt and rotation of the polyhedron and the cation displacement. In the expansion and contraction of the polyhedron degree of freedom, we use picometer scale scanning transmission electron microscopy (STEM) studies of lithium ion battery materials, which can be classified as the unit cell breathing model and Jahn-Teller effect, to illustrate the relationship between functionalities and structural origins. In the tilt and rotation of the polyhedron degree of freedom, we mainly focus on the heterointerface among functional oxides. The emergent phenomena arising from this heterointerface form most of the topics of modern condensed matter physics, such as high temperature superconductivity, magnetoresistance, antiferromagnetic, multiferroics, etc. Lattice mismatch caused polyhedron tilt and rotation alter the local symmetry of the oxides, and hence various functionalities are created. In terms of cation displacement degree of freedom, ferroelectricity is the typical example to show the structure-property relationship. The most exciting future of ferroelectric material is the potential to be the next generation memory material because of its topological property. With the help of aberration corrected (AC)-STEM, flux closure type and vortex type ferroelectric domain were directly observed at picometer accuracy. Furthermore, through in-situ mechanical and electric force studies, reverse transition from topological and normal ferroelectric state was observed, which paves the way of the application of the ferroelectric material as future memory material system. At the electronic structure level, the charge structure under the picometer-scale lattice structure is reviewed. Through the simultaneously recorded EELS data, we can know the valence state and content fluctuation of certain element. For example, in lithium ion battery materials, atomic and electronic structure changes all the time during the whole working duration. It is crucial to know the local valence state of the transition metal elements. AC-STEM and electron energy loss spectroscopy (EELS) can help us to solve this problem under one time acquisition. And future perspectives of the electron microscopy research on the orbit and spin structure with high spacial resolution are discussed, such as monochromatic EELS, electron-energy-loss magnetic circular dichroism (EMCD), electron-energy-loss magnetic linear dichroism (EMLD) and convergent beam electron diffraction (CBED). CBED is considered to be the best way to discover the orbit structure with high spacial resolution. In terms of spin structure, EMCD and EMLD are not the direct methods to detect the electron spin, which limits the resolution and signal to noise ratio of acquired results. In the future, spin polarized electron source is considered as the proper way to discover the spin structure with high spacial resolution. Nowadays, spacial resolution is not the limit of the TEM. The ability of accessing more kinds of materials, powerful in-situ method and acquiring orbit and spin structure with picometer resolution are goals for us.

Learning motifs and their hierarchies in atomic resolution microscopy

Learning Effective Molecular Models from Experimental Observables.

Structure motif centric learning framework for inorganic crystalline systems

Exploring Motifs and Their Hierarchies in Crystals Via Unsupervised Learning

Auto-resolving atomic structure at van der Waal interfaces using a generative model

A multiscale generative model to understand disorder in domain boundaries

Spies Within Metal-Organic Frameworks: Investigating Metal Centers Using Solid-State Nmr

From structure mining to unsupervised exploration of atomic octahedral networks

Direct Motif Extraction from High Resolution Crystalline STEM Images

Multiscale structure-property discovery via active learning in scanning tunneling microscopy

Machine learning-enabled tomographic imaging of chemical short-range atomic ordering

Uncovering Material Deformations via Machine Learning Combined with Four-Dimensional Scanning Transmission Electron Microscopy

Short-Range Order Structure Motifs Learned from an Atomistic Model of a Zr$_{50}$Cu$_{45}$Al$_{5}$ Metallic Glass

Encoding the atomic structure for machine learning in materials science

Picometer Scale Electron Microscopy Exploration for the Functionality Origin of Functional Materials

Uncovering Two Principles of Multivariate Hierarchical Metal–Organic Framework Synthesis via Retrosynthetic Design

Deep Learning of Atomically Resolved Scanning Transmission Electron Microscopy Images: Chemical Identification and Tracking Local Transformations

Low-dimensional metal–organic frameworks: a pathway to design, explore and tune magnetic structures

Direct Visualization of Chemically Resolved Multilayered Domains in Mixed‐Linker Metal–Organic Frameworks

Topological representations of crystalline compounds for the machine-learning prediction of materials properties

Picoscale Structure of Functional Materials Constructed by Functional Units