Research

Overview

Professor Huolin Xin's research group at UCI studies and builds next-generation energy storage and conversion devices that power our portable electronics, vehicles and our cities. In particular, we focus on the study of the fundamental failure modes and degradation mechanisms of batteries and fuel cells. The improved understanding of the existing problems and challenges of these devices allow us to design new architectures and build new materials that are better, safer, and more eco-friendly. We also aim to develop artificially intelligent TEM, that will enable statistical imaging, superresolution, and autonomous characterization.

Nature communications, 2014, 5: 3529.

Lithium-ion Battery

Energy storage devices such as lithium-ion batteries (LIBs) are considered the holy grail for the electrification of our vehicles. The physical and chemical processes at the electrode surfaces and interfaces have great impact on their performances. We focus on the development of next generation LIBs, Li-metal, and all-solid-state batteries with improved safety, long cycle life through studying the failure modes and degradation pathways at the atomic and nano- scale as well as synthesizing new materials and building new architectures that can overcome these 'fundamental' limits.

Fuel Cell and Electrocatalysts

We work on the theoretical design and synthesis of atomically dispersed catalysts with high loading and sufficient active sites to resolve the thick-electrode conundrum. We also fabricate and synthesize 3-D nanostructured electrodes, which are important for the utilization and scale-up of atomically dispersed catalysts.

Nature materials, 2013, 12(1): 81.

Artificially Intelligent TEM

Conventional EM resolution is limited by the lens aberrations, the source size, the diffraction limit, the chromaticity and so on. We aim use artificial intelligent tol dramatically enhance the performance and automation of TEM and its related spectrosocpies.

Multi-dimenionsal TEM Imaging

Electron tomography is a powerful tool for resolving the 3D complexity inside of materials at the atomic and the nano- scale. We have developed several 3D imaging techniques that can reconstruct the internal structure, composition and charge distributions inside of materials with nanoscopic resolution. These techniques are critically important for understanding the spatial heterogeneity developed in nanocatalysts and other energy materials.

ACS nano, 2018, 12(8): 7866-7874.

AlphaTomo

One of the unique problems of electron tomography is that it suffers from the severe missing wedge problem due to either limited space or unfavorable sample geometry. This ill-defined inverse problem has traditionally been mitigated by regularized inversion. However, all of these methods have hyperparameters that are problem-specific and need to be adjusted by the human operators. In this respect, machine learning approaches and specifically, deep learning, offer end-to-end solutions to produce artifact-free reconstructions.

Scientific Reports, 9, 12803 (2019)

Cryo-EM

Soft materials such as biomolecules and polymers are sensitive to electron beam irradiation which make the study of their atomic structures and bonding environment difficult. By cooling these materials down to the liquid nitrogen boiling temperature, their dose tolerance can be improved by three folds and in some cases by orders of magnitude. With the help of single particle reconstruction, we clearly depict their detailed 3D structures and their self-assembled derivative.

Nature nanotechnology, 2015, 10(7): 637.

ACS Nano, 2018, 12(8): 7866-7874.

Nature Communications, 2016, 7, 13335

In-situ TEM

In-situ observation is crucial to understand how materials evolves spatially during chemical reactions. We are interested in studying the formation and failure mechanisms of energy materials. Information distilled from these studies can inform and inspire the design and synthesis of these materials.

TEM ImageNet

http://TEMImageNet.com

Introduction of the first TEM Imagenet to the public.

AtomSegNet

A deep learning app for processing atomic resolution STEM images