Host: Prof. Anyuan Gao
Venue: TDLI Meeting Room N600
Tencent Meeting link: https://meeting.tencent.com/dm/vdqjDKJUsSuw Meeting ID: 975232008, no password
Abstract:
I will talk about the novel properties due to crystal symmetry spin-splitting antiferromagnets (AFMs) [Nat. Commun. 12, 2846 (2021)] (widely known as altermagnets now). In 2021, we propose the crystal-symmetry-paired spin-valley/momentum locking (CSVL/CSML), which is enabled by crystal symmetries intrinsically in AFMs (e.g., V2Se2O, V2Te2O, MnTe and RuO2) [Nat. Commun. 12, 2846 (2021); Phys. Rev. X 15, 021083 (2025)]. CSML enables feasible controls of AFMs by manipulating the corresponding crystal symmetry. Typically, one can use a strain field to induce net valley polarization/magnetization and use an electric field to generate a noncollinear spin current even without spin-orbit coupling. All the predictions have been confirmed in experiments [Nat. Phys. 21, 760 (2025); Nat. Phys. 21, 754 (2025)]. These properties have helped us realize the electric readout and 180o deterministic switching of the Néel order in our experimental work [Sci. Adv. 10, eadn0479 (2024); Nature 638, 645 (2025)], which can be well understood by switching symmetry in our recent proposed unified theory of all magnetic deterministic switching [arXiv:2603.29136].
Biography:
Professor Junwei Liu earned his Ph.D. from Tsinghua University in 2014 and conducted postdoctoral research at MIT from 2014 to 2017. He joined the Hong Kong University of Science and Technology (HKUST) as an Assistant Professor in 2017 and was promoted to Associate Professor in 2023. His research interests span condensed matter physics, materials science, and machine learning.
In the discovery of new phases, he successfully predicted the SnTe-type topological crystalline insulator, the WTe₂-type and TaIrTe₄-type quantum spin Hall insulators and achieved the first experimental observation of monolayer ferroelectricity in SnTe thin films. In methodological development, he has combined machine learning with effective models to create self-learning Monte Carlo methods, achieving speedups of several orders of magnitude without loss of accuracy. He also designed and realized the world’s first all-optical neural network.
His current research focuses on novel quantum phenomena arising from crystal symmetry and his related contributions include (1) theoretical proposal of crystal-symmetry-paired spin-valley locking in collinear spin-splitting antiferromagnets (also named altermagnet) and its realization and experimental verification in V2(Se, Te)2O-family materials [Nature Communications 12, 2846 (2021); Physical Review X 15, 021083 (2024); Nature Physics 21, 760-767 (2025)], which is also the first layered quasi-2D altermagnets; (2) the first electrical readout and 180° switching of the Néel order in a collinear spin-splitting antiferromagnet [Sci. Adv.10, eadn0479 (2024)]; and (3) the first demonstration of multiple Majorana zero modes in a single vortex and their hybridization [Nature633, 71–76 (2024)].
He has published more than 80 papers, including 2 in Science, 3 in Nature, 2 in Nature Physics, 3 in Nature Materials, 7 in Nature Communications, 1 in Physical Review X, 4 in Physical Review Letters, and 1 in Optica.