arXiv subjects

Chenhao Liang

Publications and source records attributed to Chenhao Liang.

7 recordsLinked to original sources

arXiv2026

Spin-Polarized Weyl Fermions and Unconventional Spin Transport

Large and electrically controllable spin splitting in nonmagnetic materials is highly desirable for condensed-matter physics and spintronics, enabling strongly spin-polarized quasiparticles and unconventional spin transport as well as their electrical control. Here, we establish a general framework for critical spin-polarized Weyl fermions in two-dimensional (2D) nonmagnetic systems. We predict the $α$-Bi-like Group-V puckered-layer family of 2D ferroelectrics as a material platform combining critical spin-Weyl fermions, giant spin splitting, and electrically switchable nonlinear and spin transport. Our calculations reveal exchange-scale spin splitting induced by spin-orbit coupling and a $p$-wave spin texture in the BiAs monolayer. Moreover, BiAs hosts four critical spin-Weyl points, each formed within a single-spin channel. The nearly quantized $\langle S_z\rangle$ is protected by an emergent $S_z$-conservation symmetry in the low-energy limit. The spin-Weyl cones also give rise to unconventional spin-transport phenomena, including a purely out-of-plane Edelstein response and a pronounced spin Nernst effect. Ferroelectric reversal further enables electrical switching of the $p$-wave spin texture and inversion-odd responses, while the tunability of these properties through composition, strain, and magnetic field is also systematically demonstrated across the entire family. These results establish both a general nonmagnetic framework and a promising material platform for exploring critical spin-Weyl physics, nonlinear Hall phenomena, and unconventional spin transport.

cond-mat.mtrl-sci↗

arXiv2026

FELT: Generating Tactile Signals from Vision for Visuo-Tactile Manipulation

Tactile sensing can improve performance in contact-rich robot manipulation. However, learning policies that use real tactile feedback typically requires demonstrations collected with tactile-equipped hardware. We present Feature-Extracted Latent Tactile (FELT), a framework that uses existing paired visuo-tactile data to learn contact-aware representations from RGB observations. FELT provides these representations to downstream policies through two interfaces: generated tactile images and contact-aware features. FELT combines a frozen visual encoder with separate query decoders for the left and right finger, reflecting the spatial organization of dual-finger tactile arrays. The two branches exchange information through gated connections, and a convolutional readout produces spatial contact representations for the downstream interfaces. For downstream manipulation, the image interface decodes these spatial representations into tactile images, allowing policies trained with either real or generated tactile observations to use generated images at deployment. Alternatively, the feature interface feeds these representations directly to the policy during both training and deployment, without real tactile observations. Experiments on four contact-rich real-robot tasks show that both interfaces achieve higher success rates than vision-only policies. Directly incorporating learned contact-aware features into the policy improves success by 15-22.5 percentage points across the four tasks.

cs.RO↗

arXiv2026

CUA-SWE: When Computer-Use Agents Meet Visual Software Engineering

Software development requires more than editing code: developers repeatedly run software, interact with its interfaces, visually inspect its behavior, and use these observations to decide what to change next and whether a change works. Existing coding agents and computer-use agents are largely studied in isolation, leaving this integrated development process underexplored. Diagnosing a runtime interaction failure requires agents to connect visual observations with the responsible code, then use the application again to verify the repair. We introduce CUA-SWE, a benchmark, environment, and evaluation pipeline for software engineering with computer use. Beyond studying how GUI feedback supports diagnosis and repair, we ask whether agents can complete software engineering tasks when required specification or operational information is available only through the running application's visual interface. CUA-SWE spans four software engineering domains and requires agents to modify code and configuration, execute commands, interact with running software, and inspect visual feedback within the same task. Each task includes deterministic, task-specific tests that verify whether the resulting software satisfies the requirements and preserves specified behavior. Our evaluation characterizes how frontier agents combine source-level execution with application screenshots and graphical interaction to produce verified software changes. We examine performance across domains and task information requirements, alongside the development behaviors associated with successful repairs. CUA-SWE provides a unified testbed for studying how agents use visual feedback and interaction to guide software engineering, with executable correctness criteria for the resulting software.

cs.AI↗

arXiv2026

Three-dimensional topological insulator feature of ternary chalcogenide Ge2Bi2Te5

The exploration of novel topological insulators (TIs) beyond binary chalcogenides has been accelerated in pursuit of exotic quantum states and device applications. Here, the layered ternary chalcogenide Ge2Bi2Te5 is identified as a three-dimensional TI. The bulk electronic structure of Ge2Bi2Te5 features a hole-type Fermi surface at Fermi level EF, which dominates the transport properties. Moreover, an unoccupied topological surface state with a Dirac point located at 290 meV above EF has been observed. Theoretical calculations confirm a bulk bandgap and a nontrivial Z2 topological invariant (000;1). The present study demonstrates that the material family of layered tetradymite-like ternary compounds is an important platform to explore exotic topological phenomena.

cond-mat.str-el↗

arXiv2025

Optical Excitations of Flat Bands Induced by Exciton Condensation in Ta$_2$Pd$_3$Te$_{5}$

We report on the charge dynamics of Ta$_2$Pd$_3$Te$_5$ using temperature-dependent optical spectroscopy with polarized light. We observe a metal-insulator transition characterized by the collapse of Drude response and the emergence of sharp and narrow absorption peaks at low temperatures. Unlike previous excitonic insulator candidates such as TiSe$_2$ and Ta$_2$NiSe$_5$, where the excitonic order is intertwined with charge density wave or structural instabilities, the sharp features in Ta$_2$Pd$_3$Te$_5$ point to intrinsic excitonic excitations associated with ultra-flat bands driven by many-body renormalization of the band structure via spontaneous exciton condensation. Our findings thus provide clear-cut optical evidence for exciton condensation in a bulk crystal and establish Ta$_2$Pd$_3$Te$_5$ as a promising platform for exploring correlated quantum phases and novel excitonic phenomena.

cond-mat.mtrl-sci↗

arXiv2025

Polarization-induced Quantum Spin Hall Insulator and Topological Devices in InAs Quantum Wells

In this work, we predict the emergence of a quantum spin Hall insulator (QSHI) in conventional semiconductors, specifically InAs quantum wells, driven by a built-in polarization field. We propose QSHI InAs quantum wells as a platform to engineer topological field effect devices. More precisely, we first present a novel topological logic device that operates without a topological phase transition. Subsequently, we design a high-performance topological transistor due to the presence of edge states. Our approach provides a potential framework for harnessing the unique features of QSHI in device design, paving the way for future topological devices.

cond-mat.mtrl-sci↗

arXiv2023

VASP2KP: kp models and Lande g-factors from ab initio calculations

The $k\cdot p$ method is significant in condensed matter physics for the compact and analytical Hamiltonian. In the presence of magnetic field, it is described by the effective Zeeman's coupling Hamiltonian with Landé $ g $-factors. Here, we develop an open-source package VASP2KP (including two parts: vasp2mat and mat2kp) to compute $k\cdot p$ parameters and Landé $g$-factors directly from the wavefunctions provided by the density functional theory (DFT) as implemented in Vienna ab initio Simulation Package (VASP). First, we develop a VASP patch vasp2mat to compute matrix representations of the generalized momentum operator $ \mathbf{\hatπ}=\mathbf{\hat{p}}+\frac{1}{2mc^2}\left(\mathbf{\hat{s}}\times\nabla V(\mathbf{r})\right) $, spin operator $\mathbf{\hat{s}}$, time reversal operator $\hat{T}$ and crystalline symmetry operators $\hat{R}$ on the DFT wavefunctions. Second, we develop a python code mat2kp to obtain the unitary transformation $U$ that rotates the degenerate DFT basis towards the standard basis, and then automatically compute the $k\cdot p$ parameters and $g$-factors. The theory and the methodology behind VASP2KP are described in detail. The matrix elements of the operators are derived comprehensively and computed correctly within the projector augmented wave method. We apply this package to some materials, e.g., Bi$_2$Se$_3$, Na$_3$Bi, Te, InAs and 1H-TMD monolayers. The obtained effective model's dispersions are in good agreement with the DFT data around the specific wave vector, and the $g$-factors are consistent with experimental data. The VASP2KP package is available at https://github.com/zjwang11/VASP2KP.

cond-mat.mtrl-sci↗