Xin He

Tellurium nanowires have demonstrated great potential for various applications, such as thermoelectric devices, optoelectronics, and biosensors, due to their excellent electronic, thermal, and optical properties. However, lacking an effective method to prepare vertically aligned and uniformly distributed tellurium nanowire arrays is a bottleneck for their practical applications. In this study, we successfully fabricated aligned tellurium nanowires (A-Te NWs) on a Pt/polyimide/glass substrate using a seedless chemical liquid-phase method. Pt square arrays were employed as the catalytic layer to define patterned selective growth regions on the polyimide/glass substrate. Such geometric confinement of growth proves highly effective in promoting the vertical ordering of A-Te NWs. Moreover, the verticality and density of A-Te NWs highly depend on the Pt pattern size. Specifically, smaller patterns result in denser and more vertical A-Te NWs due to their lower surface roughness. In addition, the Pt layer thickness and liquid-phase reaction conditions for the growth of A-Te NWs have been optimized to yield densely packed and vertically aligned A-Te NWs, with a Pt layer thickness of 3 nm, a reaction time of 6 h, a reaction temperature of 30 °C, a pH of 10, a hydrazine hydrate amount of 2 mL, and a cetyltrimethylammonium bromide concentration of 5 mM. Our work paves the way for the practical applications of tellurium nanowires.

Polar skyrmions, as noncollinear polarization textures, have attracted a considerable amount of attention so far because of their novel physical properties. Multistate switching of polar skyrmions is paramount for brain-inspired computing applications, but experimental realization remains challenging because of the inaccessibility of feasible boundary conditions. Here, we explore in situ electrical switching of CuInP2S6-based polar skyrmion bubbles with respect to resistance states on silicon substrates. Because of Cu+ mediation and the large size of CuInP2S6 skyrmions, we can electrically manipulate the boundary conditions and create a series of intermediate skyrmion states. Consequently, we acquire eight nonvolatile resistance states from a single skyrmion bubble, in which information storage density far exceeds that of the current hard drives. On the basis of these devices, we build skyrmion-based solid neurons and simulate spiking neural networks with a recognition accuracy of 98%. Our study shows the uncharted application prospects of polar skyrmions in silicon-based, high-density memory hardware and neuromorphic computing.

Whole-body intelligent locomotion systems face persistent challenges of redundant actuation and poor energy efficiency, limiting real-world deployment. Bio-inspired central pattern generators offer a promising framework for rhythmic control, yet hardware implementations struggle to match the efficiency and adaptability of biological systems. Here, we introduce an in-situ spike-malleable artificial plateau neuron integrating a bistable plateau gate with a transient threshold-switch. The neuron generates amplitude-programmable rhythmic spike bursts, achieving energy-efficient, antagonistic activation of extensors and flexors via a scalable circuit comprising two paired units (plateau gate and threshold-switch). The design leverages distributed encoding for coordinated muscle control, operating at ultra-low energy dissipation (141.37 pJ/spike). An expanded four-unit circuit enhances dynamic spike malleability, enabling parallel processing for multi-joint coordination. On a quadruped robot (Unitree Go2), these distributed circuits directly drive joint-level proportional derivative controllers using the Gaussian-filtered rhythmic spikes, enabling energy-efficient trotting without centralized computation. Critically, the system achieves stable on-ground locomotion and demonstrates adaptive gait transitions in real-world environments. Our approach merges ultra-compact hardware with bio-inspired architecture, advancing neuromorphic systems for energy-efficient autonomous robotics.

Insect-inspired polarization vision systems leverage incident light polarization as an additional dimension for weighting synaptic responses, enabling on-device discrimination of dynamic environmental stimuli and efficient visual information processing. However, existing polarization-sensitive vision devices often face limitations such as suboptimal performance, high power consumption, and poor CMOS compatibility. This work demonstrates a ReS₂/p⁺-Si polarization-sensitive synaptic junction field-effect phototransistor that utilizes the rich polarization information carried by light to perform visual preprocessing directly at the device level, thereby reducing computational burdens on downstream processors. The device exhibits energy-efficient transistor characteristics, including a field-effect mobility of 12.5 cm² V^–1 s^–1, a low subthreshold swing of 74 mV/dec, and a high on/off ratio of 10⁵. Furthermore, its polarization-tunable synaptic response facilitates polarization-dependent functionalities such as contrast enhancement, glare suppression, and fingerprint recognition, offering a path toward polarization-enhanced neuromorphic vision sensing.

Nontrivial topological polarization textures, including polar skyrmions, possess nanoscale three-dimensional polarization arrangements that exhibit unique properties and promising applications. Theoretically, these nontrivial polar textures can be created and annihilated by a flexoelectricity effect, but experimental realization, particularly on pure silicon substrates, remains elusive. Here, we show a scalable approach of leveraging silicon patterns to induce flexoelectricity in a van der Waals ferroionic CuInP₂S₆ crystal. Mediated by copper ion migration, this approach can transform conventional ferroelectric CuInP₂S₆ domains into polar skyrmions. We controllably manipulate the size of polar skyrmions by adjusting the diameter of silicon nanopatterns and the thickness of the CuInP₂S₆ samples. Such skyrmionic textures can achieve reversible resistance switching. Phase-field simulations, Raman spectroscopy, and second harmonic generation prove that the transformation to polar skyrmions is associated with flexoelectricity-induced collective interactions between electrostatic, elastic, and gradient energy. Our study provides an avenue for artificially engineering nontrivial polar textures toward silicon-based high-density memory.

Photonic memristors based on two-dimensional materials are emerging as critical components for ultrascalable, energy-efficient artificial vision systems, integrating opto-sensing, data storage and processing capabilities. However, existing devices typically exhibit narrow spectral response ranges and operate in a single mode (for example, non-volatility), limiting their applications in complex computing scenarios. Here we introduce photonic memristor arrays based on a wafer-scale hexagonal boron nitride (hBN)/silicon (Si) heterostructure. These memristors are developed via in situ, low-temperature (250 °C), large-area growth of highly homogeneous hBN films on Si-based substrates. The devices exhibit opto-reconfigurability across a broad spectral range from ultraviolet to near infrared. By adjusting the incident laser power, the device can be reconfigured between non-resistive-switching, volatile and non-volatile modes. This light-induced reconfigurability is attributed to the formation of conductive filaments through interactions between hydrogen ions and photogenerated electrons within the engineered hBN/Si heterostructures. Furthermore, the photonic memristor features a switching ratio exceeding 10⁹, retention time surpassing 40,000 s, endurance over 10⁶ cycles and thermal stability up to 300 °C. These findings provide a scalable solution for developing integrated sensing–storage–computation artificial vision systems, fully compatible with sophisticated Si-based semiconductor technologies.

Programmable and nonvolatile Schottky junctions are highly desirable for next-generation electronic and neuromorphic systems. However, conventional metal-semiconductor and even van der Waals (vdW) Schottky diodes often suffer from fixed rectifying behaviors or limited tunability. Here, we report a programmable nonvolatile ferroelectric Schottky diode based on a vdW heterojunction between semimetallic 1T'-MoTe₂ and ferroelectric α-In₂Se₃. The diode exhibits near-ideal performance, including a rectification ratio exceeding 10⁴, a leakage current down to 1 pA, and an ideality factor as low as 1.38. By switching ferroelectric polarization, the Schottky barrier can be modulated in a programmable manner, enabling reversible, nonvolatile, and multilevel rectification states. The device demonstrates polarization-dependent photoresponse and transient integrate-and-leak dynamics, closely resembling biological spiking neurons. A spiking neural network is implemented based on this behavior, achieving image recognition accuracy up to 98.4%. This work establishes programmable ferroelectric Schottky diodes as promising candidates for low-power memory, reconfigurable logic, and neuromorphic vision.

Ferroelectric polarization switching, achieved by mechanical forces, enables the storage of stress information in ferroelectrics and holds promise for human interface applications. The prevailing mechanical approach is locally induced flexoelectricity with large strain gradients. However, this approach usually requires huge mechanical forces, which greatly impede device applications. Here, we report an approach of using triboelectric effect to mechanically, reversibly switch ferroelectric polarization across α-In ₂ Se ₃ ferroelectric memristors. Through contact electrification and electrostatic induction effects, triboelectric units are used to sensitively detect mechanical forces and generate electrical voltage pulses to trigger α-In ₂ Se ₃ resistance switching. We realize multilevel resistance states under different mechanical forces, by which a neuromorphic stress system is demonstrated. Notably, we achieve the reversal of α-In ₂ Se ₃ ferroelectric polarization with a record-low mechanical force of ~10 kilopascals and even with tactile touches. Our work provides a fundamental but pragmatic strategy for creating mechanical tactile ferroelectric memory devices.