Central to this advancement is the SnS₂-based memristor, a component engineered via precision chemical vapor deposition techniques. This device harnesses the unique interplay of photogenerated carrier trapping and recombination at intrinsic defect sites within the material, yielding a tunable conductance that sensitively adapts to distinct illumination durations. Such a setup enables the memristor to instantaneously respond to rapid changes in light intensity, thus capturing fast-moving objects, while simultaneously manifesting synapse-like behavior for the analysis of slower motions, thereby accommodating a wide spectrum of temporal dynamics with remarkable flexibility.

The researchers report exceptional optoelectronic metrics for the SnS₂ memristor, including an unprecedented responsivity quantified at 208.9 amperes per watt, an external quantum efficiency soaring beyond 700%, and an impressive detectivity on the order of 10¹¹ Jones. These parameters underscore the device’s superior sensitivity and its aptitude for precise optical signal transduction, which are critical for enabling real-time, adaptive motion recognition directly at the sensory front end.

A pivotal innovation lies in the system’s ability to strategically modulate light intensity based on the velocity of the observed motion. Fast-moving entities are probed under low-intensity illumination, eliciting swift optoelectronic responses that capture transient spatial-temporal features with high fidelity. Conversely, slow-moving targets benefit from intensified light input, provoking prolonged dynamic responses analogous to biological synaptic functions, which enhance the detection of gradually evolving inputs. This dual-regime light adaptation remedies the prevalent issues of signal saturation and loss that plague standard fixed-response configurations.

The practical efficacy of this neuromorphic system was rigorously evaluated through dynamic motion recognition tasks utilizing the widely recognized Weizmann video dataset. In these trials, the in-sensor reservoir achieved flawless classification performance across diverse activities — notably running, side jumping, and walking — performed at varying velocities. Remarkably, this was accomplished with a significantly streamlined computational load, as the model required minimally 20,739 parameters, far fewer than those demanded by traditional long short-term memory (LSTM) networks.

This amalgamation of sensing, memory storage, and information processing within a singular hardware platform effectively bypasses the von Neumann bottleneck, a long-standing hurdle in conventional computing architectures characterized by separation of processing and data storage units. By unifying these operations, the SnS₂ memristor-based system delivers enhanced computational efficiency and paves the way for energy-efficient, edge-localized intelligence services—imperative for autonomous robotics and embedded AI systems operating in dynamic real-world environments.

Despite these striking results, the research team acknowledges that scaling the technology to high-resolution and more complex datasets, such as UCF101 which encompasses a wide array of human action recognition scenes, remains an open challenge. Future work will address scalability concerns and improve robustness in more cluttered, naturalistic scenes, propelling this hardware neuromorphic approach closer to broad applicability in intelligent machine vision.

The scientific rigor behind the project stems from the collaboration of several experts, including Linfeng Sun, Zhongrui Wang, and their colleagues at Beijing Institute of Technology. Their multidisciplinary efforts meld materials science, device physics, and computational neuroscience, illustrating the transformative potential of integrating optoelectronic memristors within neuromorphic frameworks.

This research was bolstered by substantial funding support, including grants from the Beijing Natural Science Foundation and China’s National Key Research & Development Plan, alongside contributions from Hong Kong’s Research Grants Council and Shenzhen’s Science and Technology Commission. Such backing underscores the strategic priority assigned to developing next-generation sensors and computing platforms.

Published in the journal Cyborg and Bionic Systems on September 30, 2025, the paper titled “Tunable Neuromorphic Computing for Dynamic Multi-Timescale Sensing in Motion Recognition” marks a significant milestone in the journey toward intelligent, adaptive sensory hardware capable of real-time environmental interaction. The DOI for this publication is 10.34133/cbsystems.0412, providing extensive technical details for further scholarly reference.

Beyond its immediate utility in motion recognition, this technology holds promise for widespread deployment in scenarios requiring rapid, on-site processing of complex sensory data under constrained power budgets—a cornerstone for the evolution of autonomous systems. By enabling hardware-level temporal dynamics tuning within a compact device, it redefines what is achievable in neuromorphic hardware design and implementation.

The implications extend toward revolutionizing edge-computing devices, potentially influencing sectors ranging from security surveillance to wearable health monitors and adaptive robotics. This SnS₂ memristor device stands as a beacon of innovation, bridging material science breakthroughs with real-world computational needs and setting the stage for a new era of smart sensors that effectively mimic biological sensing and processing strategies at previously unattainable performance levels.

Subject of Research: Neuromorphic computing with SnS₂-based in-sensor reservoir devices for dynamic motion recognition

Article Title: Tunable Neuromorphic Computing for Dynamic Multi-Timescale Sensing in Motion Recognition

News Publication Date:

The synthesis of monolayer MoS₂ begins with the meticulous preparation of the precursors. In this instance, molybdenum trioxide (MoO₃) and sulfur serve as the primary components, facilitating the growth of MoS₂ during the CVD process. The researchers executed this step by distributing a precisely measured amount of MoO₃ powder within an alumina boat, strategically positioned at the center of a single-zone tube furnace. Complementing this setup was a SiO₂ substrate, which was treated with a sodium hydroxide (NaOH) promoter to enhance the growth process. The controlled environment further included the placement of sulfur powder in a separate alumina boat, carefully positioned 17 centimeters upstream, optimizing the reaction dynamics during synthesis.

Prior to initiating the growth phase, the tube furnace underwent a purging process to remove any impurities or contaminants. Here, nitrogen gas was introduced at a flow rate of 460 standard cubic centimeters per minute (s.c.c.m.) while the furnace was heated to a temperature of 150°C. Once the environment was purged, the nitrogen flow rate was reduced, establishing an ideal backdrop for the growth of MoS₂. The MoO₃ source was subsequently heated to an impressive 720°C, a critical temperature that allows the synthesis reactions to take place efficiently. Meanwhile, sulfur stabilized at a temperature of around 230°C, ensuring that the reaction remained active yet controlled, thus yielding high-quality monolayer MoS₂.

Upon completion of the growth phase, careful attention was paid to cooling the tube back down. To maintain the integrity of the synthesized material, the sulfur source was withdrawn from the heating zone, and the nitrogen gas flow was resumed at 460 s.c.c.m. This cooling protocol was instrumental in preserving the structural characteristics of MoS₂. For the transfer of this material to target substrates, the researchers employed a polydimethylsiloxane (PDMS) dry transfer technique. This method is crucial for ensuring the integrity of the material as it transitions from one substrate to another, marking a significant step in its practical deployment.

In addition to the monolayer MoS₂, the researchers also explored the characteristics of few-layered MoS₂ and tungsten diselenide (WSe₂) samples, which were obtained through mechanical exfoliation from bulk crystals. This process utilized the blue tape method, known for its simplicity and effectiveness in producing high-quality few-layer materials. The substrates designated for these transferred materials mirrored those used for the CVD-grown samples, thereby ensuring consistency across the experimental components.

The substrates themselves played a critical role in the overall study. The research team utilized boron degenerately doped silicon substrates, which were further enhanced with layers of thermally grown SiO₂, atomic-layer-deposition-grown hafnium oxide (HfO₂), and zirconium oxide (ZrO₂). Each of these dielectric materials presented unique electrical and structural properties essential for the successful integration of MoS₂ within electronic devices. Understanding the deposition conditions and characterizations of these dielectrics remains paramount, as they influence the performance and efficiency of the overall devices.

Moreover, to investigate the electronic properties and interface characteristics of the materials, soft and hard X-ray photoelectron spectroscopy (XPS) measurements were conducted. Utilizing the advanced capabilities at beamline I09 at the Diamond Light Source in the UK, the researchers collected detailed spectral information. This data enabled them to determine the elemental composition and electronic states present within the MoS₂ samples. Notably, the high-energy analyzer employed demonstrated precision, capturing spectra that were crucial for confirming the absence of sample charging and beam damage, thereby ensuring the reliability of their results.

The binding energy scale utilized during these measurements was calibrated against the gold 4f core level, lending credibility to the data acquired. A meticulous approach was taken, which included repeated acquisitions to substantiate the findings. This careful consideration of experimental conditions signifies the researchers’ commitment to achieving high-quality results, demonstrating best practices in the synthesis and characterization of 2D materials.

The theoretical framework underpinning this research involved first-principles density functional theory (DFT) calculations. Utilizing the QuantumATK package, researchers applied the hybrid functional of Heyd-Scuseria-Ernzerhof (HSE06) to understand the electronic interactions at the MoS₂/dielectric interfaces. The study focused on various interface models, including those involving HfO₂ and ZrO₂, enabling a comprehensive investigation into their potential as substrates for MoS₂ applications. The DFT calculations supported the experimental findings, corroborating the ideal lattice parameters and structural properties of the materials.

In constructing the interface models, researchers deployed supercell strategies to capture the nuances of the interactions between the 2D TMDs and the dielectric layers. Multiple configurations were tested to optimize the alignment of the lattices, ensuring minimal strain and maximizing the quality of the heterojunctions. The results of these calculations revealed vital insights into the nature of the interactions occurring at the interfaces, highlighting the significance of van der Waals forces in stabilizing the heterostructures.

More practically, the electrical measurements of the synthesized materials were conducted using a Keithley 4200 current-voltage system. The intricate characterization allowed for the assessment of material performance under various conditions, reflecting their potential for real-world applications. Furthermore, photoluminescence (PL) and Raman spectroscopy data collected using a focused laser revealed critical vibrational modes and electronic transitions within the MoS₂ layers, indicating their viability for optoelectronic devices.

In addressing the physical topography of the samples, Atomic Force Microscopy (AFM) data imagery was captured to reveal surface characteristics and layer thicknesses. Employing the Dimension Icon device in peak-force tapping mode provided insights into nanoscale features critical for device engineering. The culmination of these efforts has positioned this research at the forefront of material science, facilitating the design of cutting-edge electronic devices that may redefine industry standards.

Through meticulous research and breakthrough techniques in synthesis and characterization, this work lays the groundwork for understanding and leveraging the attributes of monolayer and few-layer MoS₂ in contemporary technology. The advancement of 2D materials not only emphasizes their unique electronic and optical properties but also enriches the realm of nanoscale engineering, heralding a promising future for advanced electronic devices.

Subject of Research: Synthesis and characterization of monolayer MoS₂ and few-layer TMDs for advanced electronic applications.

Article Title: A clean van der Waals interface between the high-k dielectric zirconium oxide and two-dimensional molybdenum disulfide.

Article References: Yan, H., Wang, Y., Li, Y. et al. A clean van der Waals interface between the high-k dielectric zirconium oxide and two-dimensional molybdenum disulfide.
Nat Electron (2025). https://doi.org/10.1038/s41928-025-01468-1

Image Credits: AI Generated

DOI: 10.1038/s41928-025-01468-1

Keywords: MoS₂, two-dimensional materials, van der Waals interfaces, CVD, XPS, DFT, semiconductor applications.

Delving into the methodology, the team expertly utilized in situ domain engineering coupled with controllable direct chemical vapor deposition (CVD) techniques. This innovative approach allows for the careful modulation of defect structures at the domain level, a previously elusive achievement in the synthesis of homojunctions. By conducting theoretical simulations, researchers identified optimal intrinsic defect configurations, paving the way for their experimental implementation.

The CVD process was intricately designed in two distinct phases. Initially, the researchers established two types of growth domains within hexagonal WS2 samples under equilibrium conditions. This strategy not only facilitated the growth of the material but also ensured a precise control over the structural characteristics that would ultimately influence the device’s performance.

In the subsequent phase, the researchers undertook an in situ manipulation of the atomic configurations specific to each domain. This critical step aimed at engineering the electronic band structures of the resulting homojunctions, which play a pivotal role in determining their electronic properties and overall functionality. By exploiting van der Waals interactions and lateral atomic bonding, they succeeded in integrating these structures without damaging their integrity.

A notable aspect of this work is the researchers’ foresight in controlling the epitaxial growth process itself. By manipulating the precursor feeding rates of tungsten trioxide and sulfur, they created a state of equilibrium that allowed the growth rate of S-zigzag edges to match that of W-zigzag edges. This nuanced control mechanism is essential for tailoring the properties of the homojunctions, ensuring that they exhibit desired characteristics while maintaining atomic precision.

The resulting WS2 homojunctions demonstrated a remarkable array of field-effect characteristics, standing out due to their impressive overlapping lattice match and customized band alignment at the interfaces. Such precision in engineering has significant implications for the future of electronic devices, where the performance hinges on the quality of the materials used.

One of the practical applications of these synthesized structures is the development of logic inverters. The research highlights that these inverters achieved rail-to-rail operation, delivering a peak voltage gain of up to 12. Moreover, the dynamic delay measured around 135 microseconds showcases the swift response times achievable with these new materials. Most impressively, the power consumption was recorded at a mere 1.3 nanowatts, underscoring the efficiency and sustainability of these systems.

Beyond the immediate implications for electronic applications, the study provides deeper insights into the realm of two-dimensional materials. It sheds light on the significance of defect engineering within atomic layers and the effect these configurations have on the performance of low-dimensional devices. This knowledge could pave the way for future innovations in material science, culminating in the realization of advanced devices that outperform current technologies.

Additionally, the researchers have positioned their findings within the broader context of semiconductor technology. As industries worldwide seek alternatives to traditional silicon-based systems, the development of semiconductors that operate efficiently at the atomic level is becoming critical. This study represents a crucial step in that direction and ignites further exploration into the untapped potential of monolayer materials for a new generation of electronics.

Furthermore, this research emphasizes the importance of interdisciplinary collaboration, bringing together theoretical insights and practical experimentation. The successful synthesis of monolayer WS2 homojunctions is not solely a victory for material science but also reflects advancements in chemical engineering and nanotechnology, opening up avenues for educational initiatives and collaborative projects across institutions.

In conclusion, the synthesis of monolayer WS2 lateral homojunctions heralds a new era in the landscape of electronic materials. The ability to engineer defect structures and manipulate them in subsequent growth phases sets a precedent for future research endeavors aiming to push the boundaries of what is possible with two-dimensional materials. As researchers continue to explore the potential of these materials, the implications for the electronics industry are immense, promising a future where advanced, efficient, and sustainable devices become the norm.

Subject of Research: Monolayer WS2 lateral homojunctions synthesis
Article Title: USTC Reports Epitaxy Growth of Semiconducting Monolayer WS2 Lateral Homojunctions
News Publication Date: 13-Jun-2025
Web References: https://doi.org/10.1021/jacs.5c04546
References: 10.1021/jacs.5c04546
Image Credits: Image by USTC

Keywords

Monolayers, Transition Metal Dichalcogenides, Epitaxy Growth, Field-Effect Transistors, Defect Engineering, Chemical Vapor Deposition, Electronic Devices, 2D Materials.