The research centered on an innovative approach that leverages the complementary electrical and optical properties of tungsten oxyselenide (WSeOx) and graphene, two materials renowned for their exceptional characteristics at the nanoscale. Traditionally, phase modulators have struggled to balance low energy consumption, high speed, and minimal signal degradation, often resulting in efficiency bottlenecks that hinder widespread adoption in miniaturized devices. By constructing hybrid electrodes from these two-dimensional materials, the team overcame these limitations, demonstrating a leap forward in modulation fidelity and energy efficiency.

At its core, the system exploits the strong light-matter interactions intrinsic to two-dimensional semiconductors, enabling dynamic control over the phase of light waves traversing ultrathin photonic circuits. The tungsten oxyselenide layer contributes a tunable electronic environment due to its unique band structure, which facilitates modulation through strain and charge density variations. Meanwhile, graphene acts as an exceptional conductor and transparent electrode, ensuring minimal resistive losses and rapid electronic response. The synergy of these materials results in an electrode platform that enables electrically driven phase shifts without the usual penalty of optical signal attenuation.

Beyond demonstrating fundamental compatibility, the research delved into the fabrication challenges associated with integrating WSeOx and graphene at the nanoscale. Employing state-of-the-art chemical vapor deposition and transfer techniques, the team successfully engineered a smooth, defect-free interface that maintains high carrier mobility. Precise control over thickness and interfacial properties was key to optimizing the modulator’s performance, as any imperfections at the atomic layer junction could introduce scattering and dissipative effects detrimental to near-lossless operation.

The experimental setup revealed phase modulation efficiencies far surpassing those of traditional modulators, achieving a figure-of-merit that approaches the theoretical upper limit. Specifically, the devices exhibited ultra-low insertion losses and modulation depths tunable over a wide wavelength range in the visible-to-near-infrared spectrum. Such versatility accentuates their applicability in diverse photonic systems, ranging from integrated optical interconnects to programmable meta-surfaces and dynamic holography.

A critical insight emerged from thorough spectroscopic and electrical characterization of the devices, which uncovered how subtle interactions at the heterostructure interface influence the carrier dynamics and optical response. The researchers utilized advanced scanning near-field optical microscopy (SNOM) alongside electrical transport measurements to unravel the mechanisms governing the phase modulation process on a nanoscale level. These revelations pave the way for further refinement of material properties through doping and strain engineering.

The implications of this work extend well beyond the lab, with potential to revolutionize telecommunications infrastructure by reducing signal distortion and power requirements. The unprecedented combination of low loss and high-speed operation enables the development of compact, on-chip photonic components that deliver enhanced bandwidth and reduced latency, critical parameters for 5G and forthcoming 6G networks. Moreover, the tunability of these hybrid electrodes allows for real-time adaptive photonic circuits capable of responding to changing environmental or computational demands.

Further, the marriage of tungsten oxyselenide and graphene introduces pathways for embedding quantum coherent control into classical photonics. The near-lossless modulation sets the stage for integrating these components into quantum photonic devices, where preserving the coherence of quantum states over extended times is essential. The prospect of electrically controlled phase shifters operating at room temperature marks a significant milestone towards scalable quantum computing architectures and secure quantum communication channels.

From a materials science perspective, the study offers valuable insights into the design principles governing two-dimensional semiconductor heterostructures with electronic and optical multifunctionality. It underscores the importance of interfacial engineering, chemical stability, and defect passivation in realizing high-performance nanodevices. The unique properties of tungsten oxyselenide, in particular, invite further exploration of other transition metal chalcogenide oxides as potential candidates for hybrid photonic applications alongside graphene and related carbon allotropes.

To translate these laboratory successes into practical technologies, scalability and integration challenges remain focal points for ongoing research. Ensuring reproducible, wafer-scale fabrication of hybrid WSeOx/graphene electrodes compatible with existing semiconductor manufacturing is imperative for commercial viability. Concurrently, developing comprehensive modeling frameworks that capture the coupled electro-optic phenomena at play will assist in optimizing device architectures tailored for targeted applications.

The discovery also aligns with broader trends in utilizing two-dimensional materials to achieve multifunctional optoelectronic systems that combine sensing, modulation, and signal processing within minimal footprints. This integration supports the increasing demand for miniaturized and energy-efficient components essential for portable and wearable technologies, including augmented reality displays and biomedical imaging devices. The low power consumption and high speed of these modulators could dramatically extend device lifetimes and enhance user experiences.

In summary, the development of hybrid tungsten oxyselenide/graphene electrodes represents a significant breakthrough in the field of two-dimensional semiconductor phase modulators. By achieving near-lossless modulation, the researchers have unlocked new opportunities for high-performance photonic devices that are faster, smaller, and more energy efficient than their predecessors. This innovation stands as a testament to the power of materials hybridization at the atomic scale to overcome long-standing limitations in photonics and electronics.

As the pace of discovery accelerates, the fusion of novel two-dimensional materials with advanced fabrication methods promises to redefine the boundaries of optical communication and computation. The intricate interplay between electronic structure, optical properties, and interface phenomena showcased in this work will inspire a new generation of devices that harness the unique capabilities of low-dimensional systems. Consequently, we can anticipate rapid advancements in integrated photonics that will permeate diverse technological sectors globally.

Ultimately, this work embodies the convergence of material science, nanotechnology, and applied physics, delivering a platform with far-reaching implications. From enhancing global data transmission infrastructure to enabling cutting-edge quantum information systems, hybrid tungsten oxyselenide/graphene phase modulators poised at the frontier of scientific innovation may well shape the photonic world of tomorrow.

Subject of Research:
Hybrid two-dimensional semiconductor electrodes combining tungsten oxyselenide and graphene for advanced phase modulation applications.

Article Title:
Hybrid tungsten oxyselenide/graphene electrodes for near-lossless 2D semiconductor phase modulators.

Article References:
Guo, S., Lee, SG., Gong, X. et al. Hybrid tungsten oxyselenide/graphene electrodes for near-lossless 2D semiconductor phase modulators. Light Sci Appl 15, 42 (2026). https://doi.org/10.1038/s41377-025-02058-8

Image Credits: AI Generated

DOI: 10.1038/s41377-025-02058-8

The essence of this advancement lies in the unique properties of monolayer MoS₂, a material celebrated for its remarkable electrical and mechanical attributes. The minimal transmission loss of just 0.51 dB in the MoS₂ channel indicates exceptional efficiency, positioning it as a prime candidate for future microwave applications. This low loss translates into more reliable communication with reduced power wastage, aligning perfectly with the contemporary demands for energy-efficient systems. Moreover, the experimental design has granted us insights into how these two-dimensional semiconductors can be harnessed to create streamlined and cost-effective solutions for integrated microwave systems.

In terms of power consumption, the complete 16-element transmitter achieves a mere 3.2 µW, a staggering feat given the complexity of such technology. This commendable power efficiency is not just a technical milestone but also speaks to the sustainability of future communication systems. As engineers and designers grapple with balancing performance against energy use, the innovations witnessed in this study exemplify how utilizing advanced materials like MoS₂ can yield significant results. The ability to operate efficiently with minimal energy expenditure could have far-reaching consequences across various sectors, including telecommunications, aerospace, and defense.

Central to the highlights of this development is the remarkable capabilities of a 4 × 4 phased array transmitter that not only provides communication functions but also integrates radar functionalities. This dual-purpose approach could significantly enhance capabilities in both civilian and military applications, leading to diversified uses of the technology. The device boasts a bandwidth of 6 GHz, an impressive feature that enhances its utility for both high-data-rate communication and accurate detection, making it a highly versatile module. Applications in autonomous vehicles, robotics, and smart city technologies could be equally optimized using this advanced semiconductor technology.

The practical performance of the transmitter is equally noteworthy, demonstrating a beam scanning angle ranging from -35° to 35°. This feature allows for dynamic alignment and signal directing based on real-time needs, paving the way for applications that require adaptability and precision. Whether it is for radar signal reception or directing communication waves, the flexibility of this transmitter is a strong selling point. Such adaptability is crucial in environments that require successful interference rejection and effective signal management, highlighting the Symphony of technology and versatility offered by these integrated systems.

Transmission distance is another critical performance metric, with an impressive operational range of 136 meters underlinable in real-world usage scenarios. This range necessitates thorough exploration into the implications of this capability in practical applications. For instance, in smart environments where devices need to communicate over significant distances without compromising data integrity, the presented technology provides a robust solution. Additionally, considering the challenges posed by urban landscapes or interference-prone areas, this transmitter’s performance aligns well with the need for resilience in communications.

Battery life also plays a crucial role in the sustainability of any low-power technology. The integration of this transmitter with a 1,000 mAh-capacity battery results in an astounding standby time of 26 days. Such longevity minimizes the frequency of recharging and maintenance, which is particularly advantageous for applications in remote or hard-to-reach areas. Creating devices that can last longer while maintaining performance enhances user experience and operational reliability, representing a significant step forward in electronic design.

The compact dimensions of the complete board-level system, approximately 3 x 2 cm², accentuate its suitability for applications that demand minimal footprint without sacrificing functionality. In a world where miniaturization of technology has dominated trends, the ability to integrate such powerful capabilities into a small device opens the door to numerous innovative applications. This compactness motivates designers to explore embedding these transmitters in robotic systems, wearable devices, and even miniaturized aerial platforms, thus broadening the horizons for technological innovation.

The research team led by Wu, Zhu, and Dong have successfully demonstrated a prototype that illustrates the potential of monolayer MoS₂ in practical scenarios. This achievement marks a crucial development in the realm of microwaves and advanced materials. However, the journey toward commercial and widespread applications is only just beginning. As researchers continue to investigate and refine these technologies, the translational efforts required to bring them to the marketplace will rely heavily on collaboration across disciplines and sectors.

Looking ahead, it is evident that integrated two-dimensional microwave transmitters have the potential to redefine how communities engage with technology. The implications extend beyond mere enhancement of communication; they encompass collaborative platforms that foster connectivity across multiple devices and systems. Improved energy efficiency, adaptability, and practical applications stand to transform various industries, heralding a new chapter in the evolution of electronic systems. The path ahead, while promising, will undeniably involve challenges as stakeholders navigate technologies in the pursuit of scalability and applicability.

As always in the realm of scientific discovery, the dialogue between researchers and engineers will be essential to crystallizing these advancements into market-ready solutions. The implications of these findings could catalyze shifts in industry standards, reinforcing the need for low-power solutions in an interconnected world. The groundwork laid by this innovative study is more than just a technological achievement; it is a call to action for a generation of engineers and technologists poised to innovate and disrupt conventional models of communication.

Through combined scientific rigor and inventive spirit, the integrated two-dimensional microwave transmitters represent the future of efficient communication systems. By leveraging the properties of two-dimensional materials like MoS₂, researchers are not just addressing current challenges but are also paving the way for a future where communication is seamless, instantaneous, and remarkably efficient. As these technologies advance, we can expect to witness transformative changes that not only improve how we communicate but also how we interact with and perceive the world around us, reshaping relationship dynamics in an increasingly digital landscape.

In summary, the development of integrated two-dimensional microwave transmitters using monolayer MoS₂ has opened new vistas in low-power communication technologies. The impressive performance metrics and innovative design of these transmitters point to a brighter, technologically-driven future characterized by efficiency, versatility, and miniaturization. As researchers continue to unlock the potential of two-dimensional materials, the next frontier in electronic systems is set to break conventional boundaries, ushering in a new age of connectivity and communication.

Subject of Research: Development of Integrated Two-Dimensional Microwave Transmitters

Article Title: Integrated two-dimensional microwave transmitters fabricated on the wafer scale

Article References:

Wu, T., Zhu, L., Dong, X. et al. Integrated two-dimensional microwave transmitters fabricated on the wafer scale.
Nat Electron (2025). https://doi.org/10.1038/s41928-025-01452-9

Image Credits: AI Generated

DOI:

Keywords: Two-Dimensional Materials, Microwave Technology, Telecommunication, Molybdenum Disulfide, Integrated Circuits, Low Power Consumption, Advanced Electronics.

The foundation of this breakthrough lies in the principle of space-division multiplexing (SDM), a technique that capitalizes on the ability of a multimode fiber to carry multiple spatial modes simultaneously. Unlike single-mode fibers that transmit data through a single spatial channel, MMFs multiplex orthogonal transverse modes which can theoretically multiply the bandwidth capacity severalfold. However, the complex interplay between modes during propagation inside the fiber causes random mode coupling. This coupling results in intricate speckle patterns, which scramble the transmitted signals and render traditional coherent detection methods insufficient for accurate mode recovery. Consequently, the problem of mode demultiplexing remains an imposing hurdle due to the overwhelming computational demands associated with conventional digital signal processing (DSP) algorithms.

Enter the FPGA-accelerated mode decomposition engine, a solution ingeniously crafted to tackle this complexity with both speed and precision. At its core, the system utilizes a customized convolutional neural network (CNN) specifically trained on extensive synthetic datasets to predict the amplitude and relative phase of each mode directly from a single intensity image captured at the fiber output. This approach cleverly circumvents the need for coherent detection, a method traditionally demanding complex hardware setups and significant power consumption. By leveraging neural inference, mode decomposition becomes a matter of computational pattern recognition, streamlined through the power of deep learning.

Implementation of this CNN on a field-programmable gate array (FPGA) marks a vital innovation in energy-efficient and high-throughput signal processing. FPGAs offer a unique balance of configurability and low power consumption, facilitating real-time neural network inference at speeds unattainable by conventional graphics processing units (GPUs) without substantial energy costs. The research team reports an inference throughput exceeding 100 frames per second while consuming merely 2.4 watts of power. Such performance is transformative when juxtaposed with GPU-based approaches, which can draw tens of watts for similar tasks, thereby making this technology eminently suitable for practical deployment in industry- and medically relevant applications where energy and thermal budgets are severely constrained.

Experimental validation of the concept took place within a meticulously designed optical test bench. This setup combines a spatial light modulator (SLM) that synthesizes tailored superpositions of spatial modes, a precision six-axis fiber coupling stage enabling controlled mode excitation, and a high-sensitivity infrared camera that captures the resulting speckle patterns after transmission through the MMF. The captured intensity images are streamed in real time to the on-board FPGA, where the CNN inference operates to reconstruct the complex field of up to six spatial modes. Empirical results showcase reconstruction fidelities surpassing 97%, underscoring the method’s reliability and robustness under realistic experimental conditions.

A cardinal innovation underpinning this achievement is the network’s capability to resolve the notorious phase ambiguity inherent in intensity-only measurements. Typically, training neural networks solely on intensity patterns encounters an inherent global phase uncertainty since intensity measurements eliminate absolute phase information. This uncertainty often results in ambiguous outputs that compromise the physical relevance of the reconstructed fields. The team’s novel strategy involves harnessing the relative phases of higher-order modes to nullify this global phase ambiguity. This ensures the output is both unique and physically meaningful, even as the global phase drifts unpredictably, enhancing system stability and interpretability.

Beyond high-speed and high-fidelity mode recovery, the approach’s integration into an FPGA platform offers outstanding versatility in terms of system integration. The compact form factor and reconfigurability of FPGAs allow these engines to be embedded directly into optical communication hardware, medical imaging devices, or fiber sensor systems. Such embedded implementations pave the way for next-generation closed-loop adaptive optics setups, where rapid feedback on the optical field can dramatically enhance performance. Likewise, ultra-dense SDM links leveraging this technology promise to push the limits of fiber network throughput, addressing ever-escalating data demands with scalable hardware.

The implications of this research extend into several cutting-edge technology domains. For fiber-optic sensing applications, where precise phase information is critical for detecting subtle environmental changes, the FPGA-accelerated mode decomposition could enable vibration-tolerant, low-latency interrogation schemes. In biomedical optics, particularly endoscopic imaging, the method’s rapid and energy-efficient phase retrieval could enhance imaging clarity and speed, enabling real-time diagnostic procedures that were previously limited by slower computational algorithms or bulkier equipment. Consequently, the intersection of advanced photonics and embedded machine learning here represents a paradigm shift with broad societal relevance.

At its essence, this work exemplifies how deep neural networks, when thoughtfully integrated with hardware accelerators like FPGAs, can transcend traditional limitations of optical signal processing. The synergy between synthetic data-driven training and physically informed network design not only accelerates computation but also grounds the reconstructions in realistic optical physics. This balance is critical for translating laboratory successes into practical solutions capable of redefining the infrastructure of global communications and sensing architectures.

The study, entitled “FPGA-accelerated mode decomposition for multimode fiber-based communication,” was carried out by Professor Czarske’s team with co-first authors doctoral candidate Qian Zhang and graduate student Yuedi Zhang. Published in the distinguished journal Light: Advanced Manufacturing, this work not only opens pathways toward next-generation optical fiber technologies but also exemplifies the fertile convergence of photonic engineering and machine learning acceleration.

As data traffic continues its inexorable ascent, innovations like this FPGA-accelerated mode decomposition engine will be pivotal in ensuring that the communications backbone of the digital age remains agile, scalable, and efficient. With the ability to overcome longstanding computational barriers and integrate seamlessly into compact devices, this technology stands poised to transform both how we transmit information and how we interface with the optical domain in real time.

Subject of Research: FPGA-accelerated deep learning for real-time mode decomposition in multimode fiber communications

Article Title: FPGA-accelerated mode decomposition for multimode fiber-based communication

Web References:
10.37188/lam.2025.031

Image Credits: Qian Zhang, Yuedi Zhang et al.

Keywords

Multimode fiber, space-division multiplexing, mode decomposition, convolutional neural network, FPGA acceleration, optical communications, phase retrieval, deep learning, real-time signal processing, spatial light modulator, speckle pattern analysis, energy-efficient computing