A recent review in npj 2D Materials and Applications explores how two-dimensional (2D) materials are shaping the development of neuromorphic and artificial sensory devices.
With properties that mimic biological neural functions, materials like graphene and transition metal dichalcogenides (TMDs) are emerging as strong candidates for next-generation hardware in AI and robotics.
Image Credit: NicoElNino/Shutterstock.com
Why 2D Materials Matter
Neuromorphic devices aim to replicate how biological neurons process and respond to information, often using memristive components that change resistance in response to stimuli.
2D materials are well-suited for these systems thanks to their atomically thin structure, high carrier mobility, and tunable electronic properties. These traits allow for rapid signal transmission and energy-efficient operation.
Their mechanical strength and chemical stability also make them ideal for flexible, wearable technologies, which are key in building neural interfaces and sensory platforms that mimic human perception.
Highlights from Recent Research
The review highlights a wide range of recent studies demonstrating the versatility of 2D materials in neuromorphic applications.
One example involves MoS₂ heterostructures used in charge-trapping synaptic transistors. These transistors can exhibit both volatile and non-volatile memory behaviors, mimicking short-term and long-term synaptic plasticity.
In another study, a MoSe₂/Bi₂Se₃ heterostructure enabled artificial synapses to respond to varying light wavelengths and intensities, supporting optical signal processing for neuromorphic computing. Devices that process optical stimuli directly are particularly valuable for developing artificial sensory and visual systems.
Other investigations have explored electrochemical modulation in materials such as WSe₂, enabling functions similar to human senses like taste, touch, and smell. These devices typically operate at extremely low power, often in the femtojoule range, and show strong endurance and data retention, making them suitable for real-world use.
The review also covers efforts to integrate 2D-based memristors and artificial synapses into complex architectures. Materials like graphene, hexagonal boron nitride (h-BN), and WSe₂ have demonstrated essential characteristics for learning and memory tasks, including high linearity, broad dynamic range, and multi-level switching.
Beyond individual devices, the review discusses progress in building integrated neuromorphic systems that combine arrays of artificial synapses and sensors. These systems are designed to replicate human sensory experiences—vision, touch, hearing, and smell—with high accuracy and low energy use. For example, artificial visual synapses inspired by biological systems have been developed to deliver fast operation, high spatial resolution, and minimal signal interference, key traits for next-generation vision platforms.
Discussion
The review underscores that the usefulness of 2D materials in neuromorphic devices comes from their unique physical and chemical properties. Their atomic-scale thickness allows for fast charge transport and low energy consumption, key features for efficient neuromorphic computing.
In addition, their properties can be finely tuned by stacking layers or adjusting composition, enabling devices tailored to respond to specific types of sensory input such as light, pressure, sound, or chemicals. This flexibility makes them strong candidates for multifunctional artificial sensory systems that combine vision, touch, hearing, and smell in a single platform.
However, transitioning from lab-scale demonstrations to practical technologies requires overcoming several manufacturing challenges. Achieving large-area, defect-free synthesis remains a major bottleneck. Current methods, such as CVD and PVD, often introduce grain boundaries and impurities that reduce performance and device consistency. The complexity of building uniform, reliable heterostructures further complicates scalability.
Another major hurdle is integrating individual devices into larger, interconnected neural networks. Building dense arrays with stable, high-performance behavior is necessary to replicate brain-like processing. For real-world applications, such as in robotics, wearables, and autonomous systems, this level of integration must also meet commercial standards for durability and environmental stability.
To move forward, the review emphasizes the need for continued innovation in both material engineering and fabrication techniques. Refining deposition processes, improving transfer methods, and developing robust architectures will be essential. Just as important is the design of systems that not only emulate neural circuits but also scale effectively and operate reliably over time.
Ultimately, 2D materials hold strong potential to enable highly efficient, adaptable neuromorphic systems that bring artificial sensory processing closer to biological performance. Continued progress in synthesis, stability, and large-scale integration will be key to unlocking their use in next-generation AI hardware and smart sensing technologies.
Journal Reference
Ko J., et al. (2025). Two-dimensional materials for artificial sensory devices: advancing neuromorphic sensing technology. npj 2D Materials and Applications, 9, 35. DOI: 10.1038/s41699-025-00556-2,
