Trying to mimic the feeling of human touch has been a puzzle in robotics for decades. Most designs stopped at detecting pressure or shear, leaving robots with a limited sensory palette. Others used chemical detection, relying on rigid or invasive sensors unsuited to flexible, deformable systems.
How It Works
The artificial skin is built around a sponge-like network of CNT foam embedded in a flexible silicone (PDMS) film.
The CNT foam makes the sensor conductive, elastic, and sensitive to deformation. When pressed, its electrical resistance shifts, translating force into measurable signals with high precision. Optical fibers are threaded through the film/foam structure and deliver near-infrared (NIR) light, capturing its reflections and revealing the chemical fingerprints of whatever surface or tissue the robot is touching.
The researchers used Monte Carlo simulations to model how light travels through different materials, preparing the sensor for anything from fruit pulp to biological tissue. The result is a sensor that can detect pressure changes in real time while reading out molecular compositions without sacrificing flexibility or stretchability.
Download your PDF now!
From Medicine to Agriculture and More
The oe-skin’s applications range widely. In medicine, it could help robots manipulate delicate tissues without causing damage, and simultaneously monitor intraocular pressure, or even measure glucose levels. In agriculture, robotic harvesters could assess fruit ripeness, firmness, and sugar content during processing, reducing waste and improving efficiency.
In both cases, the combined tactile and chemical feedback reduces the need for multiple sensors, streamlining design while enriching information.
The CNT framework can be further functionalized with specific biomolecules to detect disease markers, while extending the optical spectrum into ultraviolet or mid-infrared would broaden chemical coverage.
The researchers also envision combining the skin with machine learning, allowing robots to sense, interpret trends, and act independently.
A Step Toward Human-Like Perception
There are still challenges that stand in the way of mass adoption of the sensor: resolution, shear-force detection, and multiplexing need refinement.
But this advance marks a step toward robots that can sense their environments with a subtlety approaching and, in some instances, surpassing human skin. By combining touch and chemistry into one system, the oe-skin may transform robotics in healthcare, agriculture, and hazardous environments alike.
Journal Reference
Dai B., et al. (2025). An optical/electronic artificial skin extends the robotic sense to molecular sensing. npj Flexible Electronics 9, 87. DOI: 10.1038/s41528-025-00431-6, https://www.nature.com/articles/s41528-025-00431-6