As transistors shrink towards atomic dimensions, even the boundary between two materials can become an engineering problem.
Researchers at National Yang Ming Chiao Tung University (NYCU), working with TSMC Corporate Research, have developed an insulating interface just 0.42 nanometres thick that could help improve transistors made from atomically thin semiconductors.
Materials such as molybdenum disulfide (MoS₂) are being explored as possible alternatives to conventional silicon because they can form semiconductor layers only an atom thick. But building other parts of a transistor on top of these delicate materials isn’t straightforward.
One particular challenge is the gate dielectric, the insulating layer that helps the gate control the movement of charge through the transistor. Making this layer thinner improves electrical control, but depositing it onto MoS₂ can introduce defects and electrical disorder that interfere with electron movement.
Advertisement
Rather than redesigning the semiconductor, the researchers concentrated on the interface itself. They deposited an ultrathin layer of aluminium onto monolayer MoS₂ and oxidised it, creating an aluminium oxide layer around 0.42 nm thick. A high-κ hafnium oxide gate dielectric was then added above it.
The aluminium oxide acts as an atomic-scale buffer. It provides a more uniform surface for the hafnium oxide while reducing unwanted interactions with the semiconductor underneath.
Using the approach, the team fabricated top-gate transistors from CVD-grown monolayer MoS₂ with an equivalent oxide thickness of around 1 nm. Devices with channels of roughly 100 nm achieved maximum transconductance of 0.45 mS/μm, alongside low leakage current and minimal hysteresis.
Importantly, the work combines extremely thin dielectric scaling with strong electrical control while maintaining carrier transport – properties that have proved difficult to achieve together in two-dimensional transistors.
Published in Nature Electronics, the research shows just how important interfaces are becoming as electronics approach atomic dimensions. Future progress may not simply depend on finding a better semiconductor. Sometimes, the crucial engineering could happen in the fraction of a nanometre between two materials.
Advertisement
Research published in Nature Electronics.