For more than a decade, atomically thin semiconductors have been seen as the most promising route beyond conventional silicon. These materials can be just a single atom thick while still offering impressive electrical properties. But one stubborn engineering problem has held them back: making the insulating gate dielectric thin enough without destroying electron flow.
Researchers at National Yang Ming Chiao Tung University (NYCU), working with TSMC Corporate Research, have now solved this problem with an elegantly thin solution — a buffer layer just 0.42 nanometers thick.
The Problem
A working transistor requires an extremely thin insulating layer called the gate dielectric. As transistors shrink, making this layer thinner improves electrical control. But adding such layers to atomically thin semiconductors disturbs the delicate interface between materials, scattering electrons and erasing performance gains.
For years, researchers faced a difficult tradeoff: they could strengthen gate control, or they could protect electron mobility. Achieving both at once proved elusive.
The Solution: Two Atoms of Aluminum Oxide
Instead of changing the semiconductor or the dielectric, the NYCU team redesigned the interface connecting them. They placed an ultrathin epitaxial aluminum layer directly onto monolayer molybdenum disulfide (MoS2), then carefully oxidized it to produce an aluminum oxide layer just 0.42 nanometers thick — roughly two atomic layers.
This engineered buffer performs two critical jobs: it creates a smooth surface for the hafnium oxide gate dielectric to grow uniformly, and it acts as an atomic shield that limits unwanted electrical interactions between the dielectric and the semiconductor.
The Results
The resulting transistors achieved an equivalent oxide thickness of roughly one nanometer, with low leakage current, minimal hysteresis, and strong electrostatic control — all while sustaining electron transport through the channel.
"When transistor components become only a few atomic layers thick, the interface is no longer simply the boundary between materials — it becomes an active part of the device," said Professor Tsung-En Lee, a corresponding author of the study.
Why It Matters
Because the researchers used CVD-grown monolayer MoS2 rather than mechanically exfoliated flakes, the approach brings the technology closer to processes suitable for wafer-scale manufacturing.
As the semiconductor industry looks beyond traditional silicon scaling, the findings suggest that future advances may depend not only on discovering new materials, but also on learning to precisely control the atomic boundaries that connect them.
The study was published in Nature Electronics.
