Semiconductors — Materials Science & Microelectronics

A Hidden Atomic Gap Could Break Next-Generation Chips

Ultrathin 2D materials look ideal for the next generation of chips — until they are stacked with insulating layers, where a microscopic gap quietly kills their performance.

Decades of progress in computing have come from squeezing more power into smaller components. As silicon approaches its physical limits, researchers have turned to two-dimensional materials — single or double atomic layers of graphene, molybdenum disulfide, and similar compounds — as candidates for the next generation of transistors. Their thinness should let devices shrink far beyond today's limits.

A team at the Vienna University of Technology found that the promise breaks down at the interface. A transistor needs more than a conducting layer: it needs an insulating oxide beneath it, separated from a gate electrode that switches the device on and off. When a 2D material is paired with that insulator, the two are held together only by weak van der Waals forces. At the atomic scale, that leaves a tiny but real separation between the layers. Electrons feel the gap, and the device's performance drops sharply.

The implication is expensive. Semiconductor firms could spend billions developing chip approaches that hit this physical wall before they even reach production. The TU Wien team argues that designers need to treat the 2D-to-insulator interface as a first-class engineering problem, not an afterthought.

The proposed fix is a "zipper material": an intermediate layer chemically engineered to bind more strongly to both the 2D conductor and the insulator, pulling the two into close contact and suppressing the gap. If the bonding can be made strong enough, the ultrathin advantage of 2D materials survives in real chips rather than only in simulations.

Knowledge takeaway: 2D materials lose their advantage when a microscopic gap forms at the insulator interface; the gap stems from weak van der Waals bonding and degrades electron performance; "zipper materials" designed to lock the layers together could preserve 2D miniaturization in real chips.