An open-access Nature paper published October 7 reports a way to control where two-dimensional semiconductor crystals start growing. The authors call the process etching-flux-mediated single-centred nucleation. Their experiments address a manufacturing problem: crystals can begin at random locations even when deposition is confined to a selected region.
The method uses oxygen released from an oxide barrier to create a sideways etching flux. That suppresses new crystal formation near the boundary of a patterned growth region, leaving its geometric center as the surviving starting point. The study separates control of the initial crystal from its subsequent growth.
The researchers demonstrated single-crystal molybdenum disulfide at the 10-micrometer scale and working field-effect transistors. They report field-effect mobility up to 117 square centimeters per volt-second. That measures how readily charge moves through the tested material; it is not a processor speed or a finished-product performance rating.
For a larger-area experiment, the team fabricated 400 patterned sites across 16 regions on a two-centimeter substrate. The paper reports a single-crystal yield exceeding 99%. This is a yield for the study's patterned crystal-growth sites, not a commercial wafer yield or proof of mass production.
The paper also describes growth on an amorphous substrate and adapting the barrier material for processing at 430 degrees Celsius. Its demonstrations include multiple transistors within a common crystal and aligned lateral semiconductor structures. These are reported research capabilities rather than features available in a shipping chip.
Limits: the study offers a possible path toward future two-dimensional electronic integration. It does not announce a factory, customer shipment, production date or replacement for current silicon chips. October 11 reporting drew renewed attention to the work, but the original paper's publication date remains October 7.