Research update | 8 October 2026 — Making an atomically thin semiconductor is only part of the challenge. For practical circuits, manufacturers also need crystals to begin growing in predictable positions. A peer-reviewed study published in Nature on 7 October 2026 reports a way to steer the starting point of molybdenum disulfide (MoS₂) crystal growth by using a nearby oxide layer that releases an etching species during deposition.
What changed?
Conventional area-selective growth can determine which patterned regions receive a semiconductor, but the first crystal nucleus may still appear at an unpredictable spot inside each region. When several nuclei appear, they can grow into separate domains and introduce grain boundaries. The new approach, called etching-flux-mediated single-centred nucleation, is designed to suppress those competing starting points.
The researchers patterned a growth region next to a hafnium oxide (HfO₂) barrier. During metal–organic chemical vapour deposition, oxygen released by the barrier creates a lateral chemical environment that discourages MoS₂ nucleation near the edges. Under suitable conditions, the centre remains the preferred place for one nucleus to survive and expand. The key advance is therefore not merely selective coating: it is control over where a crystal starts within a selected area.
What did the experiments demonstrate?
- Positioned single crystals: single-crystal MoS₂ features on roughly the 10-micrometre scale, with controlled nucleation positions.
- Reproducibility: more than 99% single-crystal yield across 400 patterned sites distributed over 16 regions of a 2-centimetre substrate in the reported experiment.
- Electrical performance: field-effect mobility as high as 117 cm² V⁻¹ s⁻¹ in the measured devices. This is a reported maximum, not a guaranteed production specification.
- Process flexibility: a related zirconium oxide barrier enabled centred growth at 430 °C; the authors also showed line-shaped crystals, multiple transistors within a larger crystal and aligned lateral heterostructures.
These measurements support the authors’ central claim that the chemical environment around a growth window can be engineered to control nucleation rather than leaving the position to chance. They do not establish that full commercial wafers or mass-produced circuits are already available.
Why this matters for semiconductor manufacturing
Two-dimensional transition-metal dichalcogenides are being investigated for transistors and other compact electronic devices. Their extremely thin channels may be useful where conventional scaling faces electrostatic or integration challenges. Yet a material that performs well in an isolated laboratory device is not automatically suitable for a densely patterned circuit. Crystal placement, orientation, consistency and compatible process temperatures all affect whether thousands or millions of devices can be fabricated predictably.
A process that reliably initiates one crystal in a chosen location could help reduce unwanted grain boundaries and simplify the alignment of semiconductor regions with contacts and neighbouring devices. The demonstrated lateral heterostructures also suggest a route to more deliberate interfaces between different 2D materials.
Possible applications
The most direct research direction is patterned 2D transistor arrays, including devices built from MoS₂. Other plausible directions include aligned 2D material junctions, integrated sensing structures and experiments that require reproducible crystal placement. These are potential applications of the manufacturing principle, not products validated by this paper.
Development stage and limitations
This is a peer-reviewed laboratory demonstration, not a commercial fabrication qualification. The reported uniformity test covers 400 sites on a 2-centimetre substrate, which is meaningful but far short of a production-scale wafer yield assessment. The oxide barrier, oxygen flux and growth conditions must be carefully matched: an etching environment strong enough to suppress unwanted nuclei can also prevent growth if poorly controlled.
Further work is needed to demonstrate consistent crystal orientation across different substrate types, large-wafer throughput, integration with contacts and dielectrics, device-to-device variability, long-term reliability and cost competitiveness. In particular, the paper notes that orientations on amorphous substrates can vary from site to site, even when each individual flake is single-crystalline.
Element Folio perspective
The wider lesson is that semiconductor fabrication may benefit from controlling the chemistry around a desired crystal as carefully as the material deposited inside it. This research moves beyond defining growth areas toward programming the first step of crystal formation—a useful distinction for anyone tracking the path from promising nanomaterials to manufacturable electronics.
Primary source
Jeongwon Park and colleagues, “Spatially deterministic nucleation of 2D semiconductors by etching flux,” Nature 658, 398–406 (published 7 October 2026). DOI: 10.1038/s41586-026-11095-1. Read the original open-access paper.
Leave a Reply