4. Next-Generation Transistors

Overview

A layered crystal thinned to a few atomic layers behaves differently from its bulk, and the growth process is where that thickness, and the physics that follows from it, is decided.

In a layered crystal the atoms bond strongly within each plane and only weakly between planes. Thinning it to a few layers changes its band structure, its dielectric screening, and the way its carriers and spins couple, so properties the bulk does not show can appear in the film.

These materials are also where thin film growth is hardest. Chalcogens are volatile, so stoichiometry drifts with substrate temperature. Weak interlayer bonding gives little driving force for registry with the substrate, so orientation has to be imposed by the growth conditions. And the surface that carries the interesting physics is the same surface that degrades in air. Growth sets the structure, stoichiometry, and defect chemistry of the film. Interface engineering sets what the film touches, and therefore what its transport looks like. We work on both.

Hybrid PLD Growth

Pulsed laser deposition transfers the composition of a target to the substrate, but a chalcogen such as Se or S desorbs from a hot substrate before it incorporates, and the film ends up metal rich. Our hybrid system supplies the metal flux by laser ablation and the chalcogen flux from a separate thermal source, so the chalcogen activity during growth is set independently of the metal arrival rate. This decouples film stoichiometry from the target and gives epitaxial films of MoS₂, VSe₂, HfSe₂, Bi₂O₂Se, and InSe on single crystal substrates, with chalcogen content as a growth variable rather than a fixed outcome.

[Figure: hybrid PLD chamber schematic, laser plume from the metal target plus chalcogen vapor from a heated source, converging on the heated substrate]

Emergent Properties

Magnetic order in VSe₂

Bulk 1T-VSe₂ is a paramagnetic metal. In epitaxial films the interlayer spacing can be modulated through the growth conditions, which changes the coupling between layers and with it the magnetic ground state. We use this handle to study how magnetic order emerges in a layered metal that does not order in the bulk.

[Figure: XRD (00l) shift vs growth condition alongside the corresponding M–H loops]

Layered semiconductors

HfSe₂ has a band gap close to that of silicon and a native oxide that is itself a high-k dielectric, but it degrades within minutes in air. Bi₂O₂Se is a layered oxyselenide in which [Bi₂O₂] and Se planes are bound electrostatically rather than by van der Waals forces, which gives it an air-stable surface and high electron mobility; its controlled oxidation forms Bi₂SeO₅, a high-k dielectric that grows directly on the channel. We study how growth and post-growth oxidation set the carrier density, mobility, and dielectric interface of these films.

Interfaces and Devices

A few-layer film is nearly all surface, so its transport is set by what touches it. We engineer metal contacts for band alignment and grow passivation layers such as amorphous BN in situ, without breaking vacuum, so the film never sees air before it is encapsulated. Field-effect transistors built on these stacks let us measure the intrinsic properties of the films, mobility, carrier type, and band alignment, and test whether they survive in a working device.

[Figure: cross-section of the a-BN/HfSe₂/Al₂O₃ stack with Bi contacts, next to a transfer curve]