Growing a layered crystal on a surface it barely sticks to. Because the bond is weak, the new layer does not have to match the spacing of the atoms underneath, so many more material combinations can be grown.
Going deeper
Van der Waals epitaxy grows a layered crystal on a surface it barely bonds to. Because nothing is stretched, lattice mismatch stops being the constraint – but orientation, nucleation and the substrate’s surface condition still decide how the film turns out.
Epitaxy without a lattice match
In conventional the film is chemically bonded to the , so every percent of lattice mismatch is stored as until it becomes cheaper to relieve it with a misfit dislocation. That is why the materials that can be grown on a given are essentially those whose nearly match it, and why so much of thin-film engineering is a search for compatible substrates.
Koma’s observation was that a layered crystal, whose surface has no , is not bound that way. It rests on the substrate through , and the potential it feels is weak and smooth. Mismatches of tens of percent then cost almost nothing, and the film relaxes to its own lattice constant almost immediately. The word epitaxy still applies because the substrate sets the film’s orientation – but only its orientation, not its spacing.
What weak bonding does not solve
The same weak potential that tolerates mismatch also aligns the film only loosely. For a three-fold-symmetric on a three-fold substrate, orientations 60° apart are often nearly degenerate in energy, so islands nucleate in both and meet where they grow together, leaving running through the film. Those boundaries are electrically active and are a leading suspect for the gap between and device performance.
A surface with nothing to bond to is also a surface that nucleates reluctantly. Islands are sparse, they take a long time to coalesce, and in practice growth often starts from whatever is not perfect about the substrate: steps, adsorbates, defects. Thermal expansion mismatch supplies a final complication – the film may be unstrained at the growth temperature and wrinkled by the time it reaches room temperature.
Relatives and how to report it
Two related techniques are often described with the same phrase. In quasi-van der Waals epitaxy the substrate does have dangling bonds, passivated by hydrogen or by a reconstructed surface, so the interaction is intermediate. In remote epitaxy a graphene interlayer is thin enough for the substrate’s electrostatic potential to reach through it, so the film is aligned by a crystal it never touches – and can then be peeled off and reused, which is the point.
Because the claim in all three cases is about orientation statistics rather than a single , the evidence has to be statistical: electron diffraction or dark-field imaging over many grains, the fraction of 0° and 60° domains, and a stated substrate preparation. A single well-oriented flake proves very little, since the weak potential that made the growth possible is also weak enough to be overridden by whatever the substrate happened to have on it that day.
For specialists
Growth of a layered crystal on a substrate to which it bonds only weakly, so large lattice mismatches are tolerated.