Atomic layer epitaxy of aluminum nitride: Unraveling the connection between hydrogen plasma and carbon contamination

Atomistic control over the growth of semiconductor thin films, such
as aluminum nitride, is a long-sought goal in materials physics. One
promising approach is plasma-assisted atomic layer epitaxy, in which
separate reactant precursors are employed to grow the cation and
anion layers in alternating deposition steps. The use of a plasma
during the growth — most often a hydrogen plasma — is now routine
and generally considered critical, but the precise role of the
plasma is not well understood. We propose a theoretical atomistic
model and elucidate its consequences using analytical rate
equations, density-functional theory, and kinetic Monte Carlo
statistical simulations. We show that using a plasma has two
important consequences, one beneficial and one detrimental. The
plasma produces atomic hydrogen in the gas phase, which is important
for removing methyl radicals left over from the aluminum precursor
molecules. But atomic hydrogen also leads to atomic carbon on the
surface and, moreover, opens a channel for trapping these carbon
atoms as impurities in the subsurface region, where they remain as
unwanted contaminants. Understanding this dual role leads us to
propose a solution for the carbon contamination problem which leaves
the main benefit of the plasma largely unaffected.