Generation of Amorphous Silica Surfaces with Controlled Roughness

Amorphous
silica (a-SiO2) surfaces, when grafted with
select metals on the active sites of the functionalized surfaces,
can act as useful heterogeneous catalysts. From a molecular modeling
perspective, one challenge has been generating a-SiO2 slab
models with controllable surface roughness to facilitate the study
of the effect of surface morphology on the material properties. Previous
computational methods either generate relatively flat surfaces or
periodically corrugated surfaces that do not mimic the full range
of potential surface roughness of the amorphous silica material. In
this work, we present a new method, inspired by the capillary fluctuation
theory of interfaces, in which rough silica slabs are generated by
cleaving a bulk amorphous sample using a cleaving plane with Fourier
components randomly generated from a Gaussian distribution. The width
of this Gaussian distribution (and thus the degree of surface roughness)
can be tuned by varying the surface roughness parameter α. Using
the van Beest, Kramer, and van Santen (BKS) force field, we create
a large number of silica slabs using cleaving surfaces of varying
roughness (α) and using two different system sizes. These surfaces
are then characterized to determine their roughness (mean-squared
displacement), density profile, and ring size distribution. This analysis
shows a higher concentration of surface defects (under-/overcoordinated
atoms and strained rings) as the surface roughness increases. To examine
the effect of the roughness on surface reactivity, we re-equilibriate
a subset of these slabs using the reactive force field ReaxFF and
then expose the slabs to water and observe the formation of surface
silanols. We observe that the rougher surfaces exhibit higher silanol
concentrations as well as bimodal acidity.