Silica
is both ubiquitous in nature and important in a wide range
of applications ranging from drug delivery to catalysis. This has
spawned significant interest in modeling silica, particularly the
amorphous solid, and its interactions with liquids, adsorbates, and
embedded active sites. One of the challenges is developing descriptions
that accurately represent the synthesized materials used in experiments
when key properties, such as the surface roughness and distribution
of silanol groups, cannot be readily measured. Here, we implement
a simple, tunable melt-cleave-quench-functionalize approach to create
amorphous silica slabs with variable characteristics. We use it to
generate 2000 atomistically distinct slab models, based on the widely
used the van Beest, Kramer, van Santen (BKS) potential, that differ
in their atomistic roughness, defect site density, ring distribution,
silanol density, and spatial arrangement of silanols. The surfaces
are demonstrated to be consistent with available experimental data
and stable within the ReaxFF bond order-based reactive force field.
These amorphous silica slab models should thus be of use in a variety
of computational studies of interfacial properties.