2290 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
groups deprotonate, with accumulation of negative charge
on surface. The latter in turn boosts the hydrogen bonds
between the water molecules and the surface. All these
informations may be complemented by AIMD calculations
which are very useful in showing details on the distribu-
tion of water molecules on surface. Figure 6(a) shows how
insights on water structuring may be obtained by moni-
toring both oxygen and hydrogen densities on the quartz
slab settled at 0 Å. The latter corresponds to the position of
the outermost oxygens belonging to the slab which is con-
sidered fully hydroxylated. The first two peaks (blue color)
display the in-/out-of-plane silanols on surface. These peaks
are quite well defined thanks to the adequate interfacial
sampling. Yet, AIMD fails in providing well defined layer-
ing structures at further away distances, because of the poor
statistical sample. Water hydrogen densities (pink profile)
are still defined within the first water layering. Beyond that,
the profiles for water oxygen and hydrogen blur with the
axial distance. The poor sampling restriction can be fixed by
employing the MLFF potential. As can be noted from the
orange profile in Figure 6(a), the ML simulation qualita-
tively recovers the explicit AIMD profile, exhibiting a pro-
nounced peak corresponding to the first adsorption layer
characterized by a highly structured water arrangement.
By increasing the order of magnitude by a factor three, the
data collected by the trajectory can design a well converged
density profile, showing water structuring in the first two
layers. The first minimum in the orange profile indicates
the poor water exchange within the first two hydration
films whereas the second minimum precedes the formation
of bulk water. Next, we employed the MLFF potential to
investigate how the addition of the cation M+ (M =Li, Na,
K) may affect the water layering as well as the relative densi-
ties. Depending on the cationic charge and size, as well as
on the amount of interfacial silanols, the presence of the
cation can redistribute the arrangement of water molecules
in the inner and outer shells. By maintaining a fully hydrox-
ylated slab, profiles of Figure 6(b) exhibit a broader and
less intense first peak increasing the size of M+, meaning
less-ordered region in the inner solvation shell, with more
distorted hydrogen bonds and a liquid-like water network.
Further, from the minimum between the two consecutive
peaks, we can qualitatively evaluate the exchange of water
molecules between shells, which in this case increases with
the size of M+. It is also possible to notice a density shift
at larger axial distance. Finally, from the second minimum
Figure 5. Evolution of the isosteric enthalpy as a function of the adsorbed amount of water onto kaolinite
calculated by means of classical (CMD) and ab initio molecular dynamics (AIMD). The onset points out
the larger achievable scale length using MLFF technique
groups deprotonate, with accumulation of negative charge
on surface. The latter in turn boosts the hydrogen bonds
between the water molecules and the surface. All these
informations may be complemented by AIMD calculations
which are very useful in showing details on the distribu-
tion of water molecules on surface. Figure 6(a) shows how
insights on water structuring may be obtained by moni-
toring both oxygen and hydrogen densities on the quartz
slab settled at 0 Å. The latter corresponds to the position of
the outermost oxygens belonging to the slab which is con-
sidered fully hydroxylated. The first two peaks (blue color)
display the in-/out-of-plane silanols on surface. These peaks
are quite well defined thanks to the adequate interfacial
sampling. Yet, AIMD fails in providing well defined layer-
ing structures at further away distances, because of the poor
statistical sample. Water hydrogen densities (pink profile)
are still defined within the first water layering. Beyond that,
the profiles for water oxygen and hydrogen blur with the
axial distance. The poor sampling restriction can be fixed by
employing the MLFF potential. As can be noted from the
orange profile in Figure 6(a), the ML simulation qualita-
tively recovers the explicit AIMD profile, exhibiting a pro-
nounced peak corresponding to the first adsorption layer
characterized by a highly structured water arrangement.
By increasing the order of magnitude by a factor three, the
data collected by the trajectory can design a well converged
density profile, showing water structuring in the first two
layers. The first minimum in the orange profile indicates
the poor water exchange within the first two hydration
films whereas the second minimum precedes the formation
of bulk water. Next, we employed the MLFF potential to
investigate how the addition of the cation M+ (M =Li, Na,
K) may affect the water layering as well as the relative densi-
ties. Depending on the cationic charge and size, as well as
on the amount of interfacial silanols, the presence of the
cation can redistribute the arrangement of water molecules
in the inner and outer shells. By maintaining a fully hydrox-
ylated slab, profiles of Figure 6(b) exhibit a broader and
less intense first peak increasing the size of M+, meaning
less-ordered region in the inner solvation shell, with more
distorted hydrogen bonds and a liquid-like water network.
Further, from the minimum between the two consecutive
peaks, we can qualitatively evaluate the exchange of water
molecules between shells, which in this case increases with
the size of M+. It is also possible to notice a density shift
at larger axial distance. Finally, from the second minimum
Figure 5. Evolution of the isosteric enthalpy as a function of the adsorbed amount of water onto kaolinite
calculated by means of classical (CMD) and ab initio molecular dynamics (AIMD). The onset points out
the larger achievable scale length using MLFF technique