A natural zeolite as a molecular adsorbent: computational study of cation distributions, surface chemistry, and pharmaceutical uptake

This doctoral thesis presents a thorough computational investigation into the atomic-scale
structure and adsorption behavior of natural clinoptilolite, a widely available and
industrially relevant zeolite. Though it is practically important in water treatment,
catalysis, and drug delivery, clinoptilolite remains underexplored at the atomistic
level, especially regarding its surface properties, cation distribution, and interaction
mechanisms with large pharmaceutical molecules. Addressing this gap, the present
work develops and applies a multi-layered, reproducible modeling framework that
combines dispersion-corrected density functional theory (DFT-D3) and classical
molecular dynamics (MD) to study aluminum placement, extra-framework cation
distribution, surface termination, and molecule–surface adsorption phenomena in
mono- and multi-cationic clinoptilolite systems.

The study begins with the generation of periodic bulk models including experimentally
guided aluminum distributions and extra-framework cation arrangements
(Na+, Ca2+, K+, and their combinations)—element-to-site assignments dictated with
DFT-D3 calculations—while strictly adhering to Loewenstein’s rule. Motivated by
the need for reproducible and accurate DFT simulations, an extensive comparison of
nine exchange-correlation functionals was conducted, leading to the selection of the
B97-D3 functional as the most robust for clinoptilolite systems. The framework also
introduces improved basis sets, especially for Ca2+, to better capture its coordination
environment. Cation site preferences were evaluated, and it turns out that Na+
favored Al at T2 and T3 and localized mainly in channels A and B, Ca2+ was
stabilized in channel B and coordinated with Al at all T sites, while K+ preferred the
narrower channel C and favored Al at T1 and T3. These observations validate known
experimental trends and offer reliable benchmarks for future modeling. Overall,
framework stability was shown to depend critically on both the Al distribution and
the identity of the charge-balancing cation.

The next part of the study focused on modeling the external surfaces of
Na–clinoptilolite and their interaction with 5-fluorouracil as a model pharmaceutical
compound. A hybrid sampling approach was employed, combining simulated annealing
(SA) and parallel tempering (PT) molecular dynamics for initial exploration of
adsorption motifs, followed by DFT-D3 refinement of selected low-energy configurations.
Adsorption energies spanned a broad and significant range (from –430.0 to
–174.4 kJ/mol), with configurations featuring exposed Na+ cations showing regularly
stronger binding. Contrary to Na+’s steric limitations in the bulk, surface-bound
Na+ was found to play an active role in anchoring 5-FU via Na–O and Na–F interactions,
further reinforced by hydrogen bonding with the framework or hydroxyl
oxygens. Among various terminations, surfaces exposing 8-membered rings (8MR)
consistently yielded stronger adsorption compared to 10MR-exposed surfaces, due to
enhanced confinement and more favorable H-bonding geometry. Cation-free surfaces
showed significantly weaker interactions, which highlights the critical role of surface
composition and local electrostatics.

Building upon these insights, the adsorption study was extended to Ca–, Na–Ca–,
and Na–Ca–K–clinoptilolite systems. Using the same hybrid MD/DFT strategy,
the final part of the investigation comprehensively broke down and examined the
competitive and cooperative behavior of multiple cations in driving molecular adsorption.
A clear energetic trend was observed: Na–Ca–K > Na > Ca > Na–Ca, and
it was indicated that adsorption strength is governed not merely by cation charge
or size, but by a combination of cooperative interactions and spatial arrangement.
Interestingly, configurations where Ca2+ alone coordinated with the adsorbate often
outperformed those with multiple cations directly involved, which suggests that the
spatial separation of cations and the resulting local electrostatic environment can
outweigh simple additive cation effects. Moreover, cations that are not coordinating
with the molecule directly such as K+ and Na+ were found to enhance adsorption
indirectly by modulating the local electrostatic environment. Al distribution asymmetry
was also shown to influence adsorption, with uneven Al placement enhancing
surface anionic character and thus improving cation anchoring and molecule binding.
Altogether, this thesis provides deep mechanistic insight, quantitative adsorption
energies, and significant structural comparisons that are fully reproducible through
the provided computational framework and accompanying scripts. The study demonstrates
that clinoptilolite surface reactivity is governed by a delicate balance of
Al–cation configuration, surface accessibility, and molecular complementarity (how
well the molecule “fits” or “matches” the surface), offering a new paradigm for tuning
zeolite-based adsorbents through cation engineering. This work not only fulfills but
also extends the original aims of the project proposal, which sets a computational
benchmark for future studies in zeolite surface modeling and bridges the gap between
structural complexity and adsorption performance in natural materials.