editorial


 

 

HUNGARIAN JOURNAL 
OF INDUSTRY AND CHEMISTRY 

VESZPRÉM 
Vol. 42(1) pp. 19–23 (2014) 

ADSORPTION AND INTERCALATION OF SMALL MOLECULES ON 
KAOLINITE FROM MOLECULAR MODELLING STUDIES 

ATTILA TÁBOROSI,1! RÓBERT KURDI,1 AND RÓBERT K. SZILÁGYI2 

1 Surfaces and Nanostructures Laboratory, Institute of Environmental Engineering, University of Pannonia,  
Veszprém, H-8201, HUNGARY 

2 Department of Chemistry and Biochemistry, Montana State University, Bozeman, MT 59717, U.S.A. 
!E-mail: taborosia@almos.uni-pannon.hu 

Kaolinite is an abundant natural material with considerable industrial potential. Despite its simple composition 
(Al2Si2O5(OH)4 and layered structure being a phyllosilicate), it is notable that only little known about the interaction of 
kaolinite sheets with small organic reagents at the molecular level. These assumed to govern intercalation, delamination, 
and then the complete exfoliation processes. A common methodology to model the molecular structure of kaolinite is the 
employment of periodic boundary conditions. The application of molecular cluster models gained importance nowadays 
by capitalizing on the availability of wide range of theoretical tools for describing structural features and reaction 
mechanisms. In this study, we present our results using theoretical methodologies and modelling strategies from literature 
that are applied for adsorption and intercalation of urea, ethylene glycol, and potassium acetate. 

Keywords: kaolinite, intercalation, adsorption, hydrogen bonding, molecular cluster model 

Introduction 

A significant hindrance in the design, optimization, and 
execution of experimental research involving clay 
minerals is the lack of our understanding of molecular 
level processes occurring at the outer (adsorption) or at 
the inner surface (intercalation) of layers of clay. 
Molecular modelling approaches have the potential to 
aid experimental design, mitigating cost of chemicals 
used, rationalizing unsuccessful trial-and-error 
experiments, and thus contributing to tailored utilization 
of clay minerals as industrial raw materials. 

Given the periodic nature of the crystalline clay 
minerals, a common approach in modelling is the use of 
periodic boundary conditions [1–4]. A competitive 
approach is based on molecular cluster models [5–9]. 
An advantage of using the latter is the availability of a 
broad range of quantum chemical tools for studying the 
local chemical environment of the site(s) of adsorption 
and/or intercalation. 

In the given study, we have utilized kaolinite, which 
is a member of the 1:1 class of clay minerals. This 
implies that the layers of this simplest phyllosilicate are 
composed of octahedral (O) and tetrahedral (T) sheets. 
The O-sheet contains hexacoordinate Al3+ ions, while 
the T-sheet has tetracoordinate Si4+ ions. The O- and T-
sheets are connected via tricoordinate oxide anions, 
while the separate OT or 1:1 layers are connected 
through hydrogen bonds (H-bonds) (Fig.1). 

Computational Methods 

The computational models were created from the unit 
cell of the experimental X-ray diffraction structure of 
crystalline kaolinite [10]. We will discuss results here 
for models for separate O- and T-sheets with 

 
Figure 1: Schematic presentation of adjacent OT layers of 

delaminated kaolinite 



 

 

20 

stoichiometry of Al6O24H30 and Si6O18H12, respectively 
(Figs.2A and C). These are the smallest possible cluster 
models for the hexagonal, honeycomb like units of the 
O- and T-sheets. In addition, two larger models of 
Al10O38H46 and Si10O29H18 were also considered 
(Figs.2B and D) for the larger intercalating agents 
(ethylene glycol, potassium acetate). The unit cell of 
kaolinite was multiplied in all crystal dimensions to 
create a super-cell. This was then truncated to the 
cluster models presented above. The charge neutrality 
was achieved by terminating dangling bonds with 
protons at the periphery of the models. The initial 
positions of the adsorbent reagents were set at the centre 
of the hexagonal honeycomb-like unit of both O- and T-
sheets and it was allowed to fully relax during 
optimizations. 

Given that the computational models shown in Fig.2 
are simplifications of the experimental crystal structure 
environment, structural constrains were applied during 
geometry optimizations. In all calculations, the atoms of 
the organic reagents, the surface hydroxide groups (s-
HO-) of the O-sheet, and the bridging oxide groups (b-
O2-) of the T-sheets were allowed to move, while the 
positions of the rest of the atoms were kept fixed. After 
structural optimizations, the positions and relative 
arrangements of the organic reagents with respect of the 
outer and inner surfaces, the shortest distances of 
possible ion/dipole interactions or H-bonds were noted. 
For computational models of T…reagent…O 
intercalation, the basal spacing was set at the beginning 
of calculations to the experimental values of 10.7, 9.5, 
11.5 Å, and 14.2 Å for urea [11], ethylene glycol [12], 
and potassium acetate in the absence or in the presence 
of water [13], respectively. 

The calculations were carried out using the B3LYP 
hybrid density functional [14,15] with LANL2DZ basis 
set [16] as implemented in Gaussian09 [17]. The 
interaction energies were not considered in this work 
due to the modest level of theory and lack of 
consideration of explicit or implicit solvation shell. The 
main focus of the given work is to establish initial 
structural models for rationalizing experimental 
observations and prepare for more advanced 
computational modeling. The optimized structures 
presented in this work were made available at 
computational.chemistry.montana.edu/SI. 

Results and Discussions 

Kaolinite-Urea Interactions 

The lowest energy, optimized structure for the adsorbed 
urea on O-sheet can be characterized by two short 
O(urea)…s-HO- distances of 1.92 and 1.95 Å, while one 
of the amide groups approaches the opposite s-HO- 
group within 1.73 Å as shown in Fig.3A. The presence 
of both donor and acceptor interactions for the same 
reagent molecule result in a tilted orientation of the 
urea. An alternative orientation in which only the 
carbonyl group is interacting with three s-HO- groups 
(see below for intercalation model) was not found to be 
a stationary structure at this level of theory. Figs.3B and 
C illustrate the two different orientations of the urea at 
the T-sheet, where the amide groups are H-bonded to 
the b-O2- groups of the same Si4+ ion (geminal) with 
NH…b-O2- distances of 2.03 and 2.12 Å or two 
different Si4+ ions (vicinal) with 2.02 and 2.08 Å 

 
Figure 2: Computational models for the hexagonal honeycomb-like units of [6Al] (A) and [10Al] (B) octahedral, [6Si] (C) 

and [10Si] (D) tetrahedral cluster models for O- and T-sheets of kaolinite, respectively 



 

 

21 

distances. The energetic difference between the 
germinal and vicinal coordination is not significant (less 
than 10 kJ mol-1) at this level of theory. 

 When the adsorption of urea considered 
simultaneously at both the O- and T-sheets, a model can 
be obtained for a plausible structure of the urea 
intercalated between two OT-layers (Fig.4). In this 
computational model, the carbonyl group of urea forms 
three H-bonds at 1.74, 1.78, and 1.83 Å distances with 
three s-HO- groups of the hexagonal O-sheet cross from 
each other. At the same time, the intercalated urea forms 
two H-bonds with the geminal b-O2- groups at 1.94 and 
2.62 Å distances. The tilted orientation of the urea is the 
consequence of the experimental basal spacing of 10.7 
Å, which was kept fixed during the structural 
optimizations. It is important to recognize that 
according to our computational model, there is not 
enough room between the expanded layers of kaolinite 
to accommodate urea perpendicular to the basal planes 
or with parallel carbonyl groups to the crystallographic 
‘c’ axis.  

Kaolinite-Ethylene Glycol Interactions 

Ethylene glycol is generally used in exchange-
intercalation as a step followed by initial dimethyl 

sulfoxide or potassium acetate intercalation. As shown 
in Fig.5A, the donor interaction between the reagent 
hydroxyl group and the Al3+ bound s-HO- is much 
shorter (1.66 Å) than those seen for urea adsorption 
(1.73 Å), while the acceptor interaction is slightly 
longer (1.99 Å). Given that the b-O2- groups of the T-
sheet can only act as H-bond acceptors with ethylene 
glycol, the reagent spans the entire hexagonal 
honeycomb-like unit of the T-sheet by forming longer 
(2.05 and 1.94 Å), thus weaker H-bond interactions than 
with the O-sheet. 

The complementarity of the O- and T-sheets with 
respect of being H-bond donor and acceptor, 
respectively, allows for an ideal orientation of the 
functional groups of ethylene glycol for maximizing the 
strength of H-bond interactions. The O-sheet forms a 
short (1.61 and 1.79 Å) interaction between the 
O(glycol)…s-HO- groups. The T-sheet accepts the HO 
groups of the ethylene glycol at 1.73 and 1.88 Å 
distances. Comparison of the urea (Fig.5) and ethylene 

 
Figure 3: Computational models for urea absorption at O- (A) and T-layers in geminal (B) and vicinal (C) arrangements 

 

 
Figure 5: Computational models for the adsorption of ethylene 

glycol at the (A) O- and (B) T-sheets 

 
Figure 4: Computational model for intercalation of urea in 

kaolinite 



 

 

22 

glycol (Fig.6) intercalated models, the network of 
interactions between the layers and reagents parallel 
well the different experimental basal spacing (10.7 vs. 
9.5 Å, respectively). While the urea opens the interlayer 
space more; however, it anchors the two layers with a 
stronger set of interactions than ethylene glycol. This 
can be correlated with the limited reactivity of 
urea/kaolinite complex in exchange intercalation; while 
ethylene glycol often provides the last step before 
complete delamination and exfoliation of clays. 

Kaolinite-Potassium Acetate Interactions 

A limitation of the in vacuo models used in this study 
greatly hinders the dissociation of the potassium ion 
from acetate and thus in both adsorption models the 
cation remained at the centre of the hexagonal 
honeycomb-like model for the O- and the T-sheets in 
the vicinity of the anion acetate. In addition, the ionized 
carboxylic group of the reagent form two short and thus 
strong H-bonding interactions at 1.65 and 1.89 Å 
distances with two germinal s-HO- groups of the O-
sheet. The shortest acetate C-H…b-O2- distance in 
Fig.7B is 2.95 Å, which indicates the presence of weak 

interactions between the methyl group and the b-O2- 
groups. In the adsorption model for the T-sheet, the 
acetate appears to be held in place by the cation at about 
2.66 and 2.68 Å.  

Some of the plausible structures for the intercalation 
of potassium acetate were obtained from simulations in 
the absence (Fig.8A) and presence (Fig.8B) of solvent 
water. Similarly to the other models the values of basal 
spacing were set to their corresponding experimental 
values of 11.5 and 14.2 Å, respectively. The 2.7 Å 
difference allows for the accommodation of a water 
molecule as shown in Fig.8B. In agreement with the 
absorption models in Fig.7, we found short H-bond 
distances between the carboxylate group and the s-HO- 
of the O-sheet at 1.59 and 1.83 Å distances. When water 

 

 
Figure 7: Computational models for the adsorption of 

potassium acetate on (A) O- and (B) T-sheets 

 
Figure 6: Computational model for ethylene glycol 

intercalation in kaolinite 

 
Figure 8: Computational models for potassium acetate intercalation in the (A) absence and (B) presence of solvent water 



 

 

23 

is present, it is likely to solvate the negatively charged 
end of the acetate ion with a characteristic -COO-…H-
OH distance of 1.63 Å. The interaction of the water and 
the O-sheet is asymmetrical with 1.77 and 2.08 Å 
O(water)…s-HO- distances. 

Given the energetic preference of the O-H…OH- H-
bonding interaction, the methyl group of the acetate will 
only interact with anionic groups of the T-sheet with 
representation C-H…b-O2- distances of 3.15 and 3.25 Å. 
The lack of a representative solvent environment is of 
less importance for the position of the potassium ion in 
the intercalation models, since the inner environment of 
this model will likely be only exposed to bulk solvent 
only to a limited extent. 

Conclusions 

We have started with a few notable preliminary results 
the establishment of a structural database for reference 
models related to the adsorption and intercalation of 
small molecules and kaolinite. We found that even a 
modest level of theory with respect of the basis set can 
provide reasonable structures. These results allow us for 
visualizing molecular structures in order to rationalize 
experimental observables and generate new ideas for 
research directions. Furthermore, these models can be 
considered as initial structures for higher-level 
calculations that are currently in progress in our 
laboratory.  

Acknowledgement 

This research was supported by the European Union and 
the State of Hungary, co-financed by the European 
Social Fund in the framework of TÁMOP 4.2.4.A/2-11-
1-2012-0001 ‘National Excellence Program’ (RKS and 
AT) and TÁMOP-4.2.2.A-11/1/KONV-2012-0071 
(RK). The authors acknowledge the helpful discussions 
with Prof. JÁNOS KRISTÓF and Prof. ERZSÉBET 
HORVÁTH. We acknowledge the use of the 
computational facilities at Montana State University, 
Bozeman, MT. 

REFERENCES 

[1] SMRCOK L., TUNEGA D., RAMIREZ-CUESTA A.J., 
IVANOV A., VALUCHOVA J.: The combined 
inelastic neutron scattering and solid-state DFT 
study of hydrogen-atoms dynamics in kaolinite-
dimethylsulfoxide intercalate, Clays Clay 
Minerals, 2010, 58(1), 52–61 

[2] MERCIER P.H.J., LE PAGE Y.: Rational ab initio 
modeling for low energy hydrogen-bonded 
phyllosilicate polytypes, Eu. J. Mineral., 2011, 
23(3), 401–407 

[3] GEATCHES D.L., JACQUET A., CLARK S.J., 
GREENWELL H.C.: Monomer adsorption on 

kaolinite: modeling the essential ingredients, J. 
Phys. Chem. C, 2012, 116(42), 22365–22374 

[4] TUNEGA D., BUCKO T., ZAOUI A.: Assessment of 
ten DFT methods in predicting structure of sheet 
silicates: Importance of dispersion corrections, J. 
Chem. Phys., 2012, 137(11), 114105 

[5] HATÓ Z., MAKÓ É., KRISTÓF T.: Water-mediated 
potassium acetate intercalation in kaolinite as 
revealed by molecular simulation, J. Mol. Mod. 
2014, 20(3), 1–10 

[6] MICHALKOVA A., ROBINSON T.L., LESZCZYNSKI J.: 
Adsorption of thymine and uracil on 1:1 clay 
mineral surfaces: comprehensive ab initio study on 
influence of sodium cation and water, PCCP, 
2011, 13(17), 7862–7881 

[7] DAWLEY M.M., SCOTT A.M., HILL F.C., 
LESZCZYNSKI J., ORLANDO T.M.: Adsorption of 
formamide on kaolinite surfaces: A combined 
infrared experimental and theoretical study, J. 
Phys. Chem. C, 2012, 116(45), 23981–23991 

[8] SCOTT A.M., DAWLEY M.M., ORLANDO T.M., 
HILL F.C., LESZCZYNSKI J.: Theoretical study of 
the roles of Na+ and water on the adsorption of 
formamide on kaolinite surfaces, J. Phys. Chem. 
C, 2012, 116(45), 23992–24005 

[9] SONG K.H., WANG X., QIAN P., ZHANG C., ZHANG 
Q.: Theoretical study of interaction of formamide 
with kaolinite, Comp. Theor. Chem., 2013, 102, 
72–80 

[10] BISH D. L.: Rietveld refinement of the kaolinite 
structure at 1.5-K, Clays Clay Minerals, 1993, 
41(6), 738–744 

[11] MAKÓ É., KRISTÓF J., HORVÁTH E., VÁGVÖLGYI 
V.: Kaolinite-urea complexes obtained by 
mechanochemical and aqueous suspension 
technics-a comparative study. J. Colloid Inter. Sci., 
2009, 330(2), 367–373 

[12] GARDOLINSKI J.E.F.C., LAGALY G.: Grafted 
organic derivatives of kaolinite: I. Synthesis, 
chemical and rheological characterization, Clay 
Minerals, 2005, 40(4), 537–546 

[13] FROST R.L., KRISTÓF J., KLOPROGGE J.T., 
HORVÁTH E.: Rehydration of Potassium Acetate-
intercalated Kaolinite at 298 K. Langmuir, 2000, 
16(12), 5402–5408 

[14] LEE C.T., YANG W.T., PARR R.G.: Development of 
the Colle-Salvetti correlation energy formula into a 
functional of the electron-density, Phys. Rev. B, 
1988, 37(2), 785–789 

[15] BECKE A.D.: Density-Functional thermochemistry 
3 The role of exact exchange, J. Chem. Phys. 
1993, 98(7), 5648–5652 

[16] DUNNING JR. T.H., HAY P.J., (ED.: SCHAEFER III 
H.F.): Modern Theoretical Chemistry: Methods of 
Electronic Structure Theory, Plenum, New York 
1976, Vol. 3, pp. 1–28 

[17] FRISCH M.J.T., TRUCKS G.W., SCHLEGEL H.B., 
SCUSERIA G.E., ROBB M.A., CHEESEMAN J.R.; et 
alas: Revision D.01 ed., Gaussian, Inc., 
Wallingford CT, 2009