DFT investigation of Montmorillonite edge surfaces stability and
their acid-basic properties
1st Carla G. Fonseca,1 2nd Viviane Vaiss,1 3rd Fernando Wypych,2 4rd Renata Diniz,3 5rd Alexandre A. Leitão1,*
1
Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Juiz de Fora, Martelos, Juiz de Fora, MG 36033-330, Brazil.
Departamento de Química, Universidade Federal do Paraná, P.O. Box 19032, 81531-980 Curitiba, PR, Brazil.
3
Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Minas Gerais, Pampulha, Belo Horizonte, MG 31270-901,
Brazil.
*Corresponding author: [email protected]
2
Keywords: Montmorillonite, clay minerals, edge surfaces, ab initio calculations, Biodiesel, heterogeneous catalysis
1. Introduction
Renewable energies make up the industrial sector with
the greatest growth in the world. The synthesis of
biodiesel from renewable biological sources, including
animal fats and vegetable oils, has received considerable
attention due to its environmental advantages and
sustainable production in comparison with fossil fuels
[1].
In general, biodiesel is produced by transesterification of
vegetable oils or animal fats or by esterification of fatty
acids with short chain alcohols, in homogeneous or
heterogeneous catalytic conditions. Heterogeneous
catalysis based on clays such as Montmorillonite (Mt)
have received considerable attention due to their
environmental compatibility, low cost, selectivity,
thermal stability and recyclability [2]. Montmorillonite is
classified as a 2:1 cationic exchanger phyllosilicate
family and their acid properties can be enhanced by acid
activation, these include: the edges of the crystals are
opened and Al3+ and M2+ cations of the octahedral sheet
are leached from the structure, causing an increase in
both surface area and pore diameter. The acid treatment
of clay minerals, besides leaching cations preferably
from the octahedral sheets, also dissolves impurities and
replaces exchangeable cations present in the interlayer
space by hydrated protons. The acid-activated Mt proved
to be reusable, besides presenting structural
improvements and maintenance of catalytic activity. It
show to be a promising catalyst for the production of
fatty acid esters by esterification of different fatty
materials [2,3]. Crystallographic planes that contribute
predominantly to Mt edge surfaces are (010) and (110)
and they are the catalytically active plans in these
applications. Although the similar stabilities and the
complicated nature of the edges have been
acknowledged, the studies are no yet conclusive. In the
present work, a comprehensive DFT study on the
stability of montmorillonite edge structures is introduced.
The aim is to carry out a survey over a range of Mt edge
geometry variations for (010) and (110) planes, estimate
the Gibbs free energy of the cleavage processes in order
to distinguish the most probable edge structures. To
assess the reactivity of these sites, we examined the acidbasic properties by adsorbing a probe molecule CO on
the Mt acid-activated surfaces. The changes in the
electronic and acid-basic properties were also
investigated by density of states calculations (DOS) and
projected density of states (PDOS). To better understand
the reaction mechanisms, the theoretical study of aspects
of the transesterification of triglycerides and
esterification of fatty acids reactions catalyzed by Mt
edge surfaces is in progress.
2. Theoretical and Experimental Details
The quantum mechanical calculations were performed
using the codes available in Quantum ESPRESSO (QE)
package [4]. Electronic structure calculations were based
on density functional theory (DFT) implemented with
periodic boundary conditions using plane wave functions
as basis set. The electronic correlation and exchange
terms were employed using GGA with the PBE
functional. The vibrational modes were obtained from
phonon calculations, which were based on the harmonic
approximation by density functional perturbation theory
(DFPT) at the Γ q point. The vibrational data were also
used to calculate both the contribution of the lattice
thermal vibration and the zero point energy (ZPE). Solid
State Nuclear Magnetic Resonance analyzes were
performed with the neat Mt and acid-activated samples
with phosphoric acid used in the work performed by
Zatta et al. [2]. 31P NMR spectroscopic measurements
were conducted on a Bruker Avance HD 300 (7.05 T),
operating at a Larmor frequency of 121,5 MHz.
3. Results and discussion
According to the basic crystallographic principles, the
most common facets are represented by the
crystallographic planes with maximal atomic density. The
lateral facets of Mt are indicated in Figure 1.
Figure 1. The (010) and (110) edge structural models for
H+(2H2O)-Mt.
In the presence of water, the stabilization of the lateral
facets is achieved through chemisorption of H 2O
molecules. The energy required to section a crystal into
two separate parts is defined as the cleavage energy.
Surface Gibbs free energy, ΔGsurf, as a function of the
water coverage of the (010) and (110) facets are 26.44
and 33.97 kJ/mol respectively. Thus, the (010) surface
was found to be the most dominant face for Mt.
Churackov et al. [5] calculated the stability of different
edge faces of pyrophyllite with DFT and the (110) face
was found to be dominant while Lavikainen et al. [6]
observed energetic equivalence of the facets, but both did
not used the lattice thermal vibration and the zero point
energy as done in our calculations. Before simulate the
acid-activated Mt models, we investigated by 31P NMR
spectroscopy the hypothesis of Zatta et al. [7] of the
incorporation of phosphate in their structure as a result of
acid activation. In all analyzes no signal of any phosphate
species was observed, indicating the absence of these
species in the structure of the acid-activated
montmorillonite. The acid-activated Mt surfaces were
simulated and the reactivity of the AlV sites was checked
by means of the CO adsorption. CO is a common Lewis
basic probe molecule used to describe the Lewis acidity.
The data confirm that the sites of the face (010) are
slightly more acid than the sites of the face (110), the
adsorption energies were -4.9 kcal/mol and -4.5 kcal/mol,
respectively. The density of states (DOS) shows that the
gap found for the Mt bulk is greater than the gap of both
edge acid-activated surfaces, 4.9 compared to 4.6 and 4.3
regarding to (010) and (110), respectively. The edge
(110) is slightly more acid. The projected density of
states (pDOS) (Figure 2), shows that the H+ sites present
in the hydroxyl groups linked to Si atoms (HO-Si sites)
on the surface and the protons present in the hydroxyl
groups (HO-AlV sites), pink line, are more acidic than the
surface AlV sites (red and blue lines) for both acidactivated surfaces. Probably these are the sites
responsible for the catalytic activity of the acid-activated
montmorillonite. The Figure 2 shows the PDOS analyzes
for the (010) face.
Figure 2. PDOS of the H and AlV from (010) edge surface. The
Fermi energy was considered as the zero of energy.
The adsorption of the methanol, acetic acid and triacetin
molecules are in progress. The triacetin molecule was
considered in this work because it is the smallest
compound in tryacilglycerol family. The aim of this step
is to evaluate the selectivity of the surfaces for adsorption
of the molecules and to provide insights about the
transesterification and esterification mechanisms.
4. Conclusions
Surface Gibbs free energy variation, ΔGsurf, as a function
of the water coverage of the (010) and (110) facets shows
that the (010) is the most dominant face for Mt. The acidactivated Mt surfaces were simulated and the reactivity
of the AlV sites were checked by means of the carbon
monoxide, CO, adsorption and the data confirm the
greater acidity of the sites of the face (010) compared to
face (110). By means of DOS and PDOS analyzes we can
show that the acid sites in the acid-activated clay are
stronger than the purified clay and the acid character of
the surfaces comes mainly from the proton sites located
in the hydroxyl groups (HO-Si and HO-AlV).
Acknowledgments
This work has been supported by CAPES, CNPq, FAPEMIG, FINEP,
Brazilian agencies and the enterprise PETROBRAS S/A (CENPES).
We also acknowledge the CENAPAD-SP computational center for the
use of its facilities by performed calculations.
References
[1] L. C. Meher, D. V. Sagar, S. N. Naik, Renew Sust Energ Rev2006,
10, 248.
[2] L. Zatta, L. P. Ramos, F. Wypych, App Clay Sci 2013, 80-81, 231.
[3] T. Wang, T. Liu, D. Wu, M. Li, J. Chen, S.Teng, J. Hzard Mater
2010, 173, 335.
[4] P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. et al.,
J. Phys. Condens. Matter 2009, 21, 395502.
[5] S. V. Churackov, J. Phys. Chem. 2006, 110, 4135.
[6] L. P. Lavikainen, J. T. Hirvi, S. Kasa, T. Schatz, T. A. Pakkanen,
Theor Chem Acc 2015, 134, 112.
[7] L. Zatta, L. P. Ramos, F. Wypych, J. Oleo Sci. 2012, 61, 597.