Kiralj R., Ferreira. M. M. C., "A Simple Quantitative Structure-Property Relationship (QSPR) Approach to Stability Constants of Metal-Crown Ether Complexes in Methanol". Águas de Lindóia, SP, Brazil, 10-15/09/2006: 10th International Conference on Chemometrics in Analytical Chemistry (CAC-2006, CAC-X), Book of Abstracts (2006) P081. Poster 081.
10th International Conference on Chemometrics in Analytical Chemistry P081
A Simple QuantitativeStructure-Property
Relationship (QSPR)
Approach to Stability
Constants of Metal-Crown Ether Complexes
in Methanol
Rudolf Kiralj*, Márcia M. C. Ferreira rudolf@iqm.unicamp.br
Laboratório de Quimiometria
Teórica e Aplicada, Instituto de Química, Universidade Estadual
de Campinas, Campinas –
SP, 13083-970 BRAZIL
Keywords: crown ethers,
PLS regression, Cambridge Structural Database
_____________________________________________________________________________________
The discovery
of crown ethers in late 60's as effective metal cation carriers
(ionophores) opened a new
area in metal chemistry.
Ionophores are used today for complexation
(in homogeneous solutions) or
solvent extraction
of alkali, earth alkaline, transition and
other metals. The search for desired
metal
selectivity of
crown ethers and other ionophores in one- or two-phase
liquid systems is interesting from
the point of
view of analytical chemistry, because related
analytical techniques are employed in vital,
agricultural, biological
and industrial processes. These search
efforts are still predominantly empirical.
This work is
a theoretical approach (QSPR: Quantitative
Structure-Property Relationships) to model
stability constants of complexes
formed by cations Ag+, Na+,
K+, Mg2+,
Ca2+, Sr2+
and Ba2+ and four crown
ethers in methanol:
18-crown-6 (18C6), dibenzo-18-crown-6
(DB18C6), dicyclohexyl-18-crown-6
(DCY18C6), and dibenzopyridino-18-crown-6
(DBPY18C6).
Measured stability
constants for the 28 complexes were from literature1.
Various atomic descriptors for
metals and ligand descriptors
for the crown ethers were collected from literature or were
calculated from
the crystal
structures of crown ether and metal
complexes retrieved from the Cambridge Structural
Database (CSD). Five descriptors
were manually selected for the final QSPR model: cation radius,
metal
electron affinity, mean
metal-O,N bond length (from CSD median values), and two more steric
descriptors.
These two descriptors included
the metal radii and metal-O,N bond lengths, macrocycle size
(average O-
O,N diametral distances
from the CSD median values), and O,N van der Waals
radii. Autoscaled data
were then used
in the Partial Least Squares (PLS)
regression to predict metal-macrocycle binding
constants in log form.
The dataset was also used in exploratory analysis (Principal Component
Analysis -
PCA and Hierarchical Cluster
Analysis - HCA).
The descriptors
are moderately correlated with the stability constants
(absolute correlation coefficients
are 0.44-0.75). Three
principal components (PCs) were used for
the PLS model (99.2% of the total
variance), resulting
in acceptable QSPR statistics: Q = 0.730,
R = 0.830, SEV = 0.46, SEC = 0.40, 7
samples with relative error
greater than 10% (10-16%). The regression
vector and correlograms show
relationships that can be
well interpreted chemically. Higher complex stability is
related to larger ions that
are tightly bound
to the macrocyle and probably to the solvent,
ions that are displaced from the O,N
macrocycle plane due to
their size and higher coordination number (more than
6), and ions with higher
electron affinity (good
electron acceptors in coordinative bonds). Small variations in the
stability constants
originates from the differences
in crown ether structures. More rigid DB18C6 and
DCY18C6 are basket-
shaped and somewhat poorer
binders. More flexible 18C6 and DCY18C6 (with cyclohexyl units
acting as
additional arcs)
are better metal binders. Ag-DBY18C6
is an exception due to
electronic effects
accounting for pyridine-Ag
bond. Crystal structures for the CSD and simple molecular modeling were
used
to visualize stereoelectronic
relationships in the complexes and other related structures.
Clustering
of the samples is almost identical in PCA (3 PCs: 99.6% of
the total variance) and HCA, and
agrees with
previously noted trends. There are seven metal clusters at
similarity index 0.85 (complete
linkage was used), among
which that for Mg is isolated, and Ba-K and
Ca-Na clusters form a group at
similarity index around
0.75. PC1 is mainly related to the cation size, whilst PC2
distinguishes well metal
groups i.e.
discriminates alkali (Na, K) from alkaline
earth (Mg, Ca, Sr, Ba) and other (Ag)
metals.
The presented
approach to model metal-crown ether
stability constants was based on available
structural data
in literature and structural databases and
not on descriptors from molecular modeling.
This methodology can be
expanded to other metal-macrocycle systems and
used in practice as the first-
aid QSPR that provides reasonable
interpretation of the obtained results.
Acknowledgment: FAPESP
__________________________________________________________________________________________________________________________
References
1 Zolgharnein
J.; Tahmasebi H.; Habibi M.; Amani S. J. Inclus. Phenom. Macrocyc. Chem.
2004,
49,
231-234.