New lipophilic 3-hydroxy-4-pyridinonate iron(III) complexes: Synthesis and EXAFS structural characterisation

Article abstract of DOI:10.1039/b509671e

New tris-iron(III) chelates of 3-hydroxy-4-pyridinone ligands derived from maltol (3-hydroxy-2-methyl-4-pyrone) or ethylmaltol (2-ethyl-3-hydroxy-4-pyrone) , including a variety of N-aryl (phenyl, 4'-tolyl, 4'-(n-butyl)phenyl, 4'-(n-hexyl)phenyl) and N-benzyl (4'-methylbenzyl, 4'-fluorobenzyl and 4'-(trifluoromethyl)benzylamine) substituents on the nitrogen atom of the pyridinone ring, have been prepared. Characterization by C,H,N elemental analysis and thermogravimetric measurements indicates that most of the complexes are obtained as hydrates of general formula ML₃·xH ₂O. Structural characterization of these difficult to crystallize lipophilic complexes has been achieved by EXAFS spectroscopy. Solutions of iron(III) complexes of maltol, ethylmaltol, 1,2-dimethyl-3-hydroxy-4-pyridinone and 1-phenyl-2-methyl-3-hydroxy-4-pyridinone in methanol-water mixtures were also examined by EXAFS. Distances from the central atom to ligand atoms, within 6 A of the metal, have been determined in the solid and solution samples and the results show that the structure observed in the powder is maintained in solution. The local structure around the metal centre, bond distances and bond angles, does not change significantly with variable lipophilicity, thus indicating that ligands may be tailored according to specific needs without altering their chelation properties. EXAFS data analysis for this set of tris-iron(III) compounds illustrates the important contribution of both intra-ligand and inter-ligand multiple scattering pathways through the metal centre to a peak observed in the FT spectrum at twice the metal ligand distance (～4 A). The present results demonstrate that EXAFS features at twice the metal-ligand distance are valuable in the assignment of molecular geometry and that location of hydration water molecules, by EXAFS analysis, is limited by the geometry of the complexes, in particular for those in which ligands containing phenyl rings are present. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b509671e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
New lipophilic 3-hydroxy-4-pyridinonate iron(III) complexes: synthesis and
EXAFS structural characterisation†
Walkiria Schlindwein,a Emma Waltham,b John Burgess,b Norman Binsted,c Ana Nunes,d
Andreia Leitee and Maria Rangel*e
Received 8th July 2005, Accepted 27th October 2005
First published as an Advance Article on the web 22nd November 2005
DOI: 10.1039/b509671e
New tris-iron(III) chelates of 3-hydroxy-4-pyridinone ligands derived from maltol
(3-hydroxy-2-methyl-4-pyrone) or ethylmaltol (2-ethyl-3-hydroxy-4-pyrone), including a variety of
N-aryl (phenyl, 4ꢀ-tolyl, 4ꢀ-(n-butyl)phenyl, 4ꢀ-(n-hexyl)phenyl) and N-benzyl (4ꢀ-methylbenzyl,
4ꢀ-fluorobenzyl and 4ꢀ-(trifluoromethyl)benzylamine) substituents on the nitrogen atom of the
pyridinone ring, have been prepared. Characterization by C,H,N elemental analysis and
thermogravimetric measurements indicates that most of the complexes are obtained as hydrates of
general formula ML3·xH2O. Structural characterization of these difficult to crystallize lipophilic
complexes has been achieved by EXAFS spectroscopy. Solutions of iron(III) complexes of maltol,
ethylmaltol, 1,2-dimethyl-3-hydroxy-4-pyridinone and 1-phenyl-2-methyl-3-hydroxy-4-pyridinone in
methanol–water mixtures were also examined by EXAFS. Distances from the central atom to ligand
˚
atoms, within 6 A of the metal, have been determined in the solid and solution samples and the results
show that the structure observed in the powder is maintained in solution. The local structure around
the metal centre, bond distances and bond angles, does not change significantly with variable
lipophilicity, thus indicating that ligands may be tailored according to specific needs without altering
their chelation properties. EXAFS data analysis for this set of tris-iron(III) compounds illustrates the
important contribution of both intra-ligand and inter-ligand multiple scattering pathways through the
˚
metal centre to a peak observed in the FT spectrum at twice the metal ligand distance (∼4 A). The
present results demonstrate that EXAFS features at twice the metal–ligand distance are valuable in the
assignment of molecular geometry and that location of hydration water molecules, by EXAFS analysis,
is limited by the geometry of the complexes, in particular for those in which ligands containing phenyl
rings are present.
the human body.22 As this synthetic approach, under appropriate
Introduction
conditions, works for almost any primary amine a very wide range
of 1-substituted-3-hydroxy-4-pyridinones can be prepared,23 from
hydrophilic when sulfonate- or carboxy-groups24 are incorporated
to very lipophilic when the substituent on the ring-nitrogen is a
long alkyl chain or an alkyl-substituted phenyl or benzyl group.
Substituted benzyl groups provide a means of varying the ligand’s
HLB while having very little effect on its donor properties and
thus on stabilities of its complexes.25 F- or CF3-for-H substitution
provides another useful approach to modifying the HLB of ligands
and complexes. The excellent chelating properties exhibited by the
3-hydroxy-4-pyridinone ligands towards M(III) and M(II) metal
ions, in conjunction with the possibility of synthesizing strongly
lipophilic derivatives, also makes this class of ligands useful in the
fields of analytical and environmental chemistry. 1-Phenyl- and 1-
(4ꢀ-tolyl)-3-hydroxy-4-pyridinones have been assessed as reagents
for the spectrophotometric determination of, inter alios, Fe(III),26,27
Ti(IV),28 V(V)29 and Nb(V).30 Some 1-aryl-3-hydroxy-4-pyridinone
derivatives have been demonstrated to be effective as extractants,
from aqueous media, for a variety of metals such as Ga(III),31
Fe(III),32 Hf(IV),33 Zr(IV),34 Nb(V),35 Ta(V)36,37 and U(VI).38 In many
cases effective metal ion separations have been achieved.
Ligands of the 3-hydroxy-4-pyrone and 3-hydroxy-4-pyridinone
types are well known, mainly by reason of their biomedical
applications.1–19 The 3-hydroxy-4-pyrones maltol and ethylmaltol
have the pharmacological advantage of being permitted food
additives20 and the closely related 3-hydroxy-4-pyridinones are
particularly attractive for pharmaceutical purposes since their
structure allows tailoring of their hydrophilic/lipophilic balance
(HLB),21 without significantly changing its chelating properties.
Variations in HLB can be achieved by simply introducing appro-
priate substituents on the endocyclic nitrogen atom of the pyridi-
none ring, the 3-hydroxy-4-pyridinones are synthesized by the
reaction of 3-hydroxy-4-pyrones withprimaryamines, thus leading
to the optimal lipophilicity for delivery or removal of metal ions in
aLeicester School of Pharmacy, De Montfort University, Leicester, UK LE1
9BH
bDepartment of Chemistry, University of Leicester, Leicester, UK LE1 7RH
c, Hampshire, UK SO51 78HT
dREQUIMTE, Departamento de Qu´ımica, Faculdade de Cieˆncias, Univer-
sidade do Porto, Portugal
eREQUIMTE, Instituto de Cieˆncias Biome´dicas de Abel Salazar, Universi-
dade do Porto, Portugal. E-mail: mcrangel@fc.up.pt
In continuance of our program of synthesis and solvation stud-
ies of complexes of substituted 1-aryl-3-hydroxy-4-pyridinones
† Electronic supplementary information (ESI) available: Colour version of
Fig. 1. See DOI: 10.1039/b509671e
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 1313–1321 | 1313
©

View Article Online
39
with fluorinated substituents such as 1-CH2CF3 or with long
In the present work we report the local structure around the
iron atom for a new set of hydrated 3-hydroxy-4-pyridinonate
iron(II) complexes using a more rigorous multiple scattering
analysis. Previously we used a restrained refinement procedure
(soft restraints).48 However, we found little deformation of the
ligand during refinement, and essentially the same results could
be obtained using constrained refinement, in which the ligand
defined by the crystal structure was moved as a rigid body. This
rather simpler approach was adopted here.
alkyl chains as phenyl substituents40 we now report the synthesis
and characterization of novel iron(III) complexes with a variety of
aryl (4ꢀ-(n-butyl)phenyl, 4ꢀ-(n-hexyl)phenyl and 4ꢀ-methoxyphenyl)
and benzyl (benzyl, 4ꢀ-methylbenzyl, 4ꢀ-fluorobenzyl and 4ꢀ-
(trifluoromethyl)benzylamine) substituents on the nitrogen atom
of the pyridinone ring. The ligands and complexes in the present
study are predominantly lipophilic and those containing long-
chain substituents exhibit high solubilities in alcohols and are
sparingly soluble or essentially insoluble in water. Consequently
the compounds are relatively unsuitable for medical use, but
suitable as metal extractants. Formulae and abbreviations are given
in Table 1.
Experimental
Materials
High lipophilicity makes crystallisation very difficult and for
the majority of the reported compounds no suitable crystals for
X-ray diffraction (XRD) were, or are likely to be, obtained. We
have therefore used EXAFS spectroscopy to gain information on
the local structure around the metal centres in the new complexes
in the solid state and for some compounds in solution. From the
chemical point of view, knowledge on the types of species formed
upon solubilization or on solvation properties is very important as
most of the compounds are active in solution. In our first study on
this class of compounds we reported the presence and location of
water molecules in the solid compounds and in solution based on
All chemicals were reagent grade and were used as received
unless otherwise specified. Sources: Fe(NO3)3·6H2O (Merck),
maltol (3-hydroxy-2-methyl-4-pyrone, Aldrich), ethylmaltol (2-
ethyl-3-hydroxy-4-pyrone, Pfizer or Aldrich), aniline (Aldrich),
4-toluidine (Aldrich), 4-anisidine (Aldrich), 4-(n-butyl)aniline
(Aldrich), 4-(n-hexyl)aniline (Aldrich), benzylamine (Sigma), 4-
methylbenzylamine (Sigma), 4-fluorobenzylamine (Aldrich or
Sigma) and 4-(trifluoromethyl)benzylamine (Sigma). Methanol
was Merck, PA grade; water was deionized. The 3-hydroxy-4-
pyridinone ligands were synthesized and purified as previously
described.49–51
˚
the fitting of a peak at about 4 A using the EXCURV92 code for
EXAFS analysis.41 In this work, the number of multiple scattering
paths was quite limited, being restricted to intra-ligand paths
with only two different scattering atoms. Subsequent work has
shown that the contribution of multiple scattering contributions
involving three or more scattering atoms, and inter-ligand paths
involving the first coordination shell are also important, and the
origin of shells at approximately twice the first shell metal–ligand
distance in the EXAFS Fourier transform (FT) has been the
object of discussion in the literature.42–47 General agreement is that
contributions from the shorter inter-ligand and all the intra-ligand
multiple scattering pathways should be included in order to assess
the significance of solvation molecules.
Physical techniques
IR spectra were recorded (KBr pellets) in the range 4000–400 cm−1
using a Perkin-Elmer 580B spectrometer. Fast atom bombardment
(+FAB) spectra were obtained using a Kratos concept double-
1
focusing mass spectrometer. H NMR spectra were recorded in
CD2Cl2 with a Bruker AX 250 MHz spectrometer. Elemental anal-
ysis were performed at the Microanalytical Laboratory, Depart-
ment of Chemistry, University of Manchester and at Butterworth
Laboratories Ltd., Teddington, UK. Thermogravimetric spectra
have been obtained on a TA Instruments Thermal Analyst 2000,
TGA 50 Thermogravimetric Analyzer.
Synthesis and characterization of metal complexes
Table 1 Formulae and abbreviation of the Fe(III) complexes studied
Iron(III) complexes of general formula, ML3·xH2O (L = 3-
hydroxy-4-pyridinonate), were prepared by a method analogous
to that established in the literature for aluminium(III) complexes of
this type.52 Aqueous or ethanolic solutions of the corresponding
metal nitrates were stirred with an ethanolic solution of the ligand
(in a 1 : 3 stoichiometric amount) and the pH was adjusted to 8
using an aqueous solution of sodium hydroxide. The solids were
filtered, re-crystallized from ethanol–water and kept over P2O5. All
compounds were characterized by C,H,N elemental analysis, mass
Pyrones
Rꢀ = CH3; Fe(ma)3
Rꢀ = C2H5; Fe(ma)3
Alkyl pyridinones
Rꢀ = CH3, Rꢀꢀ = H; Fe(mpp)3
Rꢀ = CH3, Rꢀꢀ = CH3; Fe(dmpp)3
Phenyl pyridinones
Rꢀ = CH3, Rꢀꢀ = Rꢀꢀꢀ = H; Fe(ppp)3
Rꢀ = C2H5, Rꢀꢀ = CH3, Rꢀꢀꢀ = H; Fe(pbtpp)3
Rꢀ = CH3, Rꢀꢀ = C4H9; Rꢀꢀꢀ = H; Fe(pbpp)3
Rꢀ = C2,H5, Rꢀꢀ = C4H9; Rꢀꢀꢀ = H; Fe(epbpp)3
Rꢀ = CH3, Rꢀꢀ = C9H13, Rꢀꢀꢀ = H; Fe(phpp)3
1
spectrometry, FTIR and H NMR and TGA analysis; we report
below the results obtained for the new compounds, which are all
deep red–brown in color. The number of hydration molecules was
calculated according to EA results and percentage of weight loss
determined from TGA analysis.
Benzyl pyridinones
Rꢀ = C2H5, Rꢀꢀ = CF3; Fe(etbpp)3
Rꢀ = C2H5, Rꢀꢀ = F; Fe(efbpp)3
Rꢀ = CH3, Rꢀꢀ = CH3; Fe(embpp)3
Tris(1-phenyl-3-hydroxy-2-methyl-4-pyridinonato)iron(III ),
Fe(ppp)3·2H2O. (Found: C, 62.04; H, 4.94; N, 5.92. Calc. for
C36H34N3O8Fe: C, 62.44; H, 4.95; N, 6.07%), FAB-MS: m/z 657
[ML3H]+; 457 [ML2H]+; 257 [MLH]+, IR: 1492, 1505, 1538, 1589,
3200–3600 cm−1.
1314 | Dalton Trans., 2006, 1313–1321
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Tris(1-(4ꢀ -tolyl)-3-hydroxy-2-ethyl-4-pyridinonato)iron(III),
Fe(eptpp)3·2.5H2O. (Found: C, 64.51; H, 5.92; N, 5.27. Calc. for
C42H47N3O8.5Fe: C, 64.21; H, 6.03; N, 5.35%), FAB-MS: m/z 693
[ML3H]+; 465 [ML2H]+; 237 [MLH]+ IR: 1398, 1496, 1556, 1600,
3200–3600 cm−1.
decomposition was observed after irradiation of multiple scans
(average of three scans per sample).
EXAFS data analysis
EXAFS spectra were reduced to absorption vs. photon energy and
background subtracted using the program PAXAS53 and the most
recent available version of EXCURVE (version 9.274) was used
for curve fitting.
For both Fe(ma)3 and Fe(dmpp)3 an excellent fit was obtained
using the published crystal structure,54,55 with refined values of
the energy zero Ef, and either three or four groups of Debye–
Waller terms. This fit used phase shifts calculated using a common
interstitial potential and which included Hedin–Lundqvist excited
state contributions, calculated using the default value of 1.325 eV
for the inverse lifetime of the core hole. An overall amplitude factor
(AFAC) of 1, was used throughout. Multiple scattering terms of
up to fifth order were included for paths containing up to five
Tris(1-(4ꢀ -butyl)phenyl-3-hydroxy-2-methyl-4-pyridinonato)-
iron(III), Fe(pbpp)3·H2O. (Found:C, 68.76;H, 6.88;N, 5.07. Calc.
for C48H56N3O7Fe: C, 68.34; H, 6.64; N, 5.07%). FAB-MS: m/z 825
[ML3H]+; 569 [ML2H]+; 313 [MLH]+, IR: 1468, 1508, 1541, 1589,
3200–3600 cm−1.
Tris(1-(4ꢀ -butyl)phenyl-3-hydroxy-2-ethyl-4-pyridinonato)-
iron(III), Fe(epbpp)3·H2O. (Found: C, 69.79; H, 7.06; N, 4.90.
Calc. for C51H62N3O7Fe: C, 69.16: H, 7.01; N, 4.75%). FAB-MS:
m/z 867 [ML3H]+; 597 [ML2H]+; 327 [MLH]+, IR: 1470, 1508,
1537, 1587, 3200–3600 cm−1.
Tris(1-(4ꢀ -hexyl)phenyl-3-hydroxy-2-methyl-4-pyridinonato)-
iron(III), Fe(phpp)3·H2O. (Found:C, 69.82;H, 7.49;N, 4.52. Calc.
for C54H68N3O7Fe: C, 69.90; H, 7.33; N, 4.53%). FAB-MS: m/z 909
[ML3H]+; 625 [ML2H]+; 341 [MLH]+, IR: 1468, 1508, 1541, 1590,
3200–3600 cm−1.
˚
different scattering atoms. The maximum pathlength was 11 A.
Previously we employed the restrained refinement procedure41
available in EXCURVE to analyse the two model compounds
together with the same species in solution. However, by including
in the present analysis all the inter-ligand multiple scattering paths
and in addition the intra-ligand paths involving bonded atoms, the
result obtained using unmodified crystallographic coordinates is
so good that we no longer consider this necessary, and we here
use the constrained refinement procedure. This involves moving
the ligand position and orientation in terms of the distance to
and rotation about a pivotal point within the ligand (atom A11
in Fig. 1). This preserves all internal distances, angles and torsion
angles within the ligand, whilst allowing those involving the central
atom to vary. As the ligands in question are bidentate, and have a
significant difference between the two metal oxygen distances, we
selected a point between the two oxygens as the pivotal point used
in refinements. In terms of EXCURVE parameters, this involves
creating a ‘ghost’ atom whose coordinates are defined as the
mid-point of the vector RO1–RO2, where RO1, RO2 are the starting
coordinates, taken from one or other of the model compound
Tris(1-(4ꢀ -fluoro)benzyl-3-hydroxy-2-ethyl-4-pyridinonato)-
iron(III), Fe(efbpp)3·2H2O. (Found: C, 60.63; H, 5.03; N, 5.11.
Calc. for C48H60N3O8Fe: C, 60.72; H, 5.23; N, 5.06%). FAB-MS:
m/z 795 [ML3H]+; 548 [ML2H]+; 302 [MLH]+, IR: 1585, 1535,
1505, 1490, 3200–3600 cm−1.
Tris(1-(4ꢀ -methyl)benzyl-3-hydroxy-2-ethyl-4-pyridinonato)-
iron(III), Fe(embpp)3·3H2O. (Found: C, 64.24; H, 6.33; N, 4.98.
Calc. for C48H62N3O9Fe: C, 64.58; H, 6.52; N 5.02%). FAB-MS:
m/z 783 [ML3H]+; 540 [ML2H]+; 298 [MLH]+, IR: 1590, 1535,
1510, 1495, 3200–3600 cm−1.
Tris(1-(4ꢀ-trifluoromethyl)benzyl-2-ethyl-3-hydroxy-4-pyridinonato)-
iron(III), Fe(etbpp)3·3H2O. (Found C, 55.74; H, 4.22; N, 4.37.
Calc. for C48H62N3O9Fe: C, 56.14; H, 4.38; N, 4.36%). FAB-MS:
m/z 945 [ML3H]+; 648 [ML2H]+, IR: 1610, 1535, 1535, 1490,
3200–3600 cm−1.
EXAFS data collection
EXAFS experiments were performed in the region of the Fe
K-edge using station 7.1 at the Synchrotron Radiation Source,
Daresbury Laboratory. The beam energy was 2 GeV and a typical
average stored ring current was 120 mA in multibunch mode and
high-brightness lattice configuration. The energy calibration was
done with a metal foil (Fe) before the start of the beamtime and re-
checked after each new beam. Data were acquired in the EXAFS
transmission mode, with argon-filled ion chambers using a Si(111)
double crystal monochromator with 50% harmonic rejection.
Powder samples were prepared in the form of homogeneous
pellets (∼100 lm of thickness) and diluted with boron nitride when
necessary. The solution samples used for the EXAFS experiments
were saturated solutions of the complexes (0.08 mol dm−3) in a
20 : 80% (v/v) water–methanol mixture—the complexes are more
soluble in methanol-rich mixtures than in methanol or water. The
solution samples were analyzed in 2 mm thick liquid cells. All the
experiments were carried out at room temperature and no sample
Fig. 1 Structure of Fe(dmpp)3 showing the terminology used. A11 is the
‘ghost’ atom.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 1313–1321 | 1315
©

View Article Online
structures as deduced by X-ray crystallography (Fig. 1). The angles
b and 180◦ − b subtended at the ghost atom are defined as e.g.
b0−1−1 = angle (Fe–A11–O1a); b = 90◦ when the two oxygen distances
are equal.
where Nind is the number of independent data points and p the
number of parameters. Nind is normally less than the number of
data points N, and in the case that the data from kmin to kmax are
Fourier filtered using a window rmin to rmax it is given by:
In the absence of restraints, the number of refinable parameters
in the EXAFS analysis is limited to about 10 parameters, and
we prefer to use rather fewer than this as some are strongly
correlated. This limited the complexity of the analysis. The model
compound Fe(dmpp)3 was admirable in that is has three identical
bidentate ligands, plus a hydration nearest shell of three water
Ni = 2(rmax − rmin)(kmax − kmin)/p
(2)
r
min and rmax should indicate the range in r-space actually fitted, not
just that where structure is apparent. The variable p should include
all parameters refined at any time, not just those included in the
last refinement. In the program, Nind is calculated automatically,
but may be overridden if the automatic value is inappropriate.
The parameter p must always be entered by the user. Both these
parameters should be quoted and justified along with v2 if changes
in v2 are to be used as evidence for a fit. The absolute value of v2
is not meaningful, unless actual experimental statistical errors ri
have been read-in, and used to weight the spectra.
˚
molecules at 4.2 A. It was therefore possible to analyse the
pyridinone complexes using crystallographic data as a starting
model, and refining just the ligand and water distances, plus a
single rotational angle. The fit to the EXAFS of Fe(ma)3 was
excellent when the published atomic positions was used, but
the structure consisted of three significantly different ligands at
slightly different distances. It was therefore necessary to devise
a simpler model for the pyranone complexes. The best model
was found to be one in which the nitrogen in the pyridinone
was replaced by an oxygen, and this proved superior to either
the use of an averaged or a single pyranone group taken from
the structure of Fe(ma)3. When considering the significance of the
solvent shells, it is necessary to be rigorous in applying statistical
criteria to the fitting procedure. For this reason, whenever the
The overall quality of fit can be described by the R factor:
ꢀ
REXAFS
=
Nl/ri(|vexp(k) − vtih (k)|) × 100%
(3)
i
i
and gives a meaningful indication of the quality of fit to the
EXAFS data in k-space. A value of around 20% would normally
be considered a reasonable fit, with values of 10% or less being
difficult to obtain on unfiltered data.
v2 values, ev are quoted, we used an experimental measure of
2
In the present work the values obtained for the R factor, which
are presented in Tables 2–4 are smaller than 20% for the solid
samples, with the exception of the most lipophilic compound for
which the value is 26% which is still reasonable for data obtained
at room temperature. For the solution samples the R factors are
in the range 16–26% and we believe that the main reason for
such a difference is the different solubility of the compounds.
We used the same solvent mixture for all the compounds and in
fact we did noticed that for Fe(dmpp)3, which is the most soluble
compound the value obtained for the R factor (16.00) is quite good
for solution.
data noise. This was obtained from a ten-point moving average of
the absolute value of the difference between the EXAFS function
and a wide-window back transform (a function which is available
automatically in EXCURVE). This not only includes statistical
noise, but the discrepancy in background subtraction and in
treatment of atomic effects that will normally be greater than
random effects for small-molecule data. Our approach thereby
involves a pessimistic evaluation of the data. The v2 function is
described by Lytle et al.56
ꢀ
ev = 1/(Nind − p)(Nind/N) NWi(vexp(k) − vtih (k))2
(1)
2
i
i
Table 2 Best fit parameters and (variances) for Fe(III) pyranonate and alkyl pyridinonate complexesa
Pyrones
Alkyl pyridinones
Fe(etma)3 solid Fe(etma)3 solution Fe(dmpp)3 solid Fe(dmpp)3 solution Fe(mpp)3 solid
Fe(ma)3 solid
Fe(ma)3 solution
Ef
−3.83 (0.30)
1.53 (0.00)
1.96 [2.00]
2.07 [2.07]
2.78 [2.77]
2.80 [2.77]
4.13 [4.12]
4.20 [4.20]
4.84
5.12
5.14
86.9 (0.2)
0.010 (0.000)
0.011 (0.000)
0.020 (0.003)
19.07
−3.32 (0.53)
1.53 (0.01)
1.98
2.06
2.78
2.80
4.14
4.20
4.85
5.13
5.14
87.6 (0.0)
0.014 (0.000)
0.011 (0.002)
0.029 (0.006)
26.61
−3.54 (0.34)
1.53 (0.00)
1.97
2.06
2.78
2.80
4.13
4.20
4.84
5.12
5.13
87.4 (0.8)
0.013 (0.003)
0.010 (0.000)
0.017 (0.003)
17.15
−4.17 (0.42)
1.54 (0.01)
1.98
2.06
2.79
2.80
4.14
4.20
4.85
5.13
5.14
87.5 (0.3)
0.010 (0.000)
0.010 (0.002)
0.020 (0.002)
22.00
−5.12 (0.22)
1.56 (0.00)
1.98 [1.99]
2.08 [2.04]
2.80 [2.79]
2.82 [2.80]
4.15 [4.15]
4.22 [4.19]
4.86
5.14
5.16
87.2 (0.2)
0.007 (0.000)
0.010 (0.000)
0.018 (0.005)
16.32
−4.08 (0.32)
1.54 (0.00)
1.98
2.07
2.79
2.80
4.14
4.21
4.85
5.13
5.14
87.4 (0.2)
0.006 (0.000)
0.009 (0.001)
0.017 (0.003)
16.00
−3.68 (0.34)
1.54 (0.00)
1.98
2.06
2.79
2.80
4.14
4.20
4.86
5.13
5.14
87.7 (0.7)
0.008 (0.002)
0.011 (0.001)
0.021 (0.002)
16.88
R11
R1
R2
R3
R4
R6
R7
R8
R9
R10
b0−1−1
a1
a3
a6
R(EXAFS)
2
ev
0.03
0.05
0.03
0.04
0.05
0.05
0.04
a
2
˚
˚
Ef is the edge position relative to the photoelectron wave vector; Rn and an refer to the distances (in A) and Debye–Waller terms (in 2 A ) for shell n;
2
R(EXAFS), ev and b are defined in the text. Rn refers to the shell containing atoms On, Cn and Nn in Fig. 1. The values inside the square brackets are
from the crystallographic data.54,55 The values inside the normal brackets are the errors associated with the Debye–Waller terms.
1316 | Dalton Trans., 2006, 1313–1321
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Table 3 Best fit parameters and (variances) for Fe(III) phenyl pyridinonate complexesa
Phenyl pyridinones
Fe(ppp)3 solid
Fe(ppp)3 solution
Fe(eptpp)3 solid
Fe(pbpp)3 solid
Fe(epbpp)3 solid
Fe(phpp)3 solid
Ef
−4.71 (0.39)
1.54 (0.01)
1.97
2.07
2.78
2.80
4.14
4.21
4.85
5.13
5.14
87.3 (0.6)
0.007 (0.002)
0.010 (0.002)
0.018 (0.006)
19.98
−4.27 (0.39)
1.54 (0.00)
1.98
2.06
2.79
2.80
4.14
4.20
4.85
5.13
5.14
87.5 (0.4)
0.008 (0.001)
0.013 (0.003)
0.015 (0.004)
25.51
−4.43 (0.32)
1.54 (0.00)
1.98
2.07
2.79
2.81
4.14
4.21
4.85
5.13
5.15
87.4 (0.4)
0.007 (0.001)
0.012 (0.002)
0.019 (0.004)
19.22
−4.47 (0.35)
1.54 (0.01)
1.98
2.06
2.78
2.80
4.14
4.20
4.85
5.13
5.14
87.5 (0.8)
0.007 (0.002)
0.011 (0.001)
0.021 (0.005)
18.99
−4.16 (0.39)
1.54 (0.01)
1.98
2.06
2.78
2.80
4.14
4.20
4.85
5.13
5.14
87.6 (0.2)
0.008 (0.000)
0.012 (0.001)
0.026 (0.002)
18.59
−4.09 (0.39)
1.54 (0.01)
1.98
2.06
2.78
2.80
4.14
4.20
4.85
5.13
5.14
87.5 (0.0)
0.008 (0.000)
0.011 (0.002)
0.020 (0.005)
19.21
R11
R1
R2
R3
R4
R6
R7
R8
R9
R10
b0−1−1
a1
a3
a6
R(EXAFS)
2
ev
0.05
0.05
0.05
0.09
0.11
0.09
a
2
˚
˚
Ef is the edge position relative to the photoelectron wavevector; Rn and an refer to the distances (in A) and Debye–Waller terms (in 2 A ) for shell n;
2
R(exafs), ev and b are defined in the text. Rn refers to the shell containing atoms On, Cn and Nn in Fig. 1 The values inside the normal brackets are the
errors associated with the Debye–Waller terms.
Table 4 Best fit parameters and (variances) for Fe(III) benzyl pyridinonate
water molecules, whose number has been confirmed through
thermogravimetric analysis measurements. Mass spectra showed
the molecular ion [ML3H]+ and fragmentation products [ML2H]+,
[MLH]+ and [LH]+ due to the loss of ligands thus confirming
the complexes have been isolated as tris-chelates The iron isotope
pattern could be observed indicating clearly which species were
iron-containing.
The IR spectra exhibit the characteristic bands for 3-hydroxy-
4-pyridinones (1620, 1560, 1535, 1515 cm−1) with bathochromic
shifts imposed by complexation and a broad band in the region
3400–3200 cm−1 due to hydration. Isolation of hydrated complexes
has been reported for the majority of metal ion complexes of the
3-hydroxy-4-pyridinone family of ligands.52,55–60 It is interesting to
point out that although hydration is found in all these new com-
pounds we observe that for those bearing longer tails and which
are consequently more lipophilic the number of water molecules
decreases. Comparing complexes with alkyl and aryl derivatives of
3-hydroxy-4-pyridinones the number of water molecules is lower
for the aryl group and we believe that the long lipophilic tails repel
water disfavouring the formation of a large number of hydrogen
bonds and so preventing the formation of hexagonal channels as
in alkyl analogues.
Taking into account the applications that we foresee for the
ligands studied in the present work and the fact that paramagnetic
iron(III) complexes have limited spectroscopic signatures we be-
lieve that the structural information obtained both in the solid and
solution, is more valuable for the characterization of the present
compounds than a more extensive spectroscopic study.
Structural information around the metal centre of the new
compounds was obtained by EXAFS as most of the complexes
are not likely to crystallize. The results obtained for all complexes
exhibit a common pattern which can be seen from the k3
weighted Fe K-edge EXAFS and Fourier transform spectra of
representative compounds presented in Fig. 2–4.
complexesa
Benzyl pyridinones
Fe(etbpp)3
Fe(efbpp)3
Fe(embpp)3
Ef
−4.16 (0.40)
1.54 (0.00)
2.00
2.05
2.80
2.80
4.15
4.20
4.87
5.14
5.14
88.5 (0.5)
0.008 (0.000)
0.010 (0.002)
0.016 (0.004)
20.49
−4.60 (0.35)
1.54 (0.00)
1.98
2.06
2.78
2.80
4.14
4.20
4.85
5.13
5.14
87.8 (1.0)
0.007 (0.003)
0.010 (0.002)
0.022 (0.004)
19.49
−4.78 (0.50)
1.54 (0.01)
1.99
2.06
2.79
2.80
4.15
4.20
4.86
5.13
5.14
87.9 (1.3)
0.009 (0.003)
0.016 (0.004)
0.040 (0.000)
26.19
R11
R1
R2
R3
R4
R6
R7
R8
R9
R10
b0−1−1
a1
a3
a6
R(EXAFS)
2
ev
0.05
0.06
0.05
a Ef is the edge position relative to the photoelectron wavevector; Rn and
2
˚
˚
an refer to the distances (in A) and Debye–Waller terms (in 2 A ) for
2
shell n; R(EXAFS), ev and b are defined in the text. Rn refers to the shell
containing atoms On, Cn and Nn in Fig. 1. The values inside the normal
brackets are the errors associated with the Debye–Waller terms.
Results and discussion
Novel iron(III) complexes with a considerable variety of N-
phenyl- and N-benzyl-3-hydroxy-4-pyridinone ligands have been
successfully prepared, in 60–75% yields. The compounds have been
obtained as deep red–brown powders and were characterized by
elemental analysis, mass spectrometry and IR spectroscopy. The
elemental analysis results are consistent with the formation of
tris chelate complexes with a variable number of crystallization
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 1313–1321 | 1317
©

View Article Online
Fig. 2 Fourier transform (left) and best theoretical fit (right) for: (a) Fe(ma)3 solid, (b) Fe(ma)3 solution, (c) Fe(etma)3 solid, (d) Fe(etma)3 solution. The
black line is the experimental; the grey line is the theory.
˚
observed for Fe(dmpp)3, as it is also evident from the values of the
angle b obtained for the best fit for each compound.
The values reported in Tables 2–4 do not include the presence of
water molecules although the model compound, Fe(dmpp)3, con-
In all spectra three peaks centred at about 2, 2.8 and 4 A are
clearly seen in the FT spectrum and these correspond to the
various shells of atoms detected around the metal centre. The
results obtained for all the compounds studied are reported in
Tables 2–4.
˚
tained a shell of three water molecules at 4.21 A. Its inclusion in the
calculation improved the fit, and was statistically significant when
the shell was described by two parameters (distance + Debye–
Waller factor). Including a shell in the pyridinone complexes also
improved the fit and appeared significant. However, modelling the
water in this way makes the assumption that a single shell of three
water molecules is present in the structure. Although it is likely
that water is present, (the number of water molecules in the new
The distances between the metal centre and the ligand atoms
are similar for all the compounds and their average structure may
˚
be described as: three oxygen atoms at 1.98 A, three oxygen atoms
˚
˚
at 2.06 A, three nitrogen or oxygen atoms at 5.13 A, and shells of
˚
three carbon atoms at 2.79, 2.80, 4.14, 4.20, 4.86 and 5.14 A. It is
˚
noticeable that there is a difference of ca. 0.08–0.10 A between the
two Fe–O distances of each of the bidentate ligands, as has been
1318 | Dalton Trans., 2006, 1313–1321
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Fig. 3 Fourier transform (left) and best theoretical fit (right) for: (a) Fe(dmpp)3 solid, (b) Fe(dmpp)3 solution, (c) Fe(mpp)3 solid. The black line is the
experimental; the grey line is the theory.
compounds varies between one and three), EXAFS is unable to
resolve more complex models, and the results presented are based
on k3 weighted refinements that exclude water.
Analysis of the set of MS pathways and their weighted
contributions, indicates that the MS pathways that are more
centre that is not significantly changed by the lipophilicity,
bulkiness, or electron-donating or -withdrawing properties of the
ligands. This result implies that the lipophilicity of the ligands
can be almost freely adjusted in order to get the ligand with the
most appropriate properties for the extraction process without
losing its strong chelating properties. EXAFS has also been used
to analyze some of the compounds in solution. All the complexes
are soluble in alcohols and alcohol-water mixtures, but are
sparingly soluble or insoluble in water and the structure in solution
was evaluated for methanol–water (20 : 80% (v/v)) saturated
solutions of the compounds Fe(ma)3, Fe(etma)3, Fe(dmpp)3 and
Fe(ppp)3.
The results obtained and that are reported in Tables 2–4
reveal that the solid state structure is maintained in solution
thus confirming the stability of the complexes in such solvents.
Further evidence of the similarity of the group of compounds is
seen in the edge region of the XAFS spectra. A small ‘pre-edge’
feature, typical of a distorted octahedral environment for a K-
edge, is present in each spectrum. The position varies slightly,
most notably in the gap between the feature and the main edge
is reduced in the solution, consistent with their more ‘molecular’
nature (there is a greater lowering of the interstitial potential in
solids).
˚
significant to the amplitude of the long distance peak at 4 A are
those involving atoms Fe–C4–C7 and Fe–C3–C6 in each unit and
those through the central atom O1–Fe–O2a involving two units.
For the tris chelates considered in this work the importance of
the latter contribution is expected since the O1–Fe–O2a angles (O1
and O2a being oxygen donor atoms of different ligands) observed
in the model compounds are close to 170◦.42 The presence of
these contributions shows that information on solvation shells is
limited by the geometry of the complexes, in particular for those
compounds in which ring ligands are present. Comparison of these
results with those to be obtained in a parallel study on bis chelates
with ligands of the same class should shed more light onto this
interesting problem.
Conclusions
All the solid iron(III) compounds have very similar local structure
providing a pseudo-octahedral environment around the metal
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 1313–1321 | 1319
©

View Article Online
Fig. 4 Fourier transform (left) and best theoretical fit (right) for: (a) Fe(ppp)3 solid, (b) Fe(ppp)3 solution, (c) Fe(pbpp)3 solid, (d) Fe(phpp)3 solid. The
black line is the experimental; the grey line is the theory.
5 P. S. Dobbin, R. C. Hider, A. D. Hall, P. D. Taylor, P. Sarpong, J. B.
Porter, G. Xiao and D. van der Helm, J. Med. Chem., 1993, 36, 2448–
2458.
6 G. Stearns and D. Stringer, J. Nutr., 1937, 13, 127–141.
7 E. M. Widdowson and R. A. McCance, J. Hyg., 1936, 36, 13–23.
8 Z. H. Zhang, P. M. Lyster, G. A. Webb and C. Orvig, Nucl. Med. Biol.,
1992, 19, 327–335.
Acknowledgements
Financial support from De Montfort University, The British
Council and Fundac¸a˜o para a Cieˆncia e Tecnologia, project
POCTI/QUI/56214/2004 is gratefully acknowledged.
9 J. Burgess, Transition Met. Chem., 1993, 18, 439–448.
10 M. J. Clarke, F. Zu and D. R. Frasca, Chem. Rev., 1999, 99, 2511–
2533.
References
11 C. J. Anderson and M. J. Welch, Chem. Rev., 1999, 99, 2219–2234.
12 R. B. Lauffer, Chem. Rev., 1987, 87, 901–927.
13 P. Caravan, J. J. Ellison, T. J. McMurry and R. B. Lauffer, Chem. Rev.,
1999, 99, 2293–2352.
1 R. C. Hider and A. D. Hall, Prog. Med. Chem., 1991, 28, 41–173.
2 J. B. Porter, B. L. C. Gyparaki, E. R. Huehns, P. Sarpong, V. Saez and
R. C. Hider, Blood, 1988, 72, 1497–1503.
3 J. B. Porter, E. R. Huehns and R. C. Hider, Ballie`res Clin. Haematol.,
1989, 2, 257–292.
14 K. H. Thompson, J. H. McNeill and C. Orvig, Chem. Rev., 1999, 99,
2561–2571.
4 J. B. Porter, R. C. Hider and E. R. Huehns, Sem. Hematol., 1990, 27,
95–100.
15 J. Burgess, B. de Castro, C. Oliveira, M. Rangel and W. Schlindwein,
Polyhedron, 1997, 16, 789–794.
1320 | Dalton Trans., 2006, 1313–1321
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
16 M. Rangel, Transition Met. Chem., 2001, 26, 219–223.
17 M. Rangel, A. Tamura, C. Fukushima and H. Sakurai, J. Biol. Inorg.
Chem., 2001, 6, 128–132.
18 Y. Yoshihawa, E. Ueda, K. Kawabe, H. Miyake, H. Sakurai and Y.
Kojima, Chem. Lett., 2000, 874–875.
19 I. Pashalidis and G. J. Kontoghiorghes, J. Radioanal. Nucl. Chem., 1999,
242, 181–184.
20 J. Arnarp, J. Bielawski, B.-M. Dahlin, O. Dahlman, C. R. Enzell and T.
Pettersson, Acta Chem. Scand., 1990, 44, 916–926.
21 Calculation of Drug Lipophilicity, ed. R. F. Rekker and R. Mannhold,
VCH, Weinheim, 1992.
39 J. Burgess and M. S. Patel, unpublished work.
40 J. Burgess, B. de Castro, C. Oliveira and M. Rangel, J. Chem. Res. (S),
1996, 234–235; J. Burgess, B. de Castro, C. Oliveira and M. Rangel,
J. Chem. Res. (M), 1996, 1338–1351.
41 W. Schlindwein, M. Rangel and J. Burgess, J. Synchrotron Radiat., 1999,
6, 579–581.
42 N. Binsted, A. van der Gaauw, O. M. Wilkin and N. A. Young,
J. Synchrotron Radiat., 1999, 6, 239.
43 H. Sakane, A. Munoz-Pa´ez, S. D´ıaz-Moreno, J. M. Mart´ınez, R. R.
Pappalardo and E. S. Marcos, J. Am. Chem. Soc., 1998, 120, 10397.
44 P. Lindqvist-Reis, A. Munoz-Paez, S. Diaz-Moreno, S. Pattanalk, I.
Persson and M. Sandsstrom, Inorg. Chem., 1998, 37, 6675.
45 S. D´ıaz-Moreno, J. M. Mart´ınez, A. Munoz-Pa´ez, H. Sakane and I.
Watanabe, J. Phys. Chem. A, 1998, 102, 7435.
22 G. J. Kontoghiorghes and A. May, Biol. Met., 1990, 3, 183–187.
23 B. L. Rai, L. S. Dekho, H. Khodr, Y. Jin, Z. D. Liu and R. C. Hider,
J. Med. Chem., 1998, 41, 3347–3359.
24 J. J. Molenda, M. M. Jones and M. A. Basinger, J. Med. Chem., 1994,
37, 93–98.
46 N. Young, J. Chem. Soc., Dalton Trans., 1996, 249.
47 N. Young, J. Chem. Soc., Dalton Trans., 1996, 1275.
48 N. Binsted, R. W. Strange and S. S. Hasnain, Biochemistry, 1992, 31,
12117.
25 S. Alshehri, J. Burgess and A. Shaker, Transition Met. Chem., 1998, 23,
689–691.
26 M. J. Herak, M. Janko and B. Tamhina, Mikrochim. Acta, 1973, 45,
783–795.
27 B. Tamhina and M. J. Herak, Croat. Chem. Acta, 1973, 45, 603–610.
28 B. Tamhina and V. Vojkovic, Mikrochim. Acta, 1986, 4, 137–147.
29 V. Vojkovif, B. Tamhina and M. J. Herak, Fresenius’ Z. Anal. Chem.,
1975, 276, 377–378.
30 A. G. Ivsic and B. Tamhina, Talanta, 1991, 38, 1403–1407.
31 B. Tamhina, M. J. Herak and K. Jakop, J. Less-Common Met., 1973,
33, 289–294.
32 M. J. Herak and B. Tamhina, Croat. Chem. Acta, 1973, 45, 237–241.
33 V. Vojkovic and B. Tamhina, Solvent Extr. Ion Exch., 1986, 4, 27–39.
34 V. Vojkovic and B. Tamhina, Solvent Extr. Ion Exch., 1987, 5, 245–253.
35 B. Tamhina, A. Ivsic and A. Gojmerac, Solvent Extr. Ion Exch., 1987,
5, 909–922.
49 Z. Zhang, S. J. Rettig and C. Orvig, Inorg. Chem., 1991, 30, 509.
50 M. Fa¨rber, H. Osiander and T. Severin, J. Heterocycl. Chem., 1994, 31,
947.
51 M. M. Finnegan, S. J. Rettig and C. Orvig, J. Am. Chem. Soc., 1986,
108, 5033.
52 W. O. Nelson, S. J. Rettig and C. Orvig, Inorg. Chem., 1989, 28, 3153.
53 N. Binsted, PAXAS, 1996.
54 M. T. Ahmet, C. S. Frampton and J. Silver, J. Chem. Soc., Dalton Trans.,
1988, 1159.
55 E. T. Clarke, A. E. Martell and J. Reibenspies, Inorg. Chim. Acta, 1992,
196, 177.
56 F. W. Lytle, D. E. Sayers and A. Stern, Physica B., 1989, 158, 701.
57 W. O. Nelson, T. B. Karpishin, S. J. Rettig and C. Orvig, Can. J. Chem.,
1987, 66, 123.
36 V. Vojkovic, A. G. Ivsic and B. Tamhina, Solvent Extr. Ion Exch., 1996,
14, 479–490.
58 W. O. Nelson, S. J. Rettig and C. Orvig, J. Am. Chem. Soc., 1987, 109,
4121.
37 V. Vojkovic, A. G. Ivsic and B. Tamhina, Croat. Chem. Acta, 1997, 70,
59 C. A. Matsuba, W. O. Nelson, S. J. Rettig and C. Orvig, Inorg. Chem.,
1988, 27, 3935.
667–677.
38 B. Tamhina and M. J. Herak, J. Inorg. Nucl. Chem., 1976, 38, 1505–
1507.
60 W. O. Nelson, T. B. Karpishin, S. J. Rettig and C. Orvig, Inorg. Chem.,
1988, 27, 1045.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 1313–1321 | 1321
©