Complexes of a novel multinucleating poly-β-diketonate ligand

Article abstract of DOI:10.1039/b410481a

The synthesis of the new polynucleating ligand 1,3-bis-(3-oxo-3-(2- hydroxyphenyl)-propionyl)-benzene (H₄L) is reported along with the preparation, structure and properties of its dinuclear complexes [Cu ₂(H₂L)₂(py)₂] (1), [Ni ₂(H₂L)₂(py)₄] (2), [Mn ₂(H₂L)₂(dmf)₄] (3), [Co ₂(H₂L)₂(dmf)₄] (4) and [Co ₂(H₂L)₂(MeOH)₄] (5), respectively. In complexes 1 to 5, the polydentate ligand is in its bis-deprotonated form, chelating the metals through its β-diketonate moieties. Magnetic measurements show that the metals within these molecules are maintained almost mutually independent.

Full text of DOI:10.1039/b410481a

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Complexes of a novel multinucleating poly-b-diketonate ligand†
Guillem Aromí,*a Christophe Boldron,b Patrick Gamez,*b Olivier Roubeau,c Huub Kooijman,d
Anthony L. Spek,d Helen Stoeckli-Evans,e Joan Ribasa and Jan Reedijkb
a
Departament de Química Inorgànica, Universitat de Barcelona, Diagonal 647,
08028 Barcelona, Spain. E-mail: guillem.aromi@qi.ub.es; Fax: +34 93490 7725;
Tel: +34 93 402 1264
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden University, PO Box 9502,
b
2300 RA Leiden, The Netherlands. E-mail: p.gamez@chem.leidenuniv.nl;
Fax: +31 71527 4464; Tel: +31 71527 4671
Centre de Recherche Paul Pascal-CNRS UPR8641, 115 avenue du Dr Schweitzer,
c
33600 Pessac, France
d
Bijvoet Center for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH Utrecht,
The Netherlands
e
Institut de Chimie, Université de Neuchâtel, Avenue de Bellevaux 51, CH-2007 Neuchâtel,
Switzerland
Received 9th July 2004, Accepted 15th September 2004
First published as an Advance Article on the web 29th September 2004
The synthesis of the new polynucleating ligand 1,3-bis-(3-oxo-3-(2-hydroxyphenyl)-propionyl)-benzene (H4L)
is reported along with the preparation, structure and properties of its dinuclear complexes [Cu2(H2L)2(py)2] (1),
[Ni2(H2L)2(py)4] (2), [Mn2(H2L)2(dmf)4] (3), [Co2(H2L)2(dmf)4] (4) and [Co2(H2L)2(MeOH)4] (5), respectively. In
complexes 1 to 5, the polydentate ligand is in its bis-deprotonated form, chelating the metals through its b-diketonate
moieties. Magnetic measurements show that the metals within these molecules are maintained almost mutually
independent.
The first complexes of this ligand are a series of dinuclear
Introduction
species with CuII, NiII, MnII and CoII. Their single crystal X-ray
The design and synthesis of complicated multidentate ligands
structure shows that the ligand is deprotonated preferentially
with the purpose of creating polynuclear complexes with pre-
at the diketone units to chelate the metals, the phenol
determined functions or structures is increasingly usual in
moieties retaining their protons. Therefore, a direct com-
coordination chemistry. This approach has become common
parison can be established between the series of compounds
practice in such wide contexts as bioinorganic modelling,1
presented here and the analogous series previously reported with
photochemistry,2 or that of molecular devices.3,4 In the area of
H3L,15,17,20 in terms of structure and magnetic properties.
molecular magnetism, however, this approach has only recently
started to take hold.5 Thus, with a few outstanding exceptions,6
the majority of systems relevant in the latter context continue
to be formed by serendipitous self-assembly,7 or by rational
assembly of building blocks through coordination, using rigid
cyanide as bridging ligand.8 Most of the polydentate ligands
designed for the formation of magnetic clusters are based
Scheme 1
on nitrogen donors. Some structural motifs recreated in this
Experimental
Synthesis
manner are squares,9 grids5 or linear chains10 of closely spaced
paramagnetic centers. We have been engaged for some time11 in
the synthesis of new polynucleating ligands possessing oxygen
as donating atoms, primarily as poly-b-diketone moieties, aimed
at obtaining novel structural motifs in cluster coordination
chemistry. A few complexes of poly-b-diketone ligands have been
published previously by other groups.12–14 We recently reported
the synthesis of the pentadentate ligand 1,3-bis-(3-oxo-3-phenyl-
propionyl)-2-hydroxy-5-methylbenzene (H3L, see scheme I).11
The structure of this molecule displays two b-diketone units
and one phenol group, disposing five oxygen donor atoms in
a linear array, aimed at inducing the aggregation of transition
metals in the form of short, molecular chains. The coordination
chemistry of this compound has proven very varied, leading to
an extensive series of dinuclear,15 trinuclear,16,17 octanuclear18
or one-dimensional infinite species.19 We describe here the
synthesis of a new polynucleating ligand (H4L, see scheme I)
based on two b-diketone and two phenol groups. This molecule
displays six oxygen donors organized in two arrays of three.
1,3-bis-(3-oxo-3-(2-hydroxyphenyl)–propionyl)–benzene
(H4L). 4 Å molecular sieves (2 g) were activated at 180 °C
under vacuum over 2 h in a 500 mL three-necked round-bottom
flask. After cooling down to room temperature, a mixture of
dimethyl isophthalate (5 g, 25.75 mmol) and 2-hydroxyaceto-
phenone (6.14 mL, 51.50 mmol) in 290 mL of ethylene glycol
dimethyl ether (DME) was added and stirred for 30 min. To this
mixture, was carefully added at 4 °C NaH (60% dispersion in
mineral oil, 8.24 g, 206 mmol). After completion of the addi-
tion, the reaction temperature was gradually raised to that of
DME reflux: 15 min at room temperature, 15 min at 50 °C and
30 min at 75 °C. After 20 h refluxing, the reaction mixture was
cooled down to room temperature and the precipitate formed
was collected by filtration. This solid material was stirred
in a mixture of 0.1 M HCl (200 mL) and dichloromethane
(DCM; 200 mL). After 1 h of stirring, the organic phase was
collected using a separating funnel. The aqueous phase was
further extracted twice with DCM (2 × 200 mL). The combined
organic layers were dried over Na2SO4, filtered and evaporated.
The remaining residue was redissolved in refluxing EtOH.
† Electronic supplementary information (ESI) available: Figs. S1 and
S2. See http://www.rsc.org/suppdata/dt/b4/b410481a/
3 5 8 6
D a l t o n T r a n s . , 2 0 0 4 , 3 5 8 6 – 3 5 9 2
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

View Article Online
Upon cooling, a yellow precipitate (pure H4L) formed which
was filtered and dried under vacuum. The yield was 27%. ESI
MS(>0): m/z 403, [LH]+. 1H NMR (CDCl3, 300 MHz): d 15.58
(s, 2 H), 12.01 (s, 2 H), 8.48 (s, 1 H), 8.10 (d, 2 H, J = 8 Hz), 7.82
(d, 2 H, J = 8 Hz), 7.63 (t, 1 H, J = 8 Hz), 7.50 (t, 2 H, J = 8 Hz),
7.03 (d, 2 H, J = 8 Hz), 6.96 (t, 2 H, J = 8 Hz), 6.91 (s, 2 H) ppm.
Anal. Calcd. (Found) for H4L·0.5 H2O: C, 70.07 (70.08); H, 4.62
(4.76). 13C NMR (CDCl3, 75 MHz): d 93.6, 119.7, 120.0, 125.8,
129.4, 130.1, 131.0, 135.2, 137.0, 163.4, 176.7, 196.8 ppm.
the d scale (down field shifts are positive). Field cooled mea-
surements of the magnetisation of smoothly powdered micro-
crystalline samples of (1, 25.48 mg), (2, 17.25 mg), (3, 7.10 mg)
and (4, 8.06 mg) were performed in the range 2–300 K with a
Quantum Design MPMS-7XL SQUID magnetometer with an
applied field of 1 kG. Corrections for diamagnetic contribu-
tions of the sample holder to the measured magnetization and
of the sample to the magnetic susceptibility were performed
experimentally and by using Pascal’s constants, respectively.
Elemental analyses were performed in-house on a Perkin Elmer
Series II CHNS/O Analyzer 2400, at the Servei de Microanàlisi
of CSIC, Barcelona, Spain.
[Cu2(H2L)2(py)2] (1). A mixture of Cu(OAc)2·H2O (25 mg,
0.13 mmol) and H4L (50 mg, 0.13 mmol) was dissolved in
pyridine (10 mL) and stirred for a few minutes. The green solu-
tion was layered with toluene and after 10 days, green crystals,
suitable for X-ray crystallography, were collected by filtration.
The yield was 52%. Anal. Calcd. (Found) for 1·3H2O·0.5py: C,
61.60 (61.49); H, 4.31 (3.99); N, 2.97 (3.16).
Results and discussion
The synthesis of the new multinucleating ligand H4L is an
extension of our work aimed at the preparation of magnetic
clusters of open-shell metals with structures that would
not otherwise be observed. The ligand in scheme I has two
well separated groups of oxygen donors, thereby with the
potential of gathering independent sets of metals within
the same molecule. The presence of four ionizable hydrogen
atoms of different acidities (two at the 1,3-diketones and two
at the phenol moieties) suggests the possibility of tuning the
reactivity of this ligand towards metals by modifying the type
and amount of base used in the reaction. The feasibility of
this methodology was demonstrated previously with a similar
ligand, H3L, containing two b-diketone groups as in H4L,
separated by a phenol group. In that case, reactions with the
acetate salts of divalent metals (Mn, Co, Ni and Cu) led to the
formation of dinuclear complexes where the phenol proton
was maintained on the ligand. The increase of the amount and
strength of base lead to complete deprotonation of the ligand
and formation of complexes with higher nuclearity. The reacti-
vity of H4L with M(OAc)2 salts has now been investigated for
the divalent metals CuII, NiII, MnII and CoII. In all cases very
similar results were obtained, based on the formation of the
corresponding dinuclear complexes, where both metals are
chelated and bridged by dianionic H2L2− through the b-diketone
moieties with retention of the phenolic protons. Thus, equi-
molar amounts of Cu(OAc)2 and H4L in pyridine lead to the
formation of the complex [Cu2(H2L)2(py)2] (1) according to the
reaction in eqn. (1).
[Ni2(H2L)2(py)4] (2). This complex was prepared using the
exact same procedure as above using Ni(OAc)2·4H2O (31 mg,
0.13 mmol) as metal salt. The yield was 61%. Anal. Calcd.
(Found) for 2·0.5H2O: C, 65.68 (65.63); H, 4.30 (4.30); N, 4.51
(4.68).
[Mn2(H2L)2(dmf)4] (3). A mixture of Mn(OAc)2·4H2O (15 mg,
0.06 mmol) and H4L (25 mg, 0.06 mmol) was stirred in pyridine
(8 mL). Soon after, a yellow precipitate started to form. The
mixture was left unperturbed overnight and the solid was
then collected by filtration and dissolved in DMF (10 mL).
The resulting yellow solution was layered with Et2O and after
10 days, small crystals of 3, suitable for X-ray diffraction were
collected by filtration. The overall yield was 32%. Anal. Calcd.
(Found) for 3: C, 59.90 (59.62); H, 5.03 (5.28); N, 4.66 (4.59).
[Co2(H2L)2(dmf)4] (4). This complex was prepared exactly as
above, using Co(OAc)2·4H2O (15 mg, 0.06 mmol). The overall
yield was 38%. Anal. Calcd. (Found) for 4·1.5H2O: C, 58.66
(58.45); H, 5.38 (5.16); N, 5.83 (5.80).
[Co2(H2L)2(MeOH)4] (5). A mixture of Co(OAc)2·4H2O
(31 mg, 0.06 mmol) and H4L (50 mg, 0.06 mmol) were dissolved
in methanol (20 mL) and stirred for 2–3 h. After this, an orange
precipitate had formed, which was collected by filtration. The
overall yield was 55%. Anal. Calcd. (Found) for 5: C, 59.66
(60.13); H, 4.62 (4.07).
2Cu(OAc)2 + 2H4L + 2py → [Cu2(H2L)2(py)2] + 4AcOH (1)
The molecular structure of 1 was determined by single crystal
X-ray diffraction (see below), which also provided crystallo-
graphic evidence for the identity of H4L. The analogous reac-
tion with Ni(OAc)2 led to the formation of the related complex
[Ni2(H2L)2(py)4] (2), where the preference of NiII for elongated
octahedral geometry is reflected.
Physical measurements
Crystals of complexes 1, 3, 4 and 5 were placed in the cold
nitrogen stream of a Nonius KappaCCD diffractometer on
rotating anode. Data for 2 were collected on a Stoe Mark II
Image Plate Diffraction System. Details on data collection and
structure determination are given in Table 1. Data were collected
at 150 K, using Mo Ka radiation (graphite monochromator,
k = 0.71073 Å). No absorption correction was applied. The
structures were solved by direct methods (1–4) or Patterson
methods (5), using SHELXS86,21 SHELXS9722 or DIRDIF.23
Refinement on F2 was carried out by full-matrix least-squares
techniques using SHELXL97.24 Hydrogen atom positions
were refined (all hydrogens of 2, hydroxyl hydrogens of 1, 3
and 5) or placed at calculated positions, riding on their carrier
atoms. All non-hydrogen atoms were refined with anisotropic
thermal parameters with the exception of those in the dis-
ordered coordinated solvent molecules (3 and 4).
When the procedure was repeated with Mn(OAc)2, the
formation of a fine yellow precipitate was observed. To
crystallize this complex it was redissolved in DMF, from
where, crystals of the solvated dimer [Mn2(H2L)2(dmf)4] (3)
were obtained, with a geometry very similar to 2. This com-
plex was of special interest, since among the related dinuclear
complexes previously obtained with the ligand H3L, the MnII
complex was the only one where (HL)2− exhibited syn–anti
conformation of the 1,3-diketonate groups.15 One possible
explanation is that this conformation maximized the energy
gained through dipolar contacts within the ligand. This is
no longer the case in the absence of the OH from the central
phenol group, thus, in complex 3, the syn–syn conformation
is restored. This observation supports the initially formulated
hypothesis. The exact same procedure with CoII resulted in the
formation of the dinuclear complex [Co2(H2L)2(dmf)4] (4). In
addition to 4, the related complex [Co2(H2L)2(MeOH)4] (5) was
also synthesized. The preparation of 5 is more convenient than
4 since the former is rather insoluble in MeOH and therefore
CCDC reference numbers: 244348–244352, for 1 to 5.
See http://www.rsc.org/suppdata/dt/b4/b410481a/ for crystal-
lographic data in CIF or other electronic format.
1H NMR measurements were collected on a 250 MHz Bruker
DXR 250 or a 300 MHz Bruker DPX 300 spectrometer in d6-
DMSO and d-CHCl3, respectively. The protio-solvent signals
were used as reference and chemical shifts were quoted on
D a l t o n T r a n s . , 2 0 0 4 , 3 5 8 6 – 3 5 9 2
3 5 8 7

View Article Online
3 5 8 8
D a l t o n T r a n s . , 2 0 0 4 , 3 5 8 6 – 3 5 9 2

View Article Online
the synthesis can be carried out in this solvent, from where the
product precipitates in high yield. X-ray quality single crystals
of 5 had been obtained unexpectedly from a different reaction
system before its rational synthesis was laid down.
In light of the complexes obtained from the above reactions,
it is important to emphasize the potential of the ligand H4L to
coordinate a larger number of metals, via the participation of
the donor atoms that are still protonated in complexes 1 to 5.
Presumably, this can be achieved during the course of reactions
involving larger proportions of metal and in the presence of
strong bases. These reactions are currently being explored.
Fig. 3 Molecular drawing of 3. Only the molecule located at a general
position is depicted. Only hydrogen atoms bonded to oxygen are shown.
Single-lines are hydrogen bonds. Bond and angle ranges (Å and °,
including the three independent molecules in the unit cell, see text):
Mn–O(dmf), 2.213(2) to 2.301(3); Mn–O(H2L), 2.082(2) to 2.151(2);
O(dmf)–Mn–O(dmf), 167.65(10) to 170.56(10); O–Mn–O(trans),
173.79(8) to 178.73(8); O(dmf)–Mn–O(cis), 83.05(9) to 99.69(9); O–Mn–
O(cis), 84.19(9) to 98.17(9); Mn1Mn1a, 7.5401(12); Mn2Mn2a,
7.6433(12); Mn3Mn4, 7.6092(12).
Description of structures
The structures of complexes 1, 2, 3 and 5 are represented in
Figs. 1 to 5. In Table 1 the crystallographic data for all the
complexes are displayed, whereas average selected metric
parameters of each complex are in their respective figure
legends. The structure details of complex 4 are included as ESI†
since its geometry is essentially the same as that of complex 5.
Fig. 4 Molecular drawing of 5, showing the labelling of unique non-
carbon atoms. Only hydrogen atoms bonded to oxygen are shown.
Single-lines are hydrogen bonds. Bond and angle (Å and °) ranges: Co–
O(MeOH), 2.103(2) and 2.128(2); Co–O(H2L), 2.0272(19) to 2.0591(19);
O(MeOH)–Co–O(MeOH), 176.09(9); O(MeOH)–Co–O, 87.30(8) to
92.57(9); O–Co–O(cis), 87.18(8) to 93.57(8); O–Co–O(trans), 179.05(8)
and 179.16(8); CoCo, 7.411(1).
Fig. 1 Molecular drawing of 1, showing the labelling of unique non-
carbon atoms. Only hydrogen atoms bonded to oxygen are shown.
Single-lines are hydrogen bonds. Bond and angle ranges (Å and °):
Cu–N, 2.317(2); Cu–O, 1.9300(17) to 1.9597(18); N–Cu–O; 91.86(8) to
98.74(7); O–Cu–O(cis), 85.40(8) to 92.49(7); O–Cu–O(trans), 165.90(7)
and 175.65(8); CuCu, 7.539(2).
Fig. 2 Molecular drawing of 2, showing the labelling of unique non-
carbon atoms. Only hydrogen atoms bonded to oxygen are shown.
Single-lines are hydrogen bonds. Bond and angle (Å and °) ranges:
Ni–N, 2.1015(13) and 2.1381(15); Ni–O, 1.9997(14) to 2.0251(10); N–
Ni–N, 175.70(5); N–Ni–O, 85.79(5) to 92.17(5); O–Ni–O(cis), 88.35(4)
to 94.40(4); O–Ni–O(trans), 176.24(4) to 176.54(5); NiNi, 7.3226(6).
[Cu2(H2L)2(py)2] (1). The molecular structure of 1 (Fig. 1)
consists of a centrosymmetric dimer containing two CuII ions
bridged and chelated by two H2L2− ligands through their b-
diketonate moieties, which keep the metals 7.539(2) Å apart.
The metal ions are thus equatorially coordinated by four
O-donors in the form of two opposite six-membered rings.
The square pyramidal coordination geometry around copper
is completed by axial pyridine ligands pointing to opposite
directions from the plane of the molecule. The coplanarity
of the coordination pockets within each ligand is gauged by
the angle between the corresponding idealized chelate rings,
which in 1 measures 23.45(8)°. The structure of 1 reveals the
fact that H4L is preferentially deprotonated at the methylene
positions between the carbonyl groups of the b-diketone frag-
ments, presumably because the ensuing metal–ligand coordi-
nation moiety is very stable, and also because this allows the
Fig. 5 Representation of complex 5, emphasizing the hydrogen-
bonding network within the crystal lattice, involving MeOH and Et2O
of crystallization. Hydrogen atoms not involved in hydrogen bonding
were omitted for clarity.
formation of four hydrogen bonds between the phenol groups
and their adjacent oxygen atoms, respectively (Fig. 1). The
hydrogen atoms participating in these weak interactions have
been found crystallographically, and the corresponding O–HO
distances are 2.590(3) and 2.525(3) Å. The structure of 1 can be
directly compared to that of the complex [Cu2(HL)2(dmf)2],
reported previously.20 In the latter, the solvate ligands are
dimethylformamide molecules instead of pyridine. The main
D a l t o n T r a n s . , 2 0 0 4 , 3 5 8 6 – 3 5 9 2
3 5 8 9

View Article Online
difference between both complexes is the distribution of intra-
molecular hydrogen bonds, as a consequence of the different
location and number of phenol groups within the ligand. In
the complex with H3L, the ligand has only one phenol group
separating both diketone units, which participates in one intra-
molecular hydrogen bond upon formation of the complex.
This has a small structural effect, the CuII ions being only
0.16 Å further apart than in 1. The packing of the molecules
of 1 within the crystal takes place through extensive p-stacking
between the phenyl rings from H2L2− ligands and chelate rings
involving the b-diketonate groups of adjacent molecules. The
shortest distances between the centroids of the interacting p
systems are 3.50 Å.
between the two, the angle between chelate rings seems to be
related to the nature of the terminal axial ligand (see below).
The O–HO distances of the intramolecular hydrogen bonds
are 2.533(3) and 2.519(3) Å. In the solid state, a co-operative
hydrogen bond network joins the complex molecules into an
infinite, one-dimensional chain, through the intermediacy of
solvent molecules of crystallization (Fig. 5). In this network, O4
of the ligand accepts a hydrogen bond from free methanol, while
O7 (from coordinated MeOH) acts as a proton donor to the free
methanol. Again, complex 5 has its counterpart with the ligand
H3L, this time in the form of the molecule [Co2(HL)2(py)4].25
Comparison of the above structures allows one to observe a
possible relation between the angle between the least-squares
planes through the chelate rings within one ligand and the
identity of the axial ligand, with the following ordering: pyridine
(23.45, 16.73) > dmf (8.67, 9.90, 7.15, 7.99, 6.3) > MeOH (2.54).
Inspection of these parameters for the corresponding complexes
with (HL)2− reveals that these trends are maintained in this
family of compounds. Thus these angles are 29.63 and 30.56 for
complexes [M2(HL)2(py)4] (MII = NiII and CoII),15 respectively,
and 14.59 for [Cu2(HL)2(dmf)2].20 The reasons for this trend are
currently unclear, although they could be of steric nature, given
the correlation between the angle and the size of the ligand.
[Ni2(H2L)2(py)4] (2). The molecular structure of 2 (Fig. 2)
is similar to that of 1 in that two H2L2− ligands bridge two
MII ions (here, NiII) through the chelating diketonate units.
The main difference is the fact that solvation of the vacant
coordination sites of the metals takes place now at both axial
positions, leading to the usual distorted octahedral coordina-
tion geometry of NiII. The separation between NiII ions is
7.3226(6) Å, whereas the idealized six member chelate rings
within a ligand form an angle of 16.73(5)°, considerably
smaller than in 1. The O–HO distances of the intramolecular
hydrogen bonds are 2.5273(16) and 2.5068(16) Å, respectively.
Complex 2 is structurally related to the previously reported
complex [Ni2(HL)2(py)4].15 Perhaps the most striking aspect of
this comparison is that in the complex with (HL)2−, the metal
ions are 0.6 Å further apart than in 2.
Magnetic properties
Bulk magnetic susceptibility measurements were performed for
complexes 1 to 4 in the 2 to 300 K temperature range under a
constant magnetic field of 0.1 T. The results are represented
together as vmT vs. T plots (where vm is the molar paramagnetic
susceptibility) in Fig. 6, after correction for contributions.
[Mn2(H2L)2(dmf)4] (3). The structure of 3 (Fig. 3) was quite
revealing in that it shows two MnII ions linked through chela-
tion of both metals by each of two H2L2− ligands displaying
cis–cis conformation. This is the same conformation as in
complexes 1 and 2, but is in contrast with the congener of 3
with H3L, [Mn2(HL)2(py)4], where the ligand is found in the
syn–anti conformation.15 In both complexes, MnII is in a dis-
torted octahedral geometry. However, the ligand conformation
defines whether the metal configuration is trans (complex 3, with
dmf axial ligands) or cis ([Mn2(HL)2(py)4]). The cause of the
conformation in the latter complex is likely to be the stabiliza-
tion energy provided by the additional hydrogen bonds that are
allowed within (HL)2− in this configuration. Such hydrogen
bonds are not possible with H4L, thus all complexes formed
with this ligand exhibit the cis–cis form. The determination of
the structure of 3 constituted an interesting crystallographic
problem, since the crystals were twinned and the structure
could only be refined by applying a pseudomerohedral twin
operation. The lattice contains three independent molecules of
3, two located on inversion centers and one on a general posi-
tion. Therefore, three different MnMn distances are found,
7.5401(12), 7.6433(12) and 7.6092(12) Å. Likewise, four unique
values of the angle between chelate rings within the same ligand
are found, namely, 8.67(11), 9.90(11), 7.15(11) and 7.99(11).
The intramolecular O–HO distances are all within the range
of 2.510 to 2.576 Å. The main differences between the different
molecules of 3 in the unit cell are the orientations of the axial
dmf molecules. There are no significant packing forces in the
crystal structure other than van der Waals’ interactions.
Fig. 6 Plots of vmT vs. T per molecule for complexes 1 to 4.
The value of vmT at 300 K for [Cu2(H2L)2(py)2] (1) is
0.9 cm3 K mol−1, which corresponds to the value of two
independent CuII ions (S = 1/2) with g = 2.19. The isotropic
g value, as extracted from X-band powder EPR at 77 K is
2.17 (Fig. S1, see ESI†). vmT remains constant over the whole
temperature range indicating that H2L−2 does not mediate
any detectable magnetic superexchange between the metals
of the molecule. This is in contrast to the related complex
[Cu2(H2L)2(dmf)2], where a small interaction between the
metals was found of J = −0.73 cm−1 (with the H = −2JS1S2
convention for the Heisenberg–Spin Hamiltonian).19 Presum-
ably, this is allowed by the central phenol group of (HL)−2
(see Scheme I), which forms a hydrogen bond to one oxygen
atom bound to CuII and is absent in H2L−2. This observation
underscores the feasability of modifying the magnetic properties
within transition metal assemblies by introducing small modi-
fications on polynucleating ligands.
[Co2(H2L)2(MeOH)4] (5). Complex 5 (Fig. 4) is a dinuclear
centrosymmetric complex of CoII ions bridged and equatorially
coordinated by two H2L2− ligands, in the same fashion as in the
previous complexes. The solvate axial ligands that complete
the elongated octahedral geometry around cobalt are MeOH
molecules. The intermetallic distance is 7.4106(12) Å, while
the complex shows the smallest angle of the series between
both chelate rings within each ligand, which measures 2.54(9)°.
While inspection of this pair of parameters for the different
compounds studied revealed the absence of any correlation
The complex [Ni2(H2L)2(py)4] (2) exhibits a room temperature
value of vmT of 2.32 cm3 K mol−1, which would be expected for
two uncoupled octahedral NiII centers (S = 1) with g = 2.15.
This value stays constant as the temperature is decreased and
drops abruptly only at temperatures below 20 K approximately.
3 5 9 0
D a l t o n T r a n s . , 2 0 0 4 , 3 5 8 6 – 3 5 9 2

View Article Online
Table 2 250 MHz 1H NMR data of [Co2(H2L)2(dmf)4] (4) in d6-
In light of the observations from 1, we exclude antiferro-
magnetic coupling between NiII centers as the reason for this
drop. Instead, the decrease of vmT is likely to be caused by zero
field splitting (ZFS) of the spin ground state of the individual
ions. An assessment of the magnitude of the ZFS parameter that
would cause the decrease of vmT was made by fitting the data to
an equation containing the term D.26 The best fit (Fig. S2, see
ESI†) was found for D = 6.8 cm−1.
DMSO
Resonance
d/ppm
a (5 resonances)
b (2 resonances)
32.1, 27.9, 19.0, 16.2, 14.6
29.9, 26.7
c (2 resonances)
dmf (ligand, 2 resonances)
97, 81
7.7 or 8.0,a 4.3
Consistent with the above observations, vmT for [Mn2(H2L)2-
(dmf)4] (3) remains constant at approximately 9.1 cm3 K mol−1
over the whole temperature interval studied. This value is
that calculated for two non-interacting high-spin MnII centers
(S = 5/2) with g = 2.04, which falls within the range commonly
observed for this ion. Again, these observations show that H4L is
capable of assembling two paramagnetic centers within the same
molecule, keeping them magnetically isolated from each other.
The variable temperature magnetic susceptibility of [Co2-
(H2L)2(dmf)4] (4) is markedly different than that observed for
complexes 1 to 3, as is evident from Fig. 6. The value of vmT at
300 K is 6.65 cm3 K mol−1, which is higher than expected for the
presence in one molecule of two isolated spin-only CoII centers
with S = 3/2. Such a high value is observed because, in fact, the
spin angular momentum in this ion is coupled with the orbital
angular momentum. The continuous drop of the product vmT
taking place with decreasing temperature is the consequence of
the depopulation of states of higher magnetic moment within
the term 4T1 of CoII. This drop is therefore not ascribed to any
antiferromagnetic exchange within the dinuclear complex and
leads to a value for vmT of 3.46 cm3 K mol−1 at 2 K.
a This resonance comes together with another from free solvent.
1
Fig. 7 250 MHz H NMR spectrum of [Co2(H2L)2(dmf)4] (4) in d6-
DMSO. The top of the full spectrum has been cropped for clarity. The
inset is a magnified (arbitrary intensity units) portion of the region
of lowest fields. See text for details about the labels.
Paramagnetic 1H NMR
a conclusion fuels the prospect of using these compounds in
solution as building blocks for further reactivity. For example,
larger composite entities could be grown in this manner by
exploiting the potential of the “unused”phenol moieties to bind
other substrates, such as other metals or even more complicated
building blocks. The study of the reactivity of these complexes
with various sources of lanthanide ions is currently in progress.
Complexes 1 to 5 were examined by 250 MHz 1H NMR spectro-
scopy. Of all of them, only the CoII compounds could be pre-
pared in enough concentration or provided suitable spectra.
The other complexes were either too insoluble (1 and 2) or the
resonances where too broad to be detected (3). Paramagnetic
1H NMR on synthetic coordination clusters of CoII has been
employed occasionally to investigate their properties in solu-
tion.27 Also, this technique has been very useful for obtaining
crucial structural information from metalloproteins containing
this ion or its analogues.28
Conclusions
The preparation of the first MII complexes of the new poly-
nucleating ligand H4L has revealed that the latter loses pre-
ferentially H+ from its b-diketonate units rather than from the
phenol groups. The ensuing dinuclear compounds conserve,
unlike previously observed for the related ligand H3L, the same
ligand conformation throughout complexes with CuII, NiII,
MnII and CoII, namely, syn,syn with respect to the b-diketonate
groups. In complexes 1 to 4, both paramagnetic centers are
magnetically independent from each other. Work is in progress
in order to isolate complexes of H4L with higher metal loading,
by using bases capable of ionizing the phenolic ‘OH’ moieties
of this ligand.
In Fig. 7 is the spectrum of [Co2(H2L)2(dmf)4] (4) in d6-
DMSO at room temperature, whereas in Table 2 is a list
of the chemical shifts observed. The spectrum shows nine
paramagnetically shifted and broadened peaks as expected
for complex 4 with D2h symmetry in solution. Of these, the
two most broadened and shifted (labelled as c) are assigned
to the protons lying closest to CoII, namely, the phenol and
the methine groups of the ligand. The remaining signals were
associated to the aromatic protons from H2L2−, of which the
two with apparently half intensity (labelled as b) were attributed
to the two inequivalent H atoms on the central ring. The other
five are labelled as a. Of the peaks in the diamagnetic area,
the signals from free solvents present in the system (DMSO,
H2O, Et2O and DMF) were identified by their chemical shifts
(see Fig. 7 and Table 2). Two signals in the vicinity of 8 and
4 ppm, respectively where assigned to the two inequivalent ‘CH3’
groups from bound dimethylformamide. These two groups
are magnetically inequivalent because of the double bond
character of the C–N bond in this ligand, which prevents free
rotation around this vector. Other examples of this asymmetry,
manifested by paramagnetic 1H NMR of complexes containing
this solvent-ligand have been observed.29 This assignment in
complex 4 is supported by the fact that the 1H NMR spectrum
(not shown) of the related complex 5 (which does not contain
dmf) is almost identical to that in Fig. 7, but does not show
the peaks attributed to dmf. The aldehyde proton of the ligand
dmf is expected to be too broad to be detected. These results
suggest that complexes 4 and 5 are maintained in solution with
the idealized D2h symmetry observed in the solid state. Such
Acknowledgements
The authors are thankful to the Spanish Ministerio de Ciencia
y Tecnología (“Ramón y Cajal” contract to GA and Grant
BQU2003/0539 to JR), to the Dutch WFMO (Werkgroep
Fundamenteel-Materialen Onderzoek) (PG, CB and JR) CW
(Foundation for the Chemical Sciences) and NWO (Organiza-
tion for the Scientific Research) (PG, JR, CB and ALS), to the
French CNRS (Centre National de la Recherche Scientifique),
to the Conseil Général d’Aquitaine, (OR) and to the Swiss
National Science Foundation (HS-E) for financial support.
References
1 R. H. Bode, J. E. Bol, W. L. Driessen, F. B. Hulsbergen, J. Reedijk
and A. L. Spek, Inorg. Chem., 1999, 38, 1239.
2 S. Campagna, S. Serroni, F. Puntoriero, F. Loiseau, L. De Cola,
C. L. Kleverlaan, J. Becher, A. P. Sorensen, P. Hascoat and
N. Thorup, Chem. Eur. J., 2002, 8, 4461.
D a l t o n T r a n s . , 2 0 0 4 , 3 5 8 6 – 3 5 9 2
3 5 9 1

View Article Online
3 M. Fujita, K. Umemoto, M. Yoshizawa, N. Fujita, T. Kusukawa
and K. Biradha, Chem. Commun., 2001, 509.
4 V. Balzani, A. Credi and M. Venturi, Chem. Eur. J., 2002, 8, 5524.
5 L. Zhao, Z. Q. Xu, L. K. Thompson and D. O. Miller, Polyhedron,
2001, 20, 1359.
6 E. Breuning, M. Ruben, J. M. Lehn, F. Renz, Y. Garcia,
V. Ksenofontov, P. Gütlich, E. Wegelius and K. Rissanen, Angew.
Chem., Int. Ed., 2000, 39, 2504.
7 R. E. P. Winpenny, J. Chem. Soc., Dalton Trans., 2002, 1.
8 V. Marvaud, C. Decroix, A. Scuiller, C. Guyardduhayon,
J. Vaissermann, F. Gonnet and M. Verdaguer, Chem. Eur. J., 2003,
9, 1677.
17 G. Aromí, P. Gamez, O. Roubeau, P. Carrero, H. Kooijman, A. L.
Spek, W. L. Driessen and J. Reedijk, Inorg. Chem., 2002, 41, 3673.
18 G. Aromí, P. Gamez, O. Roubeau, H. Kooijman, A. L. Spek, W. L.
Driessen and J. Reedijk, Angew. Chem., Int. Ed., 2002, 41, 1168.
19 G. Aromí, J. Ribas, P. Gamez, O. Roubeau, H. Kooijman, A. L.
Spek, S. Teat, E. MacLean, H. Stoeckli-Evans and J. Reedijk,
Chem., Eur. J., 2004 in press.
20 H. Kooijman, A. L. Spek, G. Aromí, P. Gamez, P. Carrero-Berzal,
W. L. Driessen and J. Reedijk, Acta Crystallogr., 2002, E58,
m223.
21 G. M. Sheldrick, in SHELXS86 Program for Crystal Structure
Determination, Germany, 1986.
9 D. S. Cati, J. Ribas, J. Ribas-Arino and H. Stoeckli-Evans, Inorg.
Chem., 2004, 43, 1021.
10 C. J. Matthews, S. T. Onions, G. Morata, L. J. Davis, S. L. Heath
and D. J. Price, Angew. Chem., Int. Ed., 2003, 42, 3166.
11 G. Aromí, P. Gamez, P. Carrero, W. L. Driessen and J. Reedijk,
Synth. Commun., 2003, 33, 11.
12 N. A. Bailey, D. E. Fenton, J. Lay, P. B. Roberts, J.-M. Latour and
D. Limosin, J. Chem. Soc., Dalton Trans., 1986, 2681.
13 V. A. Grillo, E. J. Seddon, C. M. Grant, G. Aromí, J. C. Bollinger,
K. Folting and G. Christou, Chem. Commun., 1997, 1561.
14 R. W. Saalfrank, N. Low, S. Trummer, G. M. Sheldrick, M. Teichert
and D. Stalke, Eur. J. Inorg. Chem., 1998, 559.
15 G. Aromí, P. Gamez, O. Roubeau, P. Carrero, H. Kooijman, A. L.
Spek, W. L. Driessen and J. Reedijk, Eur. J. Inorg. Chem., 2002, 1046.
16 G. Aromí, P. Carrero, P. Gamez, O. Roubeau, H. Kooijman, A. L.
Spek, W. L. Driessen and J. Reedijk, Angew. Chem., Int. Ed., 2001,
40, 3444.
22 G. M. Sheldrick, in SHELXS97 Program for Crystal Structure
Determination, Germany, 1997.
23 P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman,
S. García-Granda, R. O. Gould, J. M. M. Smits, and C. Smycalla,
The DIRDIF99 program system, Technical Report of the Crystallo-
graphy Laboratory, University of Nijmegen, The Netherlands,
1999.
24 G. M. Sheldrick, in SHELXS86 Program for Crystal Structure
Refinement, Germany, 1997.
25 Unpublished results.
26 O. Kahn, Molecular Magnetism, VCH, New York, 1993, p. 17.
27 G. Aromí, A. S. Batsanov, P. Christian, M. Helliwell, O. Roubeau,
G. A. Timco and R. E. P. Winpenny, Dalton Trans., 2003, 4466.
28 J. Salgado, A. P. Kalverda, R. E. M. Diederix, G. W. Canters, J. M.
Moratal, A. T. Lawler and C. Dennison, J. Biol. Inorg. Chem.,
1999, 4, 457.
29 G. Aromí, PhD Thesis, Indiana University, Bloomington, IN, 1999.
3 5 9 2
D a l t o n T r a n s . , 2 0 0 4 , 3 5 8 6 – 3 5 9 2