Self-assembled polymetallic square grids ([2 × 2] M4, [3 × 3] M9) and trigonal bipyramidal clusters (M5) - Structural and magnetic properties

Article abstract of DOI:10.1039/b602595a

New self-assembled grids and clusters are reported, with square [2 × 2] M₄ (M = Mn(ii)₄, Cu(ii)₄), trigonal-bipyramidal Mn(ii)₅, and square [3 × 3] M₉ (M = Mn(ii), Cu(ii)) examples. These are based on a series of ditopic and tritopic hydrazone ligands involving pyridine, pyrimidine and imidazole end groups. In all cases the metal centres are bridged by hydrazone oxygen atoms with large (>125°) bridge angles, leading to antiferromagnetic exchange for all the Mn systems (J = -2 to -5 cm^-1), which results in S = 0 (Mn₄), and S = 5/2 (Mn₅, Mn₉) ground states. The copper systems have a 90° alternation of the Jahn-Teller axes within the Cu₄ and Cu₈ grid rings (Cu₉), which leads to magnetic orbital orthogonality, and dominant ferromagnetic coupling. For the Cu₉ grid antiferromagnetic exchange between the ring and the central copper leads to a S = 7/2 ground state, while for the Cu₄ grids S = 4/2 ground states are observed. The magnetic data have been treated using isotropic exchange models in the cases of the Cu₄ and Cu₉ grids, and the Mn₅ clusters. However due to the enormity of a fully isotropic calculation a simplified model is used for the Mn₉ grid, in which the outer Mn₈ ring is treated as the equivalent of an isolated magnetic chain, with no coupling to the central metal ion. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b602595a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Self-assembled polymetallic square grids ([2 6 2] M , [3 6 3] M ) and
4
9
trigonal bipyramidal clusters (M )—structural and magnetic properties{{
5
a
a
a
a
Louise N. Dawe, Tareque S. M. Abedin, Timothy L. Kelly, Laurence K. Thompson,* David O. Miller,
a
a
Liang Zhao, Claire Wilson, Michael A. Leech and Judith A. K. Howard
b
b
b
Received 20th February 2006, Accepted 12th April 2006
First published as an Advance Article on the web 4th May 2006
DOI: 10.1039/b602595a
New self-assembled grids and clusters are reported, with square [2 6 2] M
4 4
(M = Mn(II) ,
Cu(II) ), trigonal-bipyramidal Mn(II) , and square [3 6 3] M (M = Mn(II), Cu(II)) examples.
4
5
9
These are based on a series of ditopic and tritopic hydrazone ligands involving pyridine,
pyrimidine and imidazole end groups. In all cases the metal centres are bridged by hydrazone
oxygen atoms with large (>125u) bridge angles, leading to antiferromagnetic exchange for all the
2
1
Mn systems (J = 22 to 25 cm ), which results in S = 0 (Mn ), and S = 5/2 (Mn , Mn ) ground
4
5
9
states. The copper systems have a 90u alternation of the Jahn–Teller axes within the Cu and Cu
4
8
grid rings (Cu
coupling. For the Cu
leads to a S = 7/2 ground state, while for the Cu
magnetic data have been treated using isotropic exchange models in the cases of the Cu
grids, and the Mn clusters. However due to the enormity of a fully isotropic calculation a
simplified model is used for the Mn grid, in which the outer Mn ring is treated as the equivalent
of an isolated magnetic chain, with no coupling to the central metal ion.
9
), which leads to magnetic orbital orthogonality, and dominant ferromagnetic
grid antiferromagnetic exchange between the ring and the central copper
grids S = 4/2 ground states are observed. The
and Cu
9
4
4
9
5
9
8
Introduction
Reaction of ligands in this class with a variety of transition
metal salts (MX , M = Mn, Co, Ni, Cu, Zn) leads to [2 6 2]
2
The ability to control molecular assembly through a pro-
grammed approach of creating the substituents with desired
features, and then allowing the products to form essentially
unassisted, offers enormous advantages over other methods of
synthesis, where reaction outcomes are frequently left largely
to chance, or the end product is the result of a complex
sequence of individual steps. Yields by such methods are
frequently very low. High yield self-assembly reactions can be
very effective for the formation of [n 6 n] poly-metallic grids,
where the donor content of the ligand pockets is designed to
match the Lewis acid nature, and geometric preference, of a
self assembled heteroleptic or homoleptic M grids in high
4
8–13
yield (Scheme 1).
The close proximity of the metal ions
leads to spin exchange coupling within the grids, with
1
–3
transition metal ion. Examples of self-assembled [2 6 2] and
4
[
3 6 3] grids have been reported based on pyridazine and
pyrimidine ligands, and in rare cases extension of such ligands
5
–7
has led to larger [4 6 4] grids.
Ditopic ligands, e.g. poap (Chart 1), built on a picolinic
hydrazone core fragment, have a bidentate and a tridentate
coordination pocket, with a hydrazone oxygen atom, which
can act as a bridge between two metal centres. The formation
of five-membered chelate rings on coordination ensures that
a roughly linear M–O–M coordination fragment results.
a
Department of Chemistry, Memorial University, St. John’s,
Newfoundland, A1B 3X7, Canada
b
Department of Chemistry, University of Durham, South Road, Durham,
UK DH1 3LE
{
This article is part of a themed issue on Molecular Magnetic
Materials.
Electronic supplementary information (ESI) available: Fig. S1: plot
{
of mmol as a function of temperature for 4; Fig. S2: magnetization data
as a function of field at 2 K for 4; Fig. S3: plot of mmol as a function of
temperature for 6. See DOI: 10.1039/b602595a
Chart 1
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2645

View Article Online
Experimental
Commercially available solvents and chemicals were used
without further purification.
Physical measurements
Infrared spectra were recorded as Nujol mulls using a Mattson
Polaris FT-IR instrument, and UV/Vis spectra were obtained
using a Cary 5E spectrometer. Micro-analyses were carried out
by Canadian Microanalytical Service, Delta, Canada. Variable
temperature magnetic data (2–300 K) were obtained using a
Quantum Design MPMS5S SQUID magnetometer using field
strengths in the range 0.1 to 5 T. Background corrections for
the sample holder assembly and diamagnetic components of
the complexes were applied. NMR measurements were taken
with a Bruker AVANCE 500 MHz spectrometer. LCMS
measurements were taken using an Agilent 1100 Series LC/
MSD in APCI mode with methanol as the solvent.
Scheme 1
anti-ferromagnetic and ferromagnetic examples. Homoleptic
1
1,14
self-assembled [M
L
5 6
] (M = Mn, Co, Zn)
clusters can also
Synthesis of ligands
form, in rare cases, with ditopic ligands in this class, when co-
ligand competition is minimized.
mpoap. 6-Methyl-2-picolinic acid was prepared by partial
29
Extension of this ligand type with an extra pocket produces
a linear tritopic entity, e.g. 2poap and related derivatives
oxidation of 2,6-lutidine with KMnO
4
,
and converted to the
30
2 4
SO ; 60%). The ethyl ester (9.9 g,
ethyl ester (abs. EtOH/c. H
(
Chart 1), with the ability to bind to three metal ions. Self-
6
0 mmol) was dissolved in methanol (30 mL) and added slowly
assembly of ligands in this class leads to homoleptic [3 6 3] M
9
5–25
to a chilled solution of hydrazine hydrate (7.5 g, 120 mmol)
in dry methanol (30 mL), and the mixture allowed to stir
overnight. The corresponding hydrazide was obtained as a
white solid, and recrystallized from benzene (yield 4.2 g, 47%).
Mp 93–94 uC. The methyl ester of imino-2-picolinic acid was
prepared in situ by reaction of 2-cyanopyridine (2.08 g,
1
grids (Scheme 1) in high yield (M = Mn, Fe, Ni, Cu, Zn).
˚
The short M–M distances (y4 A) and large M–O–M bridge
angles (>125u) in these grids lead to spin exchange with
antiferromagnetic (Mn, Fe, Ni) and ferromagnetic (Cu)
examples. The Mn grids exhibit a unique suite of reversible
9
redox processes in the range 0.5–1.5 V vs. Ag/AgCl leading to
an eight electron redox window associated with eight Mn(II)
centers oxidizing to Mn(III). The electronic bistability of these
systems, and the fact that they can be attached to gold
2
0 mmol) with sodium methoxide solution, produced by
dissolving sodium metal (0.05 g, 2.2 mmol) in dry methanol
100 mL). 6-Methyl-2-picolinic acid hydrazide (3.02 g,
0 mmol) was added to the above solution and the mixture
(
2
2
3
26
(
Au(111)) and HOPG surfaces to form self-assembled
monolayers in some cases, and probed using STM and CITS
current imaging tunneling spectroscopy) techniques, has
refluxed for 24 h. A yellow crystalline solid was obtained,
which was filtered off, washed with water, methanol and
(
diethyl ether and dried under vacuum (yield 3.8 g, 75%). Mp
1
1
clearly indicated their potential for molecular device behaviour
in the nanoscale realm.
98–200 uC. H NMR (500 MHz, DMSO-d
6
, 25 uC): 10.5 (s,
H, OH), 8.64 (d, 1H, Ar), 8.22 (d, 1H, Ar), 7.90 (m, 3H, Ar),
1
Related ditopic and tritopic ligands are under study, with
the objective of creating increased donor functionality at
strategic locations on the ligand backbone. This has been a
focus in some recent work where thioether, and sulfide R
2 3
7.50 (m, 2H, Ar), 6.95 (s, 2H, NH ), 2.60 (s, 3H, CH ). Mass
spectrum (major mass peaks, m/z): 255 (M), 239, 238, 237, 208,
2
1
163, 107, 105. IR (Nujol mull, cm ): 3376, 3312 (nNH); 1684,
1677 (nCLO); 1589, 1565 (nCLN). Anal. calcd (%) for
C H N O: C, 61.16; H, 5.13; N, 27.43. Found: C, 61.47;
2
3,26
groups (Chart 1)
have been introduced as additional sites
1
3
13
5
for grid elaboration, and for enhanced surface attachment. In
the present study some new ditopic and tritopic ligands have
been developed with imidazole and pyrimidine end groups
H, 4.71; N 27.38.
pomp. 2-Acetyl pyridine (1.8 g, 15.0 mmol) was added
dropwise to a solution of picolinic hydrazide (1.35 g,
10.0 mmol) in 15 mL of methanol. The resulting solution
was refluxed for 6 h. A noticeable cloudiness of the solution
was observed within 1 h. A white precipitate formed, which
was filtered off, washed with methanol and ether, and vacuum
dried (yield 2.0 g, 83%). Mp 195–197 uC. Mass spectrum (m/z):
(e.g. ioapm, 2pmoap), and their coordinating ability with
Mn(II) and Cu(II) salts examined. These ligands are of interest
because of the potential for additional extended coordination
ability of the imidazole and pyrimidine groups after primary
ligand coordination. This ‘complex as ligand’ approach has
been shown to lead to extended coordination successfully with
2
7,28
+
21
?
complexes of, for example, oxamate based ligands.
present study examples of [2 6 2] and [3 6 3] square grids,
and Mn trigonal bipyramidal clusters involving just primary
In the
241.3 (M ), IR (Nujol mull, cm ) 3316 (nNH), 1701 (nCO),
995 (npy). NMR (DMSO-d , ppm) 2.5 (s, CH ), 7.45 (t), 7.72
6
3
5
(t), 7.90 (t), 8.11 (t), 8.16 (t), 8.64 (d), 8.75 (d), 11.15 (s, OH).
Anal. calcd (%) for C H N O: C, 64.98; H, 5.03; N, 23.31.
ligand coordination, are reported. Structural and magnetic
properties are highlighted.
1
3
12
4
Found: C, 64.69; H, 5.10; N, 23.07.
2
646 | J. Mater. Chem., 2006, 16, 2645–2659
This journal is ß The Royal Society of Chemistry 2006

View Article Online
ioapm. 4-Methyl-5-imidazole carboxylic acid hydrazide was
Mn
4
(C12
H
10
N
5
O)
4
(N
3
)
?4H O (bulk sample): C, 40.58; H,
4 2
generated by the dropwise addition of hydrazine hydrate (85%)
3.41; N, 31.54. Found: C, 40.57; H, 3.21; N, 31.06.
(
2.34 g, 46.7 mmol) to a stirred solution of ethyl-4-methyl-5-
imidazole carboxylate (3.39g, 22.0mmol) in 60 mL of
-butanol and 25 mL of methanol. The resulting clear, yellow
[Mn (pomp–H) (H O) (CF SO )](CF SO ) ?CH OH (2).
3
4
4
2
3
3
3
3
3 3
1
pomp (0.100 g, 0.420 mmol) was added to a solution of
2
O (3.5 mL, 0.03 g mL in MeOH–H
1
solution was refluxed for 72 h. Upon cooling a tan precipitate
formed, which was collected by suction filtration and washed
with ether (84% yield). The iminoester of 2-pyrimidinecarbo-
nitrile was generated by dissolving 2-pyrimidinecarbonitrile
Mn(CF
3
SO
3
)
2
?xH
2
2
O
4 : 1) with the addition of 2 drops of Et
3
N. The resulting deep
red solution was kept for crystallization after stirring for one
hour and filtration. Orange crystals were collected after two
days (0.055 g, 26% yield). Slow diffusion of ether into a
(
(
2.35 g, 22.4 mmol) in methanol (100 mL) with sodium metal
0.10 g, 4.4 mmol). The resulting clear colourless solution was
MeOH–CH
crystals. Anal. calcd (%) for (C13
(CH OH) (H O) (bulk sample): C, 37.54; H, 3.84; N, 10.94.
Found (%): C, 37.63; H, 2.85; N, 10.91.
3
CN solution of the complex gave X-ray quality
capped with a drying tube and stirred at room temperature for
h. Acetic acid was then added until a neutral pH was
achieved. 4-Methyl-5-imidazole carboxylic acid hydrazide
2.10 g, 15 mmol) was added to the stirred solution of
-pyrimidine iminoester, which formed a clear yellow solution.
H
11
N
4
O) Mn (CF SO ) -
4
4
3
3 4
8
3
8
2
(
2
[Cu
1.00 mmol) was added to a warm solution of Cu(ClO
6H O (0.740 g, 2.00 mmol) in methanol–water (30 mL : 5 mL).
4
(mpoap–H)
4
](ClO
4
)
4
?8H
2
O
(3). mpoap (0.255 g,
A yellow precipitate formed upon stirring for 5 min. Methanol
was added until a total volume of 250 mL was achieved,
and the solution was refluxed for 16 h. Upon heating, the
precipitate dissolved. The refluxed solution was cooled to
room temperature and remained clear and yellow. The volume
was reduced under reduced pressure to 20 mL, and 100 mL of
diethyl ether was added. A yellow precipitate formed, which
was collected by suction filtration, washed with methanol and
diethyl ether, and vacuum dried (yield 3.4 g, 93%). Mp 222–
4 2
) ?
2
The resulting suspension was stirred until complete dissolution
of the ligand occurred forming a dark green solution. The
solution was filtered and allowed to stand at room tempera-
ture. X-Ray quality green crystals formed after several days
2
1
(yield 82%). IR (Nujol mull, cm ) 1664 (s), 1611 (w), 1590
2
(m), 1571 (m) (nCLN, nCLO); 1085 (s) (n ClO
Nujol mull) l (nm) 696, 800 sh. Anal. Calcd (%) for
Cu O) (ClO ?0.5H O (vacuum dried sample): C,
7.22; H, 2.94; N, 16.69. Found: C, 37.20; H, 2.90; N, 16.71.
); Vis.
4
(
+
?
21
4
(C13
H
12
N
5
4
4
)
4
2
2
25 uC. Mass spectrum (m/z): 246.1 (M ), IR (n/cm ) 3332
nNH), 1654, 1562 (nCN, nCO), 952 (npym). Anal. Calcd (%)
for C H N O?1.5MeOH?1.2H O: C, 43.86 H, 6.17 N, 31.15.
3
(
1
0
11
7
2
[Cu (ioapm–H) ](ClO ) ?3CH CN (4). ioapm (0.21 g,
4 4 4 4 3
Found: C, 43.86; H, 4.80; N, 31.31.
0
6
1
.86 mmol) was added to a warm solution of Cu(ClO
O (0.31 g, 0.84 mmol) in methanol–acetonitrile (10 mL :
0 mL). Addition of 3 drops of Et N produced a clear, dark
4 2
) ?
H
2
2
pmoap. Dimethyl-2,6-pyridinedicarboxylate (10.07 g,
1.60 mmol) was converted to the dihydrazide by reaction
with hydrazine hydrate (85%) (2.60 g, 51.9 mmol) in methanol
yield 40%). 2-Pyrimidinecarbonitrile (1.82 g, 17.3 mmol)
5
3
green coloured solution, which was stirred for an hour and
kept after filtration for crystallization. A green precipitate
formed after standing for one week. Dark green crystals,
suitable for a structural study, were obtained by recrystalliza-
tion from 50 : 50 methanol–acetonitrile (0.043 g, 12% yield).
(
was reacted with sodium methoxide solution, produced by
dissolving sodium metal (0.15 g, 6.5 mmol) in dry methanol
(
150 mL) to generate the iminoester. The solution was then
Anal. calcd (%) for Cu
dried sample): C, 28.61; H, 2.76; N, 23.38. Found (%): C,
8.63; H, 2.59; N, 23.57.
4 10 7 4 4 4 2
(C10H N O) (ClO ) ?3H O (bulk
neutralized using acetic acid, and 2,6-pyridinedicarboxylate
dihydrazide (1.52 g, 7.79 mmol) added giving a yellow mixture,
which was refluxed for 24 h. A yellow solid was obtained,
which was filtered off, washed with methanol and diethyl ether
and dried under vacuum. Yield (2.08 g, 66%). Mass spectrum
2
[
Mn
(0.100 g, 0.420 mmol) was added to a solution of excess
Mn(ClO ?6H O (0.50 g, 2.0 mmol) in a methanol–acetoni-
5 6 4 4 3 3
(pomp–H) ](ClO ) ?CH CN?1.5CH OH (5). pomp
+
21
?
m/z): 406 (M ), IR (Nujol mull, cm ): 3413, 3332, 3181
(
(
4
)
2
2
nNH); 1685, 1612 (nCLO); 1558, 1519 (nCLN). Anal. calcd (%)
trile mixture (10 mL : 10 mL), forming initially a light yellow
15 2 3
for C17H O N11?2.25CH OH: C, 48.42; H, 5.06; N, 32.27.
clear solution. Addition of 2–3 drops of Et N produced a
3
Found: C, 48.97; H, 3.71; N 32.27.
clear, deep red coloured solution, which was stirred for an
hour and kept after filtration for crystallization. Red, rod like
crystals were collected after two days (0.060 g, 38% yield).
Synthesis of complexes
8
Mn (poap–H) (N ) ]?2H O (1). poap (0.240 g, 1.00 mmol)
[
Slow diffusion of ether into a MeOH–CH CN solution of the
3
4
4
3 4
2
was added to a warm solution of Mn(ClO
4
)
2
?6H
2
O (0.360 g,
complex gave X-ray quality crystals. Anal. calcd (%) for
1
.00 mmol) in methanol–water (20 mL : 20 mL) forming an
(C H N O) Mn (ClO ) (CH CN)(CH OH) (H O)
3.5
(bulk
1
3
11
4
6
5
4 4
3
3
2
2
orange solution. A concentrated aqueous solution (2 mL)
of sodium azide (0.2 g, 1.3 mmol) was added with stirring,
and the resulting orange solution was allowed to stand
at room temperature overnight. Dark orange crystals
suitable for structural study formed (yield 70%). IR (Nujol
sample): C, 43.27; H, 3.72; 15.38. Found (%): C, 43.11; H,
3.46; N, 15.10.
[Mn (ioapm–H) ](ClO ) ?7H O (6). ioapm (0.10 g,
2
5
6
4 4
0.41 mmol) was added to a solution of Mn(ClO ) ?6H O
4
2
2
2
1
mull, cm ) 2075 (m), 2046 (m) (nN ), 1651 (m), 1541 (m)
3
(0.17 g, 0.47 mmol) in methanol–acetonitrile (10 mL : 10 mL)
forming initially a light yellow clear solution. Addition of
(
nCLN, nCLO); 1007 (w) (npy). Anal. calcd (%) for
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2647

View Article Online
2
3
drops of Et N produced a clear, deep orange coloured
solution, which was stirred for an hour and kept after filtration
for crystallization. X-Ray quality, red, prismatic crystals were
collected after two days (0.040 g, 30% yield). Anal. calcd (%)
for (C10
H
10
7
N O)
6 5 4 4 2
Mn (ClO ) ?2.5H O: C, 33.30; H, 3.00; N,
2
6.98. Found (%): C, 32.79; H, 2.67; N, 27.40.
[
9 6 3 6 2
Mn (2pmoap–2H) ](NO ) ?28H O (7). 2pmoap (0.49 g,
1
.2 mmol) was added to a solution of excess Mn(NO ) ?6H O
3
2
2
(
1.01 g, 3.5 mmol) in methanol–acetonitrile (20 mL : 20 mL)
forming initially a light yellow clear solution. This was left to
stir with gentle heating for 2 hours. It was then filtered and
solvent was removed from the filtrate to yield 0.96 g of a
yellow powder. This was dissolved in methanol–acetonitrile
(
10 mL : 10 mL) to give a clear, yellow solution. NH
4 3
CH COO
(
0.04g, 0.51 mmol) was added producing a clear, deep red
coloured solution that was stirred with gentle heating for
hour, and at room temperature for three hours. This solution
1
was filtered and allowed to stand at room temperature. X-Ray
quality, red prismatic crystals formed after three days (0.86 g,
9
6% yield). Anal. calcd (%) for (C17
O)13 (bulk dried sample): C, 34.78; H, 2.98; 28.65. Found
%): C, 34.35; H, 2.70; N, 28.94.
13 11 2 6 9 3 6
H N O ) Mn (NO ) -
(
H
2
(
[
Cu (2pmoap–2H) (2pmoap–H) ](ClO ) ?13.2H O (8).
9 2 4 4 10 2
2
pmoap (0.12 g, 0.30 mmol) was added to a warm solution
?6H O (0.13 g, 0.35 mmol) in methanol–
of Cu(ClO
4
)
2
2
acetonitrile (10 mL : 10 mL). A clear, dark green solution
formed. The solution remained clear and dark green upon the
addition of 3 drops of Et N. It was stirred with gentle heating
3
for 45 minutes and kept after filtration for crystallization.
X-Ray quality, green prismatic crystals formed after standing
at room temperature for two weeks (0.030 g, 15% yield). Anal.
calcd (%) for (C17
H
13
N
11
O
2
)
2
(C17
H
14
N
11
O
2
)
4
Cu
O)16 (bulk sample): C, 28.62; H, 2.68; N, 21.61. Found
%): C, 29.31; H, 2.34; N, 20.91. The slight discrepancy in
9 4 10
(ClO ) -
(
H
2
(
analytical data may indicate some solvent (e.g. methanol) in
the bulk sample.
Crystallography
The diffraction intensities of an orange plate-like crystal of
1
of dimensions 0.52 6 0.12 6 0.04 mm were collected
with graphite-monochromatized Mo-Ka X-radiation using a
Bruker P4/CCD diffractometer at 193(1) K to a maximum 2h
value of 52.8u. The data were corrected for Lorentz and
polarization effects. The structure was solved by direct
3
1,32
methods.
All atoms except hydrogens were refined
anisotropically. Hydrogen atoms were placed in calculated
positions with isotropic thermal parameters set to twenty
percent greater than their bonded partners at the time of their
inclusion, and were not refined. Neutral atom scattering
3
3
34,35
factors and anomalous-dispersion terms
were taken from
the usual sources. All other calculations were performed with
3
6
the teXsan crystallographic software package. The maximum
and minimum peaks on the final difference Fourier map
2
23
˚
A
corresponded to 1.98 and 20.44 e
, respectively.
Abbreviated crystal data for 1 are given in Table 1.
Diffraction intensities for 2 and 5 were collected similarly,
2
648 | J. Mater. Chem., 2006, 16, 2645–2659
This journal is ß The Royal Society of Chemistry 2006

View Article Online
and structural solutions achieved using the same methods (see
Table 1, and ESI{).
The diffraction intensities of a green prism-like crystal of 4
of dimensions 0.72 6 0.17 6 0.11 mm, were collected with
graphite-monochromatized Mo-Ka X-radiation using
a
Rigaku AFC8 Saturn CCD diffractometer at 153(1) K to a
maximum 2h value of 62.0u. The data were corrected for
Lorentz and polarization effects. The structure was solved by
3
7
38
direct methods, and expanded using Fourier techniques.
Some non-hydrogen atoms were refined anisotropically, while
others were refined isotropically. Hydrogen atoms were
introduced in calculated positions with thermal parameters
set twenty percent greater than their bonding partners, and
were refined on the riding model. Neutral atom scattering
3
3
34,35
factors and anomalous-dispersion terms
were taken from
the usual sources. All calculations were performed using the
3
9,40
CrystalStructure
crystallographic software packages,
except for refinement, which was carried out using SHELX-
3
7. Diffraction intensities for 6–8 were collected similarly,
1
9
and structural solutions achieved using the same methods (see
Fig. 1 Structural representation of complex 1 (30% probability
Table 1, and ESI{).
thermal ellipsoids).
The data collection for 3 was carried out using graphite
monochromated Mo-Ka X-radiation with a Siemens SMART
4
1
[Mn (pomp–H) (H O) (CF SO )](CF SO ) ?CH OH (2).
4 4 2 3 3 3 3 3 3 3
CCD detector diffractometer equipped with a Cryostream
The structure of the cation in 2 is shown in Fig. 2, and
important bond distances and angles are listed in Table 3.
The square [2 6 2] grid consists of two eclipsed, parallel
tetradentate pomp ligands bound on the upper and lower
N
2
flow cooling device. Series of narrow v-scans (0.3u) were
performed at several Q-settings in such a way as to cover a
hemisphere or a full sphere of data to a maximum resolution of
˚
.7 A. Cell parameters were determined and refined within
0
4
2
faces of the Mn –(m-O) core. The four vacant sites in this
4 4
SMART
using the centroid values of approximately
arrangement are in the same relative positions as in 1, and are
500 selected reflections with 2h values between 20 and 45u.
43
Raw frame data were integrated using the SAINT program.
occupied by three water molecules, and a triflate anion. Mn–
˚
Mn distances fall in the range 3.94–4.00 A, with Mn–O–Mn
The data were corrected for absorption by an empirical psi-
4
scan based method. The structure was solved using direct
4
angles in the range 128.8–129.9u. C–O distances in the ligands
˚
fall in the range 1.298–1.309 A, indicating a predominance of
2
44
methods and refined by full-matrix least squares on F . All
atoms except hydrogen were refined with anisotropic atomic
displacement parameters (ADPs). Hydrogen atoms were
placed geometrically, with ADPs set at twenty percent of their
bonded partners, and were not refined. Abbreviated crystal
data for 3 are given in Table 1.
single CO bond character. The infrared spectrum of the ligand
(
vide infra) indicates significant ketonic character in this group
2
1
,
on the basis of the observation of a nCLO band at 1700 cm
and that some ligand rehybridization has occurred on
coordination.
CCDC reference numbers 258937, 299029–299035. For
crystallographic data in CIF or other electronic format see
DOI: 10.1039/b602595a
[
4 4 4 4 2
Cu (mpoap–H) ](ClO ) ?8H O (3). The structural represen-
tation of the homoleptic tetranuclear cation in 3 is shown in
Fig. 3, and important bond distances and angles are listed in
Results and discussion
˚
Table 2 Bond distances (A) and angles (u) for 1
Structures
Mn(1)–O(4)
Mn(1)–N(24)
Mn(1)–N(21)
Mn(1)–O(1)
Mn(1)–N(20)
Mn(1)–N(1)
Mn(2)–N(3)
Mn(2)–N(27)
Mn(2)–O(2)
Mn(2)–O(1)
Mn(2)–N(6)
Mn(2)–N(5)
Mn(3)–N(13)
Mn(3)–N(8)
Mn(3)–O(3)
Mn(3)–O(2)
2.152(6)
2.183(7)
2.196(7)
2.213(5)
2.227(7)
2.241(6)
2.151(6)
2.175(7)
2.212(5)
2.218(5)
2.268(7)
2.411(6)
2.141(6)
2.151(6)
2.173(5)
2.203(5)
Mn(3)–N(11)
Mn(3)–N(10)
Mn(4)–N(30)
Mn(4)–O(3)
Mn(4)–N(18)
Mn(4)–O(4)
Mn(4)–N(15)
Mn(4)–N(16)
Mn(1)–Mn(4)
Mn(3)–Mn(4)
Mn(1)–Mn(2)
Mn(2)–Mn(3)
Mn(1)–O(1)–Mn(2)
Mn(3)–O(2)–Mn(2)
Mn(4)–O(3)–Mn(3)
Mn(1)–O(4)–Mn(4)
2.310(7)
2.385(7)
2.143(8)
2.167(5)
2.179(7)
2.204(5)
2.274(7)
2.380(8)
3.919(4)
3.882(4)
4.007(5)
3.977(4)
[
4 4 3 4 2
Mn (poap–H) (N ) ]?2H O (1). The structure of the neutral
[
2 6 2] grid is shown in Fig. 1, and important bond distances
and angles are listed in Table 2. The tetradentate ligands
are arranged in parallel eclipsed pairs on either side of the
[
Mn –(m-O) ] square core leaving four open coordination sites,
4 4
which are occupied by four terminally bound azide groups,
with one each bound to Mn(2) and Mn(4) and two in a cis-
orientation bound to Mn(1). Mn–L distances are typical of
square grids of this type, and the relative positions of the azide
129.5(2)
9
ligands are similar to those reported in previous examples.
˚
Mn–Mn distances fall in the range 3.88–4.01 A, with
128.5(2)
126.9(2)
128.2(3)
Mn–O–Mn angles in the range 126.9–129.5u.
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2649

View Article Online
˚
Table 4 Bond distances (A) and angles (u) for 3
Cu(1)–N(29)
Cu(1)–N(76)
Cu(1)–O(32)
Cu(1)–N(26)
Cu(1)–O(70)
Cu(2)–N(48)
Cu(2)–N(38)
Cu(2)–O(51)
Cu(2)–N(40)
Cu(2)–O(32)
Cu(3)–N(9)
Cu(3)–N(57)
Cu(3)–O(12)
Cu(3)–N(1)
1.9008(19)
1.9942(19)
2.0050(17)
2.056(2)
2.2441(17)
1.903(2)
1.994(2)
1.9985(17)
2.044(2)
2.2042(17)
1.901(2)
1.978(2)
Cu(3)–O(51)
Cu(4)–N(67)
Cu(4)–N(18)
Cu(4)–O(70)
Cu(4)–N(59)
Cu(4)–O(12)
Cu(1)–Cu(2)
Cu(1)–Cu(4)
Cu(2)–Cu(3)
Cu(3)–Cu(4)
Cu(3)–O(12)–Cu(4)
Cu(1)–O(32)–Cu(2)
Cu(2)–O(51)–Cu(3)
Cu(4)–O(70)–Cu(1)
2.2252(17)
1.895(2)
1.982(2)
1.9925(17)
2.038(2)
2.1999(17)
3.957(1)
3.967(2)
3.951(1)
3.943(1)
139.31(8)
140.03(8)
138.57(8)
138.78(8)
2.0047(17)
2.045(2)
˚
and 2.199–2.245 A respectively). There are no additional axial
˚
contacts to the copper centers within 3.5 A. This may be due to
Fig. 2 Structural representation of complex cation in
2 (30%
probability thermal ellipsoids).
the presence of the 6-methyl groups, which project enough to
at least partially impede the approach of e.g. a water molecule
at each copper ion center. The ligands are arranged in parallel
pairs, with separations close to the metal–metal contacts,
˚
Table 3 Bond distances (A) and angles (u) for 2.
Mn(1)–O(8)
Mn(1)–N(15)
Mn(1)–O(1)
Mn(1)–O(4)
Mn(1)–N(1)
Mn(1)–N(16)
Mn(2)–N(6)
Mn(2)–N(3)
Mn(2)–O(1)
Mn(2)–O(2)
Mn(2)–N(5)
Mn(2)–N(4)
Mn(3)–O(5)
Mn(3)–O(2)
Mn(3)–N(10)
Mn(3)–O(3)
2.171(2)
2.181(2)
2.1840(19)
2.2095(18)
2.230(2)
2.312(2)
2.175(2)
2.176(2)
2.1938(19)
2.2250(18)
2.278(2)
2.309(2)
2.166(2)
2.1791(18)
2.191(2)
2.2342(19)
Mn(3)–N(8)
Mn(3)–N(9)
Mn(4)–O(6)
Mn(4)–O(4)
Mn(4)–O(7)
Mn(4)–O(3)
Mn(4)–N(12)
Mn(4)–N(13)
Mn(1)–Mn(2)
Mn(2)–Mn(3)
Mn(3)–Mn(4)
Mn(1)–Mn(4)
Mn(1)–O(1)–Mn(2)
Mn(3)–O(2)–Mn(2)
Mn(4)–O(3)–Mn(3)
Mn(4)–O(4)–Mn(1)
2.267(2)
2.328(3)
2.139(2)
2.1605(18)
2.178(2)
2.1857(19)
2.236(2)
2.249(2)
3.948(1)
3.980(2)
3.999(1)
3.959(1)
above and below the Cu
4
pseudo-plane, but adopt opposed
anti conformations within each pair.
4 4 4 4 3
[Cu (ioapm–H) ](ClO ) ?3CH CN (4). The structure of the
homoleptic tetranuclear cationic fragment in 4 is shown in
Fig. 4, and important bond distances and angles are listed in
Table 5. The core structure is shown in Fig. 5. Four square-
pyramidal copper ions are bound to four tetradentate ligands
arranged in two parallel groups above and below the metal
128.80(9)
129.28(8)
129.55(8)
129.87(9)
Table 4. The Cu O core is typical for self assembled ‘square’
4
4
grids of this sort, with a boat conformation, Cu–Cu separa-
˚
tions in the range 3.943–3.967 A, and Cu–O–Cu angles in the
range 138.5–140.1u. The copper centers have square-pyramidal
stereochemistries, with an alternation of short (equatorial) and
˚
long (axial) Cu–O contacts around the square (1.895–2.056 A
Fig. 4 Structural representation of the complex cation in
E
POVRAY image).
4
(
˚
Table 5 Bond distances (A) and angles (u) for 4
Cu(1)–N(4)
Cu(1)–N(8)
Cu(1)–O(1)
Cu(1)–N(7)
Cu(1)–O(2)
Cu(2)–N(10)
Cu(2)–N(2)
1.912(5)
1.940(5)
1.956(4)
2.057(5)
2.318(4)
1.924(5)
1.967(5)
Cu(2)–O(2)
Cu(2)–N(13)
Cu(2)–O(1)
Cu(1)–Cu(2)
Cu(1)–Cu(29)
Cu(1)–O(1)–Cu(2)
Cu(2)–O(2)–Cu(1)
1.992(4)
2.051(5)
2.349(4)
4.035(3)
4.056(3)
139.00(19)
140.42(19)
Fig. 3 Structural representation of complex cation in
3 (30%
probability thermal ellipsoids).
2
650 | J. Mater. Chem., 2006, 16, 2645–2659
This journal is ß The Royal Society of Chemistry 2006

View Article Online
˚
Table 6 Bond distances (A) and angles (u) for 5
Mn(1)–O(5)
Mn(1)–O(1)
Mn(1)–O(6)
Mn(1)–N(20)
Mn(1)–N(24)
Mn(1)–N(1)
Mn(2)–N(6)
Mn(2)–N(3)
Mn(2)–O(1)
Mn(2)–O(2)
Mn(2)–N(4)
Mn(2)–N(5)
Mn(3)–O(4)
Mn(3)–O(2)
Mn(3)–O(3)
Mn(3)–N(9)
Mn(3)–N(8)
Mn(3)–N(13)
Mn(4)–N(11)
Mn(4)–N(18)
Mn(4)–O(5)
2.165(4)
2.168(4)
2.189(4)
2.221(5)
2.230(5)
2.247(5)
2.173(5)
2.183(5)
2.186(4)
2.187(4)
2.295(5)
2.330(5)
2.151(4)
2.170(4)
2.177(4)
2.226(5)
2.236(5)
2.249(5)
2.171(5)
2.183(5)
2.191(4)
Mn(4)–O(3)
Mn(4)–N(17)
Mn(4)–N(12)
Mn(5)–N(22)
Mn(5)–N(15)
Mn(5)–O(6)
Mn(5)–O(4)
Mn(5)–N(21)
Mn(5)–N(16)
Mn(1)–Mn(2)
Mn(2)–Mn(3)
Mn(3)–Mn(5)
Mn(1)–Mn(5)
Mn(1)–Mn(4)
Mn(3)–Mn(4)
Mn(1)–O(1)–Mn(2)
Mn(3)–O(2)–Mn(2)
Mn(3)–O(3)–Mn(4)
Mn(3)–O(4)–Mn(5)
Mn(1)–O(5)–Mn(4)
Mn(1)–O(6)–Mn(5)
2.201(4)
2.301(5)
2.311(5)
2.173(5)
2.180(5)
2.189(4)
2.207(4)
2.290(5)
2.321(5)
3.896(3)
3.943(3)
3.941(3)
3.941(2)
3.909(3)
3.991(3)
126.92(18)
129.66(18)
131.51(18)
129.41(18)
127.67(18)
128.40(18)
E
4 (POVRAY image).
Fig. 5 Core structural representation of
Arrows highlight the Jahn–Teller axes.
pseudo-plane. The ligands have anti arrangements within each
pair (unlike 1 and 2). The charge balance in 4 requires that all
ligands bear a 21 charge as usual. Cu–Cu distances fall in the
grid, poised appropriately for further coordination. The
pyrimidine and imidazole nitrogen atoms (N(5), N(3), N(11)
and N(12) and their symmetry related counterparts) form
potentially chelating donor pairs.
˚
range 4.03–4.06 A, with Cu–O–Cu angles in the range 139.0–
1
40.4u, typical of the square copper grids. The copper square
pyramids have three nitrogen atoms and an oxygen atom in the
basal plane (Cu(2), N(2), N(10), N(13), O(29); Cu(1), N(4),
N(7), N(8), O(1)) with short Cu–L distances, and a much
longer axial contact to a second bridging oxygen atom (Cu(1)–
5 6 4 4 3 3
[Mn (pomp–H) ](ClO ) ?CH CN?1.5CH OH (5). The struc-
ture of the homoleptic cation in 5 is shown in Fig. 6, and
important bond distances and angles are given in Table 6. The
core structural fragment, showing just the metal ions and
bridging oxygen atoms, is shown in Fig. 7. Six tetradentate
˚
˚
O(2) 2.318(4) A, Cu(2)–O(1) 2.349(4) A). This results in a
magnetic ground state at each copper centre (vide infra),
dx22y2
and Cu–O distances around the square core which alternate
short (equatorial), long (axial), such that the copper atoms are
linked in an orbitally strictly orthogonal fashion. The
elongated Jahn–Teller axes are shown in Fig. 5. The same
model applies to complex 3.
5 6
ligands are bound to a trigonal bipyramidal [Mn –(m-O) ] core
cluster, such that there is an exact match between the bonding
capacity of the six ligands and the coordination requirements
of the five octahedral metal ions. This self-assembled structure
11,14
has been observed before,
The non-coordinating nitrogen atoms on the imidazole and
pyrimidine groups are arranged on the outer surface of the
and is an alternative to the
[2 6 2] heteroleptic Mn
4
structure commonly observed with
this class of ligand. Anion and solvent donor competition are
regarded as the primary influences in the cluster outcome, with
the weakly coordinating perchlorate favouring the formation
of the trigonal bipyramidal cluster. Mn–L and Mn–Mn
Fig. 6 Structural representation of complex cation in
5
(30%
Fig. 7 Core structural representation of complex 5 (30% probability
probability thermal ellipsoids).
thermal ellipsoids).
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2651

View Article Online
E
Fig. 9 Core structural representation of complex 6 (POVRAY
image).
Fig. 8 Structural representation of the complex cation in
E
POVRAY image).
6
donor pairs from the pyrimidines (e.g. N(33)–N(35), N(12)–
(
N(13) and other pairs) are highlighted in Fig. 8.
˚
distances (3.89–4.00 A) are similar to those observed in the
2 6 2] square systems, with Mn–O–Mn angles in the range
26.9–131.6u.
[
Mn (2pmoap–2H) ](NO ) ?28H O (7). The structure of the
3 6
9
6
2
[
homoleptic nonanuclear Mn
9
cation in 7 is shown in Fig. 10,
1
and important bond distances and angles are listed in Table 8.
The grid comprises a [Mn –(m-O) ] core encompassed by
9 12
[
Mn (ioapm–H) ](ClO ) ?7H O (6). The structure of the
5 6 4 4 2
two groups of three parallel, roughly planar ligands arranged
5
homoleptic pentanuclear Mn cation is shown in Fig. 8, and
important bond distances and angles are listed in Table 7. Six
ligands are bound in parallel pairs around the Mn (m-O) core,
5
6
with the ligands opposed (anti) in each pair, similar to the
arrangement in 5. The orientation of the cation in Fig. 8
highlights the parallel ligand arrangement. Each ligand adopts
a 21 charge, as in 5, and so the imidazole groups are not
deprotonated. The pentanuclear core structure is shown in
˚
Fig. 9. Mn–Mn distances fall in the range 3.865–3.950 A, and
Mn–O–Mn angles fall in the range 126.5–128.7u. These
compare closely with those found in 5.
The outer surface of the cluster is comprised of six
pyrimidine and six imidazole subunits, which could be
involved in further coordination. Potentially chelating N2
˚
Table 7 Bond distances (A) and angles (u) for 6
Mn(1)–N(25)
Mn(1)–O(2)
Mn(1)–N(11)
Mn(1)–O(4)
Mn(1)–N(14)
Mn(1)–N(28)
Mn(2)–N(32)
Mn(2)–O(1)
Mn(2)–N(4)
Mn(2)–O(5)
Mn(2)–N(7)
Mn(2)–N(34)
Mn(3)–O(4)
Mn(3)–O(1)
Mn(3)–N(37)
Mn(3)–N(2)
Mn(3)–N(23)
Mn(3)–O(6)
Mn(4)–O(3)
Mn(4)–N(9)
Mn(4)–N(30)
2.147(4)
2.182(4)
2.183(5)
2.197(4)
2.392(5)
2.425(5)
2.121(5)
2.135(4)
2.139(5)
2.159(4)
2.438(5)
2.460(5)
2.185(4)
2.194(4)
2.200(5)
2.204(5)
2.206(5)
2.207(4)
2.184(4)
2.194(5)
2.197(5)
Mn(4)–O(5)
Mn(4)–N(16)
Mn(4)–O(2)
Mn(5)–N(18)
Mn(5)–N(39)
Mn(5)–O(6)
Mn(5)–O(3)
Mn(5)–N(21)
Mn(5)–N(41)
Mn(1)–Mn(4)
Mn(4)–Mn(5)
Mn(4)–Mn(2)
Mn(2)–Mn(3)
Mn(5)–Mn(3)
Mn(1)–Mn(3)
Mn(2)–O(1)–Mn(3)
Mn(1)–O(2)–Mn(4)
Mn(4)–O(3)–Mn(5)
Mn(3)–O(4)–Mn(1)
Mn(2)–O(5)–Mn(4)
Mn(5)–O(6)–Mn(3)
2.201(4)
2.201(5)
2.211(4)
2.162(5)
2.169(5)
2.172(4)
2.186(4)
2.395(5)
2.423(5)
3.944(3)
3.919(3)
3.921(3)
3.865(3)
3.946(3)
3.950(3)
126.46(17)
127.73(17)
127.50(17)
128.69(17)
128.18(17)
128.61(18)
Fig. 10 Structural representation of the complex cation in
E
POVRAY image).
7
(
˚
Table 8 Bond distances (A) and angles (u) for 7
Mn(1)–N(8)
Mn(1)–N(4)
Mn(1)–O(2)
Mn(1)–O(1)
Mn(1)–N(1)
Mn(1)–N(11)
Mn(2)–N(15)
Mn(2)–O(3)
Mn(2)–N(6)
Mn(2)–O(2)
2.135(5)
2.145(5)
2.162(4)
2.200(4)
2.280(5)
2.309(5)
2.145(4)
2.155(4)
2.183(4)
2.197(4)
Mn(2)–O(1)
Mn(2)–N(12)
Mn(3)–N(17)
Mn(3)–O(3)
Mn(1)–Mn(2)
Mn(1)–Mn(29)
Mn(2)–Mn(3)
Mn(1)–O(1)–Mn2
Mn(1)–O(2)–Mn2
Mn(2)–O(3)–Mn3
2.222(4)
2.373(5)
2.182(5)
2.199(3)
3.968(3)
3.912(3)
3.924(3)
127.64(17)
127.61(17)
128.65(15)
2
652 | J. Mater. Chem., 2006, 16, 2645–2659
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Fig. 12 Core structural representation of the complex cation in 8
E
POVRAY image). Arrows highlight the Jahn–Teller axes.
(
Fig. 11 Structural representation of the complex cation in
E
POVRAY image).
8
(
N(15), and O(3) define this as the d
x22y2
orbital plane, with
long axial contacts to O(1) and O(2). Cu(1) has short contacts
to N(1), N(4), N(10), and O(1), which again define the copper
above and below the metal pseudo-plane. All metal ions
have distorted octahedral geometries. Separations between
orbital. All axial contacts exceed 2.24 A˚ . Cu(3) has a
dx22y2
˚
the aromatic rings on the ligands are around 4 A, in
different axially compressed geometry (d
long contacts to O(3) atoms, and short bonds to N(17). This
has been observed before with other Cu grids in this class,
z
ground state), with
2
keeping with the Mn–Mn separations. Mn–N and Mn–O
distances are normal for grids in this class and the Mn–Mn
9
˚
distances (ave. 3.921 A) and Mn–O–Mn angles (ave. 127.8u)
are similar to those reported for the parent complex
and may result from the way in which the arrangement of
the six ligands organizes the central CuN O coordination
2
4
6,45
1
[
Mn (2poap–2H) ](ClO ) involving the equivalent ligand with
9 6 4 6
pocket, with four deprotonated oxygen sites.
This arrange-
1
5
terminal pyridine groups. The almost flat coordination mode
of the ligands brings the non-coordinating pyrimidine nitrogen
ment sets up an orbitally orthogonal magnetic connection
in the outer ring of eight copper ions (vide infra), which
is shown in Fig. 12 (arrows) in terms of the way in which
the elongated Jahn–Teller copper axes are arranged (90u
mutual angles). This is shown clearly in terms of the Cu–Cu
2
atom and the NH group at the end of each ligand into a
suitable binding mode for secondary metal ion coordination,
with N–C–C–N torsional angles of ,7u.
˚
distances, where Cu(1)–Cu(2) is 4.075(2) A, and Cu(1)–Cu(29)
[Cu
9
(2pmoap–2H)
2
(2pmoap–H)
4
](ClO
4
)
10?13.2H
2
O (8). The
˚
is 4.266(2) A.
grid cation structure of 8 is shown in Fig. 11, and important
bond distances and angles are given in Table 9. The grid cation
has four-fold symmetry, which is unusual among similar
The presence of ten perchlorate anions is unusual, but
matches the observation made previously with the complex
[
9 6 3 2
Cu (2poap–H) ](NO )12?9H O, which involves mono-depro-
1
6,45
copper grids already reported.
are typical of M [3 6 3] grids in general with Cu–Cu distances
in the range 4.05–4.27 A, and Cu–O–Cu angles in the range
36.5–142.5u. The core structure is shown in Fig. 12. Each
The overall grid dimensions
16
tonated ligands. As a result the cation charge is +10, which
requires that four ligands have a 21 charge and two a
9
˚
2
2 charge. The structure reveals four ligand protons (H(21)) in
1
the difference maps, which are located on non-coordinated
diazine nitrogen atoms (N(11)). The adjacent CO group
copper ion in the outer eight-membered ring has a distorted,
axially elongated octahedral geometry, typical of six-coordi-
nate copper(II). Four Cu(2) short contacts to N(6), N(12),
(
(
(
C(12)–O(2)) bridges Cu(1) and Cu(2) with long, axial contacts
˚
˚
2.299 A, 2.294 A) respectively, and has a short C–O distance
˚
1.240 A) consistent with a CLO group, and no formal charge
on the oxygen atom. In contrast the C(23)–O(3) CO group is
˚
much longer (1.307 A), signifying a single CO bond and a
˚
Table 9 Bond distances (A) and angles (u) for 8
Cu(1)–N(4)
Cu(1)–N(10)
Cu(1)–O(1)
Cu(1)–N(1)
Cu(1)–N(7)
Cu(1)–O(2)
Cu(2)–N(15)
Cu(2)–N(6)
Cu(2)–O(3)
Cu(2)–N(12)
1.920(4)
2.020(4)
2.070(4)
2.106(5)
2.248(5)
2.298(4)
1.908(4)
1.993(4)
2.077(3)
2.109(4)
Cu(2)–O(1)
Cu(2)–O(2)
Cu(3)–N(17)
Cu(3)–O(3)
Cu(1)–Cu(2)
Cu(2)–Cu(3)
Cu(1)–Cu(29)
Cu(1)–O(1)–Cu(2)
Cu(2)–O(2)–Cu(1)
Cu(2)–O(3)–Cu(3)
2.251(4)
2.294(4)
1.936(6)
2.203(3)
4.075(2)
4.053(2)
4.266(2)
141.13(18)
136.50(18)
142.53(17)
formal charge of 21 on the oxygen bridging atom. This
˚
atom connects Cu(2) and Cu(3) with distances of 2.077 A and
˚
.203 A respectively. This unusual observation is consistent
2
with long and short axial and equatorial contacts typical of six-
coordinate copper(II), where a stronger bridging connection
would clearly exist with an equatorial contact, and highlights
the subtle effects of charge balance and pH brought to bear in
reactions where ligands with ionizable protons compete with
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2653

View Article Online
4
Murray et al. deals more generally with the magnetic
8
properties of tetranuclear complexes.
0
2
Nb g2 SS (S z1)(2S z1)e
0
0
0
{E(S )=kT
x ~
M
(1{r)
0
0
{E(S )=kT
3
k(T{h)
S(2S z1)e
(
3)
2
Nb g S(Sz1)r
2
z
zTIP
3kT
The magnetic exchange model for the trigonal bipyramidal
structures observed in complexes 5 and 6 (Fig. 13) involves six
exchange pathways, which for simplicity may be regarded as
equal, based on the structures. The vector coupling scheme
for this model may be summarized using the exchange
Hamiltonian (eqn (4)), and the subsequent equations,
Hex = 2J{S ?S + S ?S + S ?S + S ?S + S ?S + S ?S ) (4)
1
3
4
3
5
3
1
2
5
2
4
2
=
2J{S
3
?(S
+ S
2J(S ?S145)
1
+ S
4
+ S
5
) + S
2
?(S
1
+ S
4
+ S
5
)}
Fig. 13 Magnetic exchange models for the polynuclear systems.
= 2J{(S
2
3
)?(S
1
+ S
4
+ S
5
)}
=
=
2
3
the polarization effects of metal cations, and their coordina-
tion requirements in solvents containing water.
2
2
2
)
2J/2(S 2 S23 2 S145
Magnetic properties
where S23 = S + S ; S = S + S ; S = S15 + S ; S = S + S145
(S = S (n = 1–5) = 5/2). The eigenvalues are E(S, S , S145) =
n 23
2
3
15
1
5
145
4
23
In a general case a square (D4h) or distorted square (D2d
)
[
2 6 2] grid can be described by an exchange coupling scheme
involving three exchange integrals (J1, J2, J3) according to the
J/2[S(S + 1) 2 S (S + 1) 2 S145(S145 + 1)]. The allowed values of
23
23
S, S23 and S145, and their appropriate energies, can then be
substituted into the van Vleck equation (eqn (3)) in the usual way
to create a susceptibility profile as a function of temperature for a
particular J value.
appropriate Hamiltonian expression (eqn (1)) (Fig. 13).
H
ex = 22J1{S
2
?S
2
3
+ S
J3{S
1
?S
?S
4
} 2 2J2{S
+ S ?S
1
?S
2
3 4
+ S ?S } 2
(
1)
1
3
2
4
}
The procedure of generating the exchange equation for such
large systems, particularly in the case of Mn(II) (S = 5/2),
presents a somewhat daunting task, and so the use of
generalized software packages which deal with the calculation
of the spin state energy spectrum for a cluster, and substitution
In the case of compounds 1–4 J3 is assumed to be zero,
because there is no cross-coupling connection, and it is
reasonable to assume that J = J1 = J2 because of similar
M–O–M angles within each square grid. Using the conven-
into the van Vleck equation in one operation, have some
49
4
6,47
tional spin-vector coupling model,
eigenvalues of eqn (1)
distinct advantages. MAGMUN4.1,
an interactive
E
can be obtained analytically for a D4h [2 6 2] grid with
J = J1 = J2 and are given by eqn (2),
Windows based program, allows direct fitting of molar
susceptibility data based on any exchange model (matrix
size limits apply) and vector coupling scheme like those
described above.
E(S9, S , S ) = 2J[S9(S9 + 1) 2 S (S + 1) 2
1
3
24
13 13
(
2)
S (S + 1)]
2
4
24
In the case of the [3 6 3] Mn(II) grids exchange between
9
adjacent Mn(II) centres in the outer ring of eight metals can be
assumed to be the same (J1), on the basis of the similarity in
Mn–O–Mn bridge angles, and coupling between the central
metal ion and its immediate neighbours would perhaps be
slightly different (J2). Therefore the full exchange Hamiltonian
where S13 = S + S ; S = S + S ; S9 = S + S + S + S . This
result is valid for arbitrary values of the spin quantum
1
3
24
2
4
1
2
3
4
numbers S (i = 1,…4).
i
Using the addition rules for spin vectors the allowed values
for S9, and S13, S24 can be obtained. By substituting the
appropriate energy terms into the modified van Vleck equation
(
(
eqn (5)) can be reduced to just two exchange integrals
Fig. 13). The difficulties in dealing with such large systems
(
eqn (3); E(S9) refers to the energy of each spin state defined by
using a vector coupling approach have been discussed in some
2
5
S9), the susceptibility values can be computed as a function of
temperature. Simple programming techniques using iterative
procedures spanning the required S9 values for fitted
values of J allow regression of the experimental data against
eqn (3), with applied corrections for paramagnetic impurity
detail elsewhere, and in this case (45 electrons) the immensity
of the matrix calculations required (largest matrix dimension
88 900) to generate the total spin state profile for a system
with two J values predicates a simpler approach. D spatial
4
symmetry can be applied to such a grid, which reduces the
largest matrix dimension to 22 210, but this is still a
challenging calculation with a PC, and even with larger
(
fraction r), TIP (temperature independent paramagnetism),
and a Weiss-like corrective term (h) to deal with small
5
0
intermolecular exchange effects. A comprehensive review by
computers. Another approach, described in previous studies,
2
654 | J. Mater. Chem., 2006, 16, 2645–2659
This journal is ß The Royal Society of Chemistry 2006

View Article Online
involves effectively isolating the outer ring from the central
2
3,25
metal ion, and treating it as a spin coupled chain,
using the
5
1
Fisher model (eqn (6), (7)), and assuming that coupling to
the central Mn(II) ions (J2) can be ignored. This was justified
on the basis that the grid has a ground state with S = 5/2 spin
in an antiferromagnetic low temperature limit (odd number of
spin centres), and that divergence from the fully isotropic
exchange model would be relatively small for small values of J.
For the [Cu(II) ] grid a full analysis of the exchange within an
9
isotropic model can theoretically be achieved using routines
within MAGMUN4.1, since the matrix dimensions are
considerably smaller.
H
ex = 2J1{S
1
?S
2
+ S
8
2
?S
3
+ S
8
3
?S
4
4 5
+ S ?S +
S ?S + S ?S + S ?S + S ?S } 2 J2{ S ?S +
5
(5)
6
6
7
7
1
2
9
S ?S + S ?S + S ?S }
4
9
6
9
8
9
Fig. 15 Plot of xmol as a function of temperature for 2 (the solid
2
line is calculated with g = 1.994(1), J = 23.02(1) cm , r = 0.034,
1
2
2
Ng b S(Sz1) (1zu)
3
TIP = 0 cm mol (10 R = 0.65).
21
2
xMn
~
(6)
(7)
3kT
(1{u)
ꢀ
ꢁ
ꢀ
ꢁ
shown in Fig. 15. The expected maximum is overshadowed by
a rising low temperature feature associated with paramagnetic
impurity (vide infra), and so appears as a shoulder. The data
were fitted to eqn (3) as before to give g = 1.994(1), J =
JS(Sz1)
kT
u~ coth
{
kT
JS(Sz1)
2
x
mol = [(8xMn + 1.094g )/(T 2 h)](1 2 r) +
(
8)
21
3
21
2
2
23.02(1) cm , r = 0.034, TIP = 0 cm mol , 10 R = 0.65.
This exchange integral compares closely with that for 1, as
would be expected. The solid line in Fig. 15 was calculated
with these parameters. A significant, but small, paramagnetic
impurity correction was required to get a good data fit for
the magnetic sample, and may reasonably be associated with
a mononuclear species, or perhaps a small amount of a
pentanuclear complex.
(
1.094g /T)r + TIP
Variable temperature magnetic data for 1 are shown in
Fig. 14 as a plot of molar susceptibility as a function of
temperature. The maximum in xmol at 22 K indicates the
presence of intramolecular antiferromagnetic exchange, clearly
the result of oxygen bridging between the Mn(II) centres.
The data were fitted to eqn (3) using MAGMUN4.1, assuming
2
1
J = J1 = J2, and J3 = 0, to give g = 2.02, J = 23.1 cm , r =
0
The plot of magnetic moment per mole versus temperature
for 3 is illustrated in Fig. 16, and shows a marked increase
3
.002, TIP = 0 cm mol , 10 R = 1.9 (R = [S(x
21
2
2
2 xcalc.) /
obs.
2
1/2
Sxobs ). The solid line in Fig. 14 was calculated with
these parameters. The exchange coupling is consistent with
]
B B
from 3.9 m at 300 K to 5.1 m at 2.0 K, typical of a system
dominated by intramolecular ferromagnetic coupling. The
experimental data were fitted to eqn (3) using the exchange
Hamiltonian given in eqn (1), assuming J = J1 = J2 (J3 = 0).
Non-linear regression of the experimental data using
other related Mn(II) [2 6 2] grids, with large Mn–O–Mn
bridge angles.
9
Complex 2 has a similar square structure with similar
dimensions. The plot of xmol as a function of temperature is
Fig. 14 Plot of xmol as a function of temperature for 1 (the solid line is
Fig. 16 Plot of mmol as a function of temperature for 3 (the solid line
21
2
1
3
calculated with g = 2.02, J = 23.1 cm , r = 0.002, TIP = 0 cm mol
21
is calculated with g = 2.134(5), J = 7.45(5) cm , h = 0 K, TIP = 240 6
10 cm mol , r = 0 (10 R = 0.81).
2
10 R = 1.9).
26
3
21
2
(
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2655

View Article Online
1
1
MAGMUN4.1 gave an excellent data fit with g = 2.134(5),
21
complex [Mn
nuclear structure.
Complex 6 shows an almost identical plot of moment (mmol
as a function of temperature (ESI{ Fig. S3), and a fit using the
5
(poap–H)
6
](ClO
4
)
4
,
which has a similar penta-
2
1
26
3
J = 7.45(5) cm , h = 0 K, TIP = 240 6 10 cm mol , r = 0
2
10 R = 0.81). The solid line in Fig. 16 was calculated
(
)
with these parameters. The structure of 3 shows that all the
2
1
connections between the copper centers are strictly orthogonal
same model gave g = 2.015, J = 22.65 cm , r = 0.0001, TIP =
3
21
2
(vide supra), which would indicate that antiferromagnetic
0 cm mol , 10 R = 0.44. The solid line in Fig. S3 was
calculated with these parameters. The slight difference in J can
be attributed to the different ligands, and the differences in
Mn–O–Mn angles.
coupling should not occur, in agreement with the observed
exchange situation.
Complex 4 shows a very similar plot of magnetic moment
per mole versus temperature (ESI{ Fig. S1) to 3, dropping very
The variable temperature magnetic data for 7 are shown in
Fig. 18 as plot of mmol as a function of temperature. The room
B B
slightly from 3.77 m at 300 K to 3.73 m at 140 K, and then
rising to 4.90 mB at 2.0 K, typical of a system dominated by
intramolecular ferromagnetic coupling. Fitting of the data to
eqn (3) using the exchange Hamiltonian given in eqn (1),
and assuming J = J1 = J2 (J3 = 0) gave an excellent data fit
temperature value of 17.0 mB is typical of the Mn(II) grids,
9
and the value at 2 K (6.7 m ) indicates that the ground state is
B
S = 5/2, in keeping with related grids in this class, and results
from intragrid antiferromagnetic exchange. The data were
21 21
fitted to eqn (8) to give g = 2.04, J1 = 24.8 cm , J2 = 0 cm ,
2
1
with g = 2.102(3), J = 7.2(5) cm , h = 20.2 K, TIP = 250 6
2
0
6
3
cm mol , r = 0 (10 R = 0.44). The solid line in Fig. S1
21
2
3
21
2
1
r = 0.003, TIP = 0 cm mol , 10 R = 2.1. The solid line in
Fig. 18 was calculated with these parameters. Despite the
slightly larger than expected g value the exchange coupling is
was calculated with these parameters. Magnetization versus
field data (ESI{ Fig. S2) show a rise in M to a saturation value
2
3,25
of 4.17 m
B
at 5.0 T. Fitting of the data to the appropriate
consistent with other Mn(II) grids,
9
and clearly indicates
Brillouin function gave g = 2.08 for T = 2 K, and S9 = 2,
clearly corroborating the presence of intramolecular ferro-
magnetic coupling.
the isotropic nature of the exchange process, with the obvious
primary dependence of the exchange on the Mn–O–Mn angles.
However it should be emphasized that this is not a rigorous
model, since J2 in unlikely to be zero, but is considered to give
a reasonable estimate of J1. In this context it is of interest to
The variable temperature magnetic data for 5 are shown in
Fig. 17 as a plot of moment per mole as a function of
temperature. The moment drops from 12.7 mB at 300 K to
note that for the mixed oxidation state grid [Mn(II)
(2poap) ](ClO 10, which involves Mn(III) centres at the
corner positions only, J1 = 28.3 cm , and J2 = 24.5 cm
5 4
Mn(III) -
5
.82 m
the expected spin only moment for a pentanuclear cluster
13.2 m ), and the drop in moment as temperature is lowered
B
at 2 K. The room temperature value is slightly less than
6
4
)
2
1
21
5
0
(
(Fig. 13). In this case a single fully isotropic calculation
(S9/E; based on one set of J1 and J2 parameters) can be
completed in about 2 days on a fast PC with 2 GB of RAM.
The J2 (Mn(II)–Mn(II)) value is comparable with coupling
constants obtained using the chain model (vide supra), and the
Mn(II)–Mn(III) exchange is somewhat stronger. This can
reasonably be associated with the minor changes in grid
dimensions, for example increased Mn(II)–O–Mn(III) bridge
angles, which occur on oxidation.
B
signifies the presence of intramolecular antiferromagnetic
exchange. The low temperature value indicates the presence
of one uncoupled Mn(II) centre in the ground state (S = 5/2),
and is associated with the odd number of spin centres in the
cluster. The magnetic data were fitted to eqn (3) using
Hamiltonian (4), within MAGMUN4.1 to give g = 2.012(2),
2
1
3
21
2
J = 23.01(2) cm , TIP = 0 cm mol , r = 0, 10 R = 0.44.
The solid line in Fig. 17 was calculated with these parameters.
The exchange integral is essentially the same as those observed
for 1 and 2, and compares closely with that reported for the
Variable temperature magnetic data for 8 are shown in
Fig. 19 as a plot of mmol as a function of temperature. The
Fig. 17 Plot of mmol as a function of temperature for 5 (the solid line
21
Fig. 18 Plot of mmol as a function of temperature for 7 (the solid line
2
1
3
is calculated with g = 2.012(2), J = 23.01(2) cm , TIP = 0 cm mol
21
is calculated with g = 2.04, J1 = 24.8 cm , J2 = 0 cm , r = 0.003,
21
,
2
r = 0 (10 R = 0.44).
3
21
TIP = 0 cm mol (10 R = 2.1).
2
2
656 | J. Mater. Chem., 2006, 16, 2645–2659
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Table 10 Magnetic analysis for compounds 1–8
Ground
state spin
(S9)
2
J/cm
1
Compound
g
1
2
3
4
5
6
7
8
(Mn
(Mn
4
square grid)
square grid)
square grid)
square grid)
trigonal bipyramid) 2.012 23.01
trigonal bipyramid) 2.015 22.65
2.02 23.1
1.994 23.02
0
0
2
2
2.5
2.5
2.5
4
(Cu
4
2.134
2.102
7.45
7.2
(Cu
4
(Mn
(Mn
(Mn
5
5
9
square grid)
square grid)
2.04 24.8 (J1)
2.28
(Cu
9
0.35 (J1), 235 (J2) 3.5
˚
Cu(1)9–Cu(2) 4.266 A) could lead to different exchange
coupling (both ferromagnetic). However, given the difficulty
of fitting magnetic data in general for ferromagnetically
coupled systems, and the fact that the difference is likely to
Fig. 19 Plot of mmol as a function of temperature for 8 (the solid line
2
1
is calculated with g = 2.28, J2 = 235 cm , J1 = 0.35 cm and TIP =
21
be small, a single J value model (Fig. 13) for the outer Cu
8
2
6
3
21
ring has been chosen.
8
00 6 10 cm mol . Inset: Plot of M/H at 2 K for 8. The solid line
The central copper, Cu(3), is unusual in that it has an axially
compressed six-coordinate geometry, with long contacts to
is calculated for g = 2.2 and S9 = 7/2 (2 K) using the appropriate
Brillouin function. The dashed line is calculated for g = 2.2, S9 =
7
6 1/2 (2 K).
z
O(3) atoms. This defines its magnetic ground state as d .
2
Since this orbital nominally has some electron density in the xy
plane (the O(3) plane), then it is reasonable to assume that
antiferromagnetic exchange could occur between Cu(3) and
the Cu(2) centres. However it would not be expected to be
large because of the limited electron density in this spatial
region. Therefore, despite the large Cu(3)–O(3)–Cu(2) angles
moment drops from 6.0 mB at 300 K to a minimum value of
.4 m at 30 K, followed by a rise to 7.0 m at 2 K. The steady,
5
B
B
but small, drop in moment signifies an antiferromagnetic
exchange component, while the sharp rise at low temperature
is indicative of a ferromagnetic exchange component also.
Modeling the exchange in 8, and related grids, using the vector
coupling approach described above (based on eqn (5)), in
which two J values are evaluated independently using an
iterative regression process, is a challenge regardless of any
matrix dimensionality constraints. However total spin state/
energy (S9/E) profiles can be calculated easily for fixed values
(
142.3u), which would otherwise have led to quite strong
exchange between copper centres, a relatively small negative
exchange coupling would be expected.
Spin state/energy profiles were set up within MAGMUN4.1
for 2J2/J1 ratios in the range 10 to 200, and the magnetic data
simulated with input for g, the J2/J1 ratio and TIP. The solid
line in Fig. 19 was obtained after many trial comparisons with
9
of J1 and J2 for a Cu(II) grid with nine S = K spin centres.
2
6
3
21
g = 2.28, J2/J1 = 235, and TIP = 800 6 10 cm mol for
the model in which J2/J1 = 100. This leads to J2 = 235 cm
This can be done using MAGMUN4.1 and set up as a ratio
J1/J2, such that for an arbitrarily chosen ratio fitting of the
magnetic data for a grid with two J values can be achieved.
Such an approach is not strictly rigorous, but can be used to
usefully interpret the exchange picture, when an idea of what
the exchange should be is revealed from the structure.
2
1
21
and J1 = 0.35 cm . Given the structural features, and the
orbital connectivity within the grid, these values are eminently
reasonable estimates of the exchange coupling. On the
assumption that the exchange situation would be dominated
by J1 in the low temperature limit, a magnetization versus
field study was carried out at 2 K. Fig. 19 (inset) shows a
smooth rise in M with increasing field up to a plateau value
of 7.5 mB at 5 T (50 000 Oe). The solid line is calculated for
g = 2.2 and S9 = 7/2 (2 K) using the appropriate Brillouin
function. The dashed line is calculated for g = 2.2, S9 = 7 6 1/2
The grid in 8, like other copper grids in this group, has nine
highly distorted six-coordinate copper(II) ions. The four-fold
symmetry apparent in 8 is helpful, in that it simplifies the
interpretation of the magnetic connectivity within the grid.
The corner copper centres (Cu(1)) have axially elongated
geometries with the long axis pointing to O(2), defining their
(
2 K). It is therefore clear that the M/H profile is more typical
magnetic ground states as d
x22y2
. The side copper centres
of a grouped set of copper(II) centres, than isolated ones,
supporting not only the coupling scheme, which should result
in a 7/2 ground state, but also the ferromagnetic interaction in
(
Cu(2)) have similar geometries, with the long axis pointing to
both O(1) and O(2). The short bonds in the equatorial plane
include the connection to O(1) for Cu(1) and O(3) for Cu(2).
Such an arrangement therefore shows that the Cu(1)–O(2),
Cu(2)–O(1) and Cu(2)–O(2) connections are orthogonal, while
the Cu(2)–O(3) connection is not. This effectively isolates
Cu(2), and its symmetry related counterparts, from the Cu(1)
centres in terms of any antiferromagnetic exchange, but would
8
the outer Cu ring. A closer match in the M/H profiles might
possibly be expected at a lower temperature.
Results for the magnetic analysis for compounds 1–8 are
summarized in Table 10.
Conclusions
allow for possible ferromagnetic coupling in the outer Cu
8
ring. One could argue that the different connections between
˚
Cu(2) and Cu(1) through O(1) and O(2) (Cu(1)–Cu(2) 4.075 A,
A series of new polynuclear square [2 6 2] and [3 6 3] grids
and trigonal-bipyramidal clusters (Mn
5
) involving some new
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2657

View Article Online
2
0 L. K. Thompson, L. Zhao, Z. Xu, D. O. Miller and W. M. Reiff,
Inorg. Chem., 2003, 42, 128.
picolinic hydrazone ligands are reported, with copper and
manganese examples. Complexes 4, 6–8 have a large number
of extra nitrogen donor sites on the outer surfaces of the grids
and clusters, providing the potential for further coordination,
an area which is under current investigation. The current
structures reveal orthogonal magnetic connections in all
copper cases, leading to intramolecular exchange dominated
by ferromagnetic components, while for the manganese
systems antiferromagnetic exchange prevails. Magnetic data
have been rationalized successfully on the basis of structural
2
1 Z. Xu, L. K. Thompson and D. O. Miller, Polyhedron, 2002, 21,
1715.
22 L. K. Thompson, Coord. Chem. Rev., 2002, 233–234, 193.
2
3 L. Zhao, Z. Xu, H. Grove, V. A. Milway, L. N. Dawe,
T. S. M. Abedin, L. K. Thompson, T. L. Kelly, R. G. Harvey,
D. O. Miller, L. Weeks, J. G. Shapter and K. J. Pope, Inorg.
Chem., 2004, 43, 3812.
4 L. K. Thompson, T. L. Kelly, L. N. Dawe, H. Grove,
M. T. Lemaire, J. A. K. Howard, E. C. Spencer, C. J. Matthews,
S. T. Onions, S. J. Coles, P. N. Horton, M. B. Hursthouse and
M. E. Light, Inorg. Chem., 2004, 43, 7605.
2
evidence. The Cu
alternation of the Jahn–Teller axes, leading to orthogonal
bridging connections, and in the case of the Cu grid an
4
and Cu
9
grid structures clearly show a 90u
25 L. K. Thompson, O. Waldmann and Z. Xu, Coord. Chem. Rev.,
005, 249, 2677.
2
2
6 V. A. Milway, S. M. T. Abedin, V. Niel, T. L. Kelly, L. N. Dawe,
S. K. Dey, D. W. Thompson, D. O. Miller, M. S. Alam, P. M u¨ ller
and L. K. Thompson, Dalton Trans., 2006, DOI: 10.1039/b515801j.
7 E. Pardo, K. Bernot, M. Julve, F. Lloret, J. Cano, R. Ruiz-Garcia,
F. S. Delgado, C. Ruiz-P e´ rez, X. Ottenwaelder and Y. Journaux,
Inorg. Chem., 2004, 43, 2768.
9
unusual bridging of carbonyl (CLO) groups between the
copper centres through long axial Cu–O connections.
2
Acknowledgements
28 E. Pardo, I. Morales-Osorio, M. Julve, F. Lloret, J. Cano, R. Ruiz-
Garcia, J. Pas a´ n, C. Ruiz-P e´ rez, X. Ottenwaelder and Y. Journaux,
Inorg. Chem., 2004, 43, 7594.
We thank the Natural Sciences and Engineering Research
Council of Canada (NSERC) (LKT) and EPSRC (JAKH) for
funding to support this research.
2
3
9 G. Black, E. Depp and B. B. Corson, J. Org. Chem., 1949, 14, 14.
0 H. B. Amin and R. Taylor, J. Chem. Soc., Perkin Trans. 2, 1979,
6
24.
3
3
1 (a) SHELX97: G. M. Sheldrick, University of G o¨ ttingen, 1997; (b)
SIR97: A. Altomare, M. Cascarano, C. Giacovazzo and
Z. Guagliardi, J. Appl. Crystallogr., 1993, 26, 343.
2 DIRDIF94: P. T. Beurskens, G. Admiraal, G. Beurskens,
W. P. Bosman, R. de Gelder, R. Israel and J. M. M. Smits,
The DIRDIF-94 program system, Technical Report of the
Crystallography Laboratory, University of Nijmegen, The
Netherlands, 1994.
References
1
J. Rojo, F. J. Romero-Salguero, J.-M. Lehn, G. Baum and
D. Fenske, Eur. J. Inorg. Chem., 1999, 1421.
2
E. Breuning, M. Ruben, J.-M. Lehn, F. Renz, Y. Garcia,
V. Ksenofontov, P. G u¨ tlich, E. Wegelius and K. Rissanen,
Angew. Chem., Int. Ed., 2000, 39, 2504.
3
3 D. T. Cromer and J. T. Waber, International Tables for X-ray
Crystallography, vol. IV, The Kynoch Press, Birmingham,
England, 1974, Table 2.2 A.
3
4
5
6
7
8
E. Breuning, U. Ziener, J.-M. Lehn, E. Wegelius and K. Rissanen,
Eur. J. Inorg. Chem., 2001, 1515.
P. N. W. Baxter, J.-M. Lehn, J. Fischer and M.-T. Youinou,
Angew. Chem., Int. Ed. Engl., 1994, 33, 2284.
A. M. Garcia, F. J. Romero-Salguero, D. M. Bassani, J.-M. Lehn,
G. Baum and D. Fenske, Chem.—Eur. J., 1999, 6, 1803.
S. T. Onions, A. M. Frankin, P. N. Horton, M. B. Hursthouse and
C. J. Matthews, Chem. Commun., 2003, 2864.
M. Barboiu, G. Vaughan, R. Graff and J.-M. Lehn, J. Am. Chem.
Soc., 2003, 125, 10257.
C. J. Matthews, K. Avery, Z. Xu, L. K. Thompson, L. Zhao,
D. O. Miller, K. Biradha, K. Poirier, M. J. Zaworotko, C. Wilson,
A. E. Goeta and J. A. K. Howard, Inorg. Chem., 1999, 38, 5266.
L. K. Thompson, C. J. Matthews, L. Zhao, Z. Xu, D. O. Miller,
C. Wilson, M. A. Leech, J. A. K. Howard, S. L. Heath,
A. G. Whittaker and R. E. P. Winpenny, J. Solid State Chem.,
3
3
4 J. A. Ibers and W. C. Hamilton, Acta Crystallogr., 1964, 17, 781.
5 D. C. Creagh and W. J. McAuley, International Tables for
Crystallography, vol. C, ed. A. J. C. Wilson, Kluwer Academic
Publishers, Boston, 1992, Table 4.2.6.8, pp. 219–222.
3
6 teXsan for Windows: Crystal Structure Analysis Package,
Molecular Structure Corporation, 1997.
7 SIR 92: A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi,
M. Burla, G. Polidori and M. Camalli, J. Appl. Crystallogr., 1994,
3
2
7, 435.
3
8 DIRDIF99: P. T. Beurskens, G. Admiraal, G. Beurskens,
W. P. Bosman, R. de Gelder, R. Israel and J. M. M. Smits, The
DIRDIF-99 program system, Technical Report of the
Crystallography, University of Nijmegen, The Netherlands, 1999.
9 CrystalStructure 3.7.0: Crystal Structure Analysis Package, Rigaku
and Rigaku/MSC, 9009, New Trails Dr., The Woodlands, TX
9
3
4
2001, 159, 308.
1
1
1
1
0 Z. Xu, L. K. Thompson, C. J. Matthews, D. O. Miller, A. E. Goeta
and J. A. K. Howard, Inorg. Chem., 2001, 40, 2446.
1 C. J. Matthews, L. K. Thompson, S. R. Parsons, Z. Xu, D. O. Miller
and S. L. Heath, Inorg. Chem., 2001, 40, 4448.
2 Z. Xu, L. K. Thompson and D. O. Miller, J. Chem. Soc., Dalton
Trans., 2002, 2462.
3 H. Grove, T. J. Kelly, L. K. Thompson, L. Zhao, Z. Xu,
T. S. M. Abedin, D. O. Miller, A. E. Goeta, C. Wilson and
J. A. K. Howard, Inorg. Chem., 2004, 43, 4278.
7
7381, USA, 2000–2005.
0 CRYSTALS Issue 10: D. J. Watkin, C. K. Prout, J. R. Carruthers
and P. W. Betteridge, Chemical Crystallography Laboratory,
Oxford, UK, 1996.
1 J. Cosier and A. M. Glazer, J. Appl. Crystallogr., 1986, 19, 105.
2 SMART, Data Collection Software, Ver. 4.050, Siemens Analytical
X-ray Instruments Inc., Madison, Wisconsin, USA.
4
4
4
4
3 SAINT, Data Reduction Software, Version 4.050, Siemens
Analytical X-ray Instruments Inc., Madison, Wisconsin, USA,
14 C. J. Matthews, Z. Xu, S. K. Mandal, L. K. Thompson,
K. Biradha, K. Poirier and M. J. Zaworotko, Chem. Commun.,
1
996.
4 G. M. Sheldrick, SHELXTL 5.04/VMS, An integrated system for
solving, refining and displaying crystal structures from diffraction
data, Siemens Analytical X-ray Instruments Inc., Madison,
Wisconsin, USA, 1995.
1999, 347.
1
1
1
1
1
5 L. Zhao, C. J. Matthews, L. K. Thompson and S. L. Heath, Chem.
Commun., 2000, 265.
6 L. Zhao, Z. Xu, L. K. Thompson, S. L. Heath, D. O. Miller and
M. Ohba, Angew. Chem., Int. Ed., 2000, 39, 3114.
7 O. Waldmann, R. Koch, S. Schromm, P. M u¨ ller, L. Zhao and
L. K. Thompson, Chem. Phys. Lett., 2000, 332, 73.
8 O. Waldmann, L. Zhao and L. K. Thompson, Phys. Rev. Lett.,
45 V. A. Milway, V. Niel, T. S. M. Abedin, Z. Xu, L. K. Thompson,
H. Grove, D. O. Miller and S. R. Parsons, Inorg. Chem., 2004, 43,
1874.
46 K. Kambe, J. Phys. Soc. Jpn., 1950, 5, 48.
47 R. L. Martin, in New Pathways in Inorganic Chemistry, ed.
E. A. V. Ebsworth, A. G. Maddock and A. G. Sharpe, Cambridge
University Press, Cambridge, 1978, ch. 9.
2
002, 88, 066401.
9 L. Zhao, Z. Xu, L. K. Thompson and D. O. Miller, Polyhedron,
001, 20, 1359.
2
48 K. S. Murray, Adv. Inorg. Chem., 1995, 43, 261–356.
2
658 | J. Mater. Chem., 2006, 16, 2645–2659
This journal is ß The Royal Society of Chemistry 2006

View Article Online
4
9 MAGMUN4.1/OW01.exe is available as a combined package
free of charge from the authors (http://www.ucs.mun.ca/ylthomp/
magmun). MAGMUN has been developed by Dr Zhiqiang
Xu (Memorial University), and OW01.exe by Dr O. Waldmann.
We do not distribute the source codes. The programs may be
used only for scientific purposes, and economic utilization is
not allowed. If either routine is used to obtain scientific
results, which are published, the origin of the programs should
be quoted.
50 O. Waldmann, H. U. G u¨ del, T. L. Kelly and L. K. Thompson,
Inorg. Chem., in press.
51 M. E. Fisher, Am. J. Phys., 1964, 32, 343.
This journal is ß The Royal Society of Chemistry 2006
J. Mater. Chem., 2006, 16, 2645–2659 | 2659