Complexes derived from hydrolytically 'unstable' hydrazone ligands - Some unexpected products

Article abstract of DOI:10.1016/j.poly.2005.02.016

Polyfunctional ligands derived from condensation reactions in which small molecules, e.g., water, an alcohol, etc., are eliminated can suffer the ravages of hydrolytic attack on reaction with transition metal ions in aqueous based solvents, or solvents containing small amounts of water. A series of such reactions is described involving hydrazone based ligands, in which hydrolyzed ligand fragments, and their resulting coordination complexes, are produced. Structures are reported to identify these products, which reveal some unexpected ligand rearrangements. Variable temperature magnetic data are discussed for some dinuclear complexes, and a DFT study is presented to rationalize an unusual ligand transformation.

Full text of DOI:10.1016/j.poly.2005.02.016

Polyhedron 24 (2005) 807–821
www.elsevier.com/locate/poly
Complexes derived from hydrolytically ꢀunstableꢁ hydrazone
ligands – Some unexpected products
a
Timothy L. Kelly , Victoria A. Milway , Hilde Grove , Virginie Niel ,
a
a
a
a
a,
a
a
Tareque S.M. Abedin , Laurence K. Thompson *, Liang Zhao , Rosemary G. Harvey ,
a
David O. Miller , Michael Leech , Andres E. Goeta , Judith A.K. Howard
b
b
b
´
a
Department of Chemistry, Memorial University, Elizabeth Avenue, St. Johnꢁs, Nfld, Canada A1B 3X7
Department of Chemistry, University of Durham, Durham DH1 3LE, UK
b
Received 15 October 2004; accepted 11 February 2005
Available online 13 April 2005
Abstract
Polyfunctional ligands derived from condensation reactions in which small molecules, e.g., water, an alcohol, etc., are eliminated
can suffer the ravages of hydrolytic attack on reaction with transition metal ions in aqueous based solvents, or solvents containing
small amounts of water. A series of such reactions is described involving hydrazone based ligands, in which hydrolyzed ligand frag-
ments, and their resulting coordination complexes, are produced. Structures are reported to identify these products, which reveal
some unexpected ligand rearrangements. Variable temperature magnetic data are discussed for some dinuclear complexes, and a
DFT study is presented to rationalize an unusual ligand transformation.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Keywords: Hydrazone ligands; Complexes; Copper; Nickel; Iron; Cobalt; Hydrolysis; Structure; Magnetism
1. Introduction
In a recent report the ligand L1 was shown to be
hydrolytically unstable in the presence of CuBr2 in
aqueous acetonitrile, and a mixed oxidation state
Cu(II)4Cu(I)2 complex of the ligand picolinamide hydra-
zone (PAH) was produced (Scheme 1) [1]. The reduction
of copper(II) was associated with the presence of acetoni-
trile and the reducing capacity of salicylaldehyde, de-
tected during the reaction. The symmetrical open chain
diazine ligands PAHAP and PZHPZ (Chart 1) react with
Fe(III) nitrate in aqueous solution to spontaneously form
the reduced dinuclear [Fe(II)2(L)3]4+ (L = PAHAP,
PZHPZ) complex ions in low yield, suggesting the forma-
tion of hydrazine based fragments, presumably resulting
from hydrolysis of an imine group, which act as metal ion
reductants. In the case of PAHAP, reaction with Fe(III)
nitrate also produces the complex ion [Fe(III)(POAP-
H)(H2O)2 (NO3)]2+ (POAP, Chart 1) as a minor product,
indicating the imine group as an initial point of hydrolytic
Ligands derived from condensation reactions, which
involve elimination of small molecules, e.g., water, or
an alcohol, are at risk of hydrolytic attack in solvents
that contain water, when reacted with transition metal
salts. Such reactions can be promoted or catalyzed by
the presence of the transition metal ion, whose high io-
nic potential can in many cases exert a significant polar-
izing effect on the hydrolytically vulnerable bonds. In
some cases the complexes which are formed contain
ꢀexpectedꢁ hydrolysis products, but in other cases some
surprising results can occur.
*
Corresponding author. Tel.: +1 709 737 8750; fax: +1 709 737 3702.
E-mail address: lthomp@mun.ca (L.K. Thompson).
0277-5387/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2005.02.016

808
T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
H
N
NH
Cu(II)
CHO
+
H2O
N
N
NHNH2
N
N
H
OH
HO
L1
PAH
Scheme 1.
The present report discusses transition metal com-
X
NH2
plexes derived from a group of extended polyfunctional
diazine ligands L2–L6 (Chart 1), in which metal ion
based ligand hydrolysis has occurred at a ꢀvulnerableꢁ li-
gand C@N site, with the formation of complexes con-
taining rearranged ligands and ligand fragments. The
ligand L3 has been reported elsewhere, along with a
number of coordination complexes, but there is no evi-
dence for ligand hydrolysis [3]. The ligand L4 undergoes
hydrolysis in the presence of Co(II) in air, and with
Cu(II) salts, but remains intact on reaction with Ni(II)
carbonate. Structures are reported for some iron, nickel
and copper complexes, and the magnetic properties of
some polynuclear copper and iron products are dis-
cussed. DFT studies are highlighted in one nickel case
where an unusual ligand transformation takes place.
N
N
N
N
R
X
PAHAP (X=CH, R=NH2), PZHPZ (X=N, R=NH2),
POAP (X=CH, R=OH)
H3C
CH3
L2
N
N
N
N
N
N
HO
H3C
OH
N
OH
L3
N
N
CH3
2. Experimental
N
OH
R
2.1. Physical measurements
R'
R'
Infrared spectra were recorded as Nujol mulls using a
Mattson Polaris FTIR instrument. Mass spectra were
obtained using a VG Micromass 7070HS spectrometer
and an Agilent 1100 series LC/MSD spectrometer. C,
H, N analyses on dried samples were carried out by
the Canadian Microanalytical Service, Delta, B.C., Can-
ada. Variable temperature magnetic data (2–300 K)
were obtained with a Quantum Design MPMS5S Squid
magnetometer operating at 0.1–5.0 T. Calibrations were
carried out with a palladium standard cylinder, and the
normal diamagnetic corrections were applied.
N
N
N
N
X
X
N
N
N
OH
OH
L4 (R=Cl, R'=Phenyl, X=CH)
L5 (R=Cl, R'=NH2, X=N)
N
2.2. Syntheses
N
NNH
2
N
N
2.2.1. L2
N
OH
OH
2-Acetyl pyridine (3.70 g, 0.0300 mol) was dissolved
in methanol (100 mL). Oxalic acid dihydrazide (1.77 g,
0.0150 mol) was added and the mixture refluxed for
12 h. A white precipitate formed, which was separated
by filtration, washed with methanol and ether and dried
under vacuum (Yield 4.4 g, 91%); m.p. dec. >260 ꢁC.
Mass spectrum (m/z) 189, 162, 134, 106, 78 (molecular
ion not observed). IR (m cmꢀ1) 1681 (C@O), 1012
(pyr). Anal. Calc. for C16H16N6O2: C, 59.25; H, 4.97;
N, 25.91. Found: C, 58.67; H, 4.94; N, 25.72%.
L6
Chart 1.
attack in the formation of the analogous acyl ligand, pre-
sumably promoted by the coordination of the highly
polarizing Fe(III) center [2].

T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
809
2.2.2. L3
Butanedione monoxime (5.04 g, 0.0498 mol) was dis-
[(C7H9N3)2Cu2(C2O4)(H2O)2(NO3)2]; C, 29.78; H, 3.44;
N, 17.36. Found: C, 29.79; H, 3.33; N, 17.67%. 2 was
produced under similar conditions using copper(II) per-
chlorate (Yield 50%). Anal. Calc. for [(C7H9N3)2Cu2-
(C2O4)](ClO4)2: C, 28.08; H, 2.65; N, 12.28. Found: C,
28.37; H, 2.63; N, 12.24%.
solved in absolute ethanol (40 mL). Picolinic acid hydra-
zide (3.32 g, 0.0242 mol) was added and the mixture
refluxed for several hours with the formation of a white
suspended solid. The fine white precipitate was filtered
off and washed with methanol (Yield 4.86 g, 91%).
Recrystallization from acetonitrile produced colorless
crystals, but crystals suitable for structural determina-
tion were obtained from the reaction of Gd(NO3)3 and
L3 in a MeCN, MeOH, H2O mixture, in which there ap-
peared to be no Gd complex formation. Mass spectrum
2.2.6. [Ni(L8)]BrÆCHCl3 (3)
NiBr2 (0.24 g, 1.0 mmol) was suspended in 50/50
CH3CN/MeOH (20 mL). L3 (0.20 g, 0.90 mmol) was
added and the mixture was heated with stirring. A clear
orange solution formed, which was filtered and allowed
to stand, forming a red orange powder (Yield 55%).
Recrystallization from CHCl3 gave red block shaped
crystals suitable for structural determination. A similar
reaction with Ni(ClO4)2 Æ6H2O and Ni(NO3)2 Æ6H2O
gave [Ni(L8)](ClO4)ÆH2O and [Ni(L8)](NO3)Æ2H2O,
respectively, as red powders (Yield 60–70%). Anal. Calc.
for [(C16H15N6O2)Ni](ClO4)ÆH2O: C, 38.55; H, 3.41; N,
16.86. Found: C, 38.27; H, 3.30; N, 16.73%. Anal. Calc.
for [(C16H15N6O2)Ni] (NO3)Æ2H2O: C, 40.08; H, 3.97;
N, 20.46. Found: C, 39.56; H, 3.29; N, 20.37%. Similar
infrared spectral bands associated with the ligand L8 are
observed in all three compounds.
(m/z) 221 (MH+), 203, 182, 124, 108, 79. IR (m cmꢀ1
)
3343 (OH), 1677 (C@O), 995 (pyr). Anal. Calc. for
C10H12N4O2: C, 54.55; H, 5.45; N, 25.45. Found: C,
54.55; H, 5.58; N, 25.41%.
2.2.3. L4
2-Benzoylpyridine (7.44 g, 41 mmol) was dissolved in
CHCl3 (220 mL). 4-Chloro-2,6-pyridine carboxylic acid
hydrazide (2.30 g, 10 mmol) was added, along with dry
methanol (30 mL), and the mixture was refluxed for
72 h. The volume of the resulting solution was reduced
by about 50%, and methanol (30 mL) was added and
heating continued until a white precipitate formed.
The precipitate was collected by suction filtration, and
washed with methanol and diethyl ether (Yield 3.6 g,
65%); m.p. 288–294 ꢁC. Mass spectrum (m/z) 560 (M),
542 (M ꢀ H2O). IR (m, cmꢀ1) 1697 (C@O), 1574
(C@N), 997 (py). Anal. Calc. for C31H22N7O2ClÆ0.5-
H2O: C, 65.49; H, 4.05; N, 17.25. Found: C, 65.25; H,
3.95; N, 17.33%.
2.2.7. [Cu2(L9)(L10)]Æ2H2OÆCH3OHÆL11 (4)
L4 (0.10 g, 0.18 mmol) was added to a solution of
copper acetate (0.18 g, 0.45 mmol) in methanol
(20 mL), forming a clear brown solution. Red brown
crystals of 4, suitable for a structural determination,
were obtained in low yield (35 mg) after allowing the
solution to evaporate to a small volume. Anal. Calc.
for
[(C19H11N4O3Cl)(C7H2NO4Cl)Cu2](C12H9N3)Æ
2.2.4. L6
7H2O: C, 44.53; H, 3.52; N, 10.94. Found: C, 44.47;
H, 2.67; N, 11.08%. The dried analytical sample con-
tains no methanol.
2,6-Pyridine carboxylic acid hydrazide (0.93 g,
5.0 mmol) was added to a solution of dipyridyl ketone
(1.84 g, 10 mmol) in absolute ethanol (50 mL), and the
mixture was refluxed for 10 h. The product formed as
a white solid during the reaction, and was filtered off,
washed with ethanol and dried in air (Yield 1.7 g,
90%). Anal. Calc. for C18H15N7O2(C29H21N9O2)0.22; C,
61.34; H, 4.14; N, 26.35. Found: C, 61.31; H, 4.09; N,
26.19%. The reaction appears to have produced a mix-
ture of mono and disubstituted ligands, with about
20% disubstituted. Separation could not be effected
cleanly, and so the mixture was used in the reaction with
Cu(NO3)2 (vide infra).
2.2.8. [Fe2(PZHPZ)3](ClO4)4 Æ2CH3CN (5)
L5 [4] (0.070 g, 0.16 mmol) was added to a degassed
solution of Fe(ClO4)2 Æ6H2O (0.122 g, 0.48 mmol) in
CH3CN (10 mL) under an Ar atmosphere. The mixture
was heated to reflux for 10 min., and then left at room
temperature. Purple crystals formed after several days
in the Ar atmosphere (Yield 50 mg), which were kept
under the mother liquor prior to structural analysis. Sol-
vent loss occurred on exposure of the bulk sample to air.
Anal. Calc. for [(C10H10N8)3Fe2](ClO4)4 Æ7H2O: C,
26.45; H, 3.26; N, 24.68. Found: C, 26.40; H, 2.69; N,
24.93% (air dried sample).
2.2.5. [Cu2(L7)2(C2O4)(H2O)2(NO3)2] (1),
[Cu2(L7)2(C2O4)](ClO4)2 (2)
L2 (0.324 g, 1.00 mmol) was added to a warm solu-
tion of Cu(NO3)2 Æ3H2O (0.483 g, 2.00 mmol) in
MeOH/H2O (10/5 mL), forming a green solution. Dark
green crystals, suitable for a structural analysis, were ob-
tained on standing the solution at room temperature
over several days (Yield 0.37 g, 45%). Anal. Calc. for
2.2.9. [Cu2(L12-2H)](NO3)2 Æ1.5H2O (6)
L6 (0.25 g, 0.50 mmol) was added to a solution of
Cu(NO3)2 Æ3H2O (0.72 g, 3.0 mmol) in MeOH/H2O
(20/10 mL) and the mixture was warmed. Green crystals
were obtained from the resulting green solution on
standing at room temperature (Yield 0.21 g, 68%). Anal.

810
T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
Calc. for (C18H11N5O3)Cu2(NO3)2 Æ1.5H2O: C, 34.68; H,
2.26; N, 15.73. Found: C, 34.50; H, 2.12; N, 16.10%.
The diffraction intensities of a colourless single crys-
tal of L3 were collected with graphite-monochromated
Cu Ka X-radiation using a Rigaku AFC6S diffractome-
ter at 299(1) K and the x–2h scan technique to a 2h va-
lue of 120.1ꢁ. The data were corrected for Lorentz and
polarization effects. The structure was solved by direct
methods [5,6]. All atoms except hydrogen atoms were re-
fined anisotropically. Hydrogen atoms were introduced
with isotropic thermal parameters set 20% greater than
those of their bonded partners at the time of their inclu-
sion. They were optimized by positional refinement.
Neutral atom scattering factors [7] and anomalous-
dispersion terms [8,9] were taken from the usual sources.
All calculations were performed with the teXsan [10]
crystallographic software package. Diffraction data
for 7 (red prism) were collected in a similar manner.
Abbreviated structural details for L3 and 7 are listed
in Table 1.
Diffraction data for a single crystal of 6 (green prism,
0.30 · 0.40 · 0.20 mm) were collected using a Bruker
SMART CCD diffractometer, equipped with an Oxford
Cryostream N2 cooling device [11], with graphite mono-
chromatized Mo Ka radiation. Cell parameters were
determined and refined using the SMART software
[12a], raw frame data were integrated using the SAINT
program [12b] and the structure was solved using direct
methods and refined by full-matrix least squares on F2
using SHELXTL [13]. Non-hydrogen atoms were refined
with anisotropic atomic displacement parameters
(adps). Hydrogen atoms bound to carbon atoms were
placed in geometrically calculated positions with isotro-
pic adps 1.2 times that of the parent atom.
2.2.10. [Ni(L4-2H)] (7)
NiCO3 (0.18 g, 1.5 mmol) was suspended in 1:1
MeCN/EtOH (20 mL). L4 (0.10 g, 018 mmol) was
added and the mixture was heated. Water (5 mL) was
added and heating continued for 15 min. with the for-
mation of a red solution. The solution was filtered to re-
move excess nickel carbonate, and allowed to stand at
room temperature. Red crystals suitable for structural
analysis formed after two days (Yield 60 mg, 55%).
2.3. X-ray crystallography
The diffraction intensities of a green prismatic crystal
of 1 were collected with graphite-monochromatized Mo
Ka X-radiation (rotating anode generator) using a Bru-
ker P4/CCD diffractometer at 193(1) K to a maximum 2h
value of 52.8ꢁ. The data were corrected for Lorentz and
polarization effects. The structure was solved by direct
methods [5,6]. All atoms except hydrogen atoms were re-
fined anisotropically. Hydrogen atoms were placed in
calculated positions with isotropic thermal parameters
set to 20% greater than their bonded partners, and were
not refined. Neutral atom scattering factors [7] and
anomalous-dispersion terms [8,9] were taken from the
usual sources. All other calculations were performed
with the teXsan [10] crystallographic software package.
Diffraction data were collected for 3–5 in a similar
manner. Abbreviated structural data for 1, 3–5 are
reported in Table 1.
Table 1
Summary of crystallographic data for 1, L3, 3–7
Compound
1
L3
3
4
5
6
7
Empirical formula
C
32H44N16
-
C10H12O2N4 C34H32Br2Cl6- C39H30N8O10Cl2Cu2 C34H36N26O16- C18H19N7O13Cu2 C31H20N7O2NiCl
Cu4O24
1286.96
0.71073
triclinic
N12Ni2O4
1162.66
0.71073
Cl4Fe2
Formula weight
˚
Wavelength (A)
220.23
1.54178
monoclinic
968.69
0.71073
monoclinic
C2/c
1318.39
0.71073
monoclinic
C2/c (#15)
17.4322(13)
15.6919(12)
20.7481(15)
90
668.48
616.7
0.71073
monoclinic
Cc
0.71703
monoclinic
Cc
Crystal system
Space group
triclinic
ꢀ
ꢀ
P1
10.1578(8)
P21/c (#14) P1 ð#2Þ
˚
a (A)
5.706(4)
25.812(4)
7.337(3)
90
7.3323(11)
11.5251(16)
13.7813(19)
71.791(3)
75.089(3)
79.892(3)
1063.3(2)
30821.816
193(2)
29.927(3)
8.8728(8)
30.906(3)
90
25.513(2)
7.1249(6)
16.8915(14)
90
14.405(3)
17.663(5)
10.689(4)
90
˚
b (A)
13.4753(11)
17.8890(14)
103.473(1)
90.963(1)
105.493(2)
2286.5(3)
1.869
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
100.19(4)
90
109.832(2)
90
99.335(2)
90
129.9970(10)
90
103.73(3)
90
3
˚
V (A )
1063.6(7)
1.375
7720.1(12)
1.667
5600.4(6)
1.563
193(2)
2352.2(3)
1.888
100(2)
2642(1)
1.550
Dcalc. (g cmꢀ3
Temperature (K)
)
193(2)
2
19.42
299(1)
4
8.31
193(2)
8
13.12
299(1)
4
8.81
Z
1
32.01
4
7.94
4
1.894
l (cmꢀ1
Reflections/parameters 8631/704
Final R1, wR2
)
977/182
0.0645, 0.200 0.056, 0.057 0.0562, 0.1385 0.0443, 0.1114
3698/280
7878/550
5733/404
0.064, 0.2049
5371/437
0.0171, 0.0465
2206/376
0.043, 0.040
(R,Rw)
(R,Rw)
P
P
P
P
R1 = [|Fo| ꢀ |Fc|]/ |Fo|, wR2 = [ [w(|Fo|2 ꢀ |Fc|2)2]/ [w(|Fo|2)2]]1/2
.
P
P
P
P
2 1/2
]
R = ||Fo| ꢀ |Fc||/ |Fo|, Rw = [ w(|Fo| ꢀ |Fc|)2/ wFo
.

T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
811
Table 2
˚
2.4. Computational details
Distances (A) and angles (ꢁ) for [Cu2(L7)2(C2O4)(H2O)2(NO3)2] (1)
Cu1–N1
Cu1–N2
Cu1–O2
Cu1–O1
Cu1–O21
Cu2–N5
Cu2–N4
Cu2–O4
Cu2–O3
Cu2–O20
Cu3–N7
Cu3–O5
Cu3–O6
Cu3–N8
Cu4–N10
Cu4–O8
Cu4–O7
Cu4–N11
Cu4–O9
1.965(5)
All calculations were carried out using GAUSSIAN-03
[14]. Ligand and complex structures were optimized
using Beckeꢁs hybrid B3LYP functional [15]. The split-
valence with polarization (SVP) basis set of Ahlrichs
and coworkers [16] was felt to provide a good compro-
mise between computational accuracy and efficiency,
and was used for all calculations, except where noted.
For all geometry optimizations, frequencies were also
calculated to ensure that the structures were genuine
minima on the potential energy hypersurface (no imag-
inary frequencies were found).
1.971(5)
1.974(4)
1.974(4)
2.346(5)
1.943(5)
1.975(5)
1.983(4)
1.983(4)
2.367(5)
1.968(5)
1.972(4)
1.973(4)
1.984(5)
1.970(5)
1.973(4)
1.985(5)
1.987(5)
2.273(4)
3. Results and discussion
3.1. Structures
N1–Cu1–N2
N1–Cu1–O2
N2–Cu1–O2
O6–Cu3–N8
N10–Cu4–O8
N10–Cu4–O7
O8–Cu4–O7
N10–Cu4–N11
O8–Cu4–N11
N1–Cu1–O1
N2–Cu1–O1
O2–Cu1–O1
N1–Cu1–O21
N2–Cu1–O21
O2–Cu1–O21
O1–Cu1–O21
N5–Cu2–N4
N5–Cu2–O4
N4–Cu2–O4
N5–Cu2–O3
N4–Cu2–O3
O4–Cu2–O3
N5–Cu2–O20
N4–Cu2–O20
O4–Cu2–O20
O3–Cu2–O20
N7–Cu3–O5
N7–Cu3–O6
O5–Cu3–O6
N7–Cu3–N8
O5–Cu3–N8
O7–Cu4–N11
N10–Cu4–O9
O8–Cu4–O9
O7–Cu4–O9
N11–Cu4–O9
82.0(2)
173.75(19)
95.80(19)
95.0(2)
3.1.1. [Cu2(L7)2(C2O4)(H2O)2(NO3)2] (1)
The molecular structure of 1 is comprised of three
similar dinuclear fragments, all involving the bidentate
ligand 2-acetylpyridine hydrazone (L7), and a bridging
oxalate. In combination they are represented by the
averaged formula above. Relevant bond distances and
angles are listed in Table 2. Two fragments are shown
in Fig. 1(a). The Cu(3) fragment has four distant nitrates
97.6(2)
168.6(2)
84.91(18)
81.2(2)
170.21(19)
96.87(19)
175.23(18)
84.84(16)
91.3(2)
bound to a ꢀflatꢁ dinuclear center, with long Cu–O
˚
2.524(4) A,
contacts
2.468(4) A). Cu–L in plane distances fall in the range
(Cu(3)–O(12)
Cu(3)–O(14)
˚
94.28(19)
94.72(17)
90.38(18)
81.7(2)
˚
˚
1.973–1.985 A, and the Cu–Cu distance is 5.149(3) A.
The bridging oxalate has a typical geometry. The
Cu(3) dinuclear subunit is hydrogen bonded to the
175.2(2)
99.62(19)
93.5(2)
˚
Cu(4) subunit (Fig. 1(a); O(15)–O(9) 2.795(5) A, O(9)–
˚
O(13) 2.786(6) A). Subunit Cu(4) has a similar geometry
with five-coordinate square pyramidal copper centers,
and with each metal having an axially bound water mol-
˚
ecule (Cu(4)–O(9) 2.273(4) A). The third dinuclear sub-
172.54(19)
84.75(16)
90.7(2)
95.0(2)
unit (Fig. 1(b)) has a similar structure with square
pyramidal copper centers, and axial contacts to oxygen
93.82(18)
90.71(18)
98.4(2)
atoms from a disordered combination of a water
˚
a nitrate (Cu(1)–O(21) 2.345(5) A,
176.7(2)
84.74(18)
81.8(2)
molecule and
˚
Cu(2)–O(20) 2.365(5) A). Other dimensions within this
dinuclear subunit are similar to those involving Cu(3)
and Cu(4).
179.3(2)
94.4(2)
99.26(18)
91.20(16)
91.82(16)
98.58(17)
3.1.2. L3
The structure of the ligand (L3) is shown in Fig. 2,
and relevant bond distances and angles are listed in
Table 3. The good refinement shows that a proton re-
sides on N(2) and that the C–O bond has predominantly
ligand as a whole, and in particular with respect to the
two methyl groups, as indicated by the torsional angles
˚
double bond character (C(6)–O(1) 1.230(5) A). The
˚
N(2)–N(3) distance (1.389(5) A) indicates this bond to
C(10)–C(8)–C(7)–C(9)
and
C(7)–N(3)–N(2)–C(6)
have mostly single bond character. One interesting
structural feature that bears on the reactivity of this
ligand surrounds the anti (trans) conformation of the
(177.4ꢁ, 178.1ꢁ, respectively), which are both very large.
DFT calculations on L3, which utilize the B3LYP func-
tional and 6-31G(d) basis set, indicate that the anti form

812
T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
Table 3
˚
(a)
N11
Distances (A) and angles (ꢁ) for L3
O7
Cu4
O8
O1–C6
O2–N4
N1–C1
N1–C5
N2–N3
N2–C6
N3–C7
N4–C8
C1–C2
C2–C3
C4–C5
C5–C6
C7–C8
C7–C9
C8–C10
1.230(5)
1.408(5)
1.336(6)
1.342(5)
1.389(5)
1.332(6)
1.297(5)
1.291(6)
1.380(7)
1.372(8)
1.372(6)
1.501(6)
1.459(6)
1.490(7)
1.507(7)
Cu4_2
N10
O13_2
O14
O9
O15
O6
N8
N7
Cu3
Cu3_2
O5
O12
O13
O21
(b)
O2
O4
C1–N1–C5
N3–N2–C6
N2–N3–C7
O2–N4–C8
N1–C1–C2
C1–C2–C3
C2–C3–C4
C3–C4–C5
N1–C5–C4
N1–C5–C6
C4–C5–C6
O1–C6–N2
O1–C6–C5
N2–C6–C5
N3–C7–C8
N3–C7–C9
C8–C7–C9
N4–C8–C7
N4–C8–C10
C7–C8–C10
116.7(5)
120.8(4)
115.4(4)
111.3(4)
123.4(5)
118.7(5)
118.7(5)
119.3(5)
123.2(5)
116.5(4)
120.3(4)
125.6(5)
120.5(5)
113.9(4)
114.9(4)
125.3(5)
119.9(5)
113.9(4)
124.7(4)
121.4(4)
N4
N5
N2
Cu2
Cu1
N1
O1
O3
O20
Fig. 1. (a) Structural representation of two dinuclear fragments in 1
(40% probability thermal ellipsoids). (b) Structural representation of a
third dinuclear fragment in 1 (40% probability thermal ellipsoids).
C9
C1
C2
N4
N1
O2
C7
C8
N2
C3
C5
C6
C4
N3
C10
O1
Fig. 2. Structural representation of L3 (40% probability thermal
ellipsoids).
a formal negative charge, consistent with the short Ni–N
distances. The second proton (H(1)) resides on pyridine
nitrogen N(1) (difference map), thus giving the complex
ion a formal charge of 1+. The bromide ion (Br(1))
balances the charge, and is hydrogen bonded to H(1)
and also to a chloroform molecule (Br(1)–H(16)
is more stable than the eclipsed form, with an energy dif-
ference of ꢁ25 kJ/mol.
˚
˚
3.1.3. [Ni(L8)]BrÆCHCl3 (3)
2.692(6) A, Br(1)–H(1) 2.467(6) A).
The molecular structure of 3 is shown in Fig. 3(a),
and relevant bond distances and angles are given in
Table 4. The obvious first point of discussion is the fact
that the starting ligand (L3) is not present, but a new
Schiff base ligand L8 is formed, based on a central bu-
tane-dione fragment. L8 acts in a tetradentate fashion
coordinating to the nickel ion via two diazine nitrogen
atoms, a pyridine nitrogen and an oxygen atom, creating
quite a strong in plane crystal field. This is strong en-
ough to cause the Ni(II) ion to be low spin (vide infra).
The NiN3O subunit is essentially planar (O(1)–Ni–N(5)
179.8(2)ꢁ, N(3)–Ni–N(6) 178.9(2)ꢁ), with very short
The flat nature of the Ni(II) ion, generated by the
strong in plane crystal field, predisposes the whole li-
gand to adopt a fairly flat overall conformation, with
the methyl groups in an unexpected cis-conformation
(see structure of L3). The molecules are arranged in a
stacked, alternating, trans-conformation in the lattice,
with pairs of molecules related by an inversion centre.
˚
There are many relatively short (<4 A) intermolecular
contacts between individual molecules, e.g., N(4) and
˚
O(1) (3.304 A) (highlighted in Fig. 3(b)), with the Ni–
Ni distances between molecules in the stack alternating
˚
as 3.663(3), and 4.743(4) A. Any axial contacts to the
nickel centres between molecules are considered too
weak to be of bonding significance. However, the inter-
molecular contacts are clearly instrumental in the paral-
lel arrangement of the molecules of 3. The complexes
˚
Ni-ligand contacts (1.828–1.887 A). Both C–O bonds
have predominantly double bond character (C(6)–O(1)
˚
˚
1.285(6 ) A, C(11)–O(2) 1.222(6) A), and since the ligand
has lost one proton, nitrogen atoms N(3) and N(5) bear

T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
813
excellent agreement. For the square-planar nickel coor-
dination sphere, the average deviation in bond lengths
˚
between the two structures was 0.02 A. The average
deviation in bond angles was 0.6ꢁ. This is comparable
to results reported by Jang et al. [17] for the structures
of iron carbonyls, where differences between B3LYP
(double-f plus polarization basis set) and X-ray crystal
˚
structures were on the order of 0.02 A and 0.5ꢁ for bond
lengths and angles, respectively. For a medium-sized
transition metal complex such as 3 (41 atoms, with 467
basis functions), such agreement is quite reasonable.
This indicates the theory and basis set used are likely
sufficient for the obtainment of good quality results.
3.1.4. [Cu2(L9)(L10)]Æ2H2OÆCH3OHÆL11 (4)
The structure of the asymmetric unit in 4 is shown in
Fig. 4, and relevant bond distances and angles are listed
in Table 5. The original ligand L4 is clearly absent, and
instead the complex contains three different ligand frag-
ments derived from L4; the mono-acid ligand (L9),
4-chloro-2,6-pyridinedicarboxylate (L10) and 3-phenyl-
triazolo[1,5-a]pyridine (L11). Cu(1) and Cu(2) are
bridged by the N–N diazine group in L9. A pyridine
nitrogen (N(4)) and a carboxylate oxygen (O(2)) and
the other pyridine nitrogen (N(1)) complete coordina-
tion to Cu(1). Cu(2) is bonded to the hydrazone oxygen
O(1) from L9, and three donors (N(5), O(4), O(7)) from
the dicarboxylate ligand L10. Cu(1) has four short in
˚
plane contacts (1.903–2.006 A), but in the extended
˚
structure a fifth axial contact (Cu(1)–O(3) 2.632(3) A)
indicates that the geometry is axially elongated square-
pyramidal (vide infra) (N(3)–Cu(1)–O(2) 162.2ꢁ,
N(4)–Cu(1)–N(1) 173.4ꢁ). Cu(2) has a distorted square
pyramidal coordination geometry within the asymmetric
Fig. 3. (a) Structural representation of 3 (40% probability thermal
ellipsoids). (b) Extended structural representation of 3 (40% probabil-
ity thermal ellipsoids). (c) Optimized geometry for 3 (B3LYP/SVP).
˚
unit, with short in plane distances (1.903–2.033 A), and
Table 4
˚
˚
a longer axial contact (N(2)–Cu(2) 2.368(3) A). How-
ever, in the extended structure (vide infra) a sixth con-
˚
tact (Cu(2)–O(6) 2.375(2) A) shows that the geometry
is actually axially elongated octahedral. There are no an-
ions in the structure indicating that both ligands have
Distances (A) and angles (ꢁ) for [Ni(L8)]BrÆCHCl3 (3)
Ni–N5
Ni–N3
Ni–O1
Ni–N6
N2–N3
N4–N5
N1–H1
1.828(5)
1.831(5)
1.865(4)
1.887(5)
1.397(7)
1.350(7)
0.80(9)
C37
C36
C35
N5–Ni–N3
N5–Ni–O1
N3–Ni–O1
N5–Ni–N6
N3–Ni–N6
O1–Ni–N6
C6–O1–Ni
C1–N1–C5
C1–N1–H1
C5–N1–H1
93.9(2)
C38
178.8(2)
85.05(19)
85.4(2)
C30
N8
N7
C29
O7
O6
N6
Cl1
C17
C28
O1
C27
C14
N3
N1
C5
178.9(2)
95.61(19)
108.1(4)
122.3(5)
120(7)
C16
C25
C24
C23
C15
C18
N4
Cu1
C1
N5
Cu2
C11
N2
C19
O3
Cl2
O2
C7
C22
C10
117(7)
O4
C4
C2
C3
C8
[Ni(L8)](ClO4)ÆH2O and [Ni(L8)](NO3)Æ2H2O are as-
sumed to have similar overall structures.
C9
O5
Of note is the fact that the DFT structure (Fig. 3(c))
and X-ray structure (Fig. 3(a)) of [Ni(L8)]Br are in
Fig. 4. Structural representation of the asymmetric unit in 4 (30%
probability thermal ellipsoids).

814
T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
Table 5
ingꢁ a molecule in close proximity to a metal ion site.
One is reminded of the unusual stabilization of a trypto-
phan moiety close to the active site in galactose oxidase,
with a possibly similar interaction with a tyrosine ligand
[18]. Also the phenyl ring of L9 is positioned above the
pyridine ring (N(5)) in L11, with short contacts between
˚
Distances (A) and angles (ꢁ) for [Cu2(L9)(L10)]Æ2H2OÆCH3OHÆL11
(4)
Cu1–N4
Cu1–N1
Cu1–N3
Cu1–O2
Cu2–N5
Cu2–O1
Cu2–O7
Cu2–O4
Cu2–N2
Cu2–O6
N2–N3
1.903(3)
1.922(3)
1.926(3)
2.005(2)
1.903(2)
1.971(2)
2.031(2)
2.033(2)
2.368(3)
2.375(2)
1.377(3)
1.359(5)
1.315(5)
˚
ring atoms (3.45–3.77 A) highlighting the nice fit of the
aromatic rings in this complex.
The overall structure of 4 reveals some further sur-
prises. The dinuclear subunit shown in Fig. 4 actually
forms a dimer through a carboxylate bridge (O(6)) link-
ing Cu(2) with its symmetry related counterpart. Also,
the tetranuclear subunits are cross-linked to form a
zig-zag chain, via symmetry related O(6) atoms
(Fig. 5), and finally the zig-zag chains are cross-linked
into a 2D grid by long axial O(3)–Cu(1) contacts
N6–N7
N7–N8
N4–Cu1–N1
N4–Cu1–N3
N1–Cu1–N3
N4–Cu1–O2
N1–Cu1–O2
N3–Cu1–O2
N5–Cu2–O1
N5–Cu2–O7
O1–Cu2–O7
N5–Cu2–O4
O1–Cu2–O4
O7–Cu2–O4
N5–Cu2–N2
O1–Cu2–N2
O7–Cu2–N2
O4–Cu2–N2
N5–Cu2–O6
O1–Cu2–O6
O7–Cu2–O6
O4–Cu2–O6
N2–Cu2–O6
173.41(11)
81.24(11)
94.16(11)
81.01(10)
103.56(10)
162.24(10)
173.00(10)
80.65(10)
99.77(9)
˚
(2.632(3) A) shown in Fig. 6. The global direct connec-
tivity between the copper atoms in all cases links the
metal ions orthogonally, thus predisposing the system
O3_6
80.07(10)
99.81(9)
160.38(9)
111.93(9)
75.06(8)
Cl1_6
Cu1_6
88.88(9)
94.98(9)
Cl2
Cu2_6
83.71(9)
89.33(8)
87.60(9)
93.84(9)
163.17(8)
Cl2_6
Cu2
Cl1
Cu1
lost two protons. This is consistent with the in plane Cu–
O and Cu–N distances. The torsional angle Cu(1)–N(3)–
N(2)–Cu(2) (178ꢁ) indicates an almost trans disposition
of the copper atoms about the N–N single bond; how-
ever, this is an orthogonal magnetic connection (vide in-
fra). The copper equatorial planes are also connected by
a longer N–C–O pathway through N(3) and O(1).
A remarkable feature of the structure is the presence
of the ꢀuncoordinatedꢁ, fused triazole molecule L11.
Fig. 4 shows that the phenyl triazole is held in place
above three rings associated with the coordinated ligand
L9. The pyridine ring is positioned above the N(1) pyr-
idine ring in L9, the phenyl ring above the chloro-pyri-
dine ring in L9 and the triazole ring above the chelate
ring involving O(2) in L9. Distances between these rings
O3
Fig. 5. Structural representation of the chain motif in
probability thermal ellipsoids).
4 (30%
N1
O2
Cu1
O3
N3
N4
N4_3
N3_3
˚
subunits are relatively short (e.g., 3.3–4.0 A) indicating
that there are significant p-stacking interactions at play,
which help to stabilize the location of the triazole mole-
N1
O3_3
Cu1_3
˚
cule. The shortest distance (3.386 A) is between N(7)
O2_3
and O(2). This is quite remarkable, possibly unprece-
dented in normal coordination chemistry, and highlights
the importance of p stacking forces as a means of ꢀhold-
Fig. 6. Extended cross linked structural representation of the chains in
4 (40% probability thermal ellipsoids).

T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
815
N8_2
O1
O4'
O2
O5
N3
C7
C13'
N4'
Cu1
N1
O6
N4
O4
Cu1'
N2'
N11_2
N7_2
Cu2'
O3
N2
C6
Cu2
O5'
N11
N4_2
N10
N5
N5'
Fe1_2
N9
N5_2
Fe1
N4
N1
Fig. 8. Structural representation of
ellipsoids).
6
(40% probability thermal
N5
N6
N3
˚
in 5 are slightly longer (1.961–1.970 A) in keeping with a
˚
slightly larger Fe–Fe separation (3.621(3) A). Torsional
angles (Fe–N–N–Fe) are very small (43.8ꢁ) indicating
an acute arrangement of Fe–N bonds about the N–N
˚
single bonds (N–N 1.411–1.413 A). The generally short
N2
Fig. 7. Structural representation of
ellipsoids).
5
(40% probability thermal
Fe–N distances are indicative of a strong ligand field
for Fe(II), consistent with the low spin nature of this
compound (vide infra).
to the absence of significant antiferromagnetic exchange
(vide infra).
3.1.5. [Fe2(PZHPZ)3](ClO4)4 Æ2CH3CN (5)
3.1.6. [Cu2(L12-2 H)](NO3)2 Æ1.5H2O (6)
The molecular structure of 5 is shown in Fig. 7, and
relevant bond distances and angles are listed in Table
6. L5 is clearly absent and the simple pyrazine amidraz-
one ligand is involved in coordination to Fe(II). Three
ligands are wrapped around the two iron(II) centers in
a spiral fashion, with three single N–N bonds connecting
the metal ions. The dinuclear complex cation is essen-
tially identical to that reported previously for the com-
plex [Fe2(PZHPZ)3](NO3)4 Æ5H2O [2], obtained from
the reaction of Fe(NO3)3 with PZHPZ (Chart 1), with
slight differences in internal dimensions. Fe–N distances
The structure of 6 is shown in Fig. 8, and relevant
bond distances and angles are listed in Table 7. The
structure reveals that the starting ligand (L6) has been
hydrolysed at the hydrazide function during the reaction
Table 7
˚
Distances (A) and angles (ꢁ) for [Cu2(L12-2H)](NO3)2 Æ1.5H2O (6)
Cu1–O2
Cu1–N2
Cu1–N1
Cu1–O1
Cu1–O4
N2–N4
1.9168(13)
1.9410(15)
2.0213(14)
2.0935(12)
2.4079(14)
1.346(2)
Table 6
Cu2–N3
Cu2–N4
Cu2–N5
Cu2–O3
Cu2–O6
1.9419(14)
1.9427(14)
1.9822(14)
2.1002(12)
2.1771(13)
˚
Distances (A) and angles (ꢁ) for [Fe2(PZHPZ)3](ClO4)4 Æ2CH3CN (5)
Fe1–N9
Fe1–N1
Fe1–N11
Fe1–N5
Fe1–N4
Fe1–N7
N4–N5
1.958(3)
1.963(3)
1.964(3)
1.964(3)
1.968(3)
1.971(3)
1.412(4)
1.409(5)
O2–Cu1–N2
O2–Cu1–N1
N2–Cu1–N1
O2–Cu1–O1
N2–Cu1–O1
N1–Cu1–O1
O2–Cu1–O4
N2–Cu1–O4
N1–Cu1–O4
O1–Cu1–O4
N3–Cu2–N4
N3–Cu2–N5
N4–Cu2–N5
N3–Cu2–O3
N4–Cu2–O3
N5–Cu2–O3
N3–Cu2–O6
N4–Cu2–O6
N5–Cu2–O6
O3–Cu2–O6
174.14(6)
103.86(5)
79.41(6)
96.49(5)
80.18(5)
159.59(5)
94.34(5)
80.35(6)
97.29(6)
79.47(5)
79.52(6)
165.47(6)
90.72(6)
78.96(5)
155.74(5)
108.03(5)
97.21(5)
107.92(6)
96.02(5)
85.81(5)
N11–N11
N9–Fe1–N1
N9–Fe1–N11
N1–Fe1–N11
N9–Fe1–N5
N1–Fe1–N5
N11–Fe1–N5
N9–Fe1–N4
N1–Fe1–N4
N11–Fe1–N4
N5–Fe1–N4
N9–Fe1–N7
N1–Fe1–N7
N11–Fe1–N7
N5–Fe1–N7
N4–Fe1–N7
95.86(12)
79.92(12)
167.86(13)
167.53(12)
95.87(12)
89.38(11)
97.49(12)
79.65(12)
89.56(12)
88.76(11)
95.19(13)
94.84(13)
96.87(12)
79.62(12)
166.62(13)

816
T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
with copper nitrate in aqueous methanol to form the
corresponding acid. The individual carboxylate ligands
result in the formation of a 1D chain, with dinuclear
subunits. Each ligand coordinates in a heptadentate
fashion binding to Cu(1) via two pyridine nitrogen
atoms, one diazine nitrogen atom and a carboxylate
oxygen, and to Cu(2) via pyridine and diazine nitrogen
atoms, and a terminal hydrazone oxygen. The square-
pyramidal Cu(2) coordination sphere is completed by
Cl1
O2
N6
O1
N4
Ni1
C19
N5
C13
N3
N1
N2
˚
an axial water molecule (Cu(2)–O(6) 2.177(2) A), and
two axial water molecules complete the Jahn–Teller dis-
˚
torted six-coordinate Cu(1) (Cu(1)–O(4) 2.408(1) A;
˚
Cu(1)–O(5) 2.480(1) A). The chain is propagated by
N7
syn–anti carboxylate bridging between the dinuclear
subunits through oxygen O(2), which binds equatorially
to both Cu(1) and Cu(2). The Cu(1)0–N(2)0–N(4)–Cu(2)
torsional angle (164.7ꢁ) indicates an almost trans dispo-
sition of the copper0 basal planes relative to the N–N
Fig. 9. Structural representation of
ellipsoids).
7 (40% probability thermal
˚
bridge (Cu(1)–Cu(2) 4.683(2) A). The syn–anti carbox-
ylate bridge leads to a much larger metal ion separation
˚
(Cu(1)–Cu(2) 5.336(2) A).
Table 8
˚
The presence of two nitrate anions per dinuclear sub-
unit indicates that the ligand has lost two protons. The
carboxylate is clearly one site of deprotonation. Estab-
lishing the other site of deprotonation depends on a
clear identification of the bond order within the C(6)–
N(2)–N(4)0–C(13)0–O(1) framework (bond distances
Distances (A) and angles (ꢁ) for [Ni(L4-2H)] (7)
Ni1–N1
Ni1–N3
Ni1–N4
Ni1–N5
N2–N3
N5–N6
1.906(5)
1.843(5)
1.822(5)
1.930(5)
1.378(7)
1.411(7)
˚
1.301, 1.346, 1.354, 1.249 A, respectively). This is diffi-
cult, and is complicated by very short Cu(1)–N(2) and
N1–Ni1–N3
N1–Ni1–N4
N1–Ni1–N5
N3–Ni1–N4
N3–Ni1–N5
N4–Ni1–N5
94.0(2)
159.4(2)
102.1(2)
83.7(3)
163.7(2)
81.8(3)
Cu(2)0–N(4)0 distances (1.941(2), 1.943(1) A, respec-
˚
tively). One could safely conclude that the charge is
delocalized within this bridging ligand framework.
3.1.7. [Ni(L4-2H)] (7)
The molecular structure of 7 is shown in Fig. 9, and
important distances and angles are listed in Table 8.
The complex consists of one full ligand (L4) bonded to
a single four coordinate Ni(II) center, via pyridine nitro-
gen atoms N(1) and N(4), and diazine nitrogen atoms
N(3) and N(5). The Ni–N distances are very short
pyridine hydrazone (L7) and oxalate on reaction with
Cu(NO3)2 in aqueous methanol indicates ligand hydro-
lysis at the oxalate imine sites rather than the acetyl
imine sites. Pre-coordination of two copper(II) ions to
the ligand (e.g., as in Scheme 2) would lead to Lewis acid
promoted nucleophilic attack by water at the oxalate
C@N sites, thus producing oxalate, which could then
simply bridge the two copper centers. The 2-acetylpyri-
dine hydrazone could then pivot about the pyridine–
copper bond to assume its coordination role as indicated
in the structure of 1. A similar hydrolytic decomposition
appears to have occurred for 2. It is of interest to note
that this ligand forms a stable tetranuclear Ni(II) com-
plex [Ni4(L2)4](BF4)4 Æ6H2O, with no evidence of ligand
decomposition [19].
˚
(1.82–1.91 A), typical of square planar Ni(II), but the
nickel geometry has a slight tetrahedral distortion
(N(3)–Ni(1)–N(5) 164ꢁ, N(1)–Ni(1)–N(4) 159ꢁ). No an-
ions are present and so the ligand bears a 2ꢀ charge.
˚
C–O distances are very short (C(13)–O(1) 1.232(7) A,
˚
C(19)–O(2) 1.222(7) A), indicating them to have double
bond character. The sites of deprotonation are, there-
fore, reasonably assigned to N(3) and N(5). There are
no close contacts of significance between individual mol-
ecules in the lattice.
The ligand in 3 (L8), produced from the reaction of
L3 with NiBr2, would be expected as the product from
the direct reaction of butane-2,3-dione and 2-picolinic
acid hydrazide. However, ligand L3 was derived from
butanedione monoxime, and was characterized by
3.2. Ligand decomposition reactions
L2 was prepared by the reaction of diethyl oxalate
with hydrazine, followed by the reaction of the dihyd-
razide with 2-acetylpyridine. The production of 2-acetyl-

T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
817
are all produced by the same hydrolytic process. The
square planar Ni(II) complex (3) was calculated to be
30 kJ molꢀ1 more stable than the initially proposed
[Ni(L3)Br] chelate (Fig. 10(a)). Since the same ligand
rearrangement was observed with several nickel salts,
it suggests that this might be the overall driving force
for the hydrolytic process.
H3C
N
L'
N
O
Cu
N
H2O
O
O
O
O
CH3
NH2
+
N
H2O
L'
-O
N
Cu
N
N
A recent report indicates that complexes of L3 with
Cr(III), Fe(III), Co(II), Ni(II), Cu(II), Zn(II), Cd(II) and
N
CH3
L7
2þ
UO2 can be produced by refluxing the ligand with the
appropriate metal salt in ethanol. No evidence for ligand
decomposition was presented, and no structures were re-
ported. It is of interest to note that a diamagnetic orange
nickel(II) complex, [Ni(L3-H)Cl], was discussed [3].
L4 is part of a growing group of tritopic ligands
based on a central 2,6-picolinic dihydrazide framework,
which form predominantly [M9(l-O)12] [3 · 3] grid
structures [20–26]. A similar ligand (L40) (Chart 1, L4;
R = H) appears to be quite stable in its reaction with
Mn(II) salts, and the complex [Mn9(L40)6](NO3)6 Æ
12H2O has been produced in a reasonable yield, and
shown to have a [Mn9(l-O)12] [3 · 3] grid structure by
single crystal X-ray diffraction [26]. However, reaction
of L4 with copper acetate in methanol in the presence
of air produces as the main product complex 4. The for-
mation of the mono-carboxylate ligand L9, and the
dicarboxylate ligand L10 clearly results from a ꢀreverse
hydrolysisꢁ reaction in which the full hydrazone ligand
L4 has undergone nucleophilic attack by water at the
2,6-pyridine dicarboxylic acid sites, rather than at the
Schiff base imine positions (Scheme 3). This is presum-
ably encouraged by the pre-coordination of L4 to
Cu(II), at least in the central pyridine cavity, although
it is highly likely that the pre-coordination involves a tri-
nuclear copper(II) complex, similar to other linear
Cu(II)3 complexes observed with this class of ligand
[25,27].
L2
H3C
CH3
Ni(II)
N
N
N N
L3
N
OH
HO
N
L8
Scheme 2.
X-ray crystallography so leaving little doubt as to its
identity (Scheme 2). In our hands L3 was found to be
fairly unreactive to transition metal salts in general,
and this was attributed in part to the anti conformation
of the ligand itself (Fig. 2), which would probably im-
pede the simultaneous coordination of the pyridine
and oxime ends of the ligand.
In order for L3 to bind to a metal ion in a chelating
fashion a significant energy expenditure (vide supra)
would be required to bring suitable donor groups into
an appropriate chelating conformation. Two reasonable
chelation isomers are possible on reaction with nickel
bromide (Fig. 10(a) and (b); Br is proposed as a co-li-
gand), of which the oxime N-bonded example (a) is
the more stable than (b) by 86 kJ molꢀ1. With this iso-
mer, polarization of both C@N bonds in L3 would oc-
cur, and so would provide a hydrolytic mechanism by
which both picolinic hydrazide and the methyl ketone
shown in Fig. 10 (c) could be released prior to reaction
to form L8. It is of interest to note that reaction of L3
with other Ni(II) salts also produced red diamagnetic
complexes of L8 (e.g., [Ni(L8)](ClO4)ÆH2O and [Ni(L8)]-
(NO3)Æ2H2O), and it is reasonable to assume that they
Perhaps the most surprising result is the formation of
the triazole derivative 3-phenyl-triazolo[1,5-a]pyridine
(L11). It is of interest to note that the 2-pyridyl analogue
3-(2-pyridyl)-triazolo[1,5-a]pyridine is produced by the
reaction of di-2-pyridyl ketone and hydrazine hydrate
in methanol–water (1:1) in the presence of air [28], or
by the oxidation of di-2-pyridyl ketone hydrazone with
nickel(II) peroxide in benzene [29]. It also appears to
have resulted from a reaction between a di-2-pyridyl
H3C
CH3
CH3
CH3
H
N
N
CH3
OH
N
CH3
N
N
N
N
N
N
N
O
O
Ni
Br
N
O
O
Ni
Br
Ni
O
Br
(a)
(b)
(c)
Fig. 10. Chelation isomers for L3 (a,b) and proposed hydrolysis product (c).

818
T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
Cl
O
N
N
N
+
L4
O
O
N
N
L9
Cl
N
N
N
+
O
O
L11
O
O
L10
Scheme 3.
ketone hydrazone and copper nitrate in air via an oxida-
tive cyclization, producing the copper(II) complex of
the triazole [28]. The hydrazone of 2-benzoyl pyridine
can also be oxidatively cyclized to the corresponding tri-
azole 3-phenyl-triazolo[1,5-a]pyridine, by reaction with
silver oxide [30]. The simultaneous formation of the
mono- and di-acid ligands (L9, L10) suggests that phe-
nyl-2-pyridyl ketone hydrazone is formed first, which
then oxidatively cyclizes under mild conditions in the
presence of air to produce L11. This process may be cat-
alyzed by the presence of Cu(II) ions.
L4 reacts with cobalt(II) nitrate in CH3CN/CH3OH
in air to form the complex [(L4)3(L9)6Co(III)6Co(II)6
(H2O)6(NO3)6] [Co(H2O)6](NO3)8(CH3CN)3(H2O)6 (8),
which contains both L4 and L9, and has been reported
previously [31]. Two views are shown in Figs. 11(a) and
(b) looking down the trigonal axis of the molecule, and
from a side view, respectively. The mono-carboxylate li-
gand L9 binds six metal ions on each trigonal face, with
three Co(2) (Co(II)) atoms bridged by syn–anti carboxy-
lates in an inner triangular array, with hydrazone oxy-
gen atoms bridging each of them to the outer Co(1)
(Co(III)) centers. The Co(1) centers are linked by the
two end N2O pockets of the tritopic ligands L4 to form
the dodecanuclear array. The thirteenth cobalt(II) ion
appears in the lattice in part to balance the overall
charge. There is no evidence to suggest the formation
of the di-acid (L10) or the triazole (L11), but again this
reaction serves to illustrate the hydrolytic instability of
the hydrazone function in this class of ligand.
The formation of the carboxylate ligand L9 from L4
on reaction with Co(NO3)2 underlines the hydrolytic
instability of the hydrazone linkages and the fact that
Co(II) is capable of mediating nucleophilic attack at
the imine site. The presence of Co(II) at the carboxylate
site is a surprise since Co(III) would be expected to be
more highly polarizing and to prefer the oxygen rich
environment generated in the internal trigonal core. In
Fig. 11. (a) Structural representation of 8 viewed down the threefold
symmetry axis (4% probability thermal ellipsoids), with phenyl rings
removed for clarity. (b) Structural representation of 8 viewed at 90ꢁ to
the threefold symmetry axis (POVRAY ꢀimage).
view of the fact that no ligand decomposition occurs
in the reaction between NiCO3 and L4, this suggests that
Co(III) is instrumental in the formation of L9.
The formation of the Fe(II) complex 5 containing the
bis-bidentate ligand PZHPZ (Chart 1) from L5
(Scheme 4) is somewhat surprising, and indicates a dif-
ferent ligand decomposition mechanism. In this case
any water involved in the hydrolysis must have arisen
mostly from the hydrated metal salt itself, since CH3CN
was used as the solvent. This type of ligand contains
amidrazone fragments at the ends, which are produced
by reaction of, e.g., a picolinic-dihydrazide with the

T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
819
0.0012
0.0010
0.0008
0.0006
0.0004
0.0002
0.0000
N
N
NH2
N
N
N
N
L5
NH2
PZHPZ
N
N
N
O
N
L6
0
50 100 150 200 250 300 350
Temp(K)
O
N
OH
L12
Fig. 12. Variable temperature magnetic data for 1. The solid line
represents the best fit result to Eq. (1) (see text for fitted parameters).
Scheme 4.
methylimidate ester derived from 2-cyanopyridine or
2-cyanopyrazine. Direct synthesis of PZHPZ can be
achieved in high yield by the reaction of 2-pyrazinamide
hydrazone with 2-pyrazine methylimidate [2]. It seems
reasonable to assume that a reverse hydrolytic step has
occurred in the reaction of L5 with a metal ion, in which
the appropriate amidrazone is produced as one product
by hydrolysis of the imine linkage adjacent to the central
pyridine ring. It seems highly unlikely that the appropri-
ate imidate would be produced as well, raising the ques-
tion as to how PZHPZ is produced. The elimination of a
hydrazine molecule between two condensing amidraz-
one moieties seems unlikely, but perhaps the formation
of an amide by metal ion promoted hydrolysis at the
peripheral imine C@N group could lead to a more sim-
ple metal ion facilitated condensation to produce
PZHPZ.
L6 undergoes a simple hydrolysis reaction at the
hydrazide carbonyl group on reaction with Cu(II) in
aqueous methanol to produce the corresponding car-
boxylic acid (Scheme 4). This presumably is effected by
pre-coordination of the Cu(II) ion to the hydrazide
group first, which would polarize the carbonyl group
and enhance electrophilic attack by water at the carbon
atom.
ture independent paramagnetism (Na) (Eq. (1)). A good
fit gave g = 2.21(1), J = ꢀ171.5(4) cmꢀ1, q = 0.0011,
h = 0 K, Na = 200 · 10ꢀ6 cm3 molꢀ1
ꢀ
ꢁ
2Ng2b2
1
vM
¼
ð1 ꢀ qÞ
kðT ꢀ hÞ ð3 þ expðꢀ2J=kTÞÞ
ꢂ
ꢃ
Ng2b2
4kT
þ
q þ Na
ð1Þ
P
P
2
1=2
ð102R ¼ 0:41; R ¼ ½ ðvobs: ꢀ vcalc:Þ = v2obs:
ꢂ
Þ. The
solid line in Fig. 12 was calculated with these parame-
ters. The large singlet–triplet splitting (343 cmꢀ1) in this
case is typical for oxalate bridged di-copper(II) com-
plexes [32]. Compound 2 gave an essentially identical
profile of susceptibility as a function of temperature,
and fitting of the data to Eq. (1) gave g = 2.24(4), J =
ꢀ172(8) cmꢀ1
,
q = 0.0075, h = ꢀ0.2 K, Na = 134 ·
10ꢀ6 cm3 molꢀ1 (102R = 1.7). This clearly indicates the
structural similarity between 1 and 2, and oxalate
bridging.
Complex 3 is diamagnetic, consistent with the square
planar structure and the strong in plane ligand field. The
corresponding nitrate and perchlorate complexes are
also diamagnetic. Complex 7 is also diamagnetic, consis-
tent with its square-planar structure. The variable tem-
perature magnetic properties of 4 show that the
magnetic moment is essentially constant between 2 and
300 K (2.9–3.2 lB/mol), indicative of essentially isolated
spin centers. This is consistent with the structure, which
shows a strictly orthogonal connection between Cu(1)
and Cu(2) through the diazine (N–N) bridge. The long
equatorial, non-orthogonal connection between Cu(1)
and Cu(2) via N(3) and O(1) clearly does not lead to
any significant spin exchange.
3.3. Magnetic properties
The variable temperature magnetic properties of 1
show a rise in susceptibility with increasing temperature
to a maximum around 300 K (Fig. 12), indicating signif-
icant intramolecular antiferromagnetic exchange. The
low temperature increase corresponds to the presence
of paramagnetic impurity. Given the structure of 1, with
essentially isolated oxalate bridged dinuclear copper(II)
species, the data were fitted to a simple exchange expres-
sion for two interacting S = 1/2 centres (Hex = ꢀ2J
{S1 ÆS2}) adapted for a paramagnetic impurity fraction
(q), Weiss-like temperature correction (h) and tempera-
Complex 5 is essentially diamagnetic from 2 to 300 K,
indicative of a low spin Fe(II) system, and identical
to the behaviour of the iso-structural complex
[Fe2(PZHPZ)3] (NO3)4 [2]. The chain structure of 6
creates an alternating bridging arrangement between

820
T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
the copper ions with N–N and syn–anti carboxylates
connecting the copper magnetic orbitals directly. The
presence of a direct trans diazine bridge with a large
Cu–N–N–Cu torsional angle (164.7ꢁ) would suggest
quite strong antiferromagnetic exchange [33], while the
syn–anti carboxylate bridge would not be expected to
generate significant exchange [34,35]. Variable tempera-
ture magnetic data (Fig. 13) show molar susceptibility
reaching a maximum at ꢁ200 K, indicating quite strong
antiferromagnetic exchange coupling between the cop-
per centres. Given the likely dominance of the diazine
bridge in the exchange process, the data were fitted to
Eq. (1), giving g = 2.20(3), 2J = ꢀ234(2) cmꢀ1, h =
1.4 K, q = 0.027, Na = 120 · 10ꢀ6 cm3 molꢀ1 (102R =
ilar effect in the present situation, in addition to the long
˚
distances of separation of the Co(II) centres (5.180 A).
4. Conclusions
Seemingly innocent ligands, based on condensed
hydrazone derivatives, have been shown to undergo
some unexpected hydrolytic decomposition reactions
in the presence of transition metal ions, with some frag-
ments re-arranging to new ligands and some being incor-
porated into novel complex structures. In one most
unusual case oxidatively cyclized 3-phenyl-triazolo [1,5a]-
pyridine (L11) is trapped in a complex structure via
extensive aromatic p interactions only, but is not coordi-
nated. In another case a novel dodecanuclear Co(II)/
Co(III) cluster is produced by the fortuitous combina-
tion of a hydrolyzed mono-carboxylate ligand with a
complete dihydrazone ligand. In most cases the products
can only be successfully characterized by an X-ray struc-
tural determination.
P
P
3.3; R = [ (vobs. ꢀ vcalc.)2/ vobs.
]
2 1/2). The solid line in
Fig. 13 is calculated with these fitted parameters. This
result is entirely consistent with trans N–N bridge con-
necting Cu(1) and Cu(2) [33]. The small positive h cor-
rection might be considered to arise from the syn–anti
carboxylate bridge, since the chains are well separated
in the extended structure.
The dodecanuclear cobalt complex 8 actually con-
tains thirteen cobalt ions (Co(III)6Co(II)7), with the
[Co(II)(H2O)6]2+ ion residing in the lattice. The variable
temperature magnetic data show a slight variation in
moment from 11.0 lB at 300 K to 10.2 lB at 100 K,
followed by a smooth decrease to 8.5 lB at 2 K. The
gradual drop in moment is indicative of possible weak
intra-molecular antiferromagnetic exchange, but be-
cause of the complexities involved with fitting magnetic
data for Co(II) systems in general no attempt has been
made to evaluate an exchange integral. The syn–anti car-
boxylate triangular bridging arrangement of the Co(II)
centres would not be expected to lead to significant cou-
pling, which is consistent with the observed data [34,35].
This arises from a mismatch of the metal orbitals in the
case of copper, which would be expected to have a sim-
5. Supplementary data
Crystallographic data for the structural analyses have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC reference numbers 237242–237247,
254194 for 1, 3–7, L3, respectively. Copies of this informa-
tion may be obtained free of charge from the Director,
CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK
(fax: +44 1223 336033; e-mail: deposit@ccdc.cam.ac.uk
or www: http://www.ccdc.cam.ac.uk).
Acknowledgements
We thank NSERC (Natural Sciences and Engineer-
ing Research Council of Canada), the Research Council
of Norway (HG), and EPSRC (JAKH) for financial
support, and Dr. R. McDonald, University of Alberta,
for structural data.
0.014
0.012
0.010
0.008
0.006
0.004
0.002
0.000
References
[1] Z. Xu, L.K. Thompson, D.O. Miller, E. Ruiz, S. Alvarez, Inorg.
Chem. 42 (2003) 1107.
[2] Z. Xu, L.K. Thompson, D.O. Miller, H.J. Clase, J.A.K. Howard,
A.E. Goeta, Inorg. Chem. 37 (1998) 3620.
[3] T.H. Rakha, Trans. Met. Chem. 24 (1999) 659.
[4] L.K. Thompson, L. Zhao, Z. Xu, D.O. Miller, W.M. Reiff,
Inorg. Chem. 24 (2003) 128.
0
50 100 150 200 250 300 350
Temp (K)
[5] (a) G.M. Sheldrick, SHELX-97, 1997;
(b) SIR-97: A. Altomare, M. Cascarano, C. Giacovazzo, A.
Guagliardi, J. Appl. Crystallogr. 26 (1993) 343.
Fig. 13. Variable temperature magnetic data for 6. The solid line
represents the best fit result to Eq. (1) (see text for fitted parameters).
[6] P.T. Beurskens, G. Admiraal, G. Beurskens, W.P. Bosman, R. de
Gelder, R. Israel, J.M.M. Smits, The DIRDIF-94 program system,

T.L. Kelly et al. / Polyhedron 24 (2005) 807–821
821
Technical Report of the Crystallography Laboratory, University
of Nijmegen, The Netherlands, 1994.
[16] A. Schaefer, H. Horn, R. Ahlrichs, J. Chem. Phys. 97 (1992) 2571.
[17] J.H. Jang, J.G. Lee, H. Lee, Y. Xie, H.F. Schaefer, J. Phys.
Chem. A 102 (1998) 5298.
[7] D.T. Cromer, J.T. WaberInternational Tables for X-ray Crystal-
lography IV, vol. IV, The Kynoch Press, Birmingham, England,
1974, Table 2.2 A.
[18] N. Ito, S.E.V. Phillips, C. Stevens, Z.B. Ogel, M.J. McPherson,
J.N. Keen, K.D.S. Yadav, P.F. Knowles, Nature 350 (1991) 87.
[19] L. Zhao, V. Niel, L.K. Thompson, Z. Xu, V.A. Milway, R.G.
Harvey, D.O. Miller, C. Wilson, M. Leech, J.A.K. Howard, S.L.
Heath, J. Chem. Soc., Dalton Trans. (2004) 1446.
[8] J.A. Ibers, W.C. Hamilton, Acta Crystallogr. 17 (1964) 781.
[9] D.C. Creagh, W.J. McAuley, in: A.J.C. Wilson (Ed.), Interna-
tional Tables for Crystallography, vol. C, Kluwer Academic
Publishers, Boston, 1992, pp. 219–222, Table 4.2.6.8.
[20] L. Zhao, C.J. Matthews, L.K. Thompson, S.L. Heath, Chem.
Commun. (2000) 265.
[10] teXsan for Windows: Crystal Structure Analysis Package, Molec-
ular Structure Corporation, 1997.
[21] L. Zhao, Z. Xu, L.K. Thompson, S.L. Heath, D.O. Miller, M.
Ohba, Angew. Chem. Int. Ed. 39 (2000) 3114.
[22] O. Waldmann, R. Koch, S. Schromm, P. Muller, L. Zhao, L.K.
¨
[11] J. Cosier, A.M. Glazer, J. Appl. Crystallogr. 19 (1986) 105.
[12] (a) Siemens. SMART Data Collection Software, Ver. 4.050; Siemens
Analytical X-ray Instruments Inc., Madison, WI, 1996;
(b) Siemens. SAINT Data Reduction Software, Version 4.050;
Siemens Analytical X-ray Instruments Inc., Madison, WI, 1996.
[13] Sheldrick, G.M., SHELXTL 5.04/VMS, An integrated system for
solving, refining and displaying crystal structures from diffraction
data. Siemens Analytical X-ray Instruments Inc., Madison, WI,
1995.
Thompson, Chem. Phys. Lett. 332 (2000) 73.
[23] O. Waldmann, L. Zhao, L.K. Thompson, Phys. Rev. Lett. 88
(2002) 066401-1-4.
[24] L. Zhao, Z. Xu, L.K. Thompson, D.O. Miller, Polyhedron 20
(2001) 1359.
[25] Z. Xu, L.K. Thompson, D.O. Miller, Polyhedron 21 (2002) 1715.
[26] L. Zhao, Z. Xu, H. Grove, V.A. Milway, L.N. Dawe, T.S.M.
Abedin, L.K. Thompson, T.L. Kelly, R.G. Hynes, D.O. Miller,
L. Weeks, J.G. Shapter, K.J. Pope, Inorg. Chem. 43 (2004) 3812.
[27] L. Zhao, L.K. Thompson, Z. Xu, D.O. Miller, D.R. Stirling, J.
Chem. Soc., Dalton Trans. (2001) 1706.
[14] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A.
Robb, J.R. Cheeseman, J.A. Montgomery Jr., T. Vreven, K.N.
Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V.
Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian,
J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Strat-
mann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W.
Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador,
J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels,
M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Ragh-
avachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S.
Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A.
Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe,
P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez,
J.A. Pople, GAUSSIAN-03, Revision B.05, Gaussian, Inc., Pittsburgh
PA, 2003.
[28] L.P. Battaglia, M. Carcelli, F. Ferraro, L. Mavilla, C. Pelizzi, G.
Pelizzi, J. Chem. Soc., Dalton Trans. (1994) 2651.
[29] H. Ogura, S. Mineo, K. Nakagawa, S. Shiba, Yakugaka Zasshi
101 (1981) 329.
[30] J.H. Boyer, R. Borgers, L.T. Wolford, J. Am. Chem. Soc. 79
(1957) 678.
[31] V.A. Milway, L.K. Thompson, D.O. Miller, Chem. Commun.
(2004) 1790.
[32] M. Julve, M. Verdaguer, A. Gleizes, M. Philoche-Levisalles,
Inorg. Chem. 23 (1984) 3808.
[33] L.K. Thompson, Z. Xu, A.E. Goeta, J.A.K. Howard, H.J. Clase,
D.O. Miller, Inorg. Chem. 37 (1998) 3217.
´
[34] E. Colacio, J.M. Domınguez-Vera, M. Ghazi, R. Kiveka¨s, M.
Klinga, J.M. Moreno, Eur. J. Inorg. Chem. (1999) 441.
[35] E. Colacio, M. Ghazi, R. Kiveka¨s, J.M. Moreno, Inorg. Chem.
39 (2000) 2882.
[15] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.