Spin crossover in iron(III) Schiff-base 1-D chain complexes

Article abstract of DOI:10.1039/b913234a

A series of iron(III) 1-D polymeric materials of the general formula [Fe(III)(Schiff-base)(L)](BPh₄).n(CH₃OH) (Schiff base = N,N′-ethylenebis(salicylaldimine) (H₂salen), N,N′-o-phenylenebis(salicylaldimine) (H₂salophen) and N,N′-ethylenebis(acety1acetone)-2,2′-imine (H₂ acen); L = bridging di-pyridyl or di-imidazole ligand, n = 0-4) and analogues therein, have been synthesised and structurally and magnetically characterised. In this series, a range of structural motifs are observed including linear 1-D chains, hydrogen-bonded chains, a 'hybrid' 1-D chain-and -dimer compound and a hydrogen-bonded dinuclear material; all exhibit extensive intermolecular interactions. The magnetic consequences of varying both the equatorial Schiff-base ligands and axial bridging ligands have been investigated. Overall, we find that, independent of the axial bridging ligand employed, the salen equatorial ligand results in a high spin character and the acen ligand results in spin crossover character, generally with a spin transition of a gradual nature. Variations in magnetic behaviour can be rationalized, in part, in terms of the C₂N₂ backbone conformation of the equatorial Schiff base ligand, which may either inhibit or allow a spin transition. The Royal Society of Chemistry 2010.

Full text of DOI:10.1039/b913234a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Spin crossover in iron(III) Schiff-base 1-D chain complexes†
Tamsyn M. Ross, Suzanne M. Neville, David S. Innes, David R. Turner, Boujemaa Moubaraki and
Keith S. Murray*
Received 3rd July 2009, Accepted 10th September 2009
First published as an Advance Article on the web 8th October 2009
DOI: 10.1039/b913234a
A series of iron(III) 1-D polymeric materials of the general formula [Fe(III)(Schiff-base)(L)]-
(BPh4).n(CH3OH) (Schiff base = N,N¢-ethylenebis(salicylaldimine) (H2salen),
N,N¢-o-phenylenebis(salicylaldimine) (H2salophen) and N,N¢-ethylenebis(acety1acetone)-2,2¢-imine
(H2 acen); L = bridging di-pyridyl or di-imidazole ligand, n = 0-4) and analogues therein, have been
synthesised and structurally and magnetically characterised. In this series, a range of structural motifs
are observed including linear 1-D chains, hydrogen-bonded chains, a ‘hybrid’ 1-D chain-and -dimer
compound and a hydrogen-bonded dinuclear material; all exhibit extensive intermolecular interactions.
The magnetic consequences of varying both the equatorial Schiff-base ligands and axial bridging
ligands have been investigated. Overall, we find that, independent of the axial bridging ligand
employed, the salen equatorial ligand results in a high spin character and the acen ligand results in spin
crossover character, generally with a spin transition of a gradual nature. Variations in magnetic
behaviour can be rationalized, in part, in terms of the C2N2 backbone conformation of the equatorial
Schiff base ligand, which may either inhibit or allow a spin transition.
been made to engineer a high degree of cooperativity in Fe(III)
Introduction
compounds, such as, for example, exploiting tridentate Schiff base
ligands suitably disposed to participate in supramolecular inter-
actions (i.e. hydrogen-bonding and p-stacking). For example, the
mononuclear compound [Fe(III)(pap)2](ClO4)·H2O (pap = bis-
2-hydroxyphenyl-(2-pyridyl)-methaneimine) displays an abrupt
Research in spin crossover (SCO) molecular materials continues
to gather pace from the points of view of possible technological
applications,1,2 nanoparticles3 and fundamental aspects.1 Polynu-
clear SCO systems are receiving much interest, no doubt spurred
on by the work of Kahn et al.4 on 1-D Fe(II) coordination
polymers of the type [Fe(II)(R-1,2,4-triazole)3]2+ that were tested
commercially as memory devices and displays.5 Fundamental
and mechanistic work covers a range of areas and, in the case
of polynuclear systems, for example, it covers such aspects as
determining the relationship between structure and spin transition
mechanism, and identifying the spin state on metal centres at
intermediate temperature plateaus in 2-step transitions.6-9 An
ongoing challenge is to develop ways of predicting the electronic
and memory (hysteretic) behaviour of metal compounds of Fe
and Co that display spin switching, while realising that many
intermolecular effects can influence spin transitions in crystalline
species.1,10
SCO with a 15 K wide thermal hysteresis loop (T 1 ↑ = 165 K
2
and T 1 ↓ = 180 K).12 Further highlighting the importance of
2
supramolecular interactions, the mononuclear Fe(III) compounds
[Fe(qsal)2]NCSe·MeOH and [Fe(qsal)2]NCSe·CH2Cl2 (qsal = N-
(8-quinolyl)-salicylaldiminate) were reported to show cooperative
spin transitions with wide thermal hysteresis loops, attributed to
extensive p-stacking in the crystal lattice.13 Another tridentate hy-
drazone ligand has recently been shown to yield an abrupt, 2-step
spin transition in the complex [Fe(mph)2](ClO4)(MeOH)0.5(H2O)0.5
(Hmph
=
2-methoxy-6-(pyridine-2-ylhydrazonomethyl)phenol
and hydrogen-bonding between adjacent mph ligands was felt to
be involved in the cooperativity mechanism.14
Schiff base ligands of the pap, qsal and mph types, and others
involving N4O2 donor sets, have proven to be very effective at
inducing SCO with Fe(III) and, more recently, for Fe(II)15-18
for which higher cooperativity and thermal hysteresis features
were noted.11,15-18 Fe(III) Schiff base compounds generally include
two tridentate N2O donor ligands, a single hexadentate N4O2
donor ligand, a pentadentate N3O2 plus a monodentate N-donor,
or, as in the present study, a tetradentate N2O2 ligand supple-
mented by two N-donor heterocyclic bases. Extensive studies
on mononuclear systems of the type [Fe(III)(Schiff-base)(L)n]+
(where the ligands are salen, salophen, acen or the Ja¨ger17 type;
n = 1 or 2) have been devoted towards gaining a fundamental
understanding of the factors governing their structure-magnetism
relations.11 More recently, photomagnetic (LIESST; ‘light induced
excited spin state trapping’) investigations have been reported12
and multifunctional (SCO/electrical conductivity) materials of
Fe(II) containing SCO systems continue to be the most
studied because of their bistability features, their abrupt spin
transitions and the tendency to often display high degrees of
cooperativity in the solid state. Comparatively fewer studies have
been devoted to Fe(III) compounds as they commonly show low
cooperativity, with the majority of such systems displaying a
gradual SCO transition.11 Nevertheless, successful attempts have
School of Chemistry, Monash University, VIC, Australia 3800.
E-mail: keith.murray@sci.monash.edu.au; Fax: +61-3-9905-4597;
Tel: +61-3-9905-4512
† Electronic supplementary information (ESI) available: Single crystal
X-ray diffraction, Fig. S1 to S12 and refinement details, magnetic
susceptibility measurements on compounds 1a-e and 2a-c, Fig. S13 to
S16. CCDC reference numbers 738866–738874 and 739042–739044. For
ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/b913234a
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 149–159 | 149
©

View Article Online
type [Fe(sal2-trien)][Ni(dmit)2]x (x = 1 or 3; sal2-trien2- = N,N¢-
bis(triethylenetetramine)salicylaldiminato; dmit2- = 2-thioxo-1,3-
dithiole-4,5-dithiolato) have received attention including the ef-
fects of polymorphism, in the x = 1 phases, and the effects
that ligand conformations and intermolecular contacts have upon
SCO.19,20
As in the sal2-trien/dmit work, above, and in other [Fe(sal2-
trien)]+ salts,21 it was found earlier, that the magnetic and SCO
behaviour of [Fe(III)(Schiff-base)(L)n]+ systems appeared to be
largely dependent on ligand orientation, specifically the confor-
mation of the C2N2 ethylene backbone of the equatorial Schiff-
base ligand, together with the orientation of the axial ligands and
the presence of supramolecular interactions between axial and
equatorial ligands (particularly for salen complexes).11,22 There
has been an occasional, early report of salen-Fe(III) polynuclear
materials such as the 1-D chain [Fe(III)(salen)(tmid)]ClO4 (where
tmid = 1,1¢-tetramethylenediimidazole).23
and other geometric and Van der Waal effects (H-bonding,
p-p) may engender chain-chain interactions and thus influence the
SCO features. Clear structure-magnetic correlations are described
herein for this series, including a dependence on the conformation
of the C2N2 backbone of the Schiff-base ligand. Importantly, the
SCO character in the acen-containing compounds is quite varied
within that series. Comparisons are made to the few early reports,
cited above, on related 1-D Fe(III) Schiff base compounds.23-26
Experimental
Physical measurements
Infrared spectra were recorded on a Bruker Equinox 55 series
FTIR spectrometer (equipped with a Specac Golden Gate dia-
mond ATR sample) in the range of 4000-500 cm-1, with samples
mounted in a diamond anvil (resolution 4.0 cm-1). CHN anal-
yses were performed by Campbell Microanalytical Laboratory,
Department of Chemistry, University of Otago, Dunedin, New
Zealand.
Another focus of the current study is to vary the in-plane
ligand as early work on the [Fe(III)(acen)(L)n]+ mononuclear
system often led to spin transitions being observed.11,24 While the
present study was in progress, reports emerged on 1-D chain com-
pounds [Fe(III)(acen)(bpp)]BPh4 and [Fe(III)(acen)(tmid)]BPh4
that displayed spin transitions at high temperatures.23,25,26 Since
variation in anion is also known to influence SCO properties,
it is notable that the incomplete LS ↔ HS spin transition in
[Fe(III)(acen)(tmid)]BPh4 was found to begin at ~240 K, while
in the earlier studied perchlorate salt, [Fe(III)(acen)(tmid)]ClO4, it
began at ~275 K.23,26 [Fe(III)(salen)L2]+ systems were also similarly
influenced by anion variation.22
Magnetic susceptibility
Data were collected using a Quantum Design MPMS 5 SQUID
magnetometer under an applied field of 1 T over the temperature
range 350–4 K. The powdered or polycrystalline samples were
placed in gel capsules and care was taken to allow long thermal
equilibration times at each temperature point.
General synthesis
Following our recent, and ongoing, studies of 1-D SCO
iron(II) compounds,6,27 we have made a thorough study of 1-
D iron(III) materials, of the Schiff base type, [Fe(III)(Schiff-
base)(L)](BPh4)·n(CH3OH), and report our findings here. We
have carried out a study of the effect of (1) variation of the
equatorial tetradentate Schiff-base ligand using salen, salophen
and acen, with (2) a concurrent variation of the axial pyridyl- or
imidazolyl- bridging ligand, L (see Fig. 1). The Fe ◊ ◊ ◊ Fe distance
across these L bridges are large and thus no intra-chain magnetic
exchange is anticipated between Fe(III) centres but mechanical
All the reagents and solvents were commercially available and
used as received. Methanol was dried by distillation over I2 and
magnesium turnings.
Ligand synthesis
The ligands 4,4¢-azopyridine (azpy), N-(4-pyridyl)-isonicotina-
mide (pin), 2,5-bis(4-pyridyl)-3,4-diaza-2,4-hexadiene (bpdh),
1,4-bis(imidazole-1-yl-methyl)benzene (bix), 4,4¢-dipyridylsulfide
(dps), salen, salophen, acen were prepared by literature
methods.32-36 The ligands bpp, tvp, bpe were purchased from
Sigma Aldrich Inc (Fig. 1). The metal complexes [Fe(salen)Cl]2,
[Fe(salophen)Cl]2 and [Fe(acen)Cl]2 were prepared by literature
methods.37
Metal-complex synthesis
[Fe(salen)(L)](BPh4)·n(CH3OH) (1a-e). Each of these com-
plexes was prepared by the same general method. To a 10 ml
solution of [Fe(salen)Cl]2 (50 mg, 0.14 mmol) in methanol was
added L (0.14 mmol). This solution was heated and stirred for
10 minutes and then filtered into a freshly prepared solution of
NaBPh4 (47.9 mg, 0.14 mmol) in 5 ml of methanol. Within hours
large crystals formed. 1a: [Fe(salen)(azpy)](BPh4)·2(CH3OH).
Orange-brown needle shaped crystals. Yield: 72 mg, 58%. IR
(cm-1): 3458(br), 3052(w), 1602(s), 1598(s), 1544(s), 1013(s). Calcd.
for C52H50BFeN6O4: C, 70.20; H, 5.66; N, 9.45; found C, 69.40;
H, 5.36; N, 9.32. 1b: [Fe(salen)(tvp)](BPh4)·2(CH3OH). Orange-
brown needle shaped crystals. Yield: 69.6 mg, 56%. IR: 3490(br),
3094(w), 1598(s), 1543(s), 1014(s). Calcd. for C54H52BFeN4O4:
Fig. 1 Equatorial Schiff base and axial, 2-connecting, bridging ligands
used in this study.
150 | Dalton Trans., 2010, 39, 149–159
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
C, 73.07; H, 5.90; N, 6.31; found C, 72.09; H, 5.45; N,
6.29. 1c: [Fe(salen)(bpe)](BPh4)·2(CH3OH). Orange-brown needle
shaped crystals. Yield: 64.7 mg, 52%. IR: 3570(br), 3052(w),
1614(s), 1545(s), 1154(s), 1019(s). Calcd. for C54H54BFeN4O4:
C, 72.90; H, 6.12; N, 6.30; found C, 72.23; H, 5.79; N,
6.23. 1d: [Fe(salen)(pin)](BPh4)·2(CH3OH). Orange-brown needle
shaped crystals. Yield: 54.5 mg, 43%. IR: 3462(br), 2903(w),
1617(s), 1594(s), 1537(s), 1092(s). Calcd. for C53H51BFeN5O5: C,
70.36; H, 5.68; N, 7.74; found C, 71.32; H, 5.29; N, 8.09. 1e:
[Fe(salen)(bpdh)](BPh4). Orange-brown needle shaped crystals.
Yield: 69 mg, 56%. IR: 3050(w), 1599(s), 1544(s), 11148(s), 1016(s).
Calcd. for C54H48BFeN6O2: C, 73.73; H, 5.50; N, 9.56; found C,
73.55; H, 5.49; N, 9.52.
overnight and filtered to leave a powder. Single crystals of 5a
were grown by the slow diffusion of a solution of [Fe(acen)Cl]2
(100 mg, 0.31 mmol) in 2 ml of methanol into a solution of
tvp (56.5 mg, 0.31 mmol) and NaBPh4 (109 mg, 0.31 mmol) in
2 ml of methanol, in a H-shaped tube. X-ray quality crystals
of 5a formed over 3-4 weeks. These were found to co-crystallise
with [tvpH](BPh4), 5a1. Sufficient 5a1 could not be collected for
independent characterisation, but a crystal structure was obtained.
(see ESI Fig. S12†).
5a. [(Fe(acen))2(m-tvp)(tvp)(tvpH)](tvpH)(BPh4)4·1.5(CH3OH).
Green powder. Yield 118 mg, 58%. IR: 3053(m), 2999(m),
1610(w), 1503(s), 1388(m), 1066(w), 732(m), 702(w). Calcd. for
C169.5H164B4Fe2N12O5.5: C, 77.89; H, 6.37; N, 6.40; found C, 79.34;
H, 6.38; N, 6.06. Calcd. for no CH3OH, C168H158B4Fe2N12O4:
C, 78.73: H, 6.17; N, 6.56. Calcd. for a 1:1 mixture of 5a
and 5a1: C205.5H194B5Fe2N14O5.5: C, 79.27; H, 6.28; N, 6.30. 5b:
[Fe(acen)(bpe)](BPh4). Green powder. Yield 171 mg, 68%. IR:
3053(m), 2993(m), 1615(w), 1501(s), 1394(m), 1064(w), 732(m),
703(w). Calcd. for C48H50BFeN4O2: C 73.76, H 6.45, N 7.17; found
C, 74.24; H, 6.01; N 6.46. 5c: [Fe(acen)(bpdh)](BPh4)·(CH3OH).
Green powder. Yield 156.3 mg (58%). IR: 3055(m), 2999(m),
1611(w), 1501(s), 1392(m), 1030(w), 732(m), 703(w). Calcd. for
C51H56BFeN6O3: C, 70.60; H, 6.51; N, 9.69; found C, 70.59;
H, 6.35; N, 9.68. 5d: [Fe(acen)(bix)](BPh4).CH3OH. Blue-green
powder. Yield 147.9 mg, 55%. IR: 3118(sh), 3052(m), 2993(m),
1580(w), 1503(s), 1400(m), 1032(w), 732(m), 706(w). Calcd. for
C51H56BFeN6O3: C, 70.60; H, 6.51; N, 9.69; found C, 70.71;
H, 6.23; N, 9.58. 5e: [Fe(acen)(bpp)](BPh4).24-26 Blue powder.
Yield 143 mg (58%). IR: 3053(m), 2989(m), 1610(w), 1500(s),
1388(m), 1066(w), 731(w), 703(m). Calcd. for C49H52BFeN4O2:
C, 73.97; H, 6.59; N, 7.04; found C, 74.09; H, 6.77; N, 6.72. 5f:
[Fe(acen)(pin)](BPh4)·3(CH3OH). Blue powder. Yield 185.9 mg
(65%). IR: 3341(sh), 3055(m), 3000(m), 1702(m), 1580(w), 1498(s),
1389(m), 1062(w), 736(m), 706(w). Calcd. for C50H59BFeN5O6: C,
67.27; H, 6.66; N, 7.85; found C, 66.84; H, 5.84; N 7.99.
[Fe(salen)(L)(CH3OH)](BPh4)·n(CH3OH) (2a-c). Each of
these complexes was prepared by the same general method. To
a 10 ml solution of [Fe(salen)Cl]2 (50 mg, 0.14 mmol) in methanol
was added L (0.14 mmol). This solution was heated and stirred
for 10 minutes and then filtered into a freshly prepared solution of
NaBPh4 (0.14 mmol) in 5 ml of methanol. Within hours large
crystals formed. 2a: [Fe(salen)(bpp)(CH3OH)](BPh4). Orange-
brown needle shaped crystals. Yield: 64.7 mg, 53%. IR: 3410(br),
3050(w), 1617(s), 1599(s), 1544(s), 1128(s), 1013(s). Calcd. for
C54H52BFeN4O3: C, 74.40; H, 6.01; N, 6.43; found C, 74.24; H,
6.01; N, 6.46. 2b: [Fe(salen)(dps)(CH3OH)](BPh4). 1 (CH3OH).
2
Orange-brown needle shaped crystals. Yield 140.8 mg, 58%. IR:
3458(br), 3052(w), 1602(s), 1598(s), 1544(s), 1013(s). Calcd. for
C51.5H48BFeN4O3.5S: C, 70.48; H, 5.51; N, 6.38; found C, 70.68;
H, 5.55; N, 6.44. 2c: [Fe(salen)(bpy)(CH3OH)](BPh4). Orange-
brown needle shaped crystals formed. Yield 67.4 mg, 58%. IR:
3458(br), 3052(w), 1602(s), 1598(s), 1544(s), 1013(s). Calcd. for
C51H46BFeN4O3: C, 73.84; H, 5.59; N, 6.75; found C, 73.25; H,
5.71; N, 6.63.
[Fe(salen)(bpdh)](BPh4)·[Fe(salen)Cl]2(PF6).CH3OH
(3a).
bpdh (33 mg, 0.14 mol) was added to a 10 ml solution of
[Fe(salen)Cl]2 (100 mg, 0.28 mmol). This solution was filtered
into a freshly prepared solution of KPF6 (50 mg, 0.28 mmol) in
5 ml of methanol. The solvent was slowly evaporated off under
a stream of dinitrogen gas and over one week dark red crystals
formed. Yield 10.2 mg, 10%. IR: 1612(s), 1543(s), 1016(s), 927(w),
824(m). Calcd. for C47H46ClBF6Fe2N8O5P: C, 51.55; H, 4.23; N,
10.20; found C, 51.34; H, 4.22; N 10.19.
Crystallographic data collection and refinement
Crystal diffraction data for all structures were collected on a
˚
Bruker APEX X8 diffractometer using Mo-Ka (l = 0.71073 A)
radiation and equipped with an Oxford Instruments nitrogen gas
cryostream. Crystals were mounted on a MiTeGen MicroMounts
fibre in a small amount of oil. Crystals were quench cooled to
123(2) K for all data sets.
[Fe(salophen)(bpdh)](BPh4)·2CH3CH2OH (4a). Bpdh (30 mg,
0.12 mmol) was added to a 10 ml solution of [Fe(salophen)Cl]2
(100 mg, 0.25 mmol) in ethanol. This mixture was heated to aid
dissolution and then filtered into a freshly prepared solution of
NaBPh4 (85 mg, 0.25 mmol) in 5 ml of ethanol. Within hours
large orange-brown needle shaped crystals formed. Yield 87.7 mg,
35%. IR: 3051(w), 1598(s), 1543(s), 1015(s), 829(s). Calcd. for
C62H60BFeN6O4: C, 73.02, H, 5.93, N, 8.24; found C, 73.40; H,
5.57; N, 9.49.
Diffraction data analysis was performed using SAINT+ within
the APEX2 software package.27 Empirical absorption corrections
were applied to all data using SADABS or TWINABS within
the APEX software package for twinned structures.27-28 The
structures were solved using SHELXS and refined against F2 using
SHELXL-97 within X-SEED.29-31 All non–hydrogen atoms in the
structures were refined anisotropically and hydrogen atoms were
generated using the riding model. Special refinement details are
included in the ESI.†
Crystal data and refinement parameters are summarised in
Table 1. CCDC reference numbers 738866–738874 and 739042–
739044 contain the supplementary crystallographic data for this
paper.†
[Fe(acen)(L)](BPh4)·n(CH3OH) (5a-f). Each of these com-
plexes was prepared by the same general method. L (0.31 mmol)
was added to a 10 ml solution of [Fe(acen)Cl]2 (100 mg, 0.31 mmol)
in methanol. This mixture was then heated to aid dissolution
and filtered into a freshly prepared solution of NaBPh4 (109 mg,
0.31 mmol) in 5 ml of methanol. The solution was stirred
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 149–159 | 151
©

View Article Online
152 | Dalton Trans., 2010, 39, 149–159
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Table 2 Selected bond lengths, angles and structural parameters
˚
Bonds (A)
1a
1b
1c
1d
1e
2a
2b
2c
3a
4a
5a (Fe1) 5a (Fe2)
Fe-N(sal)
Fe-N(sal)
Fe-O(sal)
Fe-O(sal)
Fe-N(py)
Fe-N(py)/Fe-OMe
2.106(3)
2.107(2)
1.8855(19) 1.887(2)
1.907(2)
2.225(2)
2.242(2)
2.104(3)
2.104(3)
2.103(4) 2.095(4)
2.109(3) 2.110(4)
1.886(3) 1.872(3)
1.918(3) 1.905(3)
2.196(4) 2.197(3)
2.203(4) 2.221(3)
2.1121(19) 2.116(3) 2.123(3), 2.096(3)
2.121(2) 2.123(3) 2.101(3), 2.115(3)
1.8906(15) 1.905(2) 1.890(2), 1.909(2)
1.9077(16) 1.906(2) 1.902(2), 1.897(2)
2.2001(18) 2.212(2) 2.190(3)
2.107(5)
2.110(5)
1.902(4)
1.902(4)
2.190(5)
2.113(4)
2.096(7) 2.108(6) 1.900(4) 1.891(4)
2.120(6) 2.105(6) 1.913(4) 1.907 (4)
1.894(5) 1.881(6) 1.904(3) 1.898 (3)
1.901(5) 1.910(6) 1.910(3) 1.900(3)
2.171(6) 2.181(7) 2.037(4) 2.011(4)
2.214(6) 2.214(6) 2.036(4) 2.029(4)
1.911(2)
2.193(3)
2.203(3)
2.2206(17) 2.066(2) 2.117(3), 2.091(3)
◦
Angles ( )
1a
1b
1c
1d
1e
2a
2b
2c
3a
4a
5a (Fe1) 5a (Fe2)
A
B
C
D
106.02(9) 106.51(10) 106.26(12) 105.54(13) 109.06(7) 108.67(9) 107.88(10), 107.59(10) 107.36(19) 108.1(2) 103.4(3) 86.00(14) 86.02(15)
88.65(9)
87.95(9)
78.28(9)
88.41(10) 88.45(13) 88.16(13) 87.84(8) 87.98(9) 87.60(10), 87.38(11)
87.81(10) 87.81(13) 88.92(13) 87.22(7) 86.57(9) 87.20(10), 88.19(10)
78.13(11) 78.33(13) 78.32(14) 76.82(8) 77.22(9) 77.55(11), 77.27(11)
87.67(19) 86.5(3) 88.8(3) 93.39(15) 93.57(16)
87.5(2)
77.9(2)
88.4(2) 90.7(3) 93.70(16) 94.07(16)
77.1(2) 77.0(3) 87.04(16) 86.56(16)
py-Fe-py dihedrala
11.7
15.1
24.0
21.2
48.5
—
—
—
52.0
83.6
14.8
28.2
Structural Parameters 1a
C2N2 conformation Meso
Inter-chain Fe ◊ ◊ ◊ Fe 13.476(2) 13.736(3) 13.753(2) 13.709(3) 15.668(2)
Intra-chain Fe ◊ ◊ ◊ Fe 7.792(2)
1b
1c
1d
1e
2a
2b
2c
3a
4a
5a (Fe1) 5a (Fe2)
meso
meso
meso
meso
meso
meso
—
meso
—
meso
—
6.940(2)
envelope umbrella meso
15.424(4) 15.656(6) 16.295(1) 10.689(1)
7.296(3) 10.341(1) 37.445(1), —
13.462(1)
7.775(5)
7.811(9) 7.648(7)
5.957(5) 8.840(1) 7.629(9), 7.819(9)
a Dihedral angle between Fe-py planes.
Results
Structural analysis
A summary of relevant bond lengths, angles and structural
parameters is provided in Table 2.
1a-e. The five compounds 1a-e are of the general formula
[Fe(III)(salen)(L)](BPh4)·n(MeOH) (L = 4,4¢-azpy, tvp, bpe, pin
and bpdh, Fig. 2). Compounds 1a-d all crystallize in the mon-
oclinic space group: P21/n, and 1e crystallizes in the lower
¯
symmetry triclinic space group P1. Each structure consists of
distorted octahedral Fe(III) centres coordinated equatorially by
the tetradentate Schiff base ligand salen which provides a N2O2
donor set (Table 2). These [Fe(III)(salen)]+ moieties are then
bridged by the two-connecting pyridyl-donor ligands, completing
the octahedral coordination geometry around each Fe(III) centre.
Infinite cationic 1-D linear chains are thus formed (Fig. 2(a)). The
intra- and inter-chain Fe ◊ ◊ ◊ Fe separations within each chain vary
depending on the ligand length (Table 2).
The main distinction between 1a-d and 1e is the packing
arrangement of the chains and the resultant intermolecular
interactions. For 1a-d the chains are arranged in alternating rows
of pairs which are close packed. The charge balance for the cationic
chains is provided by tetraphenylborate anions which are located
between the pairs of chains. These anions are involved in four-fold
edge-to-face phenyl embraces comprised of -CH ◊ ◊ ◊ p interactions
as illustrated in Fig. 2(b). Solvent methanol molecules are also
present in these materials and are involved in hydrogen-bonding
interactions with the salen phenolate oxygen, and are also involved
Fig. 2 Representative example of 1-D chain materials of the type
[Fe(salen)(L)](BPh4)·n(MeOH), showing (a) the structure of the infinite
1D chain [Fe(salen)(L)]+ moiety of 1a where L = azpy; and (b) the four
-
fold phenyl embrace which occurs between chains and BPh4 anions.
in -OH ◊ ◊ ◊ p interactions with the tetraphenylborate rings. For 1e
the parallel packed cationic chains show a staggering of the salen
ligand in alternate chains such that maximum pack efficiency is
achieved (Table 2). In this compound the efficient chain packing
arrangement results in the absence of solvent molecules in the
crystal lattice. The packing also comprises of pairs of chains, where
the C2N2 backbone of the salen ligands face outwards, and these
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 149–159 | 153
©

View Article Online
pairs of chains are separated by tetraphenyl borate anions. Each
tetraphenylborate anion is involved in three -CH ◊ ◊ ◊ p interactions.
2a-c. The compounds 2a-c have the general formula
[Fe(III)salen(L)(MeOH)](BPh4) (Fig. 3) where distorted
octahedral Fe(III) centres are equatorially coordinated by
salen to form [Fe(III)(salen)]+ units. These units are axially
coordinated by one pyridyl nitrogen atom from either L = bpp,
dps or bpy, and the octahedral coordination is completed by one
oxygen atom of a methanol molecule to give a N3O3 donor set.
While each of these compounds is of the same general formula,
they crystallise in different symmetries and show subtly different
intermolecular interactions. For 2a and 2c, each mononuclear
moiety interacts with adjacent molecules via hydrogen bonding
interactions between the methanol -OH and the non-bonded
pyridyl nitrogen of L; thus producing an overall structural motif
of zig-zag hydrogen bonded 1-D chains (Fig. 3(2a) and (2c)).
The chains in 2a are packed in pairs similar to that seen in
1e, whereas there is no pairing in 2c. The remaining crystal
volume is occupied by tetraphenylborate anions which for 2a are
involved in one -CH ◊ ◊ ◊ p interaction with a pyridyl group on bpp
ligand. In contrast, while 2b also consists of hydrogen-bonded
zig-zag chains, the [Fe(III)salen(L)(MeOH)]+ moieties show two
different, alternating, hydrogen-bonding arrangements along
each chain. Firstly, interactions exist between the methanol -OH
and the non-bonded pyridyl nitrogen of the bpp ligand as seen for
2a and secondly, there is one interaction from bound methanol
-OH to solvent -OH which then further hydrogen-bonds to the
non-bonded pyridyl nitrogen of the bpp (Fig. 3(2b)). The Fe(III)
centres are colinear in 2a and 2c but not in 2b.
Fig. 4 Representation of the structure of the ‘hybrid’ material 3a,
which shows pairs of 1-D chains of [Fe(III)(salen)(bpdh)]+ separated by
[Fe(III)(salen)Cl]2 dimers.
dimer starting material, occurs in alternating layers throughout
the crystal lattice, where one layer consists of a single array of
[Fe(III)(salen)Cl]2 dimers, and the other consists of one ‘pair’ of
[Fe(III)(salen)(bpdh)]+ chains. Notably, the salen ligands in the
‘paired’ rows of [Fe(III)(salen)(bpdh)]+ are staggered, as described
for 1e, although the intra-chain separation suggests the chains are
not closely packed (Table 2). The area between chains is filled with
methanol molecules and hexafluorophosphate anions which pack
between 1D chain and [Fe(III)(salen)Cl]2 dimer layers.
4a. The compound 4a, [Fe(III)(salophen)(bpdh)](BPh4)·
2EtOH, contains the salen analogue ligand salophen. The
structure consists of 1-D chains of distorted octahedral Fe(III)
centres coordinated equatorially by the bulky salophen (Table 2).
The [Fe(III)(salophen)]+ moieties are then bridged by the two-
connecting pyridyl ligand bpdh, to form infinite 1-D chains
(Fig. 5). In contrast to 1a-e, the salophen ligands are in a staggered
conformation along the chains. Close packing of these chains is
likely inhibited by the bulky phenyl groups of the salophen ligand
and as such there is a much larger intra-chain Fe ◊ ◊ ◊ Fe separation
than for the salen analogues (Table 2). Tetraphenylborate anions
are located throughout the crystal lattice and, unlike the salen
analogues, they are not involved in -CH ◊ ◊ ◊ p interactions with the
1-D chains. Furthermore, there is no pairing of chains as seen in
1a-e. Solvent ethanol is present in the crystal lattice and is involved
in a hydrogen-bonding interaction with a phenolate oxygen atom
of the salophen ligand.
Fig. 3 1-D hydrogen bonded compounds of the type [Fe(III)salen(L)-
(MeOH)](BPh4). L = bpp (2a), L = dps (2b) and, L = bpy (2c).
Fig.
5
Representation of the structure of [Fe(III)(salophen)-
(bpdh)]BPh4·2EtOH (4a).
3a. The atypical structure of 3a may be considered to
consist of two distinct portions comprised of linear 1-D
chains of [Fe(III)(salen)(bpdh)]+ and the starting material
[Fe(III)(salen)Cl]2 (Fig. 4). The 1-D chains are configured as
seen in 1e, where [Fe(III)(salen)]+ moieties are bridged by bpdh
ligands, thus completing the distorted octahedral Fe(III) N4O2
donor set (Table 1). The second portion of 3a, consisting of the
5a. The compound 5a consists of infinite linear
hydrogen bonded chains comprised of [(FeIII(acen))2(m-
tvp)(tvp)(tvpH)]2+ units and a protonated non-coordinating
tvp molecule (Fig. 6(a)). The [(FeIII(acen))2(m-tvp)(tvp)(tvpH)]2+
portion contains two Fe(III) centres both with equatorial
N2O2 donor sets provided by the acen ligand. The two Fe(III)
154 | Dalton Trans., 2010, 39, 149–159
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
and would probably vary significantly over the series owing to the
difference in ligand lengths and flexibilities (Fig. 7 (top)).
Magnetic susceptibility measurements
1a-e. Magnetic susceptibility measurements on each of the
salen-containing 1-D chain materials 1a-e revealed a high spin d5
character over all temperatures (ESI; Fig. S13†) The meff values
remained constant at ca. 5.5-5.9 mB over the temperature range
4-300 K consistent with that expected for Fe(III) in the high spin,
S = 5/2, state.
2a-c. The data on each of the hydrogen-bonded salen contain-
ing 1-D chain materials 2a-c revealed a high spin character over
all temperatures (ESI; Figs. S14-16†). The meff values remained
constant at ca. 5.8-5.9 mB over the temperature range 4-300 K
consistent with that expected for Fe(III) in the high spin state.
Fig. 6 Representation of the structure of the hydrogen bonded dinuclear
5a which (a) contains two Fe(III)(acen) centres joined by a m-tvp ligand
and capped by either tvp or Htvp. (b) These units are hydrogen-bonded
through unbound Htvp ligands into infinite 1-D chains.
centres are bridged by one tvp ligand. The remaining axial
donor set is then provided by two tvp molecules which ‘cap’
the ends of the dimer to complete the N4O2 overall donor set
(Table 2). The non-coordinating pyridyl group of one ‘capping’
tvp ligand is protonated and engages in a hydrogen bond
with the single protonated non-coordinating tvp molecule,
thus forming hydrogen-bonded 1-D chains of Fe(III) dimers
(Fig. 6(b)). The closely packed chains (Table 2) are surrounded
by tetraphenylborate anions, although there is sufficient room to
accommodate two methanol molecules. The tetraphenylborate
anions are closely associated with the protonated pyridyl groups
such that two tetraphenylborate anions form an ‘enclosure’
around the protonated pyridyl ring, and the aromatic hydrogens
on either side of the protonated pyridyl nitrogen atom engage in
-CH ◊ ◊ ◊ p bonds with these enclosing tetraphenylborate anions.
Each tetraphenylborate anion participates in only one -CH ◊ ◊ ◊ p
bond, implying that there are no inter-chain interactions.
3a. Susceptibility measurements on the ‘hybrid’ 1-D
chain/chloro-dimer revealed meff values at 300 K of ca. 8.36 mB
(5.91 mB per Fe(III)) consistent with two high spin Fe(III) ions.
Between 300 and 4 K the meff values decreased gradually to a final
value of 3.7 mB, per Fe, as expected owing to the presence of weakly
antiferromagnetically coupled [FeIII(salen)Cl]2 units that have a J
value of -7 cm-1 (Fig. 8).37 The decrease in meff is not due to SCO.
5b-f. Single crystals of the compounds 5b-f were not ob-
tained, hence their constitutions and structures were deduced
from IR spectroscopic and microanalytical measurements on
the powders and by comparison to the reported analogue
[FeIII(acen)(bpp)](BPh4), 5e (Fig. 7 (bottom)). This previously
reported coordination polymer consists of 1-D zig-zag chains of
[Fe(III)acen]+ units linked via bpp ligands.25 There are no inter-
chain or intermolecular interactions reported in the papers.24-26
Likely, the materials 5b-d and 5f also consist of 1-D chains but
the linear or zig-zag conformation of these chains is unknown
Fig. 8 Magnetic moment, per 2Fe, versus temperature over the tempera-
ture range 4-350 K for the mixed 1-D chain and dimer compound 3a.
5a. A spin transition was observed that was not quite complete
even at high temperatures (Fig. 9). At 350 K the meff value of 4.70
mB is less than the high spin value for the Fe(III) centres. Between
350 and 175 K the meff values decreased to a value of ca. 2.12
mB. Below 175 K the meff values again decreased, but gradually, to
reach a value of ca. 1.8 mB at 4 K, consistent with that expected for
5
Fe(III) in the t2g low spin state.
5b-f. Each of these acen containing 1-D Fe(III) chain com-
plexes undergo spin transitions, but with varying degrees of
completeness and occurring over different temperature ranges
(Fig. 9). For 5b the meff value at 350 K of 3.5 mB decreased gradually
over the whole temperature range to reach a lowest value of 2.08
mB at 4 K, consistent with the majority of Fe(III) sites being in
the low spin state. The meff value for 5c at 350 K of 3.45 mB
decreased gradually over the whole temperature range to reach
a minimum value of 1.85 mB at 4 K, consistent with a low spin
Fe(III) character. For 5d the meff value at 350 K of 5.0 mB is again
Fig. 7 (top) Schematic representation of 5b-f where Fe(III)(acen) units
are linked by L = bpe, bpdh, bix, bpp or pin into 1-D chains. (bottom)
The structure of [Fe(III)(acen)(bpp)](BPh4), 5e, as reported by Imatomi
et al.25,26
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 149–159 | 155
©

View Article Online
[Fe(III)(salen)]2O dimer formed.38 However, synthesis under an
inert atmosphere was not deemed necessary, as removal of the
products from the mother-liquor shortly after formation avoided
m-oxo dimer contamination. Further to this, while the complexes
2a-c were synthesised using the same method and stoichiometry to
those used for 1a-e, hydrogen-bonded 1-D chains which contain
coordinated methanol molecules formed, instead of the target 1-D
chains. Since the ligands that resulted in this hydrogen-bonded 1-
D chain structure are, in general, shorter than those used to form
1a-e, they may not sterically favour the formation of 1-D chains
of type [Fe(III)(salen)(L)]n.
Further highlighting the unpredictable nature of the 1-D chain
formation, the ‘hybrid’ compound 3a, containing both the starting
material [Fe(III)(salen)Cl]2 and the 1-D chain target molecule co-
crystallised, was obtained when the hexafluorophosphate anion
was used in place of tetraphenylborate. Notably, this compound
took 1 week to crystallise whereas the analogue 1e crystallised
within 24 h.
Also, within this study, attempts were made to synthesise a
range of 1-D chain materials using the Schiff-base ligand salophen;
however, only that with bpdh, the longest of all the pyridyl donor
ligands, was successful. It is likely that the bulky nature of the
salophen ligand requires a larger intra-chain Fe ◊ ◊ ◊ Fe separation
and thus a longer pyridyl linker ligand for formation of the
chain species. Presumably the L ligands provide weakly basic
conditions as well as coordinating properties. It proved difficult
to re-synthesise 4a after the first attempt, suggesting that subtle
factors influence its formation and isolation.
Fig. 9 Magnetic moment, per Fe, versus temperature over the temperature
range 4-350 K for the acen containing hydrogen bonded chain material
5a (this powder sample has constitution 1:1 5a:5a1) and the 1-D chain
materials 5b-f.
consistent with an intermediate high spin–low spin character for
the Fe(III) centres. Between 315 and 165 K the meff values decreased
to a value of ca. 2.75 mB. Below 165 K the meff values again decrease,
but more gradually, to reach a minimum value of ca. 2.21 mB at 4 K,
consistent with that expected for the majority of the Fe(III) sites
being in the low spin state. For 5e the meff value at 350 K of 5.7 mB is
consistent with an essentially complete high spin character for the
Fe(III) centres. Between 300 and 100 K the meff values decreased,
with T1/2 ~180 K, to a value of ca. 2.77 mB. Below 100 K the meff
values decreased again but more gradually, to reach a minimum
value of ca. 2.34 mB at 4 K, consistent with that expected for the
majority of the Fe(III) centres being in the low spin state. For 5f
the meff value at 300 K of ca. 6.06 mB is consistent with high spin
Fe(III). Between 300 and 155 K the meff values decreased gradually
and then more rapidly to 90 K, to reach ca. 2.96 mB. The T1/2
is 135 K. Below 90 K the meff values decreased again, but more
gradually, to reach a minimum value of 2.16 mB at 4 K, consistent
with low spin Fe(III) sites.
From the relative scarcity of acen-containing crystal structures
obtained in the present study, it is obvious that crystallised
products for [Fe(III)(acen)(L)]+ systems are not as readily formed
with the pyridyl/imidazolyl donor ligands as they are for
[Fe(III)(salen)(L)]+ systems. Attention may be drawn here to the
less bulky aromatic nature of the acen ligand, compared to salen
and salophen, that results in less intermolecular crystal packing
effects involving the bulky tetraphenylborate anion, and it is
thus less likely to yield crystalline products. In all cases, except
5a with tvp, only powdered samples were obtained for acen
systems and their structures have been proposed from a range
of physical techniques (see Fig. 7 (top)). Also included in this
series is the previously reported material [Fe(acen)(bpp)](BPh4)25,26
that helped in our proposed structural analysis of the remainder
of this series (Fig. 7 (bottom)). As just mentioned, 5a was the
only [Fe(III)(acen)(L)]+ compound where structural analysis could
be carried out, and crystals suitable for crystallographic study
were formed over a period of seven days under atmospheric
conditions by slow diffusion in an H-shaped tube. Structural
analysis revealed not a single molecule 1-D chain, as expected,
but a hydrogen-bonded ‘chain of dimers’; the bulk powdered
sample was confirmed to be the same by microanalysis and powder
X-ray diffraction analysis. The production of a tvpH+ unit in
this material implies an effective acidic solution is present in
the synthesis and we note that 5a co-crystallised with yellow-
Discussion
Synthesis
The aim of producing a range of 1-D chain Fe(III) complexes
containing the Schiff base ligand salen was achieved here, together
with the synthesis of a number of other, unexpected, materials.
Interestingly, all the salen-containing compounds were obtained
under the same reaction conditions, but different products re-
sulted, in a somewhat unpredictable manner. Importantly, the
complexes 1a-e were readily obtained as single crystals, often in a
matter of hours and usually within 24 h, through a one-pot mixing
of all the reactants. This suggests that the 1-D chain arrangement,
which is efficiently interlocked via -CH ◊ ◊ ◊ p interactions with the
tetraphenylborate anions, is a thermodynamically stable product.
Notably, if the 1-D chain materials were left standing in solution
after formation, rather than being removed from the mother-
liquor, large red crystals of the well known m-oxo bridged
-
plate crystals of tvpH+BPh4 , 5a1. Crystals of 5a could not be
adequately purified by hand-separation, so magnetism studies
were carried out on the powdered sample. Elemental analysis on
the powder gave good agreement between, either, a 1:1 mixture
of 5a:5a1 and the non-methanolate of 5a, with the former being
favoured.
156 | Dalton Trans., 2010, 39, 149–159
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Structural and magnetic analysis
Fe(III) centre one axially-bound imidazole is disordered over two
positions, one of which was postulated to allow a spin transition
while the other prohibited it.23 It is also possible that the linker
pyridyl donors, together with salen, provide a total ligand field
that is too weak for a spin transition to occur.
In the compound 4a, contrasting to 1a-e, the salophen ligand
has adopted the umbrella conformation, where the aromatic
moieties deviate from the FeN2O2 mean plane in the same
direction, while the carbon atoms comprising the C2N2 backbone
deviate from it very little (Fig. 10(c)). This conformation has been
observed in [Fe(III)(5-CH3O-salen)(Him)2]Cl, and is known to
produce high spin Fe(III) compounds.39 This suggests that 4a
is also a high spin compound but insufficient sample precluded
magnetic studies being made
In the hydrogen-bonded chain materials 2a-c, the N3O3 donor
set that they contain is not conducive to SCO behaviour in
[Fe(III)(salen)]+ systems. Thus they remain high spin and show
high spin Fe(III) Fe-N and Fe-O bond lengths. [(Fe(III)(salen))2(m-
azpy)(H2O)2](ClO4)2·(azpy)·(H2O) contains a similar coordination
environment to 2a-c, and was found to remain high spin over all
temperatures.40
The ‘hybrid’ salen material 3a displays a superposition
of the magnetic susceptibility of the antiferromagnetically
spin-coupled [Fe(III)(salen)Cl]2 dimer, and that due to the
[Fe(III)(salen)(bpdh)]+ chains. The corresponding meff val-
ues decrease markedly with temperature as expected for
[Fe(III)(salen)Cl]2 units being present. The chains are then high
spin d5, since the calculated meff for two high spin d5 ions is 8.36
mB (5.92 mB per Fe(III)) at 300 K and the observed value is of this
magnitude. Typical high spin Fe(III) Fe-N and Fe-O distances are
observed for the 1-D chains 123 K (Table 2). Interestingly, the salen
ligand in the 1-D chain portion of 3a has adopted the envelope
conformation (Fig. 10(a)), suggesting that a spin transition could
have occurred but, presumably the ligand field around each Fe is
not strong enough.
Salen-containing complexes. Despite all changes made to L,
with the tetradentate Schiff-base salen, a temperature dependence
in meff was not observed (1a-e). These compounds display typical
high spin d5 behaviour in which the meff values remain constant at
ca. 5.9 mB between 300 K and 4 K, with cm values correspondingly
Curie-like. The high spin behaviour was further confirmed in the
crystal structure data at 123 K of these materials where the Fe-
N and Fe-O bond lengths are typical of high spin Fe(III) values
(Table 2),11,14,20-22,26
It has previously been observed that for a spin transition to occur
in mononuclear [Fe(III)(salX)(L)2]+ systems, the C2N2 backbone
must adopt the envelope conformation; the other possible con-
formations include meso and planar (or umbrella) (Fig. 10). It
has been suggested that the envelope conformation allows a spin
transition to occur with minimal geometrical change, whereas the
meso and planar conformations constrain the salen ligand such
that is it not able to adapt to the incipient spin transition.22 The
meso and planar conformations essentially ‘lock’ the complex in
the high spin form. In the compounds 1a-e described here, the
meso conformation has been adopted whereby aromatic moieties
deviate slightly, by approximately equal distances and in opposite
directions to each other, from the FeN2O2 mean plane (Fig. 10(b)).
Fig. 10 Possible salen C2N2 backbone conformations (a) envelope
(e.g. 3a), (b) meso (e.g. 1a) and (c) planar (umbrella, e.g. 4a).
Acen containing complexes
In the present study, SCO behaviour has been observed for
the compounds 5a-f (Fig. 9) and these contain a range of
bridging L ligands. This, along with the previous literature
examples of SCO [Fe(III)(acen)]+ compounds, suggests that acen
provides a versatile platform for the synthesis of SCO materials.
Furthermore, the range of T1/2 temperatures observed over this
1-D series suggests that altering the character of the axial bridging
ligand provides an effective means to tune the properties of the
spin transition. Already reported are the three SCO 1-D chain
materials [Fe(III)(acen)(bpp)](BPh4), [Fe(III)(acen)(tmid)](BPh4)
and [Fe(III)(acen)(tdim)](ClO4),23-26 all of which undergo gradual
spin transitions that are incomplete at high temperatures. The first
of these has recently been shown, at temperatures midway between
where the HS and LS states are stabilised, to consist of alternating
HS and LS chains while the second consists of HS-LS-HS-LS units
within the chains,25,26 a situation akin to our observations on some
Fe(II) 1-D systems.6,27
Fig. 11 Observed acen ligand conformations in (a) [Fe(III)(acen)-
(Him)2](BPh4),39 (b) [Fe(III)(acen)(bpp)](BPh4)25,26 and (c) 5a.
It is worth noting that it is not solely the configuration of the
equatorial Schiff base ligand that may dictate whether or not a
spin transition is possible. In the previously reported compound
[Fe(III)(salen)(tmid)](ClO4), an incomplete spin transition is ob-
served below 100 K; crystallographic study shows that for each
Within the present series of compounds, a range of ligand
lengths, flexibilities, hydrogen-bonding and p-p stacking capabili-
ties has been included. However, as structural analysis was usually
not possible, only general structural-magnetic correlations can be
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 149–159 | 157
©

View Article Online
drawn. Compound 5f, containing the ligand pin, displays a spin
transition with the lowest transition temperature and the most
abrupt transition. This ligand is possibly the least flexible in the
current series and also contains central functional groups capable
of hydrogen-bonding. Compounds 5d-e display the next highest
transition temperatures and contain the flexible ligands bix and
bpp, where bix probably confers a zig-zag motif as seen for bpp
(Fig. 7).25,26 Lastly, 5b-c undergo spin transitions at the highest
transition temperatures and contain ligands (tvp and bpe) which
are more rigid and less likely to adopt the zig-zag motif, but, unlike
pin, are not likely to engage in intermolecular interactions. The
ligand field strength of the particular bridging ligand N-donor will
play a role in the observed transition temperature; however, the
only compound of this series to contain a stronger field imidazole
donor group, 5d, shows a spin transition with an intermediate
T1/2 temperature in relation to this series. This suggests that it
is a combination of crystal packing and ligand field effects that
contribute to the observed magnetic behaviour.
Complex 5a provides a novel example of a dinuclear Fe(III)
SCO complex where the only previous examples reported
are [(Fe2(salen)2(m-azpy)(H2O)2](ClO4)2·(azpy)·(H2O),40 and pen-
tadentate (N3O2) Schiff base Fe(III) complexes such as [Fe2((3-
MeO-sal)2-trien)2(m-tmid)](BPh4)241and others that have 4,4¢-bipy,
tvp and azpy bridges axially bound.6,40 Structural analysis on 5a re-
vealed no obvious inter-chain interactions. The Fe(III) centres are
connected directly by a rigid, linear bridging ligand. 5a undergoes
a gradual and incomplete transition which suggests that in this
case, the use of a rigid bridging ligand has not resulted in a highly
cooperative system. Whether the gradual nature of this transition
is, in part, due to a staggering of spin transition temperatures of
the two distinct Fe(III) centres, similar to that thought to occur
between independent 1D chains in [Fe(III)(acen)(bpp)]BPh4 (vide
supra), and at alternating Fe(III) centres within 1D chains of
[Fe(III)(acen)(tmid)]BPh4 is unclear.25,26
The poor cooperativity in [Fe(III)(acen)(L)](BPh4); L = bpp,
tmid, was attributed to the lack of inter-chain/intermolecular in-
teractions and the flexible nature of the bridging ligand which con-
fers a zig-zag packing motif.25,26 Since related monomeric systems
[Fe(III)(acen)(4-methylpyridine)2](BPh4) and [Fe(III)(acen)(3,4-
dimethylpyridine)2](BPh4) undergo complete, more abrupt spin
transitions,24 this suggests that better cooperativity is occurring in
these systems than in the polymeric systems mentioned above,
but structural details were lacking. Furthermore, a compari-
son between the above 1-D chain acen systems and the hys-
teretic monomeric compound [Fe(III)(pap)2](ClO4).H2O,12 that
displayed extensive p-stacking in the crystal lattice, suggests that
covalent linkages between heterocyclic axial donors are unsuited
to the transmission of cooperative interactions in these 1D systems.
Finally, a detailed categorisation of equatorial Schiff-base
ligand conformation, as described for the salen analogues, has
not been attempted previously for acen and derivatives. There
are some subtle structural aspects of interest that are apparent
here. The pertinent acen-containing compounds are the low spin
complex [Fe(III)(acen)(Him)2](BPh4)42 and the SCO complexes
[Fe(III)(acen)(bpp)](BPh4),25,26 [Fe(III)(acen)(tmid)](BPh4)26 and
5a. In comparing the relative acen ligand geometries in each
it can be seen that, as for salen, different conformations of
the C2 backbone (C1 and C2 in Fig. 11(d)) are observed.
[Fe(III)(acen)(Him)2](BPh4) shows the envelope conformation and
[Fe(III)(acen)(bpp)](BPh4), [Fe(III)(acen)(tmid)](BPh4) and 5a all
show the meso conformation. As suggested by the classification
of the acen conformation in each compound, there are notable
differences in the deviations of the imine carbon atoms (C3 and
C4, Fig. 11(d)) from the FeN2O2 mean planes in each compound.
In [Fe(III)(acen)(Him)2](BPh4) one imine carbon deviates above
the FeN2O2 mean plane and the other lies approximately
within the plane. In contrast, for [Fe(III)(acen)(bpp)](BPh4),
[Fe(III)(acen)(tmid)](BPh4) and 5a, the imine carbon atoms
deviate above and below the plane by equal amounts. From
this it appears that the acen ligand conformation may have an
opposite trend to that observed for the salen materials, where
the envelope conformation, rather than the meso conformation,
is usually prohibitive to a spin transition. For [Fe(III)(acen)L]+
systems, the greater the degree of distortion around the Fe(III)
centre (generated by the ligand geometry) the more a LS character
is attained. Furthermore, the dihedral angle formed by imidazole
ligands on opposing sides of the Fe(III) centre in the compound
[Fe(III)(acen)(Him)2](BPh4) is significantly greater than it is for
5a, and for [Fe(III)(acen)(bpp)](BPh4) at 108K, and this might
contribute to the observed spin state. Further data are required on
[Fe(III)(acen)]+ systems.
We have shown that there is demonstrable dependence of
the C2N2 backbone configuration on magnetic behaviour in
[Fe(III)(salX)]+ compounds, noted recently also in Fe(III)-sal2-
trien compounds,20,21 and the present work tends to suggest the
same for [Fe(III)(acen)]+ compounds. Interestingly, for compa-
rable [Fe(II)(O2N2)(L)2] iron(II) compounds of the Ja¨ger type,
differences in magnetic behaviour have also been linked to the mag-
nitude of the bond lengths in the five-membered ring comprising
the Fe(II) centre and C2N2 backbone of the Schiff base ligand.17,18
The recently reported SCO compound [FeL1b(py)2] was compared
with the HS [FeL2b(py)2] compound,17,18 the tetradentate ligand
L1b having an en-linked backbone while L2b a o-phen-linked
one. Comparison of bonds lengths within the five-membered
ring Fe-N-C-C-N for each of 1a-e to [Fe(III)(salen)(tmid)](ClO4)
in the HS state,23 and for 5a to [Fe(III)(acen)(bpp)](BPh4)
and Fe(III)(acen)(tmid)](BPh4) in the LS state25,26 and also
[Fe(III)(acen)(Him)2](BPh4),42 suggests that this interpretation
may not apply to [Fe(III)(Schiff-base)]+ compounds. Significant
bond length differences for compounds displaying different mag-
netic behaviour are not observed. It should be noted, however,
that for Fe-Naxial bond lengths there appears to be a considerable
difference between HS and SCO compounds; values are given for
1a-e, 3a and 5a in Table 2, while for [Fe(III)(salen)(tmid)](ClO4)
˚
in the HS form the average Fe-Naxial = 2.157(4) A.
Conclusions
This new series of 1-D chain Schiff base-containing materials
underscores a number of key points about how magnetic behaviour
correlates to ligand conformation. Magneto-structural character-
isation of new [Fe(III)(salX)(L)]+ compounds is supplied that is
supportive of the previously held notion that a spin transition
cannot occur when the C2N2 backbone of the equatorial ligand
adopts the meso conformation. For example, varying the length
and character of the bridging pyridyl ligand does not appear to
have any magnetic consequences for [Fe(III)(salen)(L)]+ systems
where the meso conformation has been adopted. Further to this,
158 | Dalton Trans., 2010, 39, 149–159
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
the compound 3a, in which salen has adopted the envelope
conformation, suggests that the total ligand field strength may be
too weak to induce a spin transition in [Fe(III)(salen)(L)]+ systems
when L is a pyridyl donor ligand. For [Fe(III)(salen)(L)]+ systems,
variation of the length and character of the bridging pyridyl ligand
does have structural consequences - the shorter and more flexible
ligands tend to form hydrogen bonded 1-D chain motifs in which
one of the axial positions of the Fe(III) coordination environment
is occupied by a methanol molecule. A series of new examples
of SCO 1-D chain materials of the type [Fe(III)(acen)(L)]+ are
presented, wherein axial ligand variation has resulted in a range of
magnetic behaviours. Comparison of the structural conformation
of the acen ligand in 5a to previously reported analogues is
suggestive of a correlation between structure and magnetism as
observed for [Fe(III)(salen)(L)]+ systems. Aims in future studies
include the expansion of the current [Fe(III)(acen)(L)]+ series
using a range of other N-donor ligands, and counter ion variation.
Also, compounds in which L is a three- or four-connecting
ligand may be investigated, in order to see the effect of intro-
ducing [Fe(III)(acen)(L)]+ SCO centres into higher dimensional
materials.
12 S. Hayami, Z.-Z. Gu, M. Shiro, Y. Einaga, A. Fujishima and O. Sato,
J. Am. Chem. Soc., 2000, 122, 7126.
13 S. Hayami, Z.-Z. Gu, H. Yoshiki, A. Fujishima and O. Sato, J. Am.
Chem. Soc., 2001, 123, 11644.
14 J. Tang, J. Sa´nchez Costa, S. Smulders, G. Molna´r, A. Bousseksou, S. J.
Teat, Y. Li, G. A. van Albada, P. Gamez and J. Reedijk, Inorg. Chem.,
2009, 48, 2128.
15 B.R. Mu¨ller, G. Leibeling and E.-G. Ja¨ger, Chem. Phys. Lett., 2000,
319, 368.
16 B. Weber, R. Tandon and D. Himsl, Z. Anorg. Allg. Chem., 2007, 633,
1159.
17 B. Weber and E.-G. Ja¨ger, Eur. J. Inorg. Chem., 2009, 465.
18 B. Weber, E. Kaps, J. Weigand, C. Carbonera, J.-F. Le´tard, K.
Achterhold and F. G. Parak, Inorg. Chem., 2008, 47, 487.
19 S. Dorbes, L. Valade, J. A. Real and C. Faulmann, Chem. Commun.,
2005, 69.
20 C. Faulmann, P. A. Szila´gyi, K. Jacob and J. Chahine, New J. Chem.,
2009, 33, 1268.
21 R. Pritchard, S. A. Barrett, C. A. Kilner and M. A. Halcrow, Dalton
Trans., 2008, 3159.
22 B. J. Kennedy, A. C. McGrath, K. S. Murray, B. W. Skelton and A. H.
White, Inorg. Chem., 1987, 26, 483.
23 M. Fukuya, M. Ohba, K. -I. Motoda, N. Matsumoto, H. Okawa and
Y. Maeda, J. Chem. Soc., Dalton Trans., 1993, 3277.
24 H. Ohshio, Y. Maeda and Y. Takashima, Y, Inorg. Chem., 1983, 22,
2684.
25 S. Imatomi, R. Kitashima, T. Hamamastu, M. Okeda, Y. Ogawa and
N. Matsumoto, Chem. Lett., 2006, 35, 502.
26 S. Imatomi, S. Hashimoto and N. Matsumoto, Eur. J. Inorg. Chem.,
2009, 721.
27 S. M. Neville, B. A. Leita, G. J. Halder, C. J. Kepert, B. Moubaraki,
J.-F. Le´tard and K. S. Murray, Chem.–Eur. J., 2008, 14, 10123.
28 SMART, SAINT, and XPREP: Area Detector control and data integra-
tion and reduction software, Bruker Analytical X-ray Instruments Inc.,
Madison, WI, 1995.
Acknowledgements
Mr. Jarry Abbas kindly contributed to preliminary syntheses. The
Australian Research Council provided a Discovery Grant for this
work.
29 G. M. Sheldrick, SADABS: Empirical Absorption Corrections Program
for Area Detector Data, University of Go¨ttingen, Germany, 1996.
30 G. M. Sheldrick, SHELXS-97: Program for crystal structure solution,
University of Go¨ttingen, Go¨ttingen, Germany, 1997.
31 L. J. Barbour, X-SEED, University of Stellenbosch, South Africa, 1999.
32 E. V. Brown and R. G. Granneman, J. Am. Chem. Soc., 1975, 97(3),
621.
33 A. J. Baer and D. H. Macartney, Inorg. Chem., 2000, 39, 1410.
34 N. Platzer, N. Goasdoue and R. Bonnaire, J. Organomet. Chem., 1978,
160, 455.
Notes and references
1 P. Gu¨tlich and H. A. Goodwin, Top. Curr. Chem., 2004, 233, 1.
2 J. -F. Le´tard, P. Guionneau and L. Goux-Capes, Top. Curr. Chem.,
2004, 235, 221.
3 T. Forestier, S. Mornet, N. Daro, T. Nishihara, S.-I Mouri, K. Tanaka,
O. Fouche´, E. Freysz and J.-F. Le´tard, Chem. Commun., 2008, 4327.
4 O. Kahn and C. J. Martinez, Science, 1998, 279, 44.
5 O. Kahn, C. J. Martinez, J. Kro¨ber, R. Claude and F. Grolie`re, Patent
EP0666561, 1995.
35 B. F. Hoskins, R. Robson and D. A. Slizys, J. Am. Chem. Soc., 1997,
119, 2952.
6 K. S. Murray, Eur. J. Inorg. Chem., 2008, 3101.
7 A. B. Gaspar, C. M. Munˇoz and J. A. Real, J. Mater. Chem., 2006, 16,
2522.
36 A. M. Comrie and J. B. Stenlake, J. Med. Chem., 1960, 2, 171.
37 M. Gerloch, J. Lewis, F. E. Mabbs and A. Richards, J. Chem. Soc. A,
1968, 112.
8 K. S. Murray and C. J. Kepert, Top. Curr. Chem., 2004, 233, 195.
9 G. J. Halder, K. W. Chapman, S. M. Neville, B. Moubaraki, K. S.
Murray, J. -F. Le´tard and C. J. Kepert, J. Am. Chem. Soc., 2008, 130,
17552.
38 K. S. Murray, Coord. Chem. Rev., 1974, 12, 1.
39 M. M. Bhadbhade and D. Srinivas, Polyhedron, 1998, 17, 2699.
40 S. Hayami, K. Inoue and Y. Maeda, Mol. Cryst. Liq. Cryst., 1999, 335,
573.
10 B. A Leita, S. M. Neville, G. J. Halder, B. Moubaraki, C. J. Kepert, J.-F.
Le´tard and K. S. Murray, Inorg. Chem., 2007, 46, 8784.
11 P. J. van Koningsbruggen, Y. Maeda and H. Oshio, Top. Curr. Chem.,
2004, 233, 259.
41 S. Imatomi, T. Sato, T. Hamamatsu, R. Kitashima and N. Matsumoto,
Bull. Chem. Soc. Jpn., 2007, 80, 2375.
42 Y. Nishida, K. Kino and S. Kida, J. Chem. Soc., Dalton Trans., 1987,
1157.
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 149–159 | 159
©