Iron(II) complexes of two amine/imine N5 chelate ligands containing a 1,4-diazepane core - To crossover or not to crossover

Article abstract of DOI:10.1002/ejic.201201149

Two new chelate ligands, 6-methyl-6-(pyridin-2-yl)-1,4-bis- [(pyridin-2-yl)methyl]-1,4-diazepane (4a) and 6-methyl-6- (pyridin-2-yl)-1,4- bis[2-(pyridin-2-yl)ethyl]-1,4-diazepane (4b), were synthesized from pyridine-derived precursors in three-step procedures. Both ligands have N ₅ donor sets consisting of two tertiary amine and three pyridyl N atoms. Complexation with FeCl₂ or FeBr₂ in MeOH followed by anion exchange with (nBu₄N)PF₆ gave the complexes [Fe(4a)X]PF₆ and [Fe(4b)X]PF₆ (X = Cl, Br) in moderate-to-good yields. The coordination geometry around the iron(II) centre, as determined by single-crystal X-ray diffraction, is strongly dis- torted octahedral for ligand 4a and more regular for ligand 4b. Magnetic susceptibility measurements show the complexes to contain high-spin iron(II) over the whole range of temperatures (2 < T < 300 K). DFT calculations for the complexes of ligands 4a and 4b reproduce the high-spin ground state and suggest that exchange of the halide ligand for ligands that exhibit some degree of π-type interaction could induce SCO behaviour. Also, calculations of the zero-field splitting (ZFS) parameters of these complexes rationalize the observed sign change on the basis of different degrees of structural distortion imparted by ligands 4a and 4b.

Full text of DOI:10.1002/ejic.201201149

FULL PAPER
DOI:10.1002/ejic.201201149
Iron(II) Complexes of Two Amine/Imine N5 Chelate
Ligands Containing a 1,4-Diazepane Core – To Crossover
or Not To Crossover
CLUSTER
ISSUE
Marc Schmidt,[a] Dennis Wiedemann,[a] Boujemaa Moubaraki,[b]
Nicholas F. Chilton,[b] Keith S. Murray,[b] Kuduva R. Vignesh,[c]
Gopalan Rajaraman,[c] and Andreas Grohmann*[a]
Keywords: Iron / N ligands / Chelates / Magnetic properties / Density functional calculations
Two new chelate ligands, 6-methyl-6-(pyridin-2-yl)-1,4-bis-
[(pyridin-2-yl)methyl]-1,4-diazepane (4a) and 6-methyl-6-
(pyridin-2-yl)-1,4-bis[2-(pyridin-2-yl)ethyl]-1,4-diazepane
(4b), were synthesized from pyridine-derived precursors in
torted octahedral for ligand 4a and more regular for ligand
4b. Magnetic susceptibility measurements show the com-
plexes to contain high-spin iron(II) over the whole range of
temperatures (2 Ͻ T Ͻ 300 K). DFT calculations for the com-
three-step procedures. Both ligands have N5 donor sets con- plexes of ligands 4a and 4b reproduce the high-spin ground
sisting of two tertiary amine and three pyridyl N atoms. Com- state and suggest that exchange of the halide ligand for li-
plexation with FeCl2 or FeBr2 in MeOH followed by anion
exchange with (nBu4N)PF6 gave the complexes [Fe(4a)X]PF6
and [Fe(4b)X]PF6 (X = Cl, Br) in moderate-to-good yields.
The coordination geometry around the iron(II) centre, as de-
termined by single-crystal X-ray diffraction, is strongly dis-
gands that exhibit some degree of π-type interaction could
induce SCO behaviour. Also, calculations of the zero-field
splitting (ZFS) parameters of these complexes rationalize the
observed sign change on the basis of different degrees of
structural distortion imparted by ligands 4a and 4b.
Introduction
redox chemistry.[7] Specifically, we attempted to tune the re-
Iron, in its coordination compounds, is fascinating for at dox potential of iron coordination modules that may be
least two reasons: the ways it interacts with dioxygen and thus obtained by varying the N(amine)/N(imine) ratio of
related species,[1,2] and the potential for spin crossover donor atoms in series of such ligands to modify the σ-do-
(SCO), especially in the +2 oxidation state.[3,4] Both aspects nor/π-acceptor properties of the ligand or by varying the
are combined in haemoglobin, in which O2 coordination/ constraints the ligand imposes on the metal ion, which re-
decoordination induces a change of spin state in the iron flects the concept of the entatic state.[8] Based on these
ion. Haem and non-haem iron complexes are crucial for ideas, we recently introduced a ligand containing a hexahy-
dioxygen processing in a multitude of metalloenzymes to dropyrimidine core and gave details of its coordination
bring about substrate oxygenation.[1,2] In this context, we chemistry with first-row transition metals, including iron.[9]
are interested in polypodal N5 ligands of high symmetry We extend this chemistry in the present contribution and
(C2v or Cs), which we have termed “square-pyramidal coor- report on the new ligands, 6-methyl-6-(pyridin-2-yl)-1,4-bis-
dination caps”.[5] In (quasi-)octahedral complexes, such li- [(pyridin-2-yl)methyl]-1,4-diazepane (4a) and 6-methyl-6-
gands leave one coordination site free for the binding of a (pyridin-2-yl)-1,4-bis[2-(pyridin-2-yl)ethyl]-1,4-diazepane
small monodentate ligand or substrate. Depending on the (4b), and their chlorido- and bromidoiron(II) complexes.
nature of this ligand, iron(II) complexes undergo tempera- The ligands vary in the number of methylene groups (one
ture-dependent spin crossover[6] or substrate-transforming vs. two) joining the tertiary amine N atoms to their pyridyl
substituents and are, therefore, expected to induce different
degrees of coordinative strain upon complexation. The
[a] Institut für Chemie, Technische Universität Berlin,
characterization includes variable-temperature magnetic
data for both types of complex (FeII/4a, FeII/4b), as well
as a theoretical study of different spin-state structures and
energies (FeII/4b). This study has also been extended to the
hypothetical (thiocyanato-κN)iron(II) complexes of both li-
gands, which are not yet synthetically accessible, to assess
their expected temperature-dependent spin behaviour.
Straße des 17. Juni 135, 10623 Berlin, Germany
Homepage: http://www.bioanorganik.tu-berlin.de/
[b] School of Chemistry, Monash University, Building 23,
Clayton, Victoria 3800, Australia
[c] Department of Chemistry, Indian Institute of Technology
Bombay,
Powai, Mumbai 400076, India
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201201149.
Eur. J. Inorg. Chem. 2013, 958–967
958
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
of the reaction is slowed drastically). The new solvent sys-
tem has major advantages: (a) increased reaction rate and
yield of 2 (Ͼ 80%) when compared with the reaction in
THF (several weeks, after which Ͻ 30% of 1 had reacted);
(b) the ditosylate 1 is practically insoluble in H2O, which
minimizes the formation of 2a; (c) simple workup to obtain
2 sufficiently pure for the subsequent synthetic step (see
Exp. Sect.). The tosyl groups were cleaved by dissolving 2
in concentrated H2SO4 and stirring the reaction mixture at
130 °C for 14 h.[12] 6-Methyl-6-(pyridin-2-yl)-1,4-diazepane
(3) was then used without further purification and reacted
with 2-(chloromethyl)pyridine and potassium carbonate in
acetonitrile[13] or 2-vinylpyridine and acetic acid in
MeOH[14] to give the N5 chelate ligands 4a and 4b, respec-
tively. The crude ligands are obtained as yellow-red oils,
which can be purified by repeated washing with hexane.
Compounds 4a and 4b are yellow oils (correct elemental
analyses), and the yields are 63% and 42%, respectively
(over three steps, based on 1).
Results and Discussion
Ligand Synthesis
The synthesis of ligands 4a and 4b proceeds in three steps
(Scheme 1) with ditosylate 1 as the starting material.[10] The
generation of 3 from 1 requires the formal introduction of
1,2-ethylenediamine (en) by nucleophilic substitution.
When this is attempted directly, the reaction is unspecific
and proceeds, at least in part, with the intermolecular cou-
pling of 1. This may be avoided by protecting the NH2 func-
tions in 1,2-ethylenediamine with p-toluenesulfonyl chloride
(TsCl).[11] Upon deprotonation of the tosylamine groups
with NaH or NaOH in tetrahydrofuran (THF), 1 is added
as a solid and the reaction mixture is subsequently heated
to reflux for several days. The suspension gradually turns
brown, and ESI mass spectrometry is compatible with a
mixture of 2 and 2a (2a is the monosubstituted side prod-
uct.)
Iron(II) Complexes
The formation of iron(II) complexes of ligands 4a and
4b by using FeCl2 or FeBr2 in methanol is straightforward.
Overall, as judged by regular probing of the mixture with The bromide salt [Fe(4b)Br]Br (5) was precipitated from
ESI-MS, the reaction of 1 is sluggish, most likely because solution as an amorphous solid by the addition of diethyl
of the limited solubility of the starting materials in THF. ether. Solid (nBu4N)PF6 was added to the MeOH solutions
Therefore, we decided to suspend the ethylenediamine dito- to precipitate polycrystalline material (yellow in all cases),
sylate in water and react it with solid NaOH at 100 °C. In and the chlorido and bromido complexes [Fe(4a)Cl]PF6 (6),
the course of one hour, this produces a colourless solution, [Fe(4a)Br]PF6 (7), [Fe(4b)Cl]PF6 (8) and [Fe(4b)Br]PF6 (9)
in which we assume the RNHTs groups to be fully depro- were thus obtained (cf. Exp. Sect.). Single crystals were
tonated. The boiling was allowed to subside, and a warm grown by recrystallization from acetonitrile. X-ray analyses
solution of 1 in toluene was added in one portion to pro- reveal isomorphous pairs (6 and 7, 8 and 9). All complexes
duce a two-phase system, which was then stirred vigorously are acetonitrile solvates in their single crystals and contain
and heated to reflux for two days. ESI-MS of the organic one solvent molecule (6/7) or half a solvent molecule (8/9).
phase indicated complete disappearance of 1 (if the two- Representative cation structures are shown in Figure 1, and
phase system is not stirred with sufficient vigour, the rate the crystallographic data is given in Table 6.
Scheme 1. Synthesis of the N5 ligands 4a and 4b: i) NaOH, water/toluene (50:50), reflux, 2 d; ii) H2SO4, 130 °C, 1 d; iii) MeCN, K2CO3,
55 °C, 3 d; iv) MeOH, acetic acid, reflux, 2 d.
Eur. J. Inorg. Chem. 2013, 958–967
959
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
one and two for 8/9, as is expected for unstrained HS
iron(II) compounds (Table 1). The average Fe–N bond
length does not vary strongly between the complexes, and
the deformation parameter Σ, which is a measure of the
deviation of all cis angles from 90°, reflects this behaviour
exactly; for the heavily distorted complexes 6 and 7, it has
high values of ca. 145°.
Figure 1. Structures of the cations in 6 (left) and 8 (right). ORTEP
representations with ellipsoids at the 50% probability level; hydro-
gen atoms, hexafluorophosphate counterions and acetonitrile mo-
lecules omitted for clarity.
Magnetic Susceptibilities
Plots of χMT vs. T for the mononuclear iron(II) com-
plexes 6 and 7 are similar in shape and that for 6 is shown
in Figure 2 (top; see Supporting Information for plots of
7). The χMT values remain independent of temperature be-
tween 300 and ca. 50 K with values of 3.45 (μeff = 5.25μB)
and 3.5 cm3 mol–1 K (μeff = 5.29μB) for 6 and 7, respectively,
which are typical of high-spin distorted octahedral FeII
compounds. A rapid decrease in χMT then occurs and it
The bond lengths in both types of complex are unexcep-
tional, and the average Fe–N bond length of ca. 2.2 Å indi-
cates high-spin iron(II) in all cases at 150(1)
K
(Table 1).[15,16] The halide ligand X1 (X = Cl or Br) is trans
to a tertiary amine N atom (N1) in all structures, although,
at least for complexes of ligand 4b, a trans arrangement
N10–Fe1–X1 would not appear precluded by structural
constraints. Indeed, structural data to be published else-
where show that ligand exchange [(chlorideǞtriflate (OTf)]
leads to a trans arrangement (N10–Fe1–OTf) in the triflato-
iron(II) complex of ligand 4b. The coordination environ-
ment in 6 and 7, in which ligand 4a provides a more con-
stricted N5 donor set, is severely distorted from (quasi-)oc-
tahedral, whereas it is more regular in 8 and 9, in which the
ligand tethers are longer (ligand 4b). We have evaluated this
distortion by using the “Continuous Symmetry Measures”
proposed by Zabrodsky, Peleg and Avnir, which “quantify
the minimal distance movement that the points of an object
have to undergo in order to be transformed into a shape of
the desired symmetry.”[17,18] The parameter S(Oh), which
ideally is zero, is greater than five for 6/7 and lies between
reaches 2.67 (μeff = 4.62μB) and 1.63 cm3 mol–1 K (μeff
=
3.61μB), respectively, which is indicative of zero-field split-
ting of ground 5A1g states. The magnetization isotherms, M
vs. H, at temperatures 2 to 20 K, are shown in Figure 2
(bottom) and do not exhibit saturation even at the lowest
temperature (2 K) and highest field (5 T), at which the M
values of 3.04 and 2.75 NAμB are much less than the S = 2
value (4.90μB) because of zero-field splitting caused by
spin–orbit coupling effects. The susceptibility and magne-
tization data for 6 and 7 were simultaneously fitted to the
1
2
2
ˆ
ˆ
ˆ
ˆ
axial spin Hamiltonian H = D(Sz – S ) + μBgS · B with
3
S = 2 by use of program PHI[20] (Figure 2 and Figure S1 in
the Supporting Information). The best-fit parameters were
determined to be as given in Table 2.
The plots of χMT vs. T for 8 and 9 are again similar in
shape. Figure 3 (top) shows the plots for 8 (see Supporting
Information for plots of 9). Similar to the behaviour of
complexes 6 and 7, the χMT values remain independent of
temperature between 300 and ca. 50 K and are
3.25 cm3 mol–1 K (μeff = 5.10 μB) as expected for high-spin
distorted octahedral FeII. There is a subsequent rapid de-
Table 1. Coordinative bond lengths [Å], cis angles [°] and distortion
parameters Σ and S(Oh) for complexes 6–9.
6 (X = Cl)
7 (X = Br) 8 (X = Cl) 9 (X = Br)
Fe1–N1
Fe1–N5
Fe1–N10
Fe1–N20
Fe1–N30
Fe1–X1
dFeN
N1–Fe1–N5
N1–Fe1–N10
N1–Fe1–N20
N1–Fe1–N30
X1–Fe1–N5
X1–Fe1–N10
X1–Fe1–N20
X1–Fe1–N30
N5–Fe1–N10
N5–Fe1–N30
2.3373(17)
2.1878(17)
2.1625(18)
2.1317(18)
2.2633(18)
2.3738(6)
2.217(4)
71.15(6)
79.03(6)
75.81(6)
107.16(6)
114.12(5)
88.58(5)
103.33(5)
85.67(5)
2.311(3)
2.187(3)
2.159(3)
2.134(3)
2.259(3)
2.5397(6)
2.210(7)
71.70(11)
79.74(11)
76.30(11)
2.298(2)
2.238(2)
2.206(2)
2.211(2)
2.250(2)
2.4135(7)
2.241(4)
71.99(8)
84.62(8)
88.51(8)
2.281(3)
2.229(3)
2.204(3)
2.209(3)
2.243(3)
2.5959(6)
2.233(7)
72.27(11)
84.91(11)
89.81(11)
96.24(11)
102.98(8)
86.00(9)
96.09(8)
92.34(8)
94.37(12)
81.31(11)
92.46(13)
92.50(12)
72.7(4)
crease in χMT and it reaches 1.2 cm3 mol–1 K (μeff
=
3.10 μB), which again indicates zero-field splitting of ground
5A1g states. The magnetization isotherms, M vs. H at tem-
peratures 2 to 20 K are shown in Figure 3 (bottom). As for
6 and 7, saturation is not reached even at the lowest tem-
perature (2 K) and highest field (5 T), at which the M value
of 2.7 NAμB is much less than the S = 2 value because of
zero-field splitting caused by spin–orbit coupling effects.
The best-fit parameters (for the susceptibility and magne-
tization data; simultaneous fit to the same axial spin Hamil-
tonian as for 6 and 7; Figures 3 and S2) were determined
to be as given in Table 2.
Although further applied-field Mössbauer and EPR
spectroscopy (parallel mode)[21] would be required to con-
firm the change in sign of D between the complexes of li-
gands 4a and 4b, the simultaneous fitting of susceptibilities
and magnetizations involved a wide exploration of param-
107.37(11) 95.68(7)
112.17(8)
88.47(8)
104.06(9)
84.55(8)
97.24(11)
72.96(11)
99.79(11)
95.18(11)
142.1(4)
5.21
104.37(6)
86.58(6)
96.19(6)
92.69(5)
93.64(8)
81.46(7)
92.79(9)
92.49(8)
74.7(3)
96.90(7)
73.01(6)
N10–Fe1–N20 99.61(7)
N20–Fe1–N30 95.19(7)
Σ/°[a]
143.1(2)
5.09
S(Oh)[b]
1.55
1.81
12
[a] Σ =
|90 – φi|; φi: angles La–Fe–Lb between cis-positioned
͚
i = 1
Fe–L bonds.[19] [b] See text.
Eur. J. Inorg. Chem. 2013, 958–967
960
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 2. Plot of χMT vs. T for 6 in a field of 0.5 T (top). Magne-
tization isotherms for 6 at temperatures 2, 3, 4, 5.5, 10 and 20 K
(bottom). The red lines are the best fits calculated by using the
parameters given in the text.
Figure 3. Plot of χMT vs. T for 8 in a field of 0.5 T (top). Magne-
tization isotherms for 8 at temperatures 2, 3, 4, 5.5, 10 and 20 K
(bottom). The red lines are the best fits calculated by using the
parameters given in the text.
Table 2. Best-fit parameters for the susceptibility and magnetiza-
tion data obtained for complexes 6–9.
which theoretical calculations of the influence of ligand
distortions upon D and E (rhombic) values have been
given.[24]
g
D [cm–1]
[Fe(4a)Cl]PF6 (6)
[Fe(4a)Br]PF6 (7)
[Fe(4b)Cl]PF6 (8)
[Fe(4b)Br]PF6 (9)
2.14
2.17
2.09
2.14
–4.50
–7.48
7.81
Theoretical Studies
13.2
Theoretical studies have been undertaken to unfold the
spin-crossover features, if any, observed in the monomeric
eter space and we are confident of the difference in sign in complexes 6–9. We acknowledge that DFT calculations, per
D. The origin of this sign change, in a parameter that relates se, relate to species in the gaseous phase and thus do not
to second order spin–orbit coupling and to distortions from include crystalline/solvent/packing effects; nevertheless,
octahedral symmetry, presumably relates to the distortions they are useful for probing spin states and spin transitions
discussed in the crystallographic section and expressed as and also to underpin the electronic structure of the complex
S(Oh) and Σ, which are markedly different for the two li- of interest. Here, calculations have been performed with the
gand systems (Table 1). Interestingly, whereas the metal–li- hybrid functional B3LYP, which tends to favour the HS
gand bond lengths in d4 MnIII Jahn–Teller distorted octahe- state. Exchange-correlation functionals incorporating less
dral species can be related to the sign of D (negative for (Ͻ 20%) exact exchange, or double-hybrid functionals in-
axial elongation),[22] this is not evident in the present struc- corporating correlation from MP2 methods, have been
tures. The presence of the “hetero” ligand, Cl or Br, might found to be superior to B3LYP calculations.[25–28] Despite
play a role here. An experiment worth pursuing on the these known issues, B3LYP in general gives structures that
negative D complexes 6 and 7 is to measure their in-phase are in good agreement with X-ray structure determinations
and out-of-phase AC susceptibilities at low temperature to and is the functional of choice for the computation of spec-
check for slow relaxation of magnetization [i.e., single mole- troscopic parameters.
cule magnetism, (SMM)], a feature recently observed in dis-
Different spin-state structures and energies were com-
torted FeII compounds with negative D values[23] and for puted to understand the spin-crossover features, g and zero-
Eur. J. Inorg. Chem. 2013, 958–967
961
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 3. Selected structural parameters computed by using the B3LYP functional (see Figure 4 for labels).
Structural parameter
6
7
8
9
X-ray HS
IS
LS
X-ray HS
IS
LS
X-ray HS
IS
LS
X-ray HS
IS
LS
Fe–N1
Fe–N2
Fe–N3
Fe–N4
2.188 2.233 2.157 2.097 2.187 2.232 2.157 2.101 2.299 2.344 2.384 2.127 2.281 2.348 2.396 2.129
2.338 2.359 2.435 2.064 2.311 2.355 2.434 2.069 2.238 2.303 2.204 2.109 2.229 2.295 2.195 2.115
2.163 2.207 2.010 2.024 2.159 2.211 2.012 2.028 2.250 2.292 2.058 2.088 2.205 2.289 2.061 2.053
2.263 2.342 2.060 2.086 2.259 2.346 2.062 2.088 2.206 2.224 2.026 2.046 2.243 2.229 2.030 2.092
2.132 2.160 2.118 1.984 2.139 2.161 2.113 1.994 2.212 2.247 2.206 2.064 2.209 2.251 2.194 2.070
Fe–N5
Fe–Cl
2.374 2.376 2.423 2.420
–
–
–
–
2.413 2.425 2.492 2.450
–
–
–
–
Fe–Br
–
–
–
–
2.540 2.553 2.613 2.611
–
–
–
–
2.596 2.586 2.670 2.616
Cl–Fe–N2
Br–Fe–N2
N5–Fe–N1
N3–Fe–N4
167.2 170.6 171.7 173.4
–
–
–
–
170.2 171.2 172.6 169.9
–
–
–
–
–
–
–
–
168.1 171.2 170.6 172.6
–
–
–
–
169.4 170.3 170.9 169.3
139.2 139.6 142.0 158.1 140.2 139.3 142.0 157.4 158.7 159.1 160.8 170.6 160.1 158.9 160.9 170.1
165.0 161.9 168.6 165.3 164.6 163.8 169.2 166.0 174.7 173.9 176.4 172.9 174.9 175.1 177.2 173.5
field splitting (zfs; D) parameters to complement the experi- and this is found to be 57.7, 59.6, 57.6 and 60.4 Jmol–1 K–1,
mental data. The computed structural parameters for 6–9 and the enthalpy change is computed to be –58.0, –59.0,
along with X-ray structural parameters are summarized in –55.3 and –57.4 kJmol–1 for 6, 7, 8 and 9, respectively.
Table 3. The structural parameters of the optimized struc- These estimated energy gaps are consistent with the experi-
tures are in general in good agreement with the X-ray struc- mental reports for FeII spin crossover compounds.[29–31]
tural parameters, and the short and long bonds observed
As spin crossover depends on enthalpy and entropy con-
within the distorted octahedron are nicely reproduced by tributions, spin-state splitting and the corresponding free-
theory. The high-spin (S = 2) structure in particular re- energy change essentially determines the temperature at
sembles the X-ray structure very closely.
which the SCO transition is observed. The plot and energies
The optimized structures (ground-state structures) of 6 of d-based orbitals of 6 and 8 are shown in Figure 5. The
and 9 along with computed energies for 6–9 are shown in energy gap between the t2g-like and eg-like orbitals for 8
Figure 4. For all four complexes, the high-spin (HS) state is and 9 is nearly 128 kJmol–1, whereas for 6 and 7 it is nearly
found to be the ground state; this is consistent with the 300 kJmol–1. This splitting energy clearly illustrates that li-
experimental observations. In addition, the low-spin (LS) gand-induced structural changes significantly alter the en-
states lie at 54.3, 54.6, 50.7 and 52.3 kJmol–1 for 6, 7, 8 and ergy levels and hence the properties of the complexes (see
9, respectively. The intermediate-spin (IS) states lie much below). The generation of the low-spin complex demands
higher in energy. As structural optimizations have been per- pairing of electrons in the dxz and dyz orbitals and both
formed, the entropy change between the ground state HS orbitals are strongly influenced by halogenido coordination
and LS spin states, ΔS = SHS – SLS, has also been estimated (Figure 5). This supports the notion that the SCO feature
can be tuned by axial ligation and ligands exhibiting mod-
Figure 5. Eigenvalue plot computed for 6 and 8. Here, α and β
represent spin-up and spin-down configurations. The energies are
Figure 4. (a) Energy level gaps [kJmol–1] computed for different
electronic configurations in 6–9. Optimized ground state structure
of (b) 6 and (c) 9. (d) Spin density plot for 8.
scaled to the lowest spin-up orbitals in each case. For orbital label-
ling, the Fe–Cl bond direction has been defined as the z axis and
one of the equatorial Fe–N(imine) directions as the x axis.
Eur. J. Inorg. Chem. 2013, 958–967
962
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
erate-to-weak π-type interactions are ideal for the observa- Table 5. Different contributions to computed ZFS parameters for
6–9.
tion of SCO features. In this context, we have modelled ax-
ial thiocyanato ligation in place of Br ligation in 9.[32] As
Complex/parameter
6
7
8
9
expected, the HS–LS gap is reduced to 38.6 kJmol–1; how-
ever, this gap is likely still too large for the observation of
Dsoc
Dss
–1.831
–0.443
–0.558
–0.468
0.056
–0.312
–0.309
0.007
2.354
0.231
0.654
0.898
0.752
0.050
2.629
0.291
0.746
0.991
0.842
0.05
SCO (see Figure S3 for the optimized structure and com- SOMO–VMO[a] αǞα –0.225
puted energy gap). A HS–LS gap of –25 to 0 kJmol–1 is
DOMO–SOMO βǞβ –0.480
SOMO–SOMO αǞβ
DOMO–VMO βǞα
–1.131
0.005
suggested as an ideal gap for the observation of SCO by
Neese et al.[26] This situation should be compared with the
results obtained for a related N5 ligand L, 2,2Ј-(pyridine-
2,6-diyl)bis(2-methylpropane-1,3-diamine).[6] Whereas the
bromido complex [FeIIBrL]Br is high spin (four unpaired
electrons) at room temperature, the carbonyl complex
[FeII(CO)L]Br2 is low spin and diamagnetic,[7] and
[FeIIL(NMe-imidazole)](OTf)2 shows SCO above 300 K.[6]
In this context, for mononuclear iron(II) complexes to be
bistable (i.e., show temperature-dependent SCO behaviour)
the ligand sphere usually consists of six N donors (mainly
N heterocycles),[33] but N4O2 donor sets are also known.[34]
DFT-computed isotropic g values for 6–9 are given in
Table 4 along with computed ZFS parameters. The com-
puted g tensors are nearly isotropic. Compared to the ex-
perimental values, the g tensors are slightly underestimated.
Although the computed ZFS parameters are severely
underestimated compared to the experimental values, the
calculations reproduce the negative and positive sign of D
observed with ligands 4a and 4b, respectively. The fact that
the ZFS parameter is underestimated by DFT is due to ne-
glect of contributions from the triplet and singlet states; this
is well-documented in the literature.[35,36] The different con-
tributions to the net D parameter are summarized in
Table 5. It is clear that Dsoc (spin–orbit) makes a significant
contribution and its sign changes as we move from com-
plexes 6/7 to 8/9. Similar behaviour is noted also for DSS
(spin–spin) contributions. The largest contributions to Dsoc
arise from DOMO (doubly-occupied molecular orbital) to
SOMO (singly-occupied MO) βǞβ excitations and SOMO
to SOMO αǞβ excitations. The sign change occurs be-
[a] VMO: virtual molecular orbital.
methylene groups in 4a), the bonds in the equatorial plane
are longer overall, and this causes less repulsion for the d
orbitals than in the structures with ligand 4a (complexes 6
and 7). This change in ligand-field strength alters the or-
bital energies, and this in turn leads to different transition
energies, and hence a sign change for the D parameter.
The spin-density plot for 8 is shown in part d of Figure 4.
A spin-density value of 3.753 has been found on the FeII
atom, which indicates a slight delocalization of the unpaired
spins to other coordinated atoms. Each of the coordinated
nitrogen atoms have spin densities of ca. 0.02–0.03, whereas
the chlorine atom has a much larger value of 0.09. For the
iron ion in 6, a spin density of 3.714 was detected and this
is significantly less compared to that of 8. As the average
metal-to-ligand bond length is shorter in 6 (Table 3), greater
delocalization takes places compared to that in 8 and 9. The
spin density distribution nicely reflects the orbital analysis
described earlier. Complex 9 shows a similar spin distribu-
tion pattern, and the bromine atom gains a spin density of
ca. 0.1 from the FeII centre.
Conclusion
The two pentadentate nitrogen ligands presented in this
cause of the difference in the energetics of the d orbitals study, 4a and 4b, together with a chlorido or bromido li-
between the two ligands sets. The intricate differences in the gand, give iron(II) complexes of the type [FeLX]+, the coor-
d-orbital energies between 6 and 8 are shown in Figure 5. dination geometries of which differ markedly in their de-
Most importantly, the dxy orbital energy is significantly grees of distortion from octahedral. The N5 donor set of
high in complex 8; this leads to small dxy–dx2–y2 (αǞβ) ligand 4a, which has the two “lateral” pyridyl substituents
transition energies and this, in turn, leads to a sign change each tethered to the ligand core by a single methylene
between 6 and 8 for the ZFS parameter. Notably, for a group, is stereochemically more constricted than in the case
Jahn–Teller compressed {MnIIICuII} complex, such orbital of 4b, in which the tethers each contain two methylene
mixing leads to a positive ZFS parameter.[37]
groups. Magnetic susceptibility measurements show all
complexes to be paramagnetic in the temperature range 2
Ͻ T Ͻ 300 K and to contain high-spin iron(II). DFT calcu-
lations have been performed on 6–9 to elucidate their elec-
tronic structures and to analyze the absence of SCO in this
type of complex. Calculations reveal that a stronger ligand
field exerted by the axial halogen ligand leads to a large d
orbital splitting and a greater HS–LS gap. Further, this
work nicely illustrates how a small modification in the che-
Table 4. B3LYP-computed g, D and E/D values for 6–9.
Complex
D [cm–1]
E/D
g (isotropic)
[Fe(4a)Cl]PF6 (6)
[Fe(4a)Br]PF6 (7)
[Fe(4b)Cl]PF6 (8)
[Fe(4b)Br]PF6 (9)
–2.280
–1.026
2.587
0.286
0.066
0.315
0.222
2.006
2.006
2.046
2.053
2.923
Apparently, in the structures with ligand 4b (8 and 9), late ligand can alter the sign of the ZFS parameter. The
which has the two “lateral” pyridyl substituents each teth- origin of the change in the sign of ZFS is rationalized by
ered to the ligand core by an ethylene group (as opposed to using computed results.
Eur. J. Inorg. Chem. 2013, 958–967
963
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Magnetic susceptibility data were collected with a Quantum Design
MPMS 5 superconducting quantum interference device (SQUID)
magnetometer under an applied field of 1 T. The samples were
placed in a gel capsule, which was secured within a plastic straw
that was attached to the sample rod. Care was taken to allow ade-
quate thermal equilibration times at each temperature point. To
achieve this, the change between temperatures (either cooling or
heating) was at a rate of 10 K per minute, and the temperature was
stabilized at particular temperatures between 2 and 70 K for 2 min
and between 70 and 400 K for 5 min. The stabilization times were
varied to find these aforementioned optimum times such that the
magnetizations did not change. To avoid crystallite torquing for
these potentially anisotropic high-spin FeII compounds, the sam-
ples were contained in Vaseline gel.[22]
Experimental Section
General: Unless noted otherwise, all reactions were carried out in
dry solvents under dry dinitrogen by using standard Schlenk tech-
niques. Basic chemicals were purchased from Aldrich and used
without further purification. IR spectra were measured as KBr
disks. Spectroscopic data were obtained with the following instru-
ments. IR spectroscopy: Nicolet MagnaSystem 750. NMR spec-
troscopy: Bruker ARX 200 and ARX 400. Elemental analyses were
carried out with a Thermo Finnigan Flash EA 1112 series analyzer.
X-ray Crystallography: Data collection was performed at 150(1) K
with an Oxford Diffraction Xcalibur S instrument equipped with
a Sapphire3 CCD detector and graphite-monochromated Mo-Kα
radiation with λ = 0.71073 Å from an Enhance X-ray source
(Table 6). Frames were integrated by using the Scale3 Abspack al-
gorithm.[38] The structures were solved by direct methods with the
program SHELXS-97[39] (7–9) or by charge-flipping methods with
Superflip[40] (6). Subsequent full-matrix least-squares refinement of
Computational Details: All calculations were performed by using
the hybrid B3LYP[43] functional with the Ahlrichs triple-ζ basis
set[44,45] as implemented in the Gaussian 09 suite of programs.[46]
Fo data was carried out by using SHELXL-97.[39] All non-hydro-
2
Calculations of the g tensors and ZFS values of complexes was
performed by using the Orca programme suite. The calculations for
D tensors were based on the X-ray structures. The DFT calcula-
tions were carried out at hybrid B3LYP levels of theory. We have
employed the unrestricted Kohn–Sham formalism (UKS) for esti-
mating DSS. The calculations employed the resolution of identity
(RI-J) algorithm for the computation of the Coulomb terms. In
our DFT calculations, the spin–orbit coupling operators are repre-
sented by an effective one electron with the spin–orbit mean field
gen atoms were refined anisotropically. Hydrogen atoms were lo-
cated on Fourier difference maps and refined by using a riding
model. The acetonitrile moieties in 8·½MeCN and 9·½MeCN are
disordered over centres of inversion. In the former compound, it
had to be treated with Platon/Squeeze,[41] whereas it could be re-
fined with an occupancy of ½ on a single position in the latter.
Molecular graphics were created with the software package OR-
TEP-3 for Windows.[42]
CCDC-902540 (for 6·MeCN), -902539 (for 7·MeCN), -902542 (for (SOMF) method, which is implemented in Orca.[47] We have used
8·½ MeCN) and -902541 (for 9·½ MeCN) contain the supplemen-
tary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via http://www.ccdc.cam.ac.uk/data_request/cif.
the coupled perturbed SOC approach (CP) rather than the Peder-
son–Khanna (PK) approach to evaluate DSOC. The spin–spin con-
tribution (DSS) was estimated by using the unrestricted natural
orbital (UNO) approach rather than the UKS approach. The spin–
Table 6. Crystallographic data for 6–9.
6·MeCN
7·MeCN
8·½ MeCN
9·½ MeCN
Empirical formula
M [gmol–1]
Crystal dimensions [mm]
Crystal description
Crystal system
Space group
a [Å]
b [Å]
c [Å]
α [°]
β [°]
C25H30ClF6FeN6P
650.82
C25H30BrF6FeN6P
695.28
C26H32.5ClF6FeN5.5
658.34
P
C26H32.5BrF6FeN5.5
702.80
P
0.05ϫ0.07ϫ0.20
pale yellow prism
monoclinic
P21/c (No. 14)
16.8407(5)
8.3178(2)
24.6954(9)
90
125.427(2)
90
2818.80(15)
4
0.06ϫ0.14ϫ0.17
pale yellow plate
monoclinic
P21/c (No. 14)
16.9098(7)
8.4160(3)
24.5552(12)
90
125.173(3)
90
2856.5(2)
4
0.05ϫ0.14ϫ0.33
pale yellow plate
triclinic
0.12ϫ0.15ϫ0.16
pale yellow spar
triclinic
¯
¯
P1 (No. 2)
P1 (No. 2)
8.4201(4)
12.6853(7)
13.7091(8)
92.301(5)
90.059(4)
96.401(4)
1453.97(14)
2
8.5704(7)
12.7299(10)
13.5549(11)
87.977(6)
89.472(7)
83.927(7)
1469.6(2)
2
γ [°]
V [Å3]
Z
ρcalc [gcm–3]
1.534
0.754
1.617
2.049
1.504
0.731
1.588
1.992
μ [mm–1]
F(000)
1336
1408
678
714
Tmin/Tmax
θmin/θmax [°]
Measured reflections
0.82089/1.00000
3.31/26.00
21417
0.83308/1.00000
3.37/26.00
12824
0.80353/1.00000
3.38/26.00
11071
0.78187/1.00000
3.35/26.00
11650
Independent reflections/Rint 5518/0.0403
5610/0.0500
4306/0.0742
5610/0/363
0.0480/0.0709
0.0843/0.0952
1.034
5713/0.0278
4781/0.0424
5711/0/353
0.0417/0.0562
0.1042/0.1092
1.057
5771/0.0398
4853/0.0553
5771/0/381
0.0499/0.0624
0.1135/0.1198
1.105
Observed reflections/Rσ
Data/restraints/parameters
R1 (observed/all data)
wR2 (observed/refined)[a]
GoF
4809/0.0398
5518/0/363
0.0383/0.0486
0.0791/0.0831
1.093
Residual density [eÅ–3]
–0.302/0.300
–0.664/0.704
–0.277/0.480
–0.537/1.035
[a] wR2 = [Σw(Fo – Fc ) /ΣwFo4]½, w = [σ2(Fo2) + (uP)2 + vP] – 1 with P = [max(Fo2,0) + 2Fc ]/3.
2
2 2
2
Eur. J. Inorg. Chem. 2013, 958–967
964
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
orbit contributions were estimated by using the DKH method im-
plemented in Orca.[47]
evaporator yielded the product as a clear orange oil; yield 2.01 g
(91%). 1H NMR (200 MHz, CDCl3, 25 °C): δ = 8.48–8.42 (m, 3
H, 11-H, 18-H, 25-H), 7.61–7.34 (m, 6 H, 13-H, 14-H, 20-H, 21-
H, 27-H, 28-H), 7.13–6.99 (m, 3 H, 12-H, 19-H, 26-H), 3.85 (br. s,
4 H, 15-H, 22-H), 3.39 (d, 2 H, 5-H, 7-H), 2.92–2.72 (m, 6 H, 2-
H, 3-H, 5-H, 7-H), 1.21 (s, 3 H, 8-H) ppm. {1H}13C NMR
(50.32 MHz, CDCl3, 25 °C): δ = 167.46 (1 C, C-9), 159.99 (2 C, C-
16, C-23), 148.94 (2C, C-18, C-25), 148.33 (1C, C-11), 136.39 (2 C,
C-20, C-27), 136.04 (1 C, C-13), 123.25 (2 C, C-21, C-28), 122.01
(1 C, C-14), 121.02 (2 C, C-19, C-26), 120.92 (1 C, C-12), 67.29 (2
C, C-5, C-7), 66.27 (2 C, C-15, C-22), 59.25 (2 C, C-2, C-3), 46.04
(1 C, C-6), 25.64 (1 C, C-8) ppm. C23H27N5 (373.49): calcd. C
73.96, H 7.29, N 18.75; found C 73.20, H 7.26, N 18.88. HRMS
(ESI): calcd. for C23H27N5 [M + H]+ 374.2339; found 374.2331.
Syntheses: The numbering schemes used for the NMR assignments
of 2, 3 and 4a/b are given in the Supporting Information.
6-Methyl-6-(pyridin-2-yl)-1,4-ditosyl-1,4-diazepane (2): Ethylenedi-
amine-1,2-ditoslylate (15.7 mmol, 5.83 g) and NaOH (36.3 mmol,
1.45 g) were dissolved in H2O (100 mL) and the mixture was heated
to reflux until a clear solution had formed. After addition of a
warm solution of 2-methyl-2-(pyridin-2-yl)propane-1,3-diyl bis(4-
methylbenzenesulfonate) (15.7 mmol, 7.52 g) in toluene (100 mL),
the emulsion was heated to reflux for 2 d. Subsequently, the warm
(not boiling) organic phase was decanted, and the aqueous phase
was washed with warm toluene (100 mL). Removal of toluene from
the combined organic phases left the crude product as an orange
solid, which was dissolved in CHCl3. The insoluble components
were removed by filtration, and the solvent of the filtrate was re-
moved under reduced pressure to give 2 as an orange resinous solid.
The product was used in subsequent syntheses without further pu-
rification; yield 6.80 g (86%). 1H NMR (200 MHz, CDCl3, 25 °C):
6-Methyl-6-(pyridin-2-yl)-1,4-bis[2-(pyridin-2-yl)ethyl]-1,4-di-
azepane (4b): To a solution of 3 (5.8 mmol, 1.10 g) in MeOH
(30 mL) was added acetic acid (14.4 mmol, 0.86 g) and 2-vinylpyr-
idine (28.8 mmol, 3.02 g). The reaction mixture was warmed to
65 °C and stirred for 2 d. The mixture was then cooled to room
temperature, and the volatiles were removed under reduced pres-
sure. The remaining red-brown oil was dissolved in H2O and care-
fully made basic with solid NaOH. Compound 4b was sub-
sequently extracted three times with toluene. The removal of tolu-
ene from the combined organic phases left the crude product as a
red oil, which was purified by repeated extraction with small
amounts of hexane. The solvent was removed in vacuo to give 4b
as a light yellow oil; yield 1.42 g (61%). 1H NMR (200 MHz,
3
δ = 8.54 (dm, JH,H = 4 Hz, 1 H, 11-H), 7.77–7.60 (m, 5 H, 13-H,
3
19-H, 23-H, 29-H, 33-H), 7.52 (d, JH,H = 8 Hz, 1 H, 14-H), 7.30–
2
7.16 (m, 5 H, 12-H, 20-H, 22-H, 30-H, 32-H), 3.77 (d, JH,H
=
14 Hz, 2 H, 5-H, 7-H), 3.61–3.44 (m, 4 H, 2-H, 3-H, 5-H, 7-H),
3.24–3.11 (m, 2 H, 2-H, 3-H), 2.41 (s, 6 H, 24-H, 34-H), 1.56 (s, 3
H, 8-H) ppm. {1H}13C NMR (50.32 MHz, CDCl3, 25 °C): δ =
163.19 (1 C, C-9), 148.26 (1 C, C-11), 143.73 (2 C, C-21, C-31),
137.31 (1 C, C-13), 135.50 (2 C, C-18, C-28), 129.96 (4 C, C-20, C-
22, C-30, C-32), 127.32 (4 C, C-19, C-23, C-29, C-33), 122.13 (1 C,
C-14), 121.35 (1 C, C-12), 58.17 (2 C, C-2, C-3), 51.23 (2 C, C-5,
C-7), 46.10 (1 C, C-6), 23.98 (1 C, C-8), 21.62 (2 C, C-24, C-34)
ppm. HRMS (ESI): calcd. for C25H29N3O4S2 [M + H]+ 500.1672;
found 500.1666.
3
CDCl3, 25 °C): δ = 8.50 (dm, JH,H = 4 Hz, 3 H, 11-H), 7.58–7.46
3
(m, 3 H, 13-H, 21-H, 29-H), 7.36 (d, JH,H = 8 Hz, 1 H, 14-H),
2
7.12–7.04 (m, 5 H, 22-H, 30-H, 12-H, 20-H, 28-H), 3.17 (d, JH,H
= 14 Hz, 2 H, 5-H, 7-H), 2.91 (br. s, 8 H, 2-H, 3-H, 15-H, 23-H),
2.81–2.68 (m, 6 H, 5-H, 7-H, 16-H, 24-H), 1.23 (s, 3 H, 8-H) ppm.
{1H}13C NMR (50.32 MHz, CDCl3, 25 °C): δ = 167.81 (1 C, C-9),
160.94 (2 C, C-17, C-25), 149.34 (2 C, C-19, C-27), 148.35 (1 C, C-
11), 136.24 (2 C, C-21, C-29), 136.02 (1 C, C-13), 123.48 (2 C, C-
22, C-30), 121.29 (1 C, C-14), 121.10 (2 C, C-20, C-28), 120.95 (1
C, C-12), 67.11 (2 C, C-5, C-7), 60.67 (2 C, C-2, C-3), 59.56 (2 C,
C-15, C-23), 45.70 (1 C, C-6), 36.30 (2 C, C-16, C-24), 25.35 (1 C,
C-8) ppm. C25H31N5 (401.55): calcd. C 74.78, H 7.78, N 17.44;
found C 74.73, H 7.92, N 17.29. HRMS (ESI): calcd. for C25H31N5
[M + H]+ 402.2652; found 402.2634.
6-Methyl-6-(pyridin-2-yl)-1,4-diazepane
(3):
Compound
2
(10.0 mmol, 5.00 g) was dissolved in H2SO4 (20 mL). The reaction
mixture was heated to 130 °C for 14 h, after which time a black
suspension had formed. The mixture was cooled to 0 °C and care-
fully basified with saturated aqueous NaOH. The aqueous solution
was extracted three times with CHCl3. The volatiles were then re-
moved from the combined organic phases with a rotary evaporator
to leave a yellow oil. The product was used in subsequent syntheses
without purification; yield 1.59 g (80%). 1H NMR (200 MHz,
3
[Fe(4b)Br]Br (5): A solution of FeBr2 (0.74 mmol, 0.16 mg) in
MeOH (3 mL) was added to a methanolic solution (3 mL) of 4b
(0.78 mmol, 0.31 mg). The solution was left to stir at room tem-
perature overnight before the volume of the solution was reduced
to ca. 1 mL with an oil-pump vacuum. The addition of Et2O
(15 mL) precipitated 5 as a yellow powder. The suspension was
centrifuged, the organic solvent was decanted, and the yellow solid
was washed with Et2O (2ϫ 10 mL) and dried in vacuo; yield 0.39 g
(81%). C25H31Br2FeN5 (617.20): calcd. C 48.65, H 5.06, N 11.35;
found C 49.19, H 5.18, N 11.31. HRMS (ESI): calcd. for
C25H31BrFeN5 [M]+ 536.1107; found 536.1096; calcd. for
C25H31FeN5 [M]2+ 228.5959; found 228.5956.
CDCl3, 25 °C): δ = 8.55 (dm, JH,H = 4 Hz, 1 H, 11-H), 7.63 (td,
4
3
3JH,H = 8, JH,H = 2 Hz, 1 H, 13-H), 7.34 (d, JH,H = 8 Hz, 1 H,
2
14-H), 7.15–7.08 (m, 1 H, 12-H), 3.41 (d, JH,H = 14 Hz, 2 H, 5-
H, 7-H), 2.96 (s, 4 H, 2-H, 3-H), 2.87 (d, JH,H = 14 Hz, 2 H, 5-
2
H, 7-H), 2.36 (br. s, 2 H, 1-H, 4-H), 1.25 (s, 3 H, 8-H) ppm.
{1H}13C NMR (50.32 MHz, CDCl3, 25 °C): δ = 167.04 (1 C, C-9),
148.90 (1 C, C-11), 136.64 (1 C, C-13), 121.31 (1 C, C-14), 120.49
(1 C, C-12), 60.20 (2 C, C-2, C-3), 52.58 (2 C, C-5,C-7), 47.47 (1
C, C-6), 24.93 (1 C, C-8) ppm. HRMS (ESI): calcd. for C11H17N3
[M + H]+ 192.1495; found 192.1490.
6-Methyl-6-(pyridin-2-yl)-1,4-bis[(pyridin-2-yl)methyl]-1,4-diazepane
(4a): Solid K2CO3 (59.04 mmol, 8.16 g) and 2-picolyl chloride hy-
drochloride (11.81 mmol, 1.94 g) were added to a solution of 3
(5.91 mmol, 1.13 g) in MeCN (70 mL, no care was taken to exclude
moisture from the solvent.) The resulting suspension was stirred at
55 °C for 3 d. and was then allowed to cool to room temperature.
The solvent was removed with a rotary evaporator, and the residue
dissolved in toluene (30 mL). Crude 4a was obtained as an orange
oil after washing the toluene phase with water and the subsequent
removal of the solvent under reduced pressure. Trituration of the
crude product with hexane and removal of the solvent with a rotary
[Fe(4a)Cl](PF6) (6): A solution of FeCl2 (0.42 mmol, 0.05 g) in
MeOH (2 mL) was added to a solution of 4a (0.44 mmol, 0.17 g)
in MeOH (2 mL), and the mixture was stirred for 1 h at room tem-
perature. The addition of solid (nBu4N)PF6 produced a yellow pre-
cipitate, which was removed by filtration and washed with Et2O.
Isothermal diffusion of Et2O into a solution of the yellow solid in
MeCN generated crystals of 6. The crystals were washed with Et2O
and dried in vacuo to give a product with the correct elemental
analysis; yield 0.22 g (84%). C23H27ClF6FeN5P (609.76): calcd. C
Eur. J. Inorg. Chem. 2013, 958–967
965
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[2] T. A. Jackson, L. Que Jr., in: Concepts and Models in Bioinor-
ganic Chemistry (Eds.: H.-B. Kraatz, N. Metzler-Nolte), Wiley-
VCH, Weinheim, Germany, 2006, p. 259.
[3] P. Gütlich, H. A. Goodwin, Top. Curr. Chem. 2004, 233, 1–47.
[4] Spin-crossover materials - properties and applications (Ed.:
M. A. Halcrow), John Wiley & Sons, Chichester, UK, 2012, in
press.
[5] A. Grohmann, Dalton Trans. 2010, 39, 1432–1440.
[6] J. Pitarch López, H. Kämpf, M. Grunert, P. Gütlich, F. W. Hei-
nemann, R. Prakash, A. Grohmann, Chem. Commun. 2006,
1718–1720.
[7] J. Pitarch López, F. W. Heinemann, R. Prakash, B. A. Hess, O.
Horner, C. Jeandey, J.-L. Oddou, J.-M. Latour, A. Grohmann,
Chem. Eur. J. 2002, 8, 5709–5722.
45.30, H 4.46, N 11.49; found C 45.30, H 4.31, N 11.13. HRMS
(ESI): calcd. for C23H27ClFeN5 [M]+ 464.1299; found 464.1288;
calcd. for C25H27FeN5 [M]2+ 214.5802; found 214.5798.
[Fe(4a)Br](PF6)·MeCN (7): A solution of FeBr2 (0.33 mmol, 0.07 g)
in MeOH (3 mL) was added to a solution of 4a (0.35 mmol,
130 mg) in MeOH to give a yellow solution, which was stirred for
12 h at room temperature. The addition of solid (nBu4N)PF6
(0.66 mmol, 0.23 g) caused the precipitation of a yellow solid. The
suspension was stirred for an additional 12 h at room temperature,
and the yellow precipitate was collected by filtration, washed with
Et2O and dissolved in MeCN. The isothermal diffusion of Et2O
into this solution produced 7 as a yellow crystalline material, which
was collected by filtration and dried in vacuo; yield 0.18 g, 79%.
C25H30BrF6FeN6P (695.26): calcd. C 43.19, H 4.35, N 12.09; found
C 43.80, H 4.56, N 11.90. HRMS (ESI): calcd. for C23H27BrFeN5
[M]+ 508.0794; found 508.0782; calcd. for C23H27FeN5 [M]2+
214.5802; found 214.5798.
[8] P. Comba, W. Schiek, Coord. Chem. Rev. 2003, 238–239, 21–
29.
[9] M. Schmidt, D. Wiedemann, A. Grohmann, Inorg. Chim. Acta
2011, 374, 514–520.
[10] S. Friedrich, M. Schubart, L. H. Gade, I. L. Scowen, J. And-
rew, M. McPartlin, Chem. Ber./Recueil 1997, 130, 1751–1759.
[11] J. Romba, D. Kuppert, B. Morgenstern, C. Neis, S. Steinhauser,
T. Weyhermüller, K. Hegetschweiler, Eur. J. Inorg. Chem. 2006,
314–328.
[12] H. Chong, K. Garmestani, D. Ma, D. E. Milenic, T.
Overstreet, M. W. Brechbiel, Med. J. Chem. 2002, 45, 3458–
3464.
[13] C. He, S. J. Lippard, Tetrahedron 2000, 56, 8245–8252.
[14] I. Blain, M. Giorgi, I. de Riggi, M. Réglier, Eur. J. Inorg. Chem.
2000, 393–398.
[15] N. Ortega-Villar, V. M. Ugalde-Saldívar, M. C. Muñoz, L. A.
Ortiz-Frade, J. G. Alvarado-Rodríguez, J. A. Real, R. Moreno-
Esparza, Inorg. Chem. 2007, 46, 7285–7293.
[16] M. Haryono, F. W. Heinemann, K. Petukhov, K. Gieb, P.
Müller, A. Grohmann, Eur. J. Inorg. Chem. 2009, 2136–2143.
[17] H. Zabrodsky, S. Peleg, D. Avnir, J. Am. Chem. Soc. 1992, 114,
7843–7851.
[18] Calculations were performed by using the online service of the
Hebrew University of Jerusalem, see: A. Zayit, M. Pinsky, H.
Elgavi, C. Dryzun, D. Avnir, Chirality 2011, 23, 17–23.
[Fe(4b)Cl](PF6) (8): FeCl2 (0.24 mmol, 0.03 g) and 4b (0.25 mmol,
0.10 g) were dissolved in MeOH (3 mL), and the mixture was
stirred at room temperature for 4 h. The addition of solid (nBu4N)-
PF6 (0.47 mmol, 0.18 g) precipitated the product as a yellow solid.
The material was collected by filtration, washed with Et2O and
dissolved in MeCN. The thermal diffusion of Et2O into this solu-
tion generated 8 as a yellow crystalline solid, which was collected
by filtration, washed with Et2O and dried in vacuo; yield 0.07 g
(44%). C25H31ClF6FeN5P (637.81): calcd. C 47.08, H 4.90, N
10.98; found C 47.56, H 5.01, N 10.73. HRMS (ESI): calcd. for
C25H31BrFeN5 [M]+ 536.1107; found 536.1096; calcd. for
C25H31FeN5 [M]2+ 228.5959; found 228.5956.
[Fe(4b)Br](PF6) (9): Solid (nBu4N)PF6 (0.4 mmol, 0.15 g) was
added to a stirred solution of 5 (0.3 mmol, 0.20 g) in MeOH
(4 mL). After a few seconds, a yellow solid started to precipitate,
and stirring of the suspension was continued overnight. Compound
9 was collected by filtration, washed with Et2O and dried under
reduced pressure; yield 0.14 g (63%). Yellow single crystals of 9
formed in a solution of the complex in MeCN after 5 d upon iso-
thermal diffusion of Et2O. C25H31BrF6FeN5P (682.26): calcd. C
44.01, H 4.58, N 10.26; found C 43.85, H 4.74, N 9.92. HRMS
(ESI): calcd. for C25H31BrFeN5 [M]+ 536.1107; found 536.1096;
calcd. for C25H31FeN5 [M]2+ 228.5959; found 228.5956.
[19]
P. Guionneau, M. Marchivie, G. Bravic, J.-F. Létard, D. Chas-
seau, J. Mater. Chem. 2002, 12, 2546–2551.
[20] N. F. Chilton, R. P. Anderson, L. D. Turner, A. Soncini, K. S.
Murray, J. Comput. Chem. 2013, in press.
[21] P. Gütlich, E. Bill, A. X. Trautwein, Mössbauer Spectroscopy
and Transition Metal Chemistry: Fundamentals and Applica-
tions, Springer-Verlag, Berlin, 2011.
[22] B. J. Kennedy, K. S. Murray, Inorg. Chem. 1985, 24, 1552–1557.
[23] D. E. Freedman, W. H. Harman, T. D. Harris, G. J. Long, C. J.
Chang, J. R. Long, J. Am. Chem. Soc. 2010, 132, 1224–1225.
[24] E. Cremades, E. Ruiz, Inorg. Chem. 2011, 50, 4016–4020.
[25] F. Neese, Coord. Chem. Rev. 2009, 253, 526–563.
[26] S. Ye, F. Neese, Inorg. Chem. 2010, 49, 772–774.
[27] C. R. Jacob, M. Reiher, Int. J. Quantum Chem. 2012, 112,
3661–3684.
Supporting Information (see footnote on the first page of this arti-
cle): χMT vs. T for 7 and 9. Numbering scheme for NMR assign-
ments.
Acknowledgments
One of the authors (A. G.) gratefully acknowledges support of this
work by the Deutsche Forschungsgemeinschaft (DFG) (grant
number GR 1247/7-1; SFB 658: Elementary Processes in Molecular
Switches on Surfaces). K. S. M. acknowledges financial support
through a Discovery Grant by the Australian Research Council
(ARC). G. R. and K. S. M. acknowledge support by the Australia-
India Strategic Research Fund (AISRF); Department of Innova-
tion, Industry, Science and Research, Canberra (Project: New Gen-
eration Molecular Magnetic Materials: Experiment and Theory;
DST/INT/AUS/P-47/11). K. R. V. is thankful to the IITB-Monash
Academy for a PhD studentship.
[28] K. P. Kepp, Coord. Chem. Rev. 2013, 257, 196.
[29]
[30]
E. König, Struct. Bonding (Berlin) 1991, 76, 51–152.
J. M. Holland, J. A. McAllister, Z. Lu, C. A. Kilner, M. Thorn-
ton-Pett, M. A. Halcrow, Chem. Commun. 2001, 577–578.
P. Gütlich, H. A. Goodwin, Top. Curr. Chem. 2004, 233, 13–
14.
A recent example in which the exchange of a halido for a thi-
ocyanato ligand leads to a complex showing temperature-de-
pendent spin crossover is: M. Bernien, D. Wiedemann, C. F.
Hermanns, A. Krüger, D. Rolf, W. Kroener, P. Müller, A.
Grohmann, W. Kuch, J. Phys. Chem. Lett. 2012, 3, 3431–3434.
K. S. Murray, C. J. Kepert, Top. Curr. Chem. 2004, 233, 195–
228.
[31]
[32]
[33]
[34]
[35]
S. Schlamp, P. Thoma, B. Weber, Eur. J. Inorg. Chem. 2012,
2759–2768.
E. Cremades, E. Ruiz, Inorg. Chem. 2011, 50, 4016–4020.
[1] B. Kräutler, B. Jaun, in: Concepts and Models in Bioinorganic
Chemistry (Eds.: H.-B. Kraatz, N. Metzler-Nolte), Wiley-VCH,
Weinheim, Germany, 2006, p. 177.
Eur. J. Inorg. Chem. 2013, 958–967
966
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[36] F. Neese, J. Am. Chem. Soc. 2006, 128, 10213–10222.
[37] N. Berg, T. N. Hooper, J. Liu, C. C. Beedle, S. K. Singh, G.
Rajaraman, S. Piligkos, S. Hill, E. K. Brechin, L. F. Jones, Dal-
ton Trans. 2013, 42, 207–216.
Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
T. Vreven, J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M.
Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Starov-
erov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell,
J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M.
Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Ad-
amo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Mar-
tin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09,
rev. A.02, Gaussian, Inc., Wallingford CT, 2009.
[38] CrysAlisPro Software System, Intelligent Data Collection and
Processing Software for Small Molecule and Protein Crystal-
lography, Agilent Technologies, Ltd., Oxford, UK, 2011.
[39] G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr.
2008, 64, 112–122.
[40] L. Palatinus, G. Chapuis, J. Appl. Crystallogr. 2007, 40, 786–
790.
[41] P. van der Sluis, A. L. Spek, Acta Crystallogr., Sect. A: Found.
Crystallogr. 1990, 46, 194–201.
[42] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565.
[43] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[44] A. Schäfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97,
2571–2577.
[47]
F. Neese, Orca, v. 2.9-35, An ab initio, DFT and Semiempirical
SCF-MO Package, Max-Planck-Institut für Chemische Ener-
giekonversion, Mülheim an der Ruhr, Germany, 2012.
Received: September 25, 2012
[45] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100,
5829–5835.
[46] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B.
Published Online: January 18, 2013
Eur. J. Inorg. Chem. 2013, 958–967
967
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim