FeIII spin-crossover complexes with photoisomerizable ligands: Experimental and theoretical studies on the ligand-driven light-induced spin change effect

Article abstract of DOI:10.1002/ejic.201101227

Conflicting results have been reported with respect to the photoinduced switching of the magnetic properties of [Fe^III(salten)]⁺ complexes [salten = 4-azaheptamethylene-1,7-bis(salicylideneiminate)] coordinated by photoisomerizable ligands. In order to address this problem, two Fe^III complexes [Fe(salten)(3-azpy)]BPh₄ (1) and [Fe(salten)(4-azpy)]BPh₄ (2) have been synthesized and characterized by various physicochemical methods (azpy = phenylazopyridine). Both 1 and 2 exhibit a low spin (S = 1/2) to high spin (HS, S = 5/2) transition in the solid state. In solution at room temperature both complexes are predominantly HS. Upon exposure to 310 (trans → cis) and 440 nm radiation (cis → trans) the free and coordinated 3- and 4-azpy ligands undergo a reversible cis-trans isomerization. For 2 a corresponding reduction of the HS fraction 2% is observed, whereas in 1 no effect is observed. Extensive DFT calculations, which employ different functionals and basis sets, explain this experimental result. The consequences of these findings with respect to the design of spin-switchable iron(III) complexes with photoactive ligands are discussed.

Full text of DOI:10.1002/ejic.201101227

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
FULL PAPER
DOI: 10.1002/ejic.201101227
III
Fe Spin-Crossover Complexes with Photoisomerizable Ligands: Experimental
and Theoretical Studies on the Ligand-Driven Light-Induced Spin Change
Effect
[
a]
[a]
[a]
[b]
Alexander Bannwarth, Sven Olaf Schmidt, Gerhard Peters, Frank D. Sönnichsen,
[a]
[b]
[a]
Wulf Thimm, Rainer Herges, and Felix Tuczek*
Keywords: Spin crossover / Photoswitching / Density functional calculations / Iron / Moessbauer spectroscopy / LD-LISC
Conflicting results have been reported with respect to the
photoinduced switching of the magnetic properties of
complexes are predominantly HS. Upon exposure to 310
(trans Ǟ cis) and 440 nm radiation (cis Ǟ trans) the free and
coordinated 3- and 4-azpy ligands undergo a reversible cis–
trans isomerization. For 2 a corresponding reduction of the
HS fraction 2% is observed, whereas in 1 no effect is ob-
served. Extensive DFT calculations, which employ different
functionals and basis sets, explain this experimental result.
The consequences of these findings with respect to the de-
sign of spin-switchable iron(III) complexes with photoactive
ligands are discussed.
III
+
[Fe (salten)] complexes [salten = 4-azaheptamethylene-1,7-
bis(salicylideneiminate)] coordinated by photoisomerizable
III
ligands. In order to address this problem, two Fe complexes
4 4
[Fe(salten)(3-azpy)]BPh (1) and [Fe(salten)(4-azpy)]BPh (2)
have been synthesized and characterized by various physico-
chemical methods (azpy = phenylazopyridine). Both 1 and 2
exhibit a low spin (S = 1/2) to high spin (HS, S = 5/2) transi-
tion in the solid state. In solution at room temperature both
Introduction
Numerous applications of the LD-LISC effect have been
reported, which underscores the importance of this concept
as a tool to switch the spin state of transition metal com-
Iron spin-crossover complexes with photoisomerizable li-
gands have attracted much attention due to their potential
application as magnetic switches.[ On irradiation, these li-
gands change their conformation (e.g. trans p cis) and, as
a consequence, their ligand-field strength. By suitable
choice of the coligands, Fe and Fe complexes with pho-
toswitchable magnetic properties can be obtained. This con-
cept is termed ligand-driven light-induced spin change
[
3,4,5,6]
plexes in solution.
For example, the LD-LISC effect
1]
has been demonstrated at room temperature by the iron(II)
complex [FeL3]2 [L = phenyl(2-pyridin-2-yl-3H-benzo-
+
[
5]
imidazol-5-yl)diazene]. Irradiation of this complex in-
II
III
duced a trans Ǟ cis isomerization and an increase of the
effective magnetic moment μ , which corresponds to a
eff
transition from LS to HS. In agreement with the initial re-
II
(LD-LISC) and was first exemplified by the Fe complex
sults obtained for [Fe(stypy) (NCS) ], this finding was as-
[2]
4
2
[
Fe(stypy) (NCS) ] (stypy = styrylpyridine). UV irradia-
4 2
cribed to a lower degree of π-electron localization, a smaller
extent of π backbonding and, thus, to a lower-ligand field
strength in the cis compared to the trans configuration of
the photoswitchable ligands.
tion of this complex induced a trans Ǟ cis isomerization of
the four ligands and, concomitantly, a transition of the Fe
II
complex from the low-spin (LS) to the high-spin (HS) state.
This finding has been attributed to the fact that the trans
ligand is planar, whereas the cis ligand is nonplanar, which
reduces the degree of conjugation within the ligand π-sys-
tem. This in turn was correlated with a reduction of π back-
bonding in the cis ligand, which renders metal–ligand
bonding in this isomer weaker than in the corresponding
trans form.
This very consistent picture is less clear for FeIII spin-
crossover complexes coordinated by similar photoswitch-
III
able ligands. The Fe complex [Fe(salten)(mepepy)]BPh4
was prepared in 1999, and the consequences of cis–trans
photoisomerization of the ligand on the spin equilibrium
were investigated in solution [salten = 4-azaheptamethyl-
ene-1,7-bis(salicylideneiminate), mepepy = 1-(pyridin-4-yl)-
[
7]
2-(N-methylpyrrol-2-yl)ethene; Scheme 1].
[Fe(salten)-
[
a] Institut für Anorganische Chemie, Christian-Albrechts-
Universität Kiel,
(
mepepy)]BPh exhibits spin-crossover behaviour in the so-
4
Max-Eyth-Str. 2, 24098 Kiel, Germany
Fax: +49-431-8801520
lid state with a temperature T1/2 (where γHS = γLS) of
ca. 300 K. In CD CN solution at room temperature, the
3
E-mail: ftuczek@ac.uni-kiel.de
[
b] Otto Diels Institut für Organische Chemie, Christian-Albrechts- complex was found to be predominantly HS (68%). Irradia-
Universität Kiel,
tion at 405 nm also induced photoisomerization of the li-
gand from trans to cis, this time, however, accompanied by
a decrease of the susceptibility (HS fraction 56%).
Otto-Hahn-Platz 4, 24098 Kiel, Germany
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201101227.
Eur. J. Inorg. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
F. Tuczek et al.
FULL PAPER
ployed to interpret the experimental findings, and the impli-
cations of the results with respect to the design of magneti-
cally switchable iron(III) complexes with photoisomerizable
ligands are discussed.
+
Scheme 1. Structure of [Fe(salten)L] , L = mepepy, stypy.
This observation was attributed to the fact that the N-
methylimidazole group of mepepy acts as a π donor, which
lowers the ligand-field strength of the ligand. Again, due to
the larger extent of π delocalization in the trans compared
to the cis form, the electronic effect was assumed to be more
pronounced in the former than the latter isomer.
III
The Fe salten complex coordinated by stypy has been Scheme 2. Structures of 1 (left) and 2 (right).
investigated with respect to its magnetic properties in solu-
tion (Scheme 1).[ This complex can be synthesized and iso-
8]
lated with coordinated stypy in the trans and cis forms;
crystal structures of both isomeric complexes have been de-
Results and Discussion
termined. It was found that both trans and cis [Fe(salten)-
Thermal Spin-Crossover in the Solid State
(stypy)] exhibit gradual spin transitions in the solid state
with T1/2 values of ca. 220 and 320 K, respectively. In this
case, the cis ligand appeared to mediate a stronger ligand
The thermal spin-crossover behaviour of 1 and 2 was first
investigated by magnetic susceptibility measurements. Plots
field than the trans ligand. In CH CN solution at room
3
of the product χ T vs. the temperature (T) are given in
M
temperature, the opposite was observed, i.e. the magnetic
moment of the cis complex (μeff = 4.22 B.M. corresponding
to 46% HS) was higher than that of the trans complex
Figure 1. Both compounds exhibit gradual spin-crossover
behaviour in the solid state with T1/2 values of about 200 K.
For 1, χ T monotonously increases from a low-tempera-
M
(μeff = 3.47 B.M. corresponding to 28% HS), which indi-
3
ture value of 0.7 to 3.5 cm K/mol at room temperature,
cates a weaker ligand-field strength of the cis compared to
the trans ligand. However, trans Ǟ cis photoisomerization
of the as-synthesized trans complex did not lead to the ex-
pected increase of the HS fraction and, as inferred from
optical absorption data, apparently left the spin equilibrium
unaffected. These contradictory results were tentatively ex-
plained by the presence of two different coordination geo-
which indicates a transition from LS to HS. The χ T vs. T
M
curve for 2 starts at a slightly higher low-temperature value
3
(
0.75 cm K/mol) and also reaches a slightly lower room-
3
temperature value [χ T(300 K) = 3.0 cm K/mol] than 1.
M
For both systems, the low-temperature χ T values are
M
larger than expected for pure S = 1/2 systems (χ T =
M
3
0
.375 cm K/mol), which might be caused by the orbital
momentum present in the T electronic groundstate of LS
Fe complexes. In contrast, the room-temperature values
+
metries of [Fe(salten)(cis-stypy)] , one in the as-synthesized
and the other in the photoexcited state, in which the latter
2
1g
III
species has the same spin-crossover properties as [Fe-
are considerably smaller than expected for pure S = 5/2
+
(
salten)(trans-stypy)] .
III
Fe systems. Although the spin transitions of both systems
To further investigate this issue, we decided to synthesize
do not appear to be complete at 300 K, the small slopes of
FeIII salten complexes coordinated by photoswitchable
phenylazopyridine (azpy) ligands. From the viewpoint of
photochemical stability these ligands should be more
favourable than their styrylpyridine analogs as stilbenes po-
tentially undergo photochemical decomposition by electro-
cyclic ring closure and oxidation reactions that are absent
the χ T vs. T curves at room temperature suggest that the
M
3
limiting values of χ T = 4.375 cm K/mol (μ = 5.9 B.M.)
M
eff
are not reached at higher temperatures. In addition, as it is
gradual, the spin transition behaviour of both systems also
appear to be incomplete. Similar observations have been
[
9,10]
for their azo counterparts.
[
The magnetic properties of
Fe(salten)(3-azpy)]+ (1) and [Fe(salten)(4-azpy)] (2,
+
Scheme 2) were investigated by magnetic susceptibility mea-
surements and Mössbauer spectroscopy. Furthermore, the
optical switching behaviour of the isolated ligands 3- and
4-azpy as well as 1 and 2 was investigated by electronic
absorption spectroscopy. Finally the consequences of trans
p cis photoisomerization of the coordinated ligands with
respect to the magnetic properties of the corresponding
III
M
Fe complexes were evaluated. DFT calculations were em- Figure 1. χ T vs. T curves of 1 (left) and 2 (right).
2
www.eurjic.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 0000, 0–0

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
FeIII Spin-Crossover Complexes with Photoisomerizable Ligands
III
made with the spin transition behaviour of other Fe spin- Magnetic Properties and Photoswitching Experiments in
crossover complexes in the solid state.[
11]
Solution
The temperature-dependent Mössbauer spectra of 1 and
are shown in Figure 2, and the corresponding fitting pa-
To investigate the photoswitching properties of the free
photoisomerizable ligands 3- and 4-azpy and their Fe
2
III
rameters are collected in Table 1. The spectra of both com-
pounds show gradual transitions from LS to HS and thus
overall reflect the thermal spin-crossover behaviour already
derived from the magnetic susceptibility data presented
above. Interesting differences between 1 and 2 are observed
with respect to their spin relaxation characteristics: the
spectra of 2 at all temperatures exhibit separate HS and LS
subspectra, indicative of slow relaxation behaviour, whereas
the spectra of 1 show only one doublet, which represents a
thermal average of the HS and LS subspectra. Conse-
quently, these Mössbauer spectra are of the “fast” relax-
complexes, irradiation experiments were performed in vari-
ous solvents (toluene, dichloromethane, ethanol, methanol,
dimethylsulfoxide, acetonitrile, hexane). Ethanol and meth-
anol were chosen as the most suitable solvents as the coor-
dination compounds are virtually insoluble in toluene and
hexane, and acetonitrile slowly exchanges with the coordi-
nated azpy ligand. The free ligands 3- and 4-azpy in their
trans ground states exhibit a πǞπ* transition at 317 and
309 nm, respectively, and an nǞπ* transition was observed
at about 460 nm for both systems (Figure 3, left). Upon ir-
radiation at 310 nm, trans Ǟ cis photoisomerization oc-
curred. The πǞπ* transitions greatly decreased in intensity,
whereas the nǞπ* transitions became somewhat more in-
tense. Moreover, a blueshift of the latter transitions by
ca. 20 nm was observed.
III
ation type, as observed for many other Fe spin-crossover
[
12]
complexes, which include those with the salten ligand.
The Mössbauer parameters given for 1 in Table 1 represent
thermal averages of those of the respective LS and HS spe-
cies, which lead to a continuous variation of isomer shifts
and quadrupole splittings with increasing temperature from
characteristic LS to HS values. From the observed line
broadenings, however, it can be concluded that spin relax-
ation is not very fast but becomes comparable to the Möss-
bauer timescale, which indicates an intermediate-to-fast re-
laxation regime.
Figure 3. Photoswitching of the ligands L (left) and the complexes
(
right). Top: L = 3-azpy, bottom: L = 4-azpy (black: not irradiated,
blue: irradiated at 310 nm, red: irradiated at 410 nm).
Similar, although not identical, results were observed in
the photoswitching experiments with 1 and 2 (Figure 3,
right). Upon irradiation at 310 nm, the πǞπ* transitions at
318 (1) and 319 nm (2) decreased in intensity by about half,
indicative of switching of the coordinated ligands from the
trans to the cis conformation. In contrast to the behaviour
of the free ligands, however, little spectral change was ob-
served in the nǞπ* region.
Figure 2. Mössbauer spectra of 1 (left) and 2 (right). Blue: LS, red:
HS.
Nevertheless, reversible photoinduced switching behav-
iour is evident from the fact that the πǞπ* bands were fully
reconstituted by irradiation at 410 nm, which induced cis
Ǟ trans isomerization of the complex-bound ligands. The
lifetimes of the photoswitched species were in the order of
hours both for the isolated ligands and complexes.
Table 1. Mössbauer fitting parameters (*: HS, LS).
δIS [mm/s]
Q
ΔE [mm/s]
T [K]
1
2
1
2
3
2
1
1
1
1
8
5
00
00
75
50
25
00
0
0.31
0.28
0.24
0.24
0.26
0.26
0.26
0.27
0.32
1.26
1.95
2.42
2.76
2.86
2.9
1.19
0.3, 0.28*
0.31, 0.29*
0.3, 0.25*
0.3, 0.28*
0.32, 0.29*
0.3, 0.26*
0.3
1.23, 2.35*
1.35, 2.68*
1.53, 2.87*
1.56, 2.91*
1.53, 2.89*
1.51, 2.9*
2.81
To quantitatively determine the magnetic properties of 1
and 2 in solution, the Evans method was employed.[ To
this end, a set of two coaxial NMR tubes that consisted of
an ordinary outer NMR tube and an inner tube with a
smaller diameter was used. The inner tube contained the
13]
2.87
2.92
Eur. J. Inorg. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
F. Tuczek et al.
FULL PAPER
reference solvent ([D ]ethanol) and 10 vol.-% tetramethylsil- was evidenced. To prevent the photoinduced decomposition
6
+
ane, and the outer tube was filled with the same mixture of the [Fe(salten)(azpy)] complexes, addition of an excess
along with the paramagnetic complex. The mass suscep- of the photoswitchable ligand in the irradiated solution was
tibility, χ , of the complex was calculated from the differ- necessary. The application of this condition and the ex-
g
ence of the chemical shifts, Δf, of the tetramethylsilane pro- clusion of water from the solution led to a very small or no
ton signals in the inner and outer tubes. The susceptibilities photoinduced change of the spin equilibrium upon expo-
were corrected for the diamagnetic susceptibilities of the sure of the complexes to UV radiation. This became appar-
solvent and the iron complex. The resulting mass suscep- ent when the irradiation experiments were performed in the
[
13]
tibilities were calculated as described in the literature.
presence of an excess of azpy and in solvents that had been
The molar susceptibility of the solvent (ethanol) was taken carefully dried prior to the experiments. To dissolve the
–
6
3
–1
to be –33.7ϫ10 cm mol , and the diamagnetic correc- complex and the excess of free ligand in the Evans experi-
tion of the complex was calculated from an increment sys- ments, [D ]MeOH had to be used as the solvent; the switch-
4
[
14]
tem.
The molar susceptibility of the complex was then ing behaviour of the ligands and the complexes in methanol
calculated from the mass susceptibility and the molecular is almost identical to that in ethanol. As evident from
weight, M, of the complexes and finally the magnetic mo- Table 3, the magnetic properties of 1 and 2 were similar to
ment, μ , was computed.
the values in ethanol. Importantly, the decrease of the HS
eff
The consequences of trans Ǟ cis photoisomerizations of fraction of 2 upon irradiation at 310 nm was only 2% and
the azpy ligands on the magnetic properties of 1 and 2 were no reduction of γHS was observed upon irradiation at
monitored by performing Evans measurements before and 310 nm for 1. It has to be concluded that the spin equilibria
after irradiation, and the results obtained are collected in of Fe(salten) complexes with coordinated azpy are not or
Table 2. With respect to the magnetic properties of the unir- only slightly influenced by the cis–trans isomerization of the
radiated complexes, 2 has a higher magnetic moment (μeff ligands.
=
5.55 B.M. corresponding to 87% HS) in ethanol solution
Table 3. Results of the Evans measurements in the presence of ex-
than 1 (μeff = 5.29 B.M. corresponding to 78% HS), in con-
trast to the solid-state measurements where the opposite
was observed. This discrepancy can, however, be attributed
cess ligand in [D
4
]MeOD.
1
c [mmol/L] Δf [Hz]
χ
M
μ
eff
% HS
to solid-state effects on the spin transition behaviour, which Not irradiated
1.189
1.189
29.270 0.0118
29.270 0.0118
5.31 79
5.31 79
have been observed and interpreted before.[
11,15]
Irrad. @ 310 nm
More im-
portantly, when the complex-bound ligands are switched
2
c [mmol/L] Δf [Hz]
χ
M
μ
eff
% HS
from trans to cis by irradiation at 310 nm, a decrease in the
Not irradiated
1.924
1.924
43.085 0.0134
42.163 0.0131
5.66 91
5.60 89
HS fraction of about 10% was observed for both 1 and 2 Irrad. @ 310 nm
cf. Table 2). However, when 1 and 2 are switched back from
(
cis to trans by irradiation at 440 nm, no concomitant in-
crease of paramagnetism to the original value of the trans
complex was observed but, instead, the paramagnetism
stayed at the lower value. In terms of the magnetic proper-
ties of 1 and 2, the consequences of the trans Ǟ cis photo- preceding sections, [Fe(salten)(4-azpy)] and [Fe(salten)(3-
switching of the azpy ligands were initially found to be azpy)] were studied by DFT calculations. For comparison,
DFT Calculations
To account for the experimental results reported in the
+
+
irreversible.
similar calculations were performed on [Fe(salten)-
+
(
stypy)] , which had been experimentally investigated be-
[
8]
Table 2. Results of the Evans measurements in [D
6
]EtOH.
fore (vide supra). First, the differences between the elec-
tronic energies of the HS and the LS states were calculated
for the trans and cis forms of the respective coordinated
ligands, and various functionals and basis sets were em-
1
c [mmol/L] Δf [Hz]
χ
M
μ
eff
% HS
Not irradiated
Irrad. @ 310 nm
3.429
3.429
67.135
61.263
0.0116 5.29
0.0106 5.05
78
70
0
ployed. Standard free reaction enthalpies, ΔG , for the tran-
2
c [mmol/L] Δf [Hz]
χ
M
μ
eff
% HS
sitions from LS to HS were evaluated for the different com-
Not irradiated
Irrad. @ 310 nm
2.509
2.509
54.086
48.300
0.0128 5.55
0.0114 5.24
87
76
plexes in their two isomeric forms. The geometries of
+
cis/trans-[Fe(salten)(stypy)]
and cis/trans-[Fe(salten)(4-
+
azpy)] are based on the single-crystal structures of the
[
8]
The presence of water was ultimately identified as the compounds. The starting geometries of cis/trans-[Fe-
origin of the photoinduced reduction of μ , which possibly (salten)(3-azpy)] were derived from their 4-azpy counter-
+
eff
led to the formation of an antiferromagnetically coupled μ- parts. Twist angles of the photoswitchable units were taken
hydroxo-bridged Fe(salten) dimer.[ Presumably formation from optimizations of the isolated ligand molecules. The
of the dimer is preceded by photoinduced loss of the azpy geometries of the complexes were optimized in the respec-
ligand and coordination of water. Similar results have been tive LS and HS electronic configurations (S = 1/2 for the
obtained in a photoacustic calorimetry study of the photo- LS and S = 5/2 for the HS states). As an example of the
16]
+
[17]
isomerization process in [Fe(salten)(mepepy)] .
In this resulting geometries, the structures of cis/trans-[Fe-
+
case, cleavage of the Fe–N bond upon irradiation at 355 nm (salten)(4-azpy)] are shown in Figure 4.
4
www.eurjic.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 0000, 0–0

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
FeIII Spin-Crossover Complexes with Photoisomerizable Ligands
0
0
0
Table 5. Standard free reaction enthalpies ΔG = ΔH – TΔS for
+
0
the transition from LS to HS for [Fe(salten)L] [kcal/mol]; Δ(ΔG )
0
0
=
ΔG (cis) – ΔG (trans).
Ligand L/config. BP86/
TZVP
TPSSH/
TZVP
B3LYP/
TZVP
OPBE/
TZVP
4
4
-stypy cis
+16.07
+15.50
0.57
+6.32
+5.74
0.58
–5.67
–5.64
–0.03
–7.51
–7.06
–0.44
-stypy trans
0
Δ(ΔG )
+
Figure 4. Geometries of [Fe(salten)(4-azpy)] in the trans (left) and
4-azpy cis
+15.46
+15.25
0.21
+5.99
+5.51
0.48
–5.56
–5.92
0.36
–7.66
–6.98
–0.68
cis (right) states derived from DFT geometry optimizations.
4-azpy trans
0
Δ(ΔG )
3
3
-azpy cis
–
–
–
+5.95
+6.03
–0.08
–5.87
–5.57
–0.30
–9.96
–9.77
–0.19
Table 4 summarizes the electronic ground states (HS or
-azpy trans
0
LS) and the energy differences, which include zero point Δ(ΔG )
energies (ZPE), to the respective excited states (LS or HS)
that result from various DFT treatments. Table 5 lists the
0
corresponding standard free reaction enthalpies, ΔG , for
were performed). These calculations predict that both stypy
and azpy complexes are in the LS state over the entire tem-
perature range, in contrast to the experimental results. The
calculation with the TZVP basis set and the TPSSH hybrid
functional, which has been found to provide proper descrip-
transitions from the LS to the HS states. In order to calcu-
0
0
late ΔG , the enthalpy changes, ΔH , and the entropy
0
changes, ΔS , for the transition from LS to HS were calcu-
lated for all complexes in their cis and trans states. The en-
tropy changes each consist of an electronic and a vi-
brational part, and calculation of the latter requires a geom-
etry optimization and a subsequent evaluation of the molec-
ular vibrations. For the investigated complexes, the stan-
dard entropy change for the transition from the LS to the
[
21]
tions of many transition metal complexes,
also gave LS
ground states, however, with smaller excitation energies to
the HS states than BP86 (8–9 kcal/mol). Correspondingly,
0
the ΔG value (ca. 6 kcal/mol) was still positive and the
–
1
–1
complexes were predicted to be in the LS state over the
entire temperature range. The TZVP basis set with a B3LYP
HS state is calculated as 10–16 calmol K (see Support-
–
1
–1
ing Information). A ΔS value of 9 calmol K has been
[
22]
III
hybrid functional,
in contrast, gave HS ground states
determined experimentally for an Fe spin-crossover com-
[
18]
with gaps of 2.5–3 kcal/mol to the excited LS state. Corre-
spondingly, all complexes were predicted to be HS over the
entire temperature range. For the OPBE functional,[ HS
ground states are obtained as well. The energy difference is
increased for the stypy, 3-azpy and 4-azpy complexes. Sim-
ilar differences with respect to the theoretical description of
plex.
The observed spin-crossover behaviour of 1 and 2
with T1/2 values of ca. 200 K would be represented by LS
ground states with energy gaps of ca. 2 to 3 kcal/mol to the
excited HS states, which corresponds to typical values
23]
III
[19]
found experimentally for Fe spin-crossover complexes.
spin-crossover complexes have already been obtained and
Table 4. Calculated ground states (LS or HS) for the [Fe(salten)-
L] complexes with the energy difference to the other spin state (in
brackets, including ZPE; all energies in kcal/mol).
[24,25,26,27,28]
discussed in the literature.
The treatment that
+
comes closest to the HS–LS energy difference of 2–3 kcal/
mol for all of the complexes is B3LYP/TZVP, although the
HS–LS gap is predicted with the wrong sign. Both BP86
and TPSSH favoured the LS state too much, i.e. they gave
too large HSLS energy gaps, whereas for all three com-
plexes, OPBE favoured the HS state too much.
Ligand L/config.
-stypy cis
-stypy trans
-azpy cis
-azpy trans
-azpy cis
-azpy trans
BP86/
TZVP
TPSSH/
TZVP
B3LYP/
TZVP
OPBE/
TZVP
4
LS
LS
(+9.43)
LS
(+9.24)
LS
(+9.31)
LS
HS
(+2.46)
HS
(+2.53)
HS
(+2.53)
HS
(+2.65)
HS
(+3.05)
HS
HS
(+3.81)
HS
(+3.76)
HS
(+4.00)
HS
(+3.88)
HS
(+6.13)
HS
(+18.93)
4
LS
The calculations also provide information with respect to
the original question of whether trans–cis isomerizations of
the complex-bound stypy and azpy ligands cause a change
to the spin equilibria of the corresponding complexes. The
spin equilibria at room temperature are determined by the
standard free reaction enthalpies (cf. Table 5). In order to
determine whether the spin equilibria are influenced by the
trans Ǟ cis isomerization of the ligands the differences
(+18.97)
4
LS
(+18.81)
4
LS
(+18.96)
(+9.14)
LS
3
–
(
+8.49)
LS
+8.49)
3
–
(
(+3.01)
(+5.90)
0
0
0
Δ(ΔG ) = ΔG (cis) – ΔG (trans) have to be considered. On
With respect to the energy gaps and the nature of the the basis of the most reliable description (B3LYP/TZVP)
0
ground states (HS or LS), the calculations provide different Δ(ΔG ) is –0.30 kcal/mol for 1 and +0.36 kcal/mol for 2
answers that depend upon the functional employed. Evi- (Table 5). So the spin equilibrium constants KHL = [HS]/
[
20]
dently, the pure exchange-correlation functional BP86 fa- [LS] would be changed at room temperature by a factor of
vours the LS state too much and resulted in a very large 1.13 for 1 and a factor of 0.87 for 2 upon switching from
excitation energy (+19 kcal/mol) to the HS state and a large, trans to cis. The HS fractions thus would go from 80% in
0
positive ΔG value (for this reason only calculations for 2 the case of 1 (cf. Table 3) to 81.9% and from 90% in the
Eur. J. Inorg. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
F. Tuczek et al.
FULL PAPER
case of 2 (cf. Table 3) to 88.7%. Thus, the effects are smaller
than 2%, which account for the experimental results.
Comparing the three systems, the effects of photoisomer-
ization of the ligand on the spin equilibria of the corre-
However, if the equilibrium constants for 1 and 2 at sponding iron(III) compounds are complex. In view of the
room temperature were in the order of unity, the calcula- experimental and theoretical results that the effects are
III
tions indicate that higher switching efficiencies could be ob- small for Fe compounds, the question arises as to why
tained. In this case the HS fractions would change from 50– much larger magnetic photoswitching efficiencies have been
II
5
3% for 1 and from 50–46.5% for 2.
observed for Fe complexes that employ such ligands (vide
[
2,6]
supra).
The answer to this question probably lies in the
II
fact that (i) due to their lower effective nuclear charge, Fe
complexes are more susceptible to π backbonding effects,
i.e. small changes in π backbonding have larger conse-
Conclusions
+
II
III
The magnetic properties of [Fe(salten)(3-azpy)] (1) and quences on Fe than Fe complexes, and (ii) large LD-
Fe(salten)(4-azpy)]+ (2) have been determined. As moni- LISC effects have often been observed for Fe complexes
II
[
[
2,6]
tored by Mössbauer spectroscopy and magnetic suscep- with two or more photoswitchable ligands.
It thus ap-
tibility measurements, both complexes exhibit a LS (S = pears that in these compounds additive (or even coopera-
/2) to HS (S = 5/2) transition in the solid state. In this tive) effects of several photoisomerizable ligands are impor-
respect 1 and 2 are similar to other monomeric and dimeric tant for the observed spin-switching of the central transition
1
+
[
Fe(salten)L] complexes coordinated to ligands with coor- metal ion. This hypothesis has to be investigated experimen-
[29]
III
dinating pyridine groups. For both 1 and 2, the spin tran- tally for Fe complexes with photoisomerizable ligands.
sition is fairly gradual and incomplete, i.e. the full magnetic More investigations along these lines are required to make
moment of the HS state is not reached even at room tem- further progress in the synthesis of magnetically switchable
perature. In alcoholic solutions at room temperature, both transition metal complexes with high photoswitching effi-
complexes are predominantly HS, as revealed by Evans ciencies,[30] for which exciting applications in the fields of
measurements. Moreover, the free and coordinated 3- and switchable magnetic resonance imaging agents or photo-
4
-azpy ligands undergo a reversible cis–trans isomerization magnetic levitation have been proposed.[31]
upon irradiation at 310 (trans Ǟ cis) and 440 nm (cis Ǟ
trans). On photoswitching in predried solvents in the pres-
ence of an excess of ligand, a small change in the HS frac-
tion (–2%) was observed for 2 and no change in the HS
fraction was observed for 1.
Experimental Section
Calculations: All input geometries were derived from single-crystal
structures of appropriate salten-type complexes. Preoptimization
was carried out at the HF/6-31g* level. Full geometric optimiza-
tions with subsequent frequency calculations were carried out using
the TZVP basis-set with the BP86, B3LYP, OPBE and TPSSH
functionals. The twist angle of the cis conformation in the respec-
tive ligands was determined by a geometry optimization at the
B3LYP/6-311g(d,p) level. All operations were performed on a NEC
DFT calculations were performed on 1, 2 and, for com-
+
parison, [Fe(salten)(stypy)] with different functionals and
basis sets. The description that comes closest to the ob-
served spin-crossover behaviour is B3LYP/TZVP. The cal-
culations indicate that the effects of the cis–trans isomeriza-
tion on the spin equilibria of 1 and 2 are small (in the order
of 1–2%). Interestingly, the calculations predict an increase SX/9 vector array using the implemented routines of the
of the HS fraction for the trans Ǟ cis isomerization of 3- Gaussian09 program package.[32]
azpy and a decrease of the HS fraction for 4-azpy. In fact,
a decrease of 2% was observed for the trans Ǟ cis isomer-
ization of 2, but due to the problems described with traces
with a coaxial pyrasil NMR microtube inlay, which contained 1%
of water in the solvent, it cannot unambiguously be decided tetramethylsilane (TMS) in [D ]ethanol and [D ]MeOH. The outer
Spectroscopy: Evans measurements were performed with a Bruker
Avance 400 instrument using a suprasil 5 mm outer NMR tube
6
4
whether this spin change is due to the dissolved complex or tube was filled with the paramagnetic complex, dissolved in the
1
3
the formation of antiferromagnetically coupled dimers. For same TMS/ethanol or TMS/methanol solution. C NMR spectra
were measured with a Bruker Avance 600 instrument. Mössbauer
1
, no change of the HS fraction was observed upon trans Ǟ
measurements were recorded with a self-assembled spectrometer
using standard transmission geometry. UV/Vis spectra were re-
corded with a Cary 5000 spectrometer.
cis isomerization of the ligand (Table 3), but the predicted
change of 2% might be too small to observe experimentally.
For [Fe(salten)(stypy)] , which has been investigated pre-
+
viously, the calculations indicate that the spin equilibrium Irradiation: Irradiation experiments were performed with a mercury
high pressure lamp. To obtain certain wavelengths, band-pass filters
were used.
should not be influenced by trans Ǟ cis photoisomerization
of the coordinated ligand, which explains the corresponding
experimental observation. The fact that for [Fe(salten)- Magnetic Measurements: The magnetic measurements were per-
+
(
stypy)] , a difference in the HS fraction between the as- formed with an automated low-temperature physical property mea-
synthesized trans and cis complexes has been observed in surement system (PPMS, Quantum Design).
solution at room temperature (28% HS for trans, 46% HS
Chemicals: All starting materials (reagent grade) were supplied by
for cis) can be attributed to ligand dissociation, which gen-
erates a larger fraction of (five-coordinate) HS complex for otherwise). Deuterated solvents for NMR measurements were pur-
the cis compared to the trans complex. chased from Deutero and used as supplied.
Sigma–Aldrich. Solvents were of technical grade (if not stated
6
www.eurjic.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 0000, 0–0

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
FeIII Spin-Crossover Complexes with Photoisomerizable Ligands
3
-Phenylazopyridine:[33] Aqueous tetramethylammonium hydroxide
solution (25 mL, 60 wt.-%) and 3-aminopyridine (2.00 g,
1.2 mmol) were dissolved in pyridine (25 mL). To this was added
N 9.38; found for 1 C 73.15, H 5.85, N 9.29; found for 2 C 73.69,
H 5.87, N 9.27.
2
Supporting Information (see footnote on the first page of this arti-
cle): Details of the DFT calculations (energies).
nitrosobenzene (3.00 g, 28.0 mmol) in pyridine (50 mL) over 60 min
at 80 °C. The solution was allowed to cool to room temp. and ex-
tracted with toluene (3ϫ50 mL). The organic extracts were washed
with brine, dried with MgSO
crude product was purified by column chromatography using silica
gel and a 2:1 mixture of cyclohexane/ethyl acetate as eluent. The
azopyridine was clearly observed as a deep red fraction on the col-
umn; yield 3.46 g (18.9 mmol, 89%). H NMR (400 MHz, CDCl
4
and the solvent was evaporated. The Acknowledgments
The authors thank the Deutsche Forschungsgemeinschaft (DFG)
(SFB 677) for funding this research.
1
3
,
4
4
TMS): δ = 9.17 (d, 1 H), 8.66 (d, J = 1.6 Hz, 1 H), 8.10 (d, J =
[
1] M.-L. Boillot, J. Zarembowitch, A. Sour, Top. Curr. Chem.
1
3
1
.6 Hz, 1 H), 7.91 (m, 2 H), 7.49 (m, 3 H), 7.40 (m, 1 H) ppm.
NMR (150 MHz, CDCl , TMS): δ = 152.5 (C ), 151.7 (C ), 147.9
), 129.2 (2C ), 126.9 (C ), 124.0 (C ), 123.1
C
2004, 234, 261.
3
t
q
[
2] C. Roux, J. Zarembowitch, B. Gallois, T. Granier, R. Claude,
(C
t
), 147.3 (C
q
), 131.7 (C
t
t
t
t
Inorg. Chem. 1994, 33, 2273.
(
2C ) ppm.
t
[3] M.-L. Boillot, C. Roux, J.-P. Audière, A. Dausse, J. Zarembow-
itch, Inorg. Chem. 1996, 35, 3975.
4
-Phenylazopyridine:[33] The synthesis was analogous to that of 3-
[
4] M.-L. Boillot, S. Chantraine, J. Zarembowitch, J.-Y. Lallem-
azopyridine but with the use of 4-aminopyridine as a starting mate-
rial; yield 3.55 g (19.4 mmol, 92%). H NMR (400 MHz, CDCl ,
3
and, J. Prunet, New J. Chem. 1999, 179.
1
[
[
5] Y. Hasegawa, S. Kume, H. Nishihara, Dalton Trans. 2009, 280.
6] Y. Hasegawa, K. Takahashi, S. Kume, H. Nishihara, Chem.
Commun. 2011, 47, 6846.
3
4
TMS): δ = 8.78(dd, J = 4.5, J = 1.6 Hz, 2 H), 7.93(m, 2 H), 7.68
(
(
1
dd, 3J = 4.5, 4J = 1.6 Hz, 2 H), 7.51(m, 3 H) ppm. 13C NMR
150 MHz, CDCl , TMS): δ = 157.2 (C ), 152.4 (C ), 151.3 (2 C ),
), 129.3 (2C ), 123.4 (2C ), 116.5(2C ) ppm.
[7] A. Sour, M.-L. Boillot, E. Rivière, P. Lesot, Eur. J. Inorg. Chem.
999, 2117.
8] S. Hirose, S. Hayami, Y. Maeda, Bull. Chem. Soc. Jpn. 2000,
3, 2059.
3
q
q
t
1
32.4 (C
t
t
t
t
[
H
2
Salten:[34] Salicylaldehyde (8.16 mL, 80.0 mmol) and bis(2-ami-
7
nopropyl)amine (5.70 mL, 40.0 mmol) were dissolved in methanol [9] L. Wang, C. Yi, H. Zou, J. Xu, W. Xu, J. Phys. Org. Chem.
(
40 mL). The mixture was connected to a Dean–Stark trap and
2009, 22, 888.
heated to reflux for 2 h. The solution was allowed to cool to room
temp., and the solvent was evaporated to yield a viscous yellow
mass. NMR analysis proved further purification to be unnecessary;
[10] U. Mazzucato, Pure Appl. Chem. 1982, 54, 1705.
[
11] M. S. Haddad, M. W. Lynch, W. D. Federer, D. N. Hendrick-
son, Inorg. Chem. 1981, 20, 123.
[12] N. Matsumoto, S. Ohta, C. Yoshimura, A. Ohyoshi, J. Chem.
Soc., Dalton Trans. 1985, 2575.
13] D. F. Evans, J. Chem. Soc. 1959, 2003.
14] G. A. Bain, J. F. Berry, J. Chem. Educ. 2008, 85, 532.
15] B. Weber, F. A. Walker, Inorg. Chem. 2007, 46, 6794.
yield 11.0 g (32.0 mmol, 81%). 1H NMR (400 MHz, CD
3
OD,
TMS): δ = 8.10 (s, 2 H), 7.30–7.18 (m, 4 H), 6.82–6.74 (m, 4 H),
[
[
[
1
3
3
.54 (t, 4 H), 3.28 (t, 1 H), 2.64 (t, 4 H), 1.83 (m, 4 H) ppm.
NMR (150 MHz, CD OD, TMS): δ = 162.4 (2C ), 132.3 (2C
31.3 (2C ), 118.9 (2C ), 117.9 (2C ), 116.9 (4C ), 55.0 (2C
2C ), 15.7 (2C ) ppm.
C
3
q
t
),
1
(
t
q
t
t
s
), 29.7 [16] J. Jullien, G. Juhasz, P. Mialane, E. Dumas, C. Mayer, J. Mar-
rot, E. Rivière, E. L. Bominaar, E. Münck, F. Sécheresse, Inorg.
Chem. 2006, 45, 6922.
s
s
[
FeIII(salten)Cl]:[34] To a solution of H
2
salten (3.70 g, 10.0 mmol)
[
[
[
17] A. Mokdad, R. W. Larsen, Inorg. Chim. Acta 2010, 363, 3338.
18] M. Sorai, Y. Maeda, H. Oshio, J. Phys. Chem. 1990, 51, 941.
19] S. Schenker, A. Hauser, R. M. Dyson, Inorg. Chem. 1996, 35,
4676.
in methanol (50 mL) was added a solution of anhydrous iron(III)
chloride (1.62 g, 10.0 mmol) in methanol (50 mL). After stirring
for 10 min at room temp., freshly distilled triethylamine (2.79 mL,
2
0.0 mmol) was added. The mixture was heated at 50 °C for 1 h
[20] J. P. Perdew, Phys. Rev. B 1986, 33, 8822.
[21] V. N. Staroverov, G. E. Scuseria, J. Tao, J. P. Perdew, J. Chem.
then allowed to cool to room temp. The black precipitate was col-
lected by filtration, washed three times with ethyl ether and dried
Phys. 2004, 120, 6898.
[
[
22] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785.
23] M. Swart, A. W. Ehlers, K. Lammertsma, Mol. Phys. 2004,
in vacuo; yield 2.70 g (6.3 mmol, 63%). C20
calcd. C 56.03, H 5.41, N 9.80, Cl 8.27; found C 56.13, H 5.48, N
0.32, Cl 8.20.
FeIII(salten)azpy]BPh
3 2
H23ClFeN O (428.7):
102, 2467.
1
[
24] H. Paulsen, L. Duelund, H. Winkler, H. Toftlund, A. X.
Trautwein, Inorg. Chem. 2001, 40, 2201.
25] S. Ye, F. Neese, Inorg. Chem. 2010, 49, 772.
26] A. Fouqueau, M. E. Casida, L. M. Lawson Daku, A. Hauser,
F. Neese, J. Chem. Phys. 2005, 122, 44110.
27] L. M. Lawson Daku, A. Vargas, A. Hauser, A. Fouqueau,
M. E. Casida, ChemPhysChem 2005, 6, 1393.
28] A. Fouqueau, S. Mer, M. E. Casida, L. M. Lawson Daku, A.
Hauser, T. Mineva, F. Neese, J. Chem. Phys. 2004, 120, 9473.
[
2
4
: To a solution of [Fe(salten)Cl] (0.98 g,
.29 mmol) in methanol (50 mL) was added the respective azopyr-
[
[
idine (0.42 g, 2.29 mmol) in methanol (5 mL) of methanol drop-
wise. The reaction was heated at 60 °C for 1 h and filtered whilst
warm. To the filtrate was added a solution of sodium tetraphen-
ylborate (0.78 g, 2.29 mmol) in methanol (10 mL). The reaction
was heated at 60 °C for 10 min. The precipitate was collected by
filtration and washed with cold methanol and ethyl ether. The halo-
gen analysis showed a significant residual chlorine content so the
crude product was redissolved in methanol (25 mL). To this was
added a solution of tetraphenylborate (0.28 g, 0.820 mmol) and
azopyridine (0.23 g, 1.26 mmol) in methanol (10 mL). The mixture
was heated at 60 °C for 1 h and stored in a freezer overnight. The
precipitate collected by filtration, washed with copious amounts of
[
[
[29] I. Nemec, R. Bo cˇ a, R. Herchel, T. Zden eˇ k, M. Gembický, W.
Linert, Monatsh. Chem. 2009, 140, 815.
30] S. Thies, H. Sell, C. Schütt, C. Bornholdt, C. Näther, F. Tuczek,
R. Herges, J. Am. Chem. Soc. 2011, 133, 16243.
31] S. Venkataramani, U. Jana, M. Dommaschk, F. D. Sönnichsen,
F. Tuczek, R. Herges, Science 2011, 331, 445.
32] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B.
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-
[
[
[
ethyl ether and dried in vacuo. [Fe(salten)(3-azpy)]BPh
.33 g (1.49 mmol, 65%). [Fe(salten)(4-azpy)]BPh (2): Yield 1.37 g
1.53 mmol, 67%). C55 52BFeN (895.7): calcd. C 73.75, H 5.85,
4
(1): Yield
1
(
4
H
6 2
O
Eur. J. Inorg. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
7

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
F. Tuczek et al.
FULL PAPER
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, Ö. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09,
revision A.1, Gaussian, Inc., Wallingford CT, 2009.
K.-S. Cheon, R. A. Cox, S.-R. Keum, E. Buncel, J. Chem. Soc.,
Perkin Trans. 2 1998, 1231.
[33]
[
34] R. Bo cˇ a, Y. Fukuda, M. Gembický, R. Herchel, R. Jaro sˇ cˇ iak,
W. Linert, F. Renz, J. Yuzurihara, Chem. Phys. Lett. 2000, 325,
411.
Received: November 3, 2011
Published Online: ᭿
8
www.eurjic.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 0000, 0–0

Job/Unit: A11227
/KAP1
Date: 26-01-12 15:37:21
Pages: 9
FeIII Spin-Crossover Complexes with Photoisomerizable Ligands
Spin-Crossover Complexes
[
(
Fe(salten)(3-azpy)]BPh
4
(1) and [Fe-
A. Bannwarth, S. O. Schmidt, G. Peters,
F. D. Sönnichsen, W. Thimm, R. Herges,
F. Tuczek* ........................................ 1–9
salten)(4-azpy)]BPh (2) have been synthe-
4
sized and characterized [salten = 4-aza-
heptamethylene-1,7-bis(salicylideneimin-
ate), azpy = phenylazopyridine]. Both exhi-
bit a low to high spin transition in the solid
state and are predominantly high spin in
solution. cis–trans isomerization of coordi-
nated azpy leaves their spin equilibria al-
most unaffected.
FeIII Spin-Crossover Complexes with Pho-
toisomerizable Ligands: Experimental and
Theoretical Studies on the Ligand-Driven
Light-Induced Spin Change Effect
Keywords: Spin crossover / Photoswitch-
ing / Density functional calculations / Iron /
Moessbauer spectroscopy / LD-LISC
Eur. J. Inorg. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
9