An inverted triplesalen ligand by a convergent synthesis and its influence on trinuclear FeIII 3 complexes

Article abstract of DOI:10.1007/s11426-012-4598-6

The inverted triplesalen ligand H₆feld^Me has been synthesized from 2,4,6-triformyl-phloroglucinol and a ketimine salen half-unit in a convergent synthesis. NMR, IR, and UV-vis spectroscopy reveal that this ligand is not in the O-protonated tautomer but in the N-protonated tautomer with substantial heteroradialene contribution. This ligand and the conventional triplesalen ligand talen^t-Bu₂ have been used to synthesize the trinuclear Fe^III complexes [(feld^Me)(FeCl) ₃] and talen^t-Bu₂(FeCl)₃]), respectively. The molecular structures of these complexes were obtained by single-crystal X-ray diffraction. Two trinuclear Fe^III complexes of [(feld^Me)(FeCl)₃] dimerize via two Fe-phenolate bonds, whereas due to steric hindrance no dimerization is observed for talen ^t-Bu₂(FeCl)₃]) . The structural data also reveal some heteroradialene contribution in the trinuclear complexes. Whereas UV-vis and Moebauer spectroscopy are not suitable to distinguish between the two complexes, FT-IR spectra show characteristic features due to the different substitution patterns of the conventional and the inverted triplesalen ligands. Another handle is provided by electrochemistry. Whereas both complexes exhibit an irreversible oxidation wave (0.94 V vs. Fc⁺/Fc for [(feld ^Me)(FeCl)₃] and 0.84 V vs. Fc⁺/Fc for talen^t-Bu₂(FeCl)₃]), which is assigned to the oxidation of the central backbone, higher potential oxidations are reversible for talen^t-Bu₂(FeCl)₃]) but irreversible for [(feld^Me)(FeCl)₃]. This is attributed to the reversible oxidation of the terminal phenolates in the di-tert-butyl substituted talen ^t-Bu₂(FeCl)₃]) in contrast to the mono-methyl-substituted phenolates in [(feld^Me)(FeCl)₃]. The magnetic properties of talen^t-Bu₂(FeCl)₃]) reveal a very small ferromagnetic coupling with significant zero-field splitting of the Fe^III S = 5/2 ions. In contrast, the dimerization of two trinuclear complexes in [(feld^Me)(FeCl)₃] results in antiferromagnetic interactions between the two phenolate-bridged Fe ^III ions, which mask the intra-trinuclear interactions transmitted by the central phloroglucinol backbone.

Full text of DOI:10.1007/s11426-012-4598-6

SCIENCE CHINA
Chemistry
June 2012 Vol.55 No.6: 951–966
doi: 10.1007/s11426-012-4598-6
• ARTICLES •
· SPECIAL TOPIC · Molecular Magnetism
An inverted triplesalen ligand by a convergent synthesis and its
influence on trinuclear FeIII complexes
3
FELDSCHER Bastian, KRICKEMEYER Erich, MOSELAGE Marc, THEIL Hubert,
HOEKE Veronika, KAISER Yvonne, STAMMLER Anja, BÖGGE Hartmut &
GLASER Thorsten*
Lehrstuhl für Anorganische Chemie I, Fakultät für Chemie, Universität Bielefeld, Universitätsstr. 25, D-33615 Bielefeld, Germany
Received February 2, 2012; accepted April 8, 2012; published online May 11, 2012
The inverted triplesalen ligand H6feldMe has been synthesized from 2,4,6-triformyl-phloroglucinol and a ketimine salen
half-unit in a convergent synthesis. NMR, IR, and UV-vis spectroscopy reveal that this ligand is not in the O-protonated tau-
tomer but in the N-protonated tautomer with substantial heteroradialene contribution. This ligand and the conventional triple-
salen ligand H6talent-Bu have been used to synthesize the trinuclear FeIII complexes [(feldMe)(FeCl)3] and [(talent-Bu2)(FeCl)3],
2
respectively. The molecular structures of these complexes were obtained by single-crystal X-ray diffraction. Two trinuclear
FeIII complexes of [(feldMe)(FeCl)3] dimerize via two Fe-phenolate bonds, whereas due to steric hindrance no dimerization is
observed for [(talent-Bu2)(FeCl)3]. The structural data also reveal some heteroradialene contribution in the trinuclear complexes.
Whereas UV-vis and Mößbauer spectroscopy are not suitable to distinguish between the two complexes, FT-IR spectra show
characteristic features due to the different substitution patterns of the conventional and the inverted triplesalen ligands. Another
handle is provided by electrochemistry. Whereas both complexes exhibit an irreversible oxidation wave (0.94 V vs. Fc+/Fc for
[(feldMe)(FeCl)3] and 0.84 V vs. Fc+/Fc for [(talent-Bu2)(FeCl)3]), which is assigned to the oxidation of the central backbone,
higher potential oxidations are reversible for [(talent-Bu2)(FeCl)3]) but irreversible for [(feldMe)(FeCl)3]. This is attributed to the
reversible oxidation of the terminal phenolates in the di-tert-butyl substituted [(talent-Bu2)(FeCl)3] in contrast to the
mono-methyl-substituted phenolates in [(feldMe)(FeCl)3]. The magnetic properties of [(talent-Bu2)(FeCl)3] reveal a very small
ferromagnetic coupling with significant zero-field splitting of the FeIII S = 5/2 ions. In contrast, the dimerization of two trinu-
clear complexes in [(feldMe)(FeCl)3] results in antiferromagnetic interactions between the two phenolate-bridged FeIII ions,
which mask the intra-trinuclear interactions transmitted by the central phloroglucinol backbone.
magnetic properties, iron complexes, spin-polarization, triplesalen
[3]. We then designed the triplesalen ligand system H6talenR,
1 Introduction
which combines the phloroglucinol bridging unit for
high-spin ground states by ferromagnetic couplings with a
salen-like coordination environment, which would introduce
magnetic anisotropy in order to fulfill the necessary re-
quirements for single-molecule magnets [4–6]. Trinuclear
CuII and VIV=O complexes of the triplesalen ligands also
In order to establish ferromagnetic interactions by the spin-
polarization mechanism [1, 2], we have developed in the
last decade trinucleating ligand systems based on
phloroglucinol (1,3,5-trihydroxybenzene) (Scheme 1). The
first ligand, H3L, was used to synthesize trinuclear CuII
complexes, which indeed exhibit ferromagnetic interactions
exhibit ferromagnetic interactions [4, 6, 7]. Structural anal-
ysis revealed that the trinuclear complexes [(talent-Bu
)
2
{Mt(solv)n}3]m+ showed a bowl-shaped molecular structure,
*Corresponding author (email: tglaser@uni-bielefeld.de)
© Science China Press and Springer-Verlag Berlin Heidelberg 2012
chem.scichina.com www.springerlink.com

952
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
Scheme 1
of which we took advantage by reaction with a hexacyano-
metallate [Mc(CN)6]3, which leads to the formation of hep-
triplesalen ligand system H6feldMe (Scheme 1(b)), which is
inverted with regard to the position of the imine and
ketimine units. We present a detailed comparison of the
properties of the triplesalen and the inverted triplesalen lig-
ands as well as their trinuclear FeIII complexes [(tal-
ent-Bu2)(FeCl)3] and [(feldMe)(FeCl)3]. The only character-
ized triplesalen FeIII complex was obtained with a chiral
derivative of the triplesalen ligand [18], for which depend-
ing on the diastereomeric form of the ligand used, an un-
coupled or a very weak antiferromagnetic coupling (J =
0.14 cm1) has been established.
tanuclear
complexes
[Mt6Mc]n+
([{(talent-Bu2)Mt3}2
{Mc(CN)6}]3+): [MnIII6CrIII]3+ [8], [MnIII6FeIII]3+ [9],
[MnIII6CoIII]3+ [10], and [MnIII6MnIII]3+ [11]. [MnIII6CrIII]3+
and [MnIII6MnIII]3+ exhibit single-molecule magnet behav-
ior. However, the analysis of the magnetic properties re-
vealed that the coupling of the three MnIII ions in the trinu-
clear triplesalen subunits is antiferromagnetic and not, as
intended and as observed in the CuII and VIV=O compounds,
ferromagnetic [8, 12]. We identified a reason for this unex-
pected coupling by the NMR spectroscopic characterization
of the ligand H3felden (Scheme 1(b)). In conjunction with
the work by MacLachlan [13, 14] and later other authors
[15], the analysis revealed that the central phloroglucinol
backbone is not in the O-protonated tautomeric form (I and
II, Scheme 1(c)), but in the N-protonated form (III and IV)
[16, 17]. The keto-enamine resonance structure IV prevails
in the description of the electronic structure, which resem-
bles a heteroradialene form. As radialenes are not fully de-
localized  systems but only cross-conjugated systems, this
electronic structure description should not transfer the
spin-polarization mechanism as efficiently as the intended
enol-imine form.
2 Experimental section
2.1 Preparation of compounds
Solvents and starting materials were of the highest commer-
cially available purity and used as received. 2,4,6-
Triformylphloroglucinol (6) was prepared according to the
reported procedure [13]. H6talent-Bu was synthesized as
2
described previously [5].
2.1.1 2-(1-(2-Amino-2-methylpropylimino)etyhl)-4-methyl-
phenol (5)
In this study, we established for the first time the inverted
2-Hydroxy-5-methylacetophenone (1.020 g, 6.79 mmol)

Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
953
and 1,2-diamino-2-methylpropane (1.133 g, 12.85 mmol,
1.9 eq.) were dissolved in methanol (5 mL) and stirred for
24 h at ambient temperature. The volatiles were removed
under vacuum, yielding a yellow oil which crystallizes over
time. Yield: 1.490 g (6.76 mmol, 99%). 1H NMR (500 MHz;
CDCl3)  = 16.19 (s, OH), 7.32 (d, 4J = 2.0 Hz, CArH), 7.10
0.52 mmol) in acetonitrile (18 mL). To the resulting solu-
tion triethylamine (87 mg, 0.86 mmol) dissolved in acetoni-
trile (2 mL) was added and the mixture was heated to reflux
for 20 min. The solution was filtered and the product crys-
tallizes over four days as purple plates which were washed
subsequently with acetonitrile and diethylether. Yield: 49
mg (0.045 mmol; 27%). MALDI-MS (positive ion mode)
3
4
3
(dd, J = 8.5 Hz, J = 2.0 Hz, CArH), 6.84 (d, J = 8.5 Hz,
CArH), 3.43 (s, CH2), 2.34 (s, CH3C=N), 2.29 (s, CArCH3),
1.52 (s, NH2), 1.25 (s, C(CH3)2). 13C NMR (125.75 MHz;
CDCl3)  = 172.1 (s, C=N), 161.4 (s, COH), 133.4 (s, CAr),
128.0 (s, CAr), 126.0 (s, CArCH3), 118.9 (s, CArCC=N),
118.3 (s, CAr), 62.4 (s, CH2), 50.3 (s, C(CH3)2), 28.8 (s,
C(CH3)2), 20.7 (s, CH3C=N), 14.7 (s, CArCH3). ESI-MS
m/z: 1085 [(feldMe)(FeCl)3]+, 1048 [(feldMe)(FeCl)2Fe]+,
−1
1013 [(feldMe)(FeCl)Fe2]+. IR: (KBr) v /cm : 2967m,

2920m, 2870m, 1578s, 1535s, 1503m, 1468w, 1406m,
1389m, 1356m, 1333w, 1308m, 1250w, 1225m, 1188m,
1159m, 1138m, 1078w, 997w, 908w, 824m, 802m, 662w,
629m, 584m, 536m, 521m, 488w. Elemental analysis:
Found: C 52.28, H 5.15, N 7.98. Calc. for [(feldMe)
(FeCl)3]∙0.5CH3CN∙1H2O (C49H57.5N6.5O7Fe3Cl3): C 52.39,
H 5.16, N 8.10.
(positive ion mode) m/z: 221 [M+H]+, 440 [2M+H]+. IR:
−1

(KBr) v /cm : 3036w, 2966m, 1620s, 1580m, 1501m,
1329m, 1294m, 1232m, 1205m, 1063m, 827m, 652m.
Elemental analysis: Found: C 70.97, H 8.99, N 12.63. Calc.
for C13H20N2O: C 70.87, H 9.15, N 12.72.
2.1.4 [(talent-Bu2)(FeCl)3]
H6talent-Bu (3.000 g, 2.70 mmol) and FeCl3∙6H2O (2.220 g,
2
8.21 mmol) were dissolved in methanol (250 mL) and
stirred at ambient temperature for ten minutes. To the re-
sulting purple mixture was added a solution of triethylamine
(1.500 g, 14.82 mmol, 5.5 eq) in methanol (50 mL) and
stirred another 2.5 h. The solution has been filtered and the
filtrate was evaporated to dryness, yielding a black solid
(5.750 g) which can be recrystallized via slow evaporation
of a solution in dimethylformamide. Yield: 2.700 g (1.98
mmol, 72%, hexagonal plates, suitable for single-crystal
X-ray diffraction). MALDI-MS (positive ion mode) m/z:
2.1.2 H6feldMe
2,4,6-Triformylphloroglucinol (6) (74 mg, 0.35 mmol) and
2-(1-(2-amino-2-methylpropylimino)ethyl)-4-methylphenol
(5) (232 mg, 1.05 mmol, 3 eq) were suspended in
dichloromethane (10 mL) and stirred at ambient
temperature for 12 h. The resulting yellow solution was
dried over sodium sulfate and the volatiles were removed
under vacuum, yielding the product as a yellow solid. Yield:
206 mg (0.25 mmol; 72%). 1H NMR (300 MHz; CDCl3)  =
15.16 (s, OH); 12.00, 11.90, 11.68, 11.66 (d, J= 13 Hz, all
NH Cs and C3h); 8.45, 8.40, 8.36, 8.29 (d, J= 13 Hz, all
CArCHCN Cs and C3h), 7.32 (m, CArH), 7.11 (m, CArH), 6.84
(m, CArH); 3.63, 3.62, 3.61, 3.47 (s, all CH2 Cs and C3h),
2.62, 2.38, 2.33, 2.31, 2.29, 2.28 (s, all N=CCH3 and
CArCH3, Cs and C3h), 1.52, 1.50, 1.29 (s, all C(CH3)2, Cs and
C3h). 13C NMR (125.75 MHz; CDCl3)  = 187.5, 184.9,
182.3 (all C=O), 173.5, 173.1, 173.0 (all C=N); 161.2,
160.1 (both COH); 153.9, 153.1, 153.0 (all CAr=C), 133.3
(CAr), 128.2 (CAr), 126.3 (CArC=N), 119.1, 118.7 (CArCH3),
118.0, 117.8, 117.7 (CAr), 104.7, 104.5 (both C=CHNH),
60.6, 60.5, 60.0 (all CH2N), 56.6, 56.5, 56.4 (all C(CH3)2),
27.6, 25.0 (both C(CH3)2), 20.6 (CArCH3), 14.9, 14.6 (both
C(CH3)C=N). ESI-MS (positive ion mode) m/z: 817
1379 [(talent-Bu2)(FeCl)3]+, 1342 [(talent-Bu2)(FeCl)2Fe]+. IR:
−1

(KBr) v /cm : 2951m, 2866m, 1672s, 1616s, 1560s,
1541m, 1485s, 1433m, 1410m, 1389m, 1364m, 1344w,
1308w, 1273m, 1254m, 1194w, 1175w, 1157w, 1145w,
1090w, 1063w, 1024w, 845m, 818m, 781w, 748w, 561w,
544m. Elemental Analysis: Found: C 58.96, H 7.30, N 7.11.
Calc. for [(talent-Bu2)(FeCl)3]∙2DMF (C75H110N8O8Fe3Cl3): C
59.05, H 7.27, N 7.34.
2.2 X-Ray crystallography
Crystal data for [(feldMe)(FeCl)3]·4CH3CN: M = 2333.96 g
mol1, C104H120Cl6Fe6N16O12, triclinic, space group P 1 , a =
11.383(4), b = 13.468(4), c = 19.061(6) Å,  = 104.016(12),
 = 96.947(13),  = 107.091(13)°, V = 2651.2(15) Å3, Z = 1,
 = 1.462 g/cm3,  = 1.018 mm1 , F (000) = 1210, Crystal
size = 0.21 × 0.18 × 0.11 mm3. A total of 43100 reflections
(2.25 <  < 25.00°) were collected of which 8940 reflec-
tions were unique (R(int)= 0.0397). Final R = 0.0705 for
5998 reflections with I0 >2(I0), R = 0.1187 for all reflec-
tions; max/min residual electron density 0.827 and 0.683
e Å3.
Me
+
Me
+
−1

[H6feld +H] , 839 [H6feld +Na] . IR: (KBr) v /cm :
2972m, 2922m, 2872w, 1601s, 1547m, 1501m, 1452m,
1393w, 1373w, 1325m, 1294m, 1234m, 1202m, 1177m,
1069m, 1017m, 824m, 781m, 654w, 523w. Elemental
analysis: Found: C 67.42, H 7.49, N 9.98. Calc. for
H6feldMe∙0.5CH2Cl2 (C48.5H61N6O6Cl): C 67.78, H 7.15, N
9.78.
Crystal data for [(talent-Bu2)(FeCl)3]·4.5DMF: M =
1708.35 g mol1, C82.50H127.50Cl3Fe3N10.50O10.50, triclinic,
2.1.3 [(feldMe)(FeCl)3]
A solution of H6feldMe (141 mg, 0.17 mmol) in acetonitrile
(10 mL) was added dropwise to a solution of FeCl3 (84 mg,
space group P 1 , a = 17.1253(15), b = 17.3279(15), c =

954
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
NO3/CH3CN. All potentials are referenced to the ferroce-
nium/ferrocene (Fc+/Fc) couple used as an internal standard.
The electrochemical cell was connected to an EG&G poten-
tiostat/galvanostat (model 273 A). Temperature-dependent
magnetic susceptibilities were measured by using a SQUID
magnetometer (MPMS XL-7 EC, Quantum Design) at 1 T
(2–290 K). For calculations of the molar magnetic suscepti-
bilities, M, the measured susceptibilities were corrected for
the underlying diamagnetism of the sample holder and the
sample by using tabulated Pascal’s constants. The JulX
program package was used for spin-Hamiltonian simula-
tions and fitting of the data by a full-matrix diagonalization
approach [21]. 57Fe Mössbauer spectra were recorded on an
alternating constant-acceleration spectrometer. The sample
temperature was maintained constant in a bath cryostat
(Wissel MBBC-HE0106). 57Co/Rh was used as the radiation
source. Isomer shifts were determined relative to -iron at
room temperature.
17.7747(16) Å,  = 96.598(4),  = 103.752(4),  =
112.287(4)°, V = 4615.6(7) Å3, Z = 2,  = 1.229 g/cm3,  =
0.610 mm1 , F(000) = 1818, Crystal size = 0.25 × 0.24 ×
0.12 mm3. A total of 90926 reflections (2.38 <  < 25.00°)
were collected of which 16091 reflections were unique
(R(int)= 0.0518). Final R = 0.0676 for 11651 reflections
with I0 > 2(I0), R = 0.0977 for all reflections; max/min
residual electron density 1.198 and 0.695 e Å3
Crystals of [(feldMe)(FeCl)3]·4CH3CN and [(talent-Bu
.
2
)
(FeCl)3]·4.5DMF were removed from the mother liquor,
coated with paraffin oil and immediately cooled to 100(2) K
on a Bruker Kappa APEX II diffractometer (four circle go-
niometer, Mo-K radiation, focusing graphite monochrom-
ator). Empirical absorption corrections using equivalent
reflections were performed with the program SADABS
2008/1 [19]. The structure was solved with SHELXS-97 [20]
and refined using SHELXL-97 [20]; structure graphics
were performed with DIAMOND 2.1 from K. Brandenburg,
Crystal Impact GbR, Bonn, Germany, 2001. [(feldMe)
(FeCl)3]·4CH3CN shows a 95%/5% disorder of the complex,
which could, however, only be resolved for the Fe and Cl
positions and can hence not be clearly interpreted. In [(tal-
ent-Bu2)(FeCl)3]·4.5DMF one tert-Bu-group of the complex
shows rotational disorder and is refined disordered on two
positions.
3 Results and discussion
3.1 Synthesis
Symmetrical salen-type ligands can easily be obtained by
reacting a diamine component, e.g. ethylenediamine, with
two equivalents of derivatives of salicylaldehyde or
2′-hydroxyacetophenone. The preparation of asymmetrical
salen-type ligands with two different kinds of phenols on
each side is not straightforward [22] and requires the prepa-
ration of a monocondensed diamine derivative, a so called
CCDC-863623 and 863624 contain the supplementary
crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic
Data Centre (12 Union Road, Cambridge CB21EZ) via
www.ccdc.cam.ac.uk/data_request/cif.
salen half-unit [23], which can undergo
a second
Schiff-base reaction. Our triplesalen ligands H6talenR
(Scheme 1(a)) contain three asymmetrically substituted eth-
ylendiamine units and therefore two principal reaction
routes can be formulated. The triplesalen ligands of the first
generation were synthesized in a divergent fashion, which
means that the half-unit contains the phloroglucinol subunit
(Scheme 2). This triplesalen half-unit 2 is synthesized in the
reaction of 2,4,6-triacetyl-1,3,5-trihydroxybenzol (triketone
1) with 1,2-diamino-2-methylpropane [24]. In this reaction
we utilize the observation of Elias and coworker that the
sterically less crowded amino function of this ethylenedia-
mine derivative does react with aldehydes as well as with
ketones whereas the more crowded amino function only
reacts with aldehydes [25]. As a consequence only one
amine function of the diamine reacts chemoselectively with
the triketone 1. In the second step, the remaining amino
function can be reacted with different derivatives of salic-
ylaldehyde [5, 24]. As an example, the reaction of the tri-
plesalen half-unit 2 with 3,5-di-tert-butyl-salicylaldehyde (3)
yields the ligand H6talent-Bu2 [5].
2.3 Other physical measurements
Infrared spectra (400–4000 cm1) of solid samples were
recorded on a Shimadzu FT-IR-8400S spectrometer as KBr
disks. UV-vis-NIR absorption spectra of the solutions were
measured on a Shimadzu UV-3101PC spectrophotometer in
the range of 200–1200 nm at ambient temperatures. ESI
mass spectra were recorded on a Bruker Esquire 3000 ion
trap mass spectrometer equipped with a standard ESI source.
MALDI TOF mass spectra were recorded with a VoyagerTM
DE instrument (PE Biosystems GmbH, Weierstadt, Ger-
many) mounted with a 1.2 m flight tube. Ionisation was
achieved using a LTB nitrogen laser MNL 100 (337 nm
beam wavelength, 3 ns pulse width, 3 Hz repetition rate).
The instrument Default Calibration was used for calibrating
the mass axis. 1H and 13C NMR spectra were measured on a
Bruker DRX500 or a Bruker Avance III 300 spectrometer
using the solvent as an internal standard. The electrochemi-
cal experiments were performed on Ar-flushed CH3CN so-
lutions containing 0.1 M [NBu4]PF6 in a classical three
electrode cell. The working electrode was either a platinum
or a glassy carbon electrode, the counter electrode a plati-
num wire and the reference electrode was Ag/0.01 M Ag
The second possible route for the preparation of triple-
salen ligands follows a convergent approach, where in the
first step a salen ketimine half-unit is required which reacts
in a second step with a tris-carbonyl derivative of phlorog-

Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
955
Scheme 2
lucinol. This approach has recently been realized for the
synthesis of triplesalophen ligands [26, 27]. Here, we apply
this convergent approach for the preparation of an inverted
triplesalen ligand. The reaction of two equivalents of
1,2-diamino-2-methylpropane with 2-hydroxy-5-methyl-
acetophenone (4) yielded chemoselectively the salen
ketimine half-unit 5 (Scheme 2). Only the less sterically
crowded amine reacted with the ketone function of the hy-
droxyacetophenone and after the excess of diamine had
been removed in vacuum the half-unit 5 could be obtained
as a yellow solid. The inverted triplesalen ligand H6feldMe
was synthesized by reacting three equivalents of this
half-unit with MacLachlan’s trialdehyde 6 (2,4,6-triformyl-
1,3,5-trihydroxybenzene) [13] (Scheme 2), analogously to
the previously reported ligands H3felden [16], H3felddien
[17] and H6baron [26, 27]. The success of the synthesis
has been confirmed by NMR, FT-IR, MS and elemental
analysis.
H6feldMe with FeCl3 yielded a solid of better solubility. A
solution of the ligand in acetonitrile was dropped to a solu-
tion of FeCl3 in the same solvent in the presence of tri-
ethylamine which afforded a purple solution. This solution
was filtered and upon standing at room temperature small
plates crystallized which have been identified as [(feldMe)
(FeCl)3]. The analogous iron compound of the ligand
H6talent-Bu can be obtained via the reaction of H6talent-Bu
2
2
with FeCl3·6H2O and triethylamine in methanol. The reac-
tion mixture was filtered and evaporated to dryness. The
crude product was recrystallized from dimethylformamide
to yield analytically pure [(talent-Bu2)(FeCl)3]. The identity
and purity of both complexes have been confirmed by
FT-IR, MS, elemental analysis and single-crystal X-ray
diffraction.
We have been particularly interested whether it would
also be possible to synthesize heptanuclear complexes of the
kind [Mt6Mc]n+ with the new ligand H6feldMe. The prepara-
tion of this class of complexes is well established for the
ligand H6talent-Bu2, and various heptanuclear complexes
with interesting properties could be obtained [2, 8, 9, 10].
However, all our attempts to obtain the analogous [Mt6Mc]n+
with the ligand H6feldMe failed. We assume that two reasons
may account for this result: (1) The trinuclear complexes of
H6talent-Bu2 seem to have some driving force for the dimeri-
The reactions of the ligand H6feldMe with NiII and CuII
metal salts afforded substances which showed poor solubil-
ity, so that no pure compounds could be obtained. This is
clearly an effect of the alkyl substituent, since the same
trend in differences of the solubility can be observed going
from the trinuclear complexes of the ligands H6talen to the
complexes of H6talent-Bu [5, 6]. However, the reaction of
2

956
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
zation of two [(talent-Bu2){M(solv)n}3]m+ units. For example,
two bowl-shaped [(talent-Bu2)Cu3] complexes can encapsu-
late two CH2Cl2 molecules [6]. The close inspection of the
heptanuclear aggregates in addition to the comparison of
tetranuclear [MnIII3CrIII]3+ [28] indicate that van der Waals
type interactions between the two triplesalen building
blocks account significantly to the driving force for dimeri-
zation [9, 28]. The terminal phenols of the ligand H6feldMe
are substituted only with one methyl group in para postion,
whereas H6talent-Bu2 exhibits two tert-butyl groups in ortho
and para-position. This might significantly reduce the driv-
ing force for the dimerization of two triplesalen units. (2)
We found the surprising result that the ligand H6feldMe hy-
drolyzes during the complexation reaction, evidenced by
severe amounts of 2-hydroxy-5-methylacetophenone (4) in
the reaction solution. This indicates that the terminal
ketimine function in H6feldMe is less stable against hydroly-
sis compared to the terminal aldimine function in
H6talent-Bu2. This made the procedure for the formation of
heptanuclear complexes, which has been established in our
group, inapplicable.
1
Figure 1 (a) H NMR spectrum of the ligand H6feldMe measured in
1
CDCl3; (b) simplification of the H NMR spectrum via substitution of NH
with deuterium oxide.
3.2 Spectroscopic characterization
Scheme 3
3.2.1 NMR spectroscopy on the ligands
In agreement with the formulation of the ligand H6feldMe in
its N-protonated heteroradialene form (Scheme 1(c), IV),
while the carbonyl function which is involved in two hy-
drogen bonds appears at 182.3 ppm. The third carbonyl
function of the Cs isomer with one hydrogen bond finds
itself in a comparable environment as the three equivalent
carbonyl functions of the Cs isomer. Therefore these two
carbonyl functions appear with 184.9 ppm at the same shift.
This assignment has been confirmed via HMQC and HMBC
experiments and agrees with the interpretation of the proton
NMR spectrum that the ligand has to be described as its
pure keto-enamine tautomer, since this chemical shift is
more characteristic for a C=O double bond than a C–OH
single bond [29].
1
the H NMR spectrum of the ligand H6feldMe exhibits four
doublets in the regions 11.5–10.5 ppm for the NH and
8.5–8.2 ppm for the vinylic protons (Figure 1). One set of
doublets represents the identical protons of the C3h isomer
and the additional set of three doublets can be assigned to
the three protons of the Cs isomer (Scheme 3). As discussed
for the ligands H3felden [16] and H3felddien [17], the NMR
spectra can be notably simplified via the exchange of the
acidic NH-protons with deuterium by addition of D2O to a
NMR sample (Figure 1(b)). After the exchange the doublets
between 8.5 and 8.2 ppm of the vinylic protons collapse to
four singlets. The three equivalent protons of the C3h isomer
can be assigned to the singlet at 8.35 ppm and the three sin-
glets at 8.44, 8.39 and 8.28 ppm to the three protons of the
particular arms of the Cs isomer. The ratio of the two iso-
mers in CDCl3 solution can be dedicated to be ca. 1:1.8
(C3h/Cs) and differs from a statistically ratio of 1:3 (C3h/Cs)
in favor of the more symmetric isomer probably due to the
maximization of the number of NH-ketone hydrogen bonds
in the C3h without competition [14]. The spectra show in the
range of 23–75 °C no temperature-dependence and no coa-
lescence.
The ligand H6talent-Bu2 in comparison shows also com-
plicated NMR spectra with multiple signals as well in the
proton as in the carbon NMR spectra [5]. However, the sig-
nals in the proton NMR are not as characteristic as in the
spectrum of H6feldMe since instead of the vinylic protons in
H6feldMe, which act as a good spectroscopic probe, the lig-
and H6talent-Bu2 exhibits methyl groups, which are not as
clearly influenced from the tautomeric variations of the
central unit. More characteristic are the chemical shifts of
the carbonyl functions in the 13C NMR spectra of the ligand.
Analogously to the ligands H6feldMe, H3felden [16] and
H3felddien [17], the ligand H6talent-Bu exhibits not well
The 13C NMR spectra of the ligand show four signals
between 188 and 182 ppm, which can be assigned to the
carbonyl resonances of the two isomers. The singlet with a
value of 187.5 ppm corresponds to the carbonyl function of
the Cs isomer which exhibits no hydrogen bonds (Scheme 3)
2
resolved signals in the region around 186 ppm and multiple
signals for the carbon atoms close to the central ring [5].
This is a clear indication that the ligand H6talent-Bu2 also has
to be described as the N-protonated heteroradialene tauto-

Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
957
mer. This has already been discussed in detail by us [16,
17].
The chemical shift of the CO resonance in 13C NMR
spectra has also been used to calculate tautomeric equilib-
rium constants (KT = [NH]/[OH]) [30, 31]. A shift in the
equilibrium from the OH tautomer to the NH tautomer
would produce a deshielding of this carbon which would
resonate as a consequence at a lower field. However, for the
calculation of the equilibrium constant it is necessary to
know the chemical shift of the extreme tautomeric struc-
tures. For example this has been done by Gawinecki et al.
who calculated equilibrium constants for benzo-annulated
N-salicylidenannilines and used values of 150 ppm for the
pure OH-tautomer and of 183 ppm for the pure NH-tauto-
mer [30]. Since the chemical shifts of our ligands are even
more shifted towards lower field it is apparent that the equi-
librium is almost completely on the side of the
NH-tautomer.
3.2.2 Infrared spectroscopy
The infrared spectra of the triplesalen ligands [5] and their
NiII [5] and CuII complexes [6] were difficult to understand
and to assign, although in salen-type ligands the assignment
of characteristic bands is usually straightforward. In this
respect, the C=N stretching vibration of an aldimine unit is
usually observed around 1635 cm1 and the C=N stretching
vibration of a ketimine group around 1615 cm1. These vi-
brations are shifted to lower energy by 10–20 cm1 upon
coordination to a metal ion. In the triplesalen ligand
Figure 2 FT-IR spectra of (a) the ligands H6feldMe and H6talent-Bu2 and (b)
of the complexes [(feldMe)(FeCl)3] and [(talent-Bu2)(FeCl)3].
the keto functions are assigned to the band at 1547 cm1,
while the band at 1452 cm1 is an overlay of several CH2
deformation bands and other highly delocalized nonspecific
vibrational modes.
H6talent-Bu a C=N stretch of the terminal aldimine group is
2
Despite the C=N stretching frequency of the terminal al-
clearly visible (Figure 2(a)), while no analogous band for a
ketimine is observed. However, a very intense broad band
appears around 1535 cm1, which is too low to be assigned
to a ketimine C=N stretching frequency. Interestingly, this
band can also be found in the corresponding trinuclear NiII
complexes and is the most resonance-enhanced band upon
irradiation at 24300 cm1 [5].
dimine units at 1629 cm1, the IR spectrum of H6talent-Bu
2
exhibits a strong band at 1529 cm1 with a strong shoulder
around 1590 cm1. An envelope of vibrational bands be-
tween 1450 and 1400 cm1 is also observed. This variation
by going from the trialdehyde-derived heteroradialene
H6feldMe to the triketone-derived heteroradialene H6talent-Bu
2
In the light of the heteroradialene nature of the extended
triplesalen ligands, we have performed an experimental and
computational IR study, which provides more understand-
ing for the special infrared spectroscopic features [32]. The
existence of both the triplesalen and the inverted triplesalen
allows us to test these theories. H6feldMe exhibits a very
strong band at 1601 cm1 with a shoulder around 1620 cm1
(Figure 2(a)). This latter shoulder and a medium band at
1501 cm1 are attributed to the terminal ketimine units and
assigned to the C=N and C=C stretching frequencies, re-
spectively. On the other hand, the intense band at 1601 cm1
together with the bands at 1547 and 1452 cm1 are a char-
acteristic signature of extended phloroglucinol ligands de-
rived from the trialdehyde 6. Interestingly, the highest band
at 1601 cm1 is not as usually assigned to the C=N stretch
but to the C=C stretch of the exocyclic doublebonds of the
heteroradialene backbone. The C=O stretching vibrations of
is predicted by the DFT calculations, which exhibit a re-
versed order in intensity for the C=C exocyclic stretches
and the C=O stretches. This leads to the assignment of the
strong band at 1535 cm1 in H6talent-Bu to normal modes
2
with strong C=O stretching character coupled with some
C–N stretching character and the shoulder around 1590
cm1 to the exocyclic C=C stretching vibrations. The enve-
lope of bands between 1450 and 1400 cm1 are assigned to
CH2 deformations and other highly delocalized vibrations
over the whole ligand. Thus, the triplesalen ligand
H6talent-Bu and the inverted triplesalen ligand H6feldMe ex-
2
hibit characteristic signatures in their FT-IR spectra.
The complex [(feldMe)(FeCl)3] exhibits a strong band at
1578 cm1 with a weak shoulder at 1600 cm1 (Figure 2(b)),
where the latter is attributed to the terminal C=N ketimine
stretching vibration. On the other hand, the strong bands at
1578 and 1503 cm1 seem to be the signature of the central

958
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
phloroglucinol-based unit. In [(talent-Bu2)(FeCl)3] the band at
CuII complexes of H6talent-Bu [5, 6]. We have applied sev-
2
1672 cm1 corresponds to the DMF molecules which are
included into the crystals of this compound in the process of
crystallization. The band at 1616 cm1 is assigned to the
terminal C=N aldimine stretching frequency, which is
shifted from 1629 cm1 in the ligand as expected. The sig-
natures for the central unit are thus at 1560 and 1485 cm1.
Again, the same trend is observed as in the ligand: the vi-
brations of the exocyclic C=C and C=O motifs of the trial-
dehyde-based complexes absorb at higher energies (1578/
1503 cm1) compared to the triketone-based complex (1560/
1485 cm1). In this respect, FT-IR spectroscopy represents a
simple tool to characterize triplesalen ligands and complex-
es as well as inverted triplesalen ligands and their complex-
es and demonstrates the presence of severe contribution of a
heteroradialene resonance formulation.
eral parameters for a quantitative description of the ligand
folding in our studies on the trinuclear triplesalen complex-
es [5, 6, 18] which have also been determined for [(feldMe)
(FeCl)3] and [(talent-Bu2)(FeCl)3] and are shown in Table 3.
While the regular ligand folding in the NiII and CuII com-
plexes of H6talent-Bu2 results in the bowl-shaped molecular
structure, the iron complexes of H6talent-Bu2 as well as of
H6feldMe show more distorted structures with no clear trend
in the parameters. The structure of [(feldMe)(FeCl)3] appears
to be almost flat with angles between the best planes
through the N2O2 donor set and the central phloroglucinol
ring () of Fe1 = 4.9°, Fe3 = 11.5 and Fe2 = 13.7°. This
correlates with Fe2 having the largest distance to the best
plane through the central phloroglucinol backbone (dFe2
=
0.40 Å, dFe1 =  0.21 Å, dFe3= 0.24 Å). This observed irreg-
ular distortion in [(feldMe)(FeCl)3] is even more pronounced
in [(talent-Bu2)(FeCl)3]. Here the angle has values of Fe1
=
3.2.3 Structural characterization
The crystal structures of the complexes [(feldMe)
(FeCl)3]·4CH3CN and [(talent-Bu2)(FeCl)3]·4.5DMF were
determined by single-crystal X-ray diffraction studies and
the molecular structures are displayed in Figure 3. Each
structure contains one neutral trinuclear complex in the
asymmetric unit in which one six-fold deprotonated ligand
coordinates three FeCl-moieties (Figure 3(a)). However, the
overall molecular structure of [(feldMe)(FeCl)3] reveals that
two trinuclear complexes are bridged via an additional bond
between a coordinated oxygen atom (O22) of one trinuclear
complex with an iron atom (Fe2) of another trinuclear com-
plex (Figure 3(b)) forming one hexanuclear complex. It is
interesting to note that a comparable preordering occurs in
[(talent-Bu2)(FeCl)3], but the Fe1···O12 distance of 5.03 Å is
too long for a bond probably due to repulsive interactions of
the tert-butyl groups of the ligands.
18.3°, Fe3 = 28.0° and Fe2 = 45.3° and the iron ions have
distances of dFe1= 0.87 Å, dFe2 =1.01 Å and dFe3 = 1.32 Å
to the best plane.
Selected interatomic distances and bond angels are sum-
marized in Tables 1 and 2. The mean Fe–N and Fe–Cl bond
lengths for [(talent-Bu2)(FeCl)3] are 2.08 and 2.23 Å, respec-
tively. The mean Fe–N bond length of [(feldMe)(FeCl)3] of
2.08 Å has the same value as in [(talent-Bu2)(FeCl)3] while
the Fe–Cl bond lengths are slightly longer (dFe1 = dFe3
2.25 Å, dFe2 = 2.29 Å).
=
Not only in the ligands [16, 17, 27] but also in the metal
complexes [16, 17] of the triplesalen ligands the heteroradi-
alene nature of the backbone prevails. One indication for
this in the iron complexes of H6feldMe and H6talent-Bu2 are
the mean C–C bond lengths of the central phloroglucinol
backbone. The mean C–C bond length in [(feldMe)(FeCl)3]
of 1.43 Å is clearly larger than the mean bond lengths of
phloroglucinol (1.39 Å) [34] or [(salen)(FeCl)] (1.38 Å)
[35], even though some contributions of keto-enol reso-
nance structures in phloroglucinol [36] and keto-enamine
resonance structures in [(salen)(FeCl)] have to be taken into
account. It is interesting to compare these values with the
mean C–C bond length of [(talent-Bu2)(FeCl)3] of 1.42 Å
which is slightly shorter. This might be a hint of a less
strong contribution of a radialene-like resonance structure in
the complex of [(talent-Bu2)(FeCl)3]. This interpretation cor-
relates with the mean C=O bond lengths of the central
phloroglucinol ring which are longer in [(talent-Bu2)(FeCl)3]
(1.32 Å) than the corresponding bond length in [(feldMe)
(FeCl)3] (1.30 Å) and therefore the double bond character is
less pronounced. The terminal Fe–O bonds are in average
slightly shorter than the central Fe–O bonds (1.89 Å versus
1.91 Å in [(talent-Bu2)(FeCl)3] and 1.89 Å versus 1.93 Å
[(feldMe)(FeCl)3]), demonstrating the slightly lower coordi-
nation ability of the central oxygen donor atoms that have
severe ketone-character.
In the structure of [(talent-Bu2)(FeCl)3] all iron ions are in
a slightly distorted square-pyramidal coordination environ-
ment ( values [33]: Fe1 = 0.17, Fe2 = 0.10, Fe3 = 0.08)
coordinated by two phenolate and two imine donors. This
also holds for the structure of [(feldMe)(FeCl)3] (Fe1 = 0.23,

Fe3 = 0.30) with exception of Fe2, which is in an octahedral
coordination environment coordinated by two imine donors
and two phenolate donors where the terminal donor acts as a
bridging donor to a second trinuclear complex. The iron
ions which are in the square-pyramidal environment are
extruded out of the best plane through the N2O2 donor at-
oms towards the apical chloride ions with Fe–Cl distances
of 2.78 Å for Fe1 and Fe2, and 2.79 Å for Fe3 in [(tal-
ent-Bu2)(FeCl)3] and 2.73 and 2.75 Å for Fe1 and Fe3 in
[(feldMe)(FeCl)3]. In both structures two of the FeCl moie-
ties ions are located on opposite sides of the plane through
the central ring.
Trinuclear triplesalen complexes are known to exhibit
various degrees of ligand folding which results in an overall
bowl-shaped molecular structure for the trinuclear NiII and

Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
959
Figure 3 Molecular structures of [(feldMe)(FeCl)3] (left) and [(talent-Bu2)(FeCl)3] (right). (a) The asymmetric units of both structures consist of one trinuclear
complex; (b) overall molecular structures with the interactions of two molecules of [(feldMe)(FeCl)3] (left) and [(talent-Bu2)(FeCl)3] (right).
29500 cm1 with  = 38 × 103 M1 cm1 and 33400 cm1
3.2.4 Electronic absorption spectroscopy
with = 32 × 103 M1 cm1. The spectra of the ligand
The electronic absorption spectra of the ligands H6feldMe
and H6talent-Bu2 as well as of the corresponding iron com-
plexes [(feldMe)(FeCl)3] and [(talent-Bu2)(FeCl)3] are shown
in Figure 4. The spectra of H3felden and H2salen are in-
cluded for comparison. H2salen exhibits one absorption at
32000 cm1 involving the imine group (= 8 × 103 M1
cm1) and one above 39000 cm1 (= 20 × 103 M1 cm1)
which has been assigned to the phenolic chromophores [37].
The ligand H6feldMe exhibits strong absorption maxima at
H6talent-Bu and also of the ligand H3felden show almost
2
identical absorption features in this region. Therefore, these
characteristic absorptions can neither be assigned to a
threefold superposition of the electronic structure of a salen
ligand nor to the terminal phenol-imine units in the triple-
salen ligands, but must be due to a common motif in the
electronic structure of all three ligands, which is the central
heteroradialene backbone. The ligands H6feldMe and

960
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
Table 1 Selected interatomic distances (Å) and angles (°) for [(feldMe)(FeCl)3]2∙4CH3CN
Fe1-O11
Fe1-O12
Fe1-N11
Fe1-N12
Fe1-Cl1
Fe2-O21
Fe2-O22
Fe2-N21
Fe2-N22
Fe2-Cl2
Fe3-O31
Fe3-O32
Fe3-N31
Fe3-N32
Fe3-Cl3
C1-O11
C3-O21
C5-O31
N11-C(11
N21-C21
N31-C31
C1-C2
1.927(4)
1.854(5)
2.046(5)
2.105(5)
2.252(3)
1.938(5)
1.974(4)
2.089(6)
2.112(5)
2.286(3)
1.933(5)
1.854(6)
2.059(5)
2.090(6)
2.241(3)
1.313(7)
1.287(8)
1.309(8)
1.316(8)
1.294(8)
1.303(8)
1.423(9)
1.427(8)
1.417(9)
1.435(9)
1.426(9)
1.431(9)
1.427(9)
1.432(9)
1.410(9)
O11-Fe1-N11
O11-Fe1-N12
O11-Fe1-Cl1
O11-Fe1-O12
O12-Fe1-N11
O12-Fe1-N12
O12-Fe1-Cl1
N11-Fe1-N12
N11-Fe1-Cl1
N12-Fe1-Cl1
O21-Fe2-N21
O21-Fe2-N22
O21-Fe2-Cl2
O21-Fe2-O22
O22-Fe2-N21
O22-Fe2-N22
O22-Fe2-Cl2
N21-Fe2-N22
N21-Fe2-Cl2
N22-Fe2-Cl2
O31-Fe3-N31
O31-Fe3-N32
O31-Fe3-Cl3
O31-Fe3-O32
O32-Fe3-N31
O32-Fe3-N32
O32-Fe3-Cl3
N31-Fe3-N32
N31-Fe3-Cl3
N32-Fe3-Cl3
C1-O11-Fe1
C3-O21-Fe2
C5-O31-Fe3
87.59(19)
156.9(2)
104.09(18)
92.8(2)
143.4(3)
86.3(2)
112.9(2)
79.7(2)
102.38(18)
97.47(19)
88.2(2)
166.9(2)
92.0(2)
105.60(19)
157.1(2)
84.31(19)
97.01(17)
79.8(2)
100.7(2)
95.3(2)
87.1(2)
157.6(3)
103.90(17)
94.1(3)
C2-C3
C3-C4
C4-C5
139.6(3)
85.7(3)
C5-C6
C1-C6
109.4(3)
78.9(2)
C2-C11
C4-C21
C6-C31
109.4(2)
97.2(2)
132.4(4)
132.3(4)
134.9(4)
H3felden and therefore consistent with the assignment of
these transition to the terminal phenol-imine chromophores.
The spectra of the complexes [(feldMe)(FeCl)3] and [(tal-
ent-Bu2)(FeCl)3] show very similar absorptions besides a few
differences in the intensities of the maxima above 30000
cm1. Although we do not intend to assign the transitions
above 25000 cm1, it is apparent that the band structures for
the heteroradialene backbone in the ligand are not reflected
directly in the spectra of the complexes. This implies a per-
turbation of the heteroradialene central backbone by coor-
dination to FeIII. The coordination of FeIII is clearly visible
by the strong band at ~19000 cm1 which is typical for a
FeIII-phenolate chromophore and is assigned to a phenolic
oxygen(p) to FeIII(d*) LMCT transition [18, 38].
Figure 4 Electronic absorption spectra of H6feldMe, H6talent-Bu2, H3felden,
H2salen, and [(feldMe)(FeCl)3] measured in CH3CN, and [(talent-Bu2)(FeCl)3]
measured in CH2Cl2.
3.2.5 Mößbauer
Mößbauer spectra of [(talent-Bu2)(FeCl)3] and [(feldMe
)
(FeCl)3] at 80 K are shown in Figure 5. Both spectra exhibit
a single asymmetrical quadrupole doublet with isomeric
shifts () and quadrupole splittings (|∆EQ|) of  = 0.43 mm
s1, |∆EQ| = 1.43 mm s1 for [(talent-Bu2)(FeCl)3], and  =
H6talent-Bu show moreover another strong absorption at
2
39300 cm1 with  = 28 × 103 M1 cm1 and 38000 cm1
with  = 24 × 103 M1 cm1, respectively, which is absent in

Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
961
Table 2 Selected interatomic distances (Å) and angles (°) for [(talent-Bu2)(FeCl)3] 4.5 C3H7NO
Fe1-O11
Fe1-O12
Fe1-N11
Fe1-N12
Fe1-Cl1
Fe2-O21
Fe2-O22
Fe2-N21
Fe2-N22
Fe2-Cl2
Fe3-O31
Fe3-O32
Fe3-N31
Fe3-N32
Fe3-Cl3
C1-O11
C3-O21
C5-O31
N11-C11
N21-C21
N31-C31
C1-C2
1.927(3)
1.897(3)
2.084(4)
2.096(4)
2.2304(14)
1.892(3)
1.878(3)
2.074(4)
2.076(4)
2.2436(17)
1.910(3)
1.897(3)
2.080(4)
2.072(4)
2.2301(15)
1.318(5)
1.313(5)
1.329(5)
1.306(6)
1.301(6)
1.317(6)
1.414(6)
1.426(6)
1.417(6)
1.429(7)
1.412(6)
1.426(6)
1.467(6)
1.460(6)
1.466(6)
O11-Fe1-N11
O11-Fe1-N12
O11-Fe1-Cl1
O11-Fe1-O12
O12-Fe1-N11
O12-Fe1-N12
O12-Fe1-Cl1
N11-Fe1-N12
N11-Fe1-Cl1
N12-Fe1-Cl1
O21-Fe2-N21
O21-Fe2-N22
O21-Fe2-Cl2
O21-Fe2-O22
O22-Fe2-N21
O22-Fe2-N22
O22-Fe2-Cl2
N21-Fe2-N22
N21-Fe2-Cl2
N22-Fe2-Cl2
O31-Fe3-N31
O31-Fe3-N32
O31-Fe3-Cl3
O31-Fe3-O32
O32-Fe3-N31
O32-Fe3-N32
O32-Fe3-Cl3
N31-Fe3-N32
N31-Fe3-Cl3
N32-Fe3-Cl3
C1-O11-Fe1
C3-O21-Fe2
C5-O31-Fe3
85.12(14)
141.63(14)
105.04(10)
93.13(13)
151.74(14)
86.36(14)
105.11(11)
77.99(14)
102.56(11)
112.13(11)
85.39(14)
144.01(17)
110.04(12)
92.20(14)
150.14(19)
87.13(16)
107.82(13)
77.90(17)
100.90(14)
104.34(13)
83.82(15)
143.57(17)
106.68(11)
92.56(13)
148.38(15)
86.69(14)
107.99(11)
78.27(15)
103.10(12)
108.15(12)
123.2(3)
C2-C3
C3-C4
C4-C5
C5-C6
C1-C6
C2-C11
C4-C21
C6-C31
134.2(3)
120.9(3)
tive potentials [(talent-Bu2)(FeCl)3] exhibits three reduction
0.43 mm s1, |∆EQ| = 1.18 mm s1 for [(feldMe)(FeCl)3]. The
three distinct FeIII sites in both complexes are not resolved.
However, these parameters are typical for ferric high-spin
ions (Si = 5/2) in a square-pyramidal coordination environ-
ment of N2O2FeCl salen complexes [35, 39].
waves at Ep,red = 1.00, 1.14, and 1.24 V, again with sig-
nificant backcurrents now for reoxidation. The occurrence
of an irreversible oxidation in [(talent-Bu2)(FeCl)3] followed
by several almost chemically reversible oxidation implies
that these processes occur in different parts of the molecule.
The comparison to the CVs of [(feldMe)(FeCl)3] allow a
tentative assignment of these electron transfer processes.
The first irreversible oxidation must be associated with the
central heteroradialene backbone as it appears in both com-
plexes. The slightly lower potential in [(talent-Bu2)(FeCl)3] is
caused by the stronger electron donating capability of the
di-tert-butyl-phenolates in [(talent-Bu2)(FeCl)3] compared to
the methyl-phenolates in [(feldMe)(FeCl)3]. The higher po-
tential oxidation processes are associated with the terminal
phenolates as it is well known that coordinated 2,4-di-tert-
butyl phenolates can be reversible oxidized to the corre-
sponding phenoxy radicals while methyl substituents only
lead to irreversible processes [40].
3.3 Electrochemistry
Cyclic and square-wave voltammograms of [(feldMe)(FeCl)3]
and [(talent-Bu2)(FeCl)3] have been recorded in acetonitrile
solutions containing 0.1 M [NBu4]PF6 as supporting elec-
trolyte and representative examples of the CVs are provided
in Figure 6. The CV of [(feldMe)(FeCl)3] exhibits two irre-
versible oxidation waves at Ep,ox = 0.94 V and Ep,ox = 1.12 V
(all potentials are referenced vs. Fc+/Fc) and two irreversi-
ble reduction waves at Ep,red = 1.05 V and Ep,red = 1.33 V.
The CV of [(talent-Bu2)(FeCl)3] shows one irreversible oxi-
dation wave at Ep,ox = 0.84 V, which causes a second wave
at Ep,red = 0.40 V followed by some unresolved oxidation
waves with maxima at Ep,ox = 1.01 V and Ep,ox = 1.09 V
which exhibits significant reductive back currents. At nega-
Interestingly, the oxidations of [(talent-Bu2)(FeCl)3] are
anodically shifted in comparison to [(talent-Bu2)Ni3] (re-

962
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
Table 3 Selected structural properties of the iron complexes [(talent-Bu
)
2
(FeCl)3] and [(feldMe)(FeCl)3]
[(talent-Bu2)(FeCl)3]
[(feldMe)(FeCl)3]
0.21
0.40
d
a)(Å)
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
0.87
1.01
1.32
18.3
45.3
28.0
17.1
10.8
4.0
0.24
b)(°)
c)(°)
4.9
13.7
11.5
9.1
40.2
10.2
 d)(°)
24.0
55.9
32.0
49.3
9.0
11.2
43.7
21.7
terminal e)(°)
27.7
16.1
55.1
16.6
20.3
36.1
0.17
0.10
0.08
15.3
central f)(°)
24.3
52.7
18.0
g)


0.23
n.a.
0.30
a) d is the shortest distance of an iron ion from the best plane formed by
the six carbon atoms of the central benzene ring of the phloroglucinol
backbone. A negative value corresponds to a displacement to the other side
of the plane; b)  is the angle between the best planes of (1) N2O2 and (2)
benzene of the central phloroglucinol backbone; c)  is the angle between
the best planes of (1) N2O2 and (2) benzene of the terminal phenolate; d) 
is the angle between the best planes of (1) benzene of the central
phloroglucinol backbone and (2) benzene of the terminal phenolate; e) bent
angle = 180°<) (Fe-X1central-X2central); f) bent angle = 180°<) (Fe-
X1term-X2term); g) values[33] for the deviation from square-pyramidal (= 0)
to trigonal-bipyramidal ( = 1).
Figure 5 57Fe Mößbauer spectra of (a) [(feldMe)(FeCl)3] and (b) [(tal-
ent-Bu2)(FeCl)3] at 80 K.
versible 1 e transfer of +0.22 V and reversible 2 e transfer
at +0.55 V)[5] and [(talent-Bu2)Cu3] (reversible 1 e transfer at
+0.26 V and reversible 2 e transfer at +0.59 V) [6]. The
higher potential in [(talent-Bu2)(FeCl)3] reflects the higher
electron density demand of FeIII compared to NiII and CuII
(stronger Lewis acidity). The irreversibility of the first oxi-
dation in [(talent-Bu2)(FeCl)3] might be attributed to the de-
localization of the -radical character from the heteroradi-
alene backbone to the FeIII ions by its empty t2g orbitals
which is not feasible in CuII and NiII.
Figure 6 CVs of (a) [(feldMe)(FeCl)3] recorded at a platinum working
electrode and (b) [(talent-Bu2)(FeCl)3] recorded at a glassy carbon working
electrode at 20°C in 0.1 M [NBu4]PF6 CH3CN solutions at a scan rate of
200 mV s1.
3.4 Magnetochemistry
Temperature-dependent measurements of the magnetic sus-
ceptibility (SQUID, 2–290 K) at different magnetic fields of
samples of [(feldMe)(FeCl)3] and [(talent-Bu2)(FeCl)3] were
performed. The temperature-dependence of the effective
magnetic moments, eff, at 1 T differs significantly for the
two complexes (Figure 7) and can be related to structural
differences.
290 K is eff = 9.43 B (Figure 7(a)) which is smaller than
the expected value of eff for three non-interacting high-spin
FeIII (S = 5/2) ions (eff = 10.25 B, g = 2.00). By decreasing
the temperature,eff decreases slightly to 50 K from where a
The effective magnetic moment of [(feldMe)(FeCl)3] at

Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
963
stronger decrease is apparent. This is typical of a weak an-
tiferromagnetic coupling. In contrast, eff is 10.29 B at 290
K for [(talent-Bu2)(FeCl)3] (Figure 7(b)) and increases slight-
ly by lowering the temperature which indicates weak ferro-
magnetic interactions.
The dimerization of mononuclear salen FeIIICl complexes
in the solid state by coordination of a phenolate oxygen at-
om of one molecule to the FeIII ion of a second molecule, is
a well known phenomenon [41, 42] and results in antifer-
romagnetic coupling constants J in the range of 7 to 8
cm1 [42]. The dimerization of two trinuclear complexes
[(feldMe)(FeCl)3] by analogous Fe-OPh bonds is thus most
likely the origin of the antiferromagnetic interactions as this
behavior is not observed in [(talent-Bu2)(FeCl)3] which does
not dimerize. As the intra-trinuclear coupling is very small
(vide infra) this inter-trinuclear interaction masks the in-
tra-trinuclear coupling. A simulation for the dimerized hex-
anuclear unit [(feldMe)(FeCl)3]2 is highly overparameterized
and therefore we refrain from simulating these data.
However, we tried to simulate the temperature-depend-
ence of eff for [(talent-Bu2)(FeCl)3] with the appropriate
spin-Hamiltonian (1) for an equilateral high-spin FeIII3 (Si =
5/2) triangle by a full-matrix diagonalization approach in-
cluding Heisenberg-Dirac-van Vleck (HDvV) exchange and
Zeeman interactions using the program package JulX [21].
3
H   2J S S  S S  S S

g  S B
(1)

3 


Figure 7 Temperature-dependence of the effective magnetic moment, µeff
at 1 T of (a) [(feldMe)(FeCl)3] and (b) [(talent-Bu2)(FeCl)3]. The solid line in
(b) is a simulation with J = 0.026 cm1, D = 2.1 cm1, w = 0 K, g = 1.998.
,

1
2
2
3
1
i
B
i
i1
The value of µeff increases only very slightly from 10.25
µB at 290 K to a maximum of 10.31 µB at 25 K and then
shows a strong decrease to a value of 7.14 µB at 2 K. Three
origins are known for such a decrease of eff at low temper-
ature: 1. zero-field splitting, 2. intermolecular antiferro-
magnetic interactions, and 3. saturation effects, which are
taken into account in the applied program package. Inter-
molecular interactions are modelled by a Curie-Weiss cor-
rection (w). The very slight increase of µeff requires a very
small value for the ferromagnetic coupling. Pure saturation
effects cannot account for the steep decrease at low temper-
atures. Indeed, a very sizeable zero-field splitting or inter-
molecular antiferromagnetic interaction must be considered.
As a larger intermolecular interaction as a perturbation to a
smaller intramolecular interaction seems to be inappropriate,
we performed simulations using a sizeable zero-field split-
ting of the FeIII high-spin ions. Several parameter sets of J,
D, and w provide almost as good reproduction of the ex-
perimental data. For example, assuming a coupling constant
of 0.037 cm1 requires a zero-field splitting of 3.1 cm1,
while a ferromagnetic coupling of 0.026 cm1 requires a
zero-field splitting of 2.1 cm1. Even with the same cou-
pling constant, several parameter sets of D and wprovide
almost the same simulation. Thus the three parameters are
highly correlated. The simulation in Figure 7(b) corresponds
to the parameter set J = 0.026 cm1, D = 2.1 cm1, w = 0
K, and g = 1.998. However, we would like to stress that this
is not a unique data set and from the multitude of simula-
tions we estimate the following range of values: J = 0.02 ±
0.01 cm1, D = 1.5 ± 1.0 cm1, and W= 0.05 ± 0.05 K.
Irrespective of the exact values, the conclusion of this anal-
ysis is that the FeIII ions in [(talent-Bu2)(FeCl)3] are very
weakly ferromagnetically coupled and the zero-field split-
ting of the FeIII high-spin ions has a sizeable contribution.
This result is in agreement to that of a related trinuclear FeIII
complex of a chiral derivative of the triplesalen ligand [18].
4 Conclusion
We have synthesized the new ligand H6feldMe, which repre-
sents the first inverted triplesalen ligand. This ligand has
been prepared not according to our established divergent
protocol for triplesalen ligands, but according to our new
convergent protocol as for the triplesalophen ligands. In this
class of ligands the ketimine and aldimine functions of the
asymmetric salen subunits are inverted so that the central
ring is substituted with the aldimine functions, which tau-
tomerize to the N-protonated radialene-like structure, and
the terminal rings are substituted with the ketimine func-

964
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
Munno G, Julve M, Cano J, Ruiz R, Caneschi A. Spin polarization
and ferromagnetism in two-dimensional sheetlike cobalt(II) polymers:
[Co(L)2(NCS)2] (L = pyrimidine or pyrazine). Angew Chem Int Ed,
1998, 37: 135–138; Fernández I, Ruiz R, Faus J, Julve M, Lloret F,
Cano J, Ottenwaelder X, Journaux Y, Munoz C. Ferromagnetic cou-
pling through spin polarization in a dinuclear copper(II) metallacy-
clophane. Angew Chem Int Ed, 2001, 40: 3039–3042; Pardo E,
Ruiz-García R, Cano J, Ottenwaelder X, Lescouezec R, Journaux Y,
Lloret F, Julve M. Ligand design for multidimensional magnetic ma-
terials: A metallosupramolecular perspective. Dalton Trans, 2008,
2780–2805; Glaser T, Theil H, Heidemeier M. Spin-polarization in
1,3,5-trihydroxybenzene-bridged first-row transition metal complexes.
C R Chim, 2008, 11: 1121–1136; Plaul D, Geibig D, Görls H, Plass
W. Mono- and dinuclear copper(II) complexes with meta-phenylene-
bridging ligands: Synthesis, structure and magnetic properties. Poly-
hedron, 2009, 28: 1982–1990
tions. We were interested whether this inversion would have
any effects on the formation and the properties of the hep-
tanuclear compounds of the kind [Mt6Mc]n+. Interestingly, it
proved to be impossible to prepare such compounds with
this ligand, probably due to two reasons: (1) Supramolecu-
lar van der Waals type interactions between the two triple-
salen ligands have severe contributions to the driving force
of the formation of heptanuclear compounds. Due to the
different substitution pattern of H6feldMe in comparison to
H6talent-Bu2 this driving force is significantly reduced; (2) we
found the surprising result that the terminal ketimine func-
tion in H6feldMe appeared to be less stable against hydrolysis
compared to the aldimine function in H6talent-Bu2. This made
the established procedure for the formation of heptanuclear
complexes inapplicable.
2
3
Glaser T. Rational design of single-molecule magnets: A supramo-
lecular approach. Chem Commun, 2011, 47: 116–130
Herein, we presented two FeIII trinuclear complexes of
the new ligand H6feldMe and of the well established ligand
H6talent-Bu2. Two trinuclear FeIII complexes of [(feldMe)
(FeCl)3] dimerize via two Fe-phenolate bonds, whereas due
to steric hindrance no dimerization is observed for [(tal-
ent-Bu2)(FeCl)3]. The structural data also reveal some heter-
oradialene contribution in the trinuclear complexes. How-
ever, it seems that this contribution is less pronounced in
[(talent-Bu2)(FeCl)3]. The differences of the two complexes
are clearly visible in the electrochemical and IR spectro-
scopic properties while UV-vis and Mößbauer spectroscopy
are not suited to differentiate between the two complexes.
The magnetic properties of [(talent-Bu2)(FeCl)3] reveal that
the three FeIII (S = 5/2) ions are very weakly ferromagneti-
cally coupled and show a significant zero-field splitting.
The intra-trinuclear interactions of the FeIII ions in [(feldMe)
(FeCl)3] are masked due to antiferromagnetic interactions
between two phenolate-bridged FeIII ions, which is a result
of the dimerization of two trinuclear complexes.
Glaser T, Gerenkamp M, Fröhlich R. Targeted synthesis of ferro-
magnetically coupled complexes with modified 1,3,5-trihy-
droxybenzene ligands. Angew Chem Int Ed, 2002, 41: 3823–3825
Glaser T, Heidemeier M, Grimme S, Bill E. Targeted ferromagnetic
coupling in a trinuclear copper(II) complex: Analysis of the St = 3/2
spin ground state. Inorg Chem, 2004, 43: 5192–5194
Glaser T, Heidemeier M, Fröhlich R, Hildebrandt P, Bothe E, Bill E.
Trinuclear nickel complexes with triplesalen-ligands: Simultaneous
occurrence of mixed valence and valence tautomerism in the oxidized
species. Inorg Chem, 2005, 44: 5467–5482
Glaser T, Heidemeier M, Strautmann JBH, Bögge H, Stammler A,
Krickemeyer E, Huenerbein R, Grimme S, Bothe E, Bill E. Trinucle-
ar copper complexes with triplesalen ligands: Geometric and elec-
tronic effects on the ferromagnetic coupling via the spin-polarization
mechanism. Chem Eur J, 2007, 13: 9191–9206
Theil H, Freiherr von Richthofen C-G, Stammler A, Bögge H, Glaser
T. Ferromagnetic coupling by the spin-polarization mechanism in a
trinuclear viv triplesalen complex. Inorg Chim Acta, 2008, 361:
916–924
4
5
6
7
8
Glaser T, Heidemeier M, Weyhermüller T, Hoffmann R-D, Rupp H,
Müller P. Property-oriented rational design of single-molecule mag-
nets: A C3-symmetric Mn6Cr complex based on three molecular
building blocks with a spin ground state of St = 21/2. Angew Chem Int
Ed, 2006, 45: 6033–6037
9
Glaser T, Heidemeier M, Krickemeyer E, Bögge H, Stammler A,
Fröhlich R, Bill E, Schnack J. Exchange interactions and zero-field
splittings in C3-symmetric MnIII6FeIII: using molecular recognition
for the construction of a series of high spin complexes based on the
triplesalen ligand. Inorg Chem, 2009, 48: 607–620
This work was supported by the Deutsche Forschungsgemeinschaft
(FOR945 ‘Nanomagnets: from Synthesis via Interactions with Surfaces to
Function’), the Fonds der Chemischen Industrie, and Bielefeld University.
10 Krickemeyer E, Hoeke V, Stammler A, Bögge H, Schnack J, Glaser
T. Synthesis and characterization of the heptanuclear [MnIII6CoIII]3+
triplesalen complex: Evidence for exchange pathways involving
low-spin CoIII. Naturforsch Z, 2010, 65b: 295–303
11 Hoeke V, Gieb K, Heidemeier M, Krickemeyer E, Müller P, Chibo-
taru LF, Stammler A, Bögge H, Schnack J, Schröder C, Glaser T.
Manuscript in preparation.
1
McConnell HM. Ferromagnetism in solid free radicals. J Chem Phys,
1963, 39: 1910; Longuet-Higgins HC. Some studies in molecular or-
bital theory I. Resonance structures and molecular orbitals in unsatu-
rated hydrocarbons. J Chem Phys, 1950, 18: 265–274; Iwamura H.
High-spin organic molecules and spin alignment in organic molecular
assemblies. Adv Phys Org Chem, 1990, 26: 179–253; Cano J, Ruiz E,
Alvarez S, Verdaguer M. Spin density distribution in transition metal
complexes: Some thoughts and hints. Comments Inorg Chem, 1998,
20: 27–56; Dougherty DA. Spin control in organic molecules. Acc
Chem Res, 1991, 24: 88–94; Karafiloglou P. Through bond interac-
tion of two radical centers: Analysis of the spin polarization and re-
lated mechanisms in linear  diradicals. J Chem Phys, 1985, 82:
3728–3740; Ovchinnikov AA. Multiplicity of the ground state of
large alternant organic molecules with conjugated bonds. Theoret
Chim Acta, 1978, 47: 297–304; Glaser T, Lügger T, Fröhlich R.
Synthesis, crystal structures, and magnetic properties of a mono- and
a dinuclear copper(II) complex of the 2,4,6-tris(2-pyridyl)-1,3,5-
triazine ligand. Eur J Inorg Chem, 2004, 394–400; Lloret F, De
12 Glaser T, Heidemeier M, Fröhlich R. The first trinuclear manganese
triplesalen complex: Synthesis, structural, and magnetic characteriza-
tion of [(talenNO2){MnIII(DMSO)2}3](ClO4)3. C R Chim, 2007, 10:
71–78; Glaser T, Heidemeier M, Theil H, Stammler A, Bögge H,
Schnack J. A MnIII triplesalen-based 1D pearl necklace: Exchange
interactions and zero-field splittings in a C3-symmetric MnIII com-
6
plex. Dalton Trans, 2010, 39: 192–199
13 Chong JH, Sauer M, Patrick BO, MacLachlan MJ. Highly stable ke-
to-enamine salicylideneanilines. Org Lett, 2003, 5: 3823–3826
14 Sauer M, Yeung C, Chong JH, Patrick BO, MacLachlan MJ.
N-Salicylideneanilines: Tautomers for formation of hydrogen-
bonded capsules, clefts, and chains. J Org Chem, 2006, 71: 775–788

Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
965
15 Yelamaggad CV, Achalkumar AS, Rao DSS, Prasad SK.
Self-assembly of C3h and Cs symmetric keto-enamine forms of tris
(N-salicyl-ideneanilines) into columnar phases: A new family of
discotic liquid crystals. J Am Chem Soc, 2004, 126: 6506–6507; Rid-
dle JA, Bollinger JC, Lee D. Escape from a nonporous solid: Me-
chanically coupled biconcave molecules. Angew Chem Int Ed, 2005,
44: 6689–6693; Riddle JA, Lathrop SP, Bollinger JC, Lee D. Schiff
base route to stackable pseudo-triphenylenes: stereoelectronic control
TM, Kauppinen R. Influence of bond fixation in benzo-annulated
N-salicylideneanilines and their ortho-C(O)X derivatives (X = CH3,
NH2, OCH3) on tautomeric equilibria in solution. J Org Chem, 2007,
72: 5598–5607
31 Alarcón SH, Olivieri AC, Sanz D, Claramunt RM, Elguero J. Sub-
stituent and solvent effects on the proton transfer equilibrium in anils
and azo derivatives of naphthol. Multinuclear NMR study and theo-
retical calculations. J Mol Struct, 2004, 705: 1–9
of assembly and luminescence.
J Am Chem Soc, 2006, 128:
32 Freiherr von Richthofen C-G, Feldscher B, Lippert K, Stammler A,
Bögge H, Glaser T. Manuscript in preparation.
10986–10987; Lim Y-K, Wallace S, Bollinger JC, Chen X, Lee D.
Triferrocenes built on a C3-symmetric ligand platform: Entry to re-
dox-active pseudo-triphenylenes via chelation-driven stereoselection
of triple schiff bases. Inorg Chem, 2007, 46: 1694–1703
33 Addison AW, Rao TN, Reedijk J, van Rijn J, Verschoor GC. Synthe-
sis, structure, and spectroscopic properties of copper(II) compounds
containing nitrogen-sulfur donor ligands: The crystal and molecular
structure of aqua[1,7-bis(N-methylbenzimidazol-2'-yl)-2,6-dithiahe-
ptane]copper(II) perchlorate. J Chem Soc, Dalton Trans, 1984,
1349–1356
34 Gorbitz CH, Kaboli M, Read ML, Vestli K. Benzene-1,3,5-triol at
105 K. Acta Crystallogr, 2008, 64: o2023
35 Shyu HL, Wei HH, Lee GH, Wang Y. Structure, magnetic properties
and epoxidation activity of iron(III) salicylaldimine complexes. Dal-
ton Trans, 2000, 915–918
16 Feldscher B, Stammler A, Bögge H, Glaser T. Synthesis and charac-
terization of a trinuclear CuII complex bridged by an extended
3
phloroglucinol-ligand: implications for a rational enhancement of
ferromagnetic interactions. Dalton Trans, 2010, 39: 11675–11685
17 Feldscher B, Stammler A, Bögge H, Glaser T. Synthesis and charac-
terization of trinuclear square-planar NiII and CuII complexes of an
3
3
extended phloroglucinol ligand: Experimental evidence for the rela-
tive contributions of benzene-like and radialene-like resonance struc-
tures. Polyhedron, 2011, 30: 3038–3047
36 Lohrie M, Knoche W. Dissociation and keto-enol tautomerism of
phloroglucinol and its anions in aqueous solution. J Am Chem Soc,
1993, 115: 919–924
18 Mukherjee C, Stammler A, Bögge H, Glaser T. Do trinuclear triple-
salen complexes exhibit cooperative effects? Synthesis, characteriza-
tion, and enantioselective catalytic sulfoxidation by chiral trinuclear
FeIII triplesalen complexes. Chem Eur J, 2010, 16: 10137–10149
19 Sheldrick GM. SADABS 2008/1 SADABS 2008, Göttingen: Univer-
sity of Göttingen,
20 Sheldrick GM. Acta Crystallogr. 2008, A64: 112–122
21 The program package JulX was used for spin-Hamiltonian simula-
tions and fittings of the data by a full-matrix diagonalization ap-
proach (Bill E, unpublished results)
22 Hernández-Molina R, Mederos A. Comprehensive Coordination
Chemistry II, Vol. 1 (Hrsg: McCleverty JA, Meyer TJ). Oxford:
Elsevier, Ltd., 2004, 411–446
23 Lopez J, Liang S, Bu XR. Unsymmetric chiral salen Schiff bases: A
new chiral ligand pool from bis-schiff bases containing two different
salicylaldehyde units. Tetrahedron Lett, 1998, 39: 4199–4202;
Janssen KBM, Laquierra I, Dehaen W, Parton RF, Vankelecom IFJ,
Jacobs PA. A dimeric form of Jacobsen’s catalyst for improved re-
tention in a polydimethylsiloxane membrane. Tetrahedron: Asym-
metry, 1997, 8: 3481–3487
24 Glaser T, Heidemeier M, Lügger T. The novel triplesalen ligand
bridges three NiII-salen subunits in a meta-phenylene linkage. Dalton
Trans, 2003, 2381–2383
25 Böttcher A, Elias H, Eisenmann B, Hilms E, Huber A, Kniep R, Röhr
C. A novel synthetic approach to asymmetric saslen, dihydrosalen,
and tetrahydrosalen ligands–structures and O2-activating properties of
their nickel(II) and cobalt(II) complexes. Z Naturforsch B, 1994, 49:
1089–1100
26 Freiherr von Richthofen C-G, Stammler A, Bögge H, Glaser T. From
triplesalen to triplesalophen: Ferromagnetic interactions via
spin-polarization in a trinuclear NiII triplesalophen complex. Eur J
Inorg Chem, 2011, 49–52
27 Freiherr von Richthofen C-G, Stammler A, Bögge H, Glaser T.
Probing the radialene-character in triplesalophen ligands by spectro-
scopic and structural analysis. J Org Chem, 2012, 77: 1435–1448
28 Freiherr von Richthofen C-G, Stammler A, Bögge H, DeGroot MW,
Long JR, Glaser T. Synthesis, structure, and magnetic characteriza-
tion of a C3-symmetric MnIII3CrIII assembly: Molecular recognition
between a trinuclear MnIII triplesalen complex and a fac-triscyano
CrIII complex. Inorg Chem, 2009, 48: 10165–10176
37 Bosnich B. An interpretation of the circular dichroism and electronic
spectra of salicylaldimine complexes of square-coplanar diamagnetic
nickel(II). J Am Chem Soc, 1968, 90: 627–632; Crawford SM. The
ultra-violet and visible spectra of some transition metal chelates with
N,N'-bis-(o-hydroxybenzylidene)ethylenediamine and N,N'-bis-(o-
hydroxybenzylidene)-o-phenylenediamine and related compounds.
Spectrochim Acta, 1963, 19: 255–270; Di Bella S, Fragala I, Ledoux
I, Diaz-Garcia MA, Marks TJ. Synthesis, characterization, optical
spectroscopic, electronic structure, and second-order nonlinear opti-
cal (NLO) properties of
a novel class of donor−Acceptor
bis(salicylaldiminato)nickel(II) Schiff base NLO chromophores. J
Am Chem Soc,1997, 119: 9550–9557
38 Lubben M, Meetsma A, Bolhuis F v, Feringa BL. Synthesis, crystal
and molecular structures, UV/Vis spectroscopy and electrochemical
properties of two iron(III) phenolate complexes. Inorg Chim Acta,
1994, 215: 123–129; Gaber BP, Miskowsk V, Spiro TG. Resonance
raman-scattering from iron(III)-transferrin and copper(II)-transferrin
and an Iron(III) model compound: Spectroscopic interpretation of
transferrin binding-site. J Am Chem Soc, 1974, 96: 6868–6873;
Flassbeck C, Wieghardt K. Synthesis of N-phenolate-functionalized
macrocycles of 1,4,7-triazacyclononane and of 1-Oxa-4,7-diazacy-
clononane and their coordination chemistry with Iron(III). Z Anorg
Allg Chem, 1992, 608: 60–68
39 Berry KJ, Clark PE, Murray KS, Raston CL, White AH. Structure,
magnetism, and mossbauer spectrum of the 5-coordinate complex
chlorobis(N-methylbenzothiohydroxamato)Iron(III). Inorg Chem,
1983, 22: 3928–3934
40 Adam B, Bill E, Bothe E, Goerdt B, Haselhorst G, Hildenbrand K,
Sokolowski A, Steenken S, Weyhermüller T, Wieghardt K. Phenoxyl
radical complexes of gallium, scandium, iron and manganese. Chem
Eur J, 1997, 3: 308–319; Strautmann JBH, Walleck S, Bögge H,
Stammler A, Glaser T. A tailor-made ligand to mimic the active site
of diiron enzymes: an air-oxidized high-valent FeIII h.s.(μ-O)2FeIV h.s.
species. Chem Commun, 2011, 47: 695–697; Strautmann JBH, Frei-
herr von Richthofen C-G, DeBeer George S, Bothe E, Bill E, Glaser
T. Highly oxidized diiron complexes: Generation, spectroscopy, and
stabilities. Chem Commun, 2009, 2637–2639; Strautmann JBH, De-
Beer George S, Bothe E, Bill E, Weyhermüller T, Stammler A,
Bögge H, Glaser T. Molecular and electronic structures of mononu-
clear iron complexes using strongly electron-donating ligands and
their oxidized forms. Inorg Chem, 2008, 47: 6804–6824
29 Claramunt RM, López C, Santa María MD, Sanz D, Elguero J. The
use of NMR spectroscopy to study tautomerism. Prog Nucl Magn
Reson Spectrosc, 2006, 49: 169–206
30 Gawinecki R, Kuczek A, Kolehmainen E, Ośmiałowski B, Krygowsi
41 Gerloch M, Mabbs FE. The crystal and molecular structure of chlo-

966
Feldscher B, et al. Sci China Chem June (2012) Vol.55 No.6
ro-(NN′-bis-salicylidene-ethylenediamine)iron(III) as
a
hexa-
dimeric complex. Indian J Chem Sec A, 2007, 46A: 1947–1950; Hao
L, Mu C, Kong B. Bis{[mu]-2,2'- [ethane-1,2-diylbis (nitrilomethyl-
idyne)]diphenolato}bis[(thiocyanato-[kappa]N)iron(III)]. Acta Crys-
tallogr, 2008, E64: m1034-m1042
co-ordinate dimer. J Chem Soc A, 1967, 1900–1908; Geiß A,
Vahrenkamp H. Cyanide-bridged arrays of 2, 3 and 4 metal atoms
based on salene-iron complexes-syntheses, structures and metal-metal
interactions. Eur J Inorg Chem, 1999, 1793–1803; Resce JL, Fanning
JC. Structure of the Fe(salen)ONO2 dimer, a ferric complex with a
unidentate nitrate ligand. Acta Crystallogr, 1987, C43: 2100–2104;
You ZL, Zhu HL. A dinuclear Schiff base iron(III) complex with the
ligand N,N'-bis (2-oxidophenylmethyleneimino) propane-1,2-diamine.
Acta Cryst, 2004, E60: m1046–m1048; Roy P, Dhara K, Chakraborty
J, Nethaji M, Banerjee P. Synthesis and crystal structure of an iron(II)
42 Reiff WM, Long GJ, Baker WA. Nature of the spin states in some
binuclear iron(III) complexes. J Am Chem Soc, 1968, 90: 6347–6351;
Gerloch M, Mabbs FE, Richards A. The preparation and magnetic
properties of some Schiff base-iron(III) halide complexes. J Chem
Soc A, 1968, 112–116; Shapira Y, Liu MT, Foner S, Dubé CE, Boni-
tatebus PJ. Magnetization steps in [Fe(salen)Cl](2). Phys Rev B, 1999,
59: 1046–1054