Do trinuclear triplesalen complexes exhibit cooperative effects? Synthesis, characterization, and enantioselective catalytic sulfoxidation by chiral trinuclear FeIII triplesalen complexes

Article abstract of DOI:10.1002/chem.201000923

The chiral triplesalen ligand H6chand provides three chiral salen ligand compartments in a meta-phenylene arrangement by a phloroglucinol backbone. The two diastereomeric versions H₆chand ^RR and H ₆chand ^rac have been used to synthesize the enantiomerically pure chiral complex [(FeCl)₃-(chand ^RR)] (3 ^RR) and the racemic complex [(FeCl)₃(chand ^rac)] (3 ^rac). The molecular structure of the free ligand H₆chand ^rac exhibits at the terminal donor sides the O-protonated phenol-imine tautomer and at the central donor sides the N-protonated keto-enamine tautomer. The trinuclear complexes are comprised of five-coordinate square-pyramidal Fe ^III ions with a chloride at the axial positions. The crystal structure of 3 ^rac exhibits collinear chiral channels of ～11 A in diameter making up 33.6 % of the volume of the crystals, whereas the crystal structure of 3 ^RR exhibits voids of 560 A ³. Moessbauer spectroscopy demonstrates the presence of Fe ^III high-spin ions. UV/Vis spectroscopy is in accordance with a large delocalized system in the central backbone evidenced by strong low-energy shifts of the imine π-π ^* transitions relative to that of the terminal units. Magnetic measurements reveal weak intramolecular exchange interactions but strong magnetic anisotropies of the Fe ^III ions. Complexes 3 ^rac and 3 ^RR are good catalysts for the sulfoxidation of sulfides providing very good yields and high selectivities with 3 ^RR being enantioselective. A comparison of 3 ^RR and [FeCl-(salen')] provides higher yields and selectivities but lower enantiomeric excess values (ee values) for 3 ^RR relative to [FeCl(salen')]. The low ee values of 3 ^RR appeared to be connected to a strong ligand folding in 3 ^RR, opening access to the catalytically active high-valent Fe-O species. The higher selectivity is assigned to a cooperative stabilization of the catalytically active high-valent Fe-O species through the phloroglucinol backbone in the trinuclear complexes.

Full text of DOI:10.1002/chem.201000923

FULL PAPER
DOI: 10.1002/chem.201000923
Do Trinuclear Triplesalen Complexes Exhibit Cooperative Effects?
Synthesis, Characterization, and Enantioselective Catalytic Sulfoxidation by
Chiral Trinuclear FeIII Triplesalen Complexes
Chandan Mukherjee, Anja Stammler, Hartmut Bçgge, and Thorsten Glaser*[a]
Abstract: The chiral triplesalen ligand
H6chand provides three chiral salen
ligand compartments in a meta-phenyl-
ene arrangement by a phloroglucinol
backbone. The two diastereomeric ver-
sions H6chandRR and H6chandrac have
been used to synthesize the enantio-
merically pure chiral complex [(FeCl)3-
structure of 3rac exhibits collinear chiral
channels of ~11 ꢁ in diameter making
up 33.6% of the volume of the crystals,
whereas the crystal structure of 3RR ex-
hibits voids of 560 ꢁ3. Mçssbauer spec-
troscopy demonstrates the presence of
FeIII high-spin ions. UV/Vis spectrosco-
py is in accordance with a large delo-
calized system in the central backbone
evidenced by strong low-energy shifts
of the imine p–p* transitions relative
to that of the terminal units. Magnetic
measurements reveal weak intramolec-
ular exchange interactions but strong
magnetic anisotropies of the FeIII ions.
Complexes 3rac and 3RR are good cata-
lysts for the sulfoxidation of sulfides
providing very good yields and high se-
lectivities with 3RR being enantioselec-
tive. A comparison of 3RR and [FeCl-
AHCTUNGTRENN(GUN salen’)] provides higher yields and se-
lectivities but lower enantiomeric
A
excess values (ee values) for 3RR rela-
T
tive to [FeClACTHNGUTERN(UNG salen’)]. The low ee
lecular structure of the free ligand
H6chandrac exhibits at the terminal
donor sides the O-protonated phenol–
imine tautomer and at the central
donor sides the N-protonated keto–en-
amine tautomer. The trinuclear com-
plexes are comprised of five-coordinate
square-pyramidal FeIII ions with a chlo-
ride at the axial positions. The crystal
values of 3RR appeared to be connected
to a strong ligand folding in 3RR, open-
ing access to the catalytically active
high-valent Fe–O species. The higher
selectivity is assigned to a cooperative
stabilization of the catalytically active
high-valent Fe–O species through the
phloroglucinol backbone in the trinu-
clear complexes.
Keywords: homogeneous catalysis ·
iron · magnetic properties · N,O
ligands · stereoselective catalysis
Introduction
prochiral sulfides catalyzed by enzymes or chiral metal com-
plexes is of interest due to their simplicity, low toxicity, and
economical benignity.
Enantioselective oxidation of prochiral sulfides to sulfox-
ides with chiral complexes of titanium,[10,11,20–22] vanadi-
um,[15,23,24] and manganese[7,12,25,26] as catalysts has been ex-
tensively investigated. In contrast, chiral iron complexes are
less studied as asymmetric catalysts for this reaction. In
2003, the asymmetric sulfoxidation of prochiral sulfides with
H2O2 as an external oxidant was reported by an in situ reac-
The synthesis of enantiomerically pure sulfoxides by the oxi-
dation of prochiral sulfides has drawn considerable attention
in the last decade.[1] Enantiomerically pure sulfoxides have
been used as both building blocks and versatile chiral con-
trollers for the synthesis of natural products and biologically
active compounds.[1–5] Several methods for the synthesis of
enantiopure sulfoxides were described,[6–19] including Ander-
senꢀs method,[6] in which Grignard reagents have been used
with chiral sulfinates. The direct asymmetric oxidation of
tion using [FeACHTNUTRGNEUNG(acac)3] (acac=acetylacetonate) and a chiral
ligand as the catalyst.[27] The catalytic properties of the non-
chiral parent iron salen complex for sulfoxidation were re-
ported in 2002.[28] Application of the chiral salen derivative,
that is, by using the Jacobsen ligand H2salen’,[29] showed that
the chiral iron complex catalyzes the asymmetric oxidative
formation of sulfoxides from prochiral sulfides in the pres-
ence of the external oxidant PhIO.[30]
[a] C. Mukherjee, A. Stammler, H. Bçgge, T. Glaser
Lehrstuhl fꢂr Anorganische Chemie I
Fakultꢃt fꢂr Chemie, Universitꢃt Bielefeld
Universitꢃtsstr. 25, 33615 Bielefeld (Germany)
Fax : (+49)521-106-6003
E-mail: thorsten.glaser@uniso-bielefeld.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000923.
Chem. Eur. J. 2010, 16, 10137 – 10149
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
10137

ligand H6chandrac
,
its racemic trinuclear FeIII complex
(chandrac)], and the chiral trinuclear FeIII complex
3A
(chandRR)]. The catalytic properties of the com-
[(FeCl)
[(FeCl) CTHNUGTRENNUNG
3ACHTUNGTRENNUNG
plexes for the sulfoxidation of sulfides, either asymmetric or
not, in the presence of PhIO or NaOCl as external oxidants
have been investigated. Furthermore, comparison of the cat-
alytic efficiencies of the trinuclear [(FeCl)
plex to the mononuclear [FeCl(salen’)] complex allows us to
(chandRR)] com-
3ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
evaluate the presence of cooperative effects in the trinuclear
complex.
Experimental Section
Materials: All reagents were obtained from commercial sources and used
as supplied, unless otherwise noted. 2,4,6-Triacetyl-1,3,5-trihydroxyben-
zene (1) was prepared according to the reported procedure.[67] The syn-
In our ongoing efforts[31–47] for a rational synthesis of new
single-molecule magnets (SMMs) and asymmetric oxidation
catalysts, we have developed the chiral triplesalen ligand
H6chand in the all R,R form, H6chandRR, and in the all S,S
form, H6chandSS,[44] which we envision as the C3-symmetric
trinuclear extension of the Jacobsen ligand H2salen’.[29] The
three chiral subunits of the ligand H6chand are connected
with each other via a phloroglucinol (1,3,5-trihydroxyben-
zene) bridging unit. Hence, obeying the spin-polarization
mechanism,[48–52] ferromagnetic interactions between the
three metal salen units may lead to high-spin ground
states.[53–62] Trinuclear CuII triplesalen complexes indeed
show a high-spin ground state with ferromagnetic coupling
between three CuII centers.[35,39] Besides that, we have been
able to synthesize several heptanuclear complexes M6M’n+
with M3M’M3 topology by using two trinuclear triplesalen
complexes as molecular building blocks and hexacyanomet-
thesis of 2,4,6-tris{1-
droxybenzene (2RR
aldimino)cyclohexylimino]ethyl}-1,3,5-trihydroxybenzene
ACHTUNGTRENNUNG
)
ACHTUNGTRENNUNG
N
)
were reported previously.[44] Enantiomerically pure trans-(R,R)-1,2-cyclo-
hexanediamine was obtained by an established kinetic resolution.[68] The
purity of thioanisole, benzyl phenyl sulfide, 4-bromo thioanisole, and
methyl p-tolyl sulfide was checked by 1H NMR spectroscopy and gas
chromatography (GC). Iodosylbenzene was prepared by the hydrolysis of
iodobenzene diacetate and was kept in the dark with a coating of alumi-
num foil.
2,4,6-Tris[1-(2-aminocyclohexylimino)ethyl]-1,3,5-trihydroxybenzene
(2rac): 2,4,6-Triacetyl-1,3,5-trihydroxybenzene (1) (0.520 g, 2.063 mmol)
was added to a solution of trans-(ꢀ)-1,2-cyclohexanediamine (1.425 g,
12.480 mmol) in EtOH (20 mL) and the resulting mixture was stirred at
room temperature for 20 h during which time the color of the solution
changed to deep red. The solution was filtered and evaporated under
vacuum at 508C to yield a red oil. Et2O (100 mL) was added to the red
oil and the mixture was stirred for 1h. The suspension was filtered to
remove
a light-yellow solid and the filtrate was evaporated under
ACHTUNGTRENNUNG
allate as the bridging unit.[37,42,47] Among them,
vacuum at 508C to yield again a red oil. This oil was dissolved in H2O
(30 mL) and extracted with CH2Cl2 (100 mL). The CH2Cl2 solution was
further washed with water (2ꢅ25 mL) and dried over Na2SO4. Upon
evaporation of CH2Cl2, a sticky yellow oil was obtained. Yield: 0.67 g
(60%); IR (KBr): n˜ =3431 (m), 3370 (m), 3293 (m), 2930 (s), 2857 (s),
1533 (s), 1458 (m), 1449 (m), 1420 (m), 1373 (m), 1340 (m), 1321 (m),
1267 (w), 1244 (w), 866 (w), 845 (w), 820 cmꢁ1 (w); 1H NMR
(500.05 MHz, CDCl3): d=1.15–1.44 (m, 12H), 1.65–1.96 (m, 18H), 2.56–
2.64 (3 s, 9H), 2.74–2.90 (m, 3H), 3.25–3.43 (m, 3H), 13.11–14.39,
18.53 ppm (brs, 3H); 13C NMR (125.77 MHz, CDCl3): d=18.5, 19.1, 19.5,
24.4, 24.5, 24.6, 24.7, 24.8, 30.4, 31.5, 32.3, 32.4, 32.7, 34.1, 34.2, 54.7, 54.8,
55.0, 60.0, 60.3, 60.8, 105.8, 170.1, 173.0, 183.5, 184.7, 185.6 ppm; MS-ESI
(+ve, CH2Cl2): m/z: 541.4 [M+H]+; elemental analysis calcd (%) for
C30H48N6O3·1CH2Cl2·0.5Et2O: C 59.80, H 8.36, N 12.68; found: C 60.00,
H 8.11, N 12.77.
[{MnIII3(talentBu )} {CrIII(CN)6}]
2ACHTUNGTRENNUNG ACHTUNTGREN(NUGN BPh4)3 (H6talen=2,4,6-tris{1-
2
[2-(3,5-di-tert-butylsalicylaldimino)-2-methylpropylimi-
no]ethyl}-1,3,5-trihydroxybenzene) behaves as a SMM.[37]
Trinuclear triplesalen complexes exhibit bowl-shaped mo-
lecular structures.[35,39] Hence, we thought trinuclear triplesa-
len complexes of the chiral ligand H6chand would likely pro-
vide a chiral pocket that should be ideally suited for asym-
metric catalysis. Furthermore, a mechanism was established
for the asymmetric nucleophilic ring opening of epoxides by
chromium salen complexes, involving catalytic activation of
both the nucleophile and the electrophile in a bimetallic
rate-determining step.[63–65] By using covalently linked dinu-
clear salen complexes, not only the intramolecular pathway
but also an intermolecular pathway was enhanced. This ob-
servation “indicates that dimer reacts more rapidly with
dimer, than does monomer with monomer and suggests that
the design of covalently linked systems bearing three or
more metal salen units may be worthwhile”.[66]
2,4,6-Tris{1-[2-(3,5-di-tert-butylsalicylaldimino)cyclohexylimino]ethyl}-
1,3,5-trihydroxybenzene (H6chandrac): A solution of half-unit 2rac (0.546 g,
1.00 mmol) and 3,5-di-tert-butylsalicylaldehyde (0.710 g, 3.03 mmol) in
MeOH (15 mL) was stirred for 20 h, during which time a yellow precipi-
tate formed. This filtrate was isolated by filtration, washed with MeOH,
and dried under air. Yield: 0.78 g (66%); IR (KBr): n˜ =3460 (w), 2951
(s), 2862 (m), 1628 (s), 1593 (m), 1539 (s), 1454 (s), 1441 (s), 1392 (m),
1362 (m), 1346 (m), 1273 (m), 1252 (m), 1242 (m), 1172 (m), 982 (w), 878
(w), 829 (w), 771 cmꢁ1 (w); 1H NMR (500.15 MHz, CDCl3): d=1.12–1.29
(m, 27H), 1.45–2.15 (m, 51H), 2.46–2.66 (m, 9H), 3.18–3.35 (m, 3H),
3.69–3.94 (m, 3H), 6.95–7.06 (m, 3H), 7.35–7.38 (m, 3H), 8.31–8.50 (m,
3H), 13.35–15.40, 18.34–18.56 ppm (m, 3H); 13C NMR (125.77 MHz,
CDCl3): d=18.4, 18.5, 18.9, 19.0, 19.2, 19.4, 23.6, 23.7, 24.1, 24.2, 24.3,
29.2, 29.3, 29.4, 30.4, 31.0, 31.2, 31.3, 31.7, 31.8, 32.1, 33.1, 33.2, 33.8, 33.9,
34.8, 34.8, 56.4, 57.0, 57.1, 57.3, 72.2, 72.3, 72.6, 72.7, 99,9, 105.4, 105.6,
Recently, we reported the successful synthesis and charac-
terization of the chiral triplesalen ligands H6chandRR and
H6chandSS, and their trinuclear MnIII complexes in conjuga-
tion with their catalytic activity for the asymmetric epoxida-
tion of unfunctionalized olefins.[44] Herein, we report the
synthesis and characterization of the racemic form of the
10138
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10137 – 10149

Chiral Trinuclear FeIII Triplesalen Complexes
FULL PAPER
105.7, 106.2, 117.3, 117.4, 117.5, 117.6, 117.7, 125.9, 126.0, 126.5, 126.8,
127.0, 127.1, 136.1, 136.2, 136.3, 136.4, 139.7, 139.8, 140.0, 157.6, 157.7,
157.8, 165.7, 166.5, 166.6, 166.7, 170.0, 170.1, 172.4, 172.4, 172.7, 172.9,
179.7, 183.3, 183.4, 184.2, 184.4, 185.0, 185.3, 185.6 ppm; MS-ESI (+ve,
CH2Cl2): m/z: 1190.8 [M+H]+; elemental analysis calcd (%) for
C75H108N6O6: C 75.72, H 9.15, N 7.06; found: C 75.60, H 9.05, N 7.03.
w at 0.38 scan width in three runs with 606, 435, and 230 frames (f=0,
88, and 1808; detector distance of 5 cm). An empirical absorption correc-
tion by using equivalent reflections was performed with the program
SADABS 2.10.[69] The structures were solved with the program SHELXS-
97[70] and refined by using SHELXL-97.[70]
As the crystals of H6chandrac·2CH3CN·MeOH were very small, the struc-
ACHTUNGTRENNUNG[(FeCl)3ACHTUNGTRENNUNG
(chandrac)] (3rac): Anhydrous FeCl3 (0.152 g, 0.94 mmol), the
ꢁ
ꢁ
ture could only be refined up to 2q=46. However, the O H and N H
ligand H6chandrac (0.350 g, 0.294 mmol), and Et3N (0.1 mL) were added
to a solvent mixture of CH3CN (25 mL) and MeOH (10 mL). The result-
ing deep-purple solution was heated at reflux for 15 min. Stirring for ad-
ditional 15 h caused the deposition of a purple microcrystalline solid that
was isolated by filtration, washed with CH3CN, and recrystallized from
CH3CN/CH2Cl2. Yield: 0.212 g (50%); IR (KBr): n˜ =2951 (s), 2866 (s),
1626 (s), 1537 (s), 1487 (s), 1435 (m), 1389 (m), 1348 (w), 1254 (s), 1238
(m), 1175 (w), 1049 (w), 822 (w), 598 (w), 544 cmꢁ1 (w); MALDI-TOF-
MS: m/z: 1457.0 [M]+, 1421.9 [MꢁCl]+; elemental analysis calcd (%) for
C75H102Cl3Fe3N6O6·1.5H2O: C 60.67, H 7.13, N 5.66; found: C 60.66, H
7.03, N 5.75.
hydrogen atoms of H6chandrac could be located in a difference Fourier
synthesis and freely be refined with U(H)=1.2UACHTNUTRGNEUNG(O/N). The measured
crystal of 3rac turned out to be a racemic twin with a ratio of 0.67/0.33 for
the two components. The unit cell of 3rac contains a huge void of approxi-
mately 4900 ꢁ3. A refinement with a “squeezed”[71] data set, however, re-
sulted in a nearly identical R value and positional and thermal parame-
ters. Thus the original data set was retained. The unit cell of 3RR contains
two voids of approximately 560 ꢁ3. No solvent molecule could be located
in these voids. Crystal data and further details concerning the crystal
structure determination are summarized in Table 1.
ACHTUNGTRENNUNG[(FeCl)3ACHTUNGTRENNUNG A solution
(chandRR)] (3RR):
Table 1. Crystallographic data for H6chandrac·2CH3CN·MeOH, 3rac, and 3RR
.
of the ligand H6chandRR (0.800 g,
0.67 mmol) in CH3CN (30 mL) was
treated with anhydrous FeCl3 (0.350 g,
2.16 mmol) and Et3N (0.4 mL). The re-
sulting deep-purple solution was
heated at reflux for 15 min. Stirring
for an additional 15 h at room temper-
Compound
H6chandrac·2CH3CN·MeOH
3rac
3RR
chem. Formula
Fw
C75H108N6O6·2CH3CN·MeOH
1303.82
C75H102Cl3Fe3N6O6
1457.53
C75H102Cl3Fe3N6O6
1457.53
173(2)
T [K]
173(2)
173(2)
space group
a [ꢁ]
b [ꢁ]
P21/n
P61
30.1782(18)
–
P212121
16.2391(6)
19.4933(7)
25.6868(10)
106.280(1)
7805.2(5)
15.2409(9)
20.6903(12)
26.4118(16)
ature caused the deposition of
a
purple microcrystalline solid that was
isolated by filtration, washed with
CH3CN, and recrystallized from
MeOH/CH3CN/CH2Cl2. Yield: 0.392 g
(40%); IR (KBr): n˜ =2952 (s), 2864
(m), 1626 (s), 1537 (s), 1479 (s), 1435
(s), 1389 (m), 1362 (m), 1348 (m),
1254 (s), 1209 (m), 1173 (m), 1049 (w),
822 (w), 598 (w), 542 cmꢁ1 (w); MS-
ESI (+ve, CH2Cl2): m/z: 1458.5
[M+H]+, 1420.6 [MꢁCl]+; MALDI-
TOF-MS: m/z: 1457.1 [M]+, 1422.0
c [ꢁ]
18.9099(17)
b [8]
V [ꢁ3]
Z
14914.4(18)
6
0.974
0.71073; 0.553
0.30ꢅ0.20ꢅ0.10
67060; 24.00
15395 (R
5519
805
0.952
0.0869; 0.1748
0.33(3)
0.760, ꢁ0.251
8328.7(9)
4
1.162
0.71073; 0.660
0.50ꢅ0.40ꢅ0.30
42975; 24.99
4
1calcd [gcmꢁ3
]
1.110
l [ꢁ]; m
N
]
0.71073; 0.071
0.40ꢅ0.14ꢅ0.12
33307; 23.00
10845 (R
6668
870
crystal size [mm]
coll. reflns; qmax
unique reflns
G
E
14631 (R
12776
859
ACHTUNGTREN(NUNG int)=0.0398)
obs. reflns (I>2s(I))
Parameters
GooF[a] on F2
1.039
0.0758; 0.2027
1.068
[MꢁCl]+, 1385.9 [Mꢁ2Cl]+; [a]2D0
=
R1[b]; wR2[c] (I>2s(I))
abs. struct. parameter[d]
max/min residuals [eꢁꢁ3
2
0.0442; 0.0916
0.010(11)
0.427, ꢁ0.246
2
ꢁ650.0 (ꢀ5.0) (c=0.02 in CHCl3); ele-
mental analysis calcd (%) for
C75H102Cl3Fe3N6O6: C 61.80, H 7.05, N
5.76; found: C 61.84, H 6.97, N 5.74.
]
0.853, ꢁ0.446
2
[a] GooF=[S[w
which w=1/[s CAHTUNGRTENNGUN
G
G
G
(Fo )2]]1/2
ACHTUNGRETNNUNG , in
2(Fo )+(aP)2 +bP], P=(Fo2 +2Fc2)/3. [d] See reference [104].
2
General procedure for catalytic reac-
tions:
0.005 mmol) was dissolved in CH2Cl2
(2 mL). The sulfide substrate
Complex
3
(7.30 mg,
(0.1 mmol) and iodosylbenzene (49.0 mg, 0. 223 mmol) were added to
this solution and the mixture was stirred at room temperature for 15 h.
The solvent was evaporated and the residue treated with CH3CN
(0.2 mL) and passed through a short SiO2 column. The column was
washed with CH3CN (10 mL). The resulting solution was evaporated and
1H NMR spectra were recorded after dissolving the residue in CDCl3
(0.6 mL). The conversion was calculated from the area of the sulfide,
CCDC-777337–777339 contain the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
Other physical measurements: FTIR spectra were recorded on a Shimad-
zu 8300 spectrometer as KBr pellets. 1H and 13C NMR spectra were re-
corded on a Bruker DRX500 by using the solvent as an internal standard.
ESI Mass spectra were obtained on a Bruker Esquire 3000 mass spec-
trometer. MALDI TOF mass spectra were recorded with a Voyager DE
instrument. Elemental analyses were carried out in the Microanalysis
Laboratory, Bielefeld University. GC analyses were performed on a Shi-
madzu GC2010 and GCMS analysis on a Shimadzu QP2010S. The enan-
tiomeric excess (ee) values were determined by HPLC on a chiral station-
ary phase (Chiralcel OD-H, 254 nm, 208C, hexane/iPrOH 9:1 or 99:1 (4-
bromo thioanisole), flow rate=0.5 or 1.0 mLminꢁ1 (4-bromo thioani-
sole).
1
sulfoxide, and sulfone signals obtained in the H NMR spectrum by using
the formula %conversion=100ꢅ([product])/([substrate]+[product]).
GC and GCMS measurements were also performed to identify the prod-
ucts.
X-ray crystallography: Yellow crystals of H6chandrac·2CH3CN·MeOH
were grown by slow evaporation of a MeOH/CH3CN/CH2Cl2 solvent
mixture. Dark-brown crystals of 3rac were grown by slow evaporation of a
CH3CN/CH2Cl2 solvent mixture and black crystals of 3RR by slow evapo-
ration of a MeOH/CH3CN/CH2Cl2 solvent mixture.
The single crystals were removed from the mother liquor, coated with
paraffin oil and immediately cooled to 173(2) K on a Bruker AXS
SMART diffractometer (three circle goniometer with 1 K CCD detector,
Moka radiation, graphite monochromator. Hemisphere data collection in
Magnetic susceptibility data were measured from powder samples of
solid material in the temperature range 2–290 K by using a SQUID mag-
netometer (Quantum Design MPMS XL-7 EC) with a field of 1.0 T. The
Chem. Eur. J. 2010, 16, 10137 – 10149
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10139

T. Glaser et al.
experimental data were corrected for underlying diamagnetism by the
use of tabulated Pascalꢀs constants.
three equivalents of 3,5-di-tert-butylsalicylaldehyde provides
the corresponding racemic ligand H6chandrac
.
57Fe Mçssbauer spectra were recorded on an alternating constant-acceler-
ation spectrometer. The minimal line width was 0.24 mmsꢁ1 full width at
half height. 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 a-iron at room temper-
ature.
The reaction of 1 with racemic trans-1,2-cyclohexanedi-
amine (R,R and S,S enantiomers) can provide principally
four different isomers of half unit 2: (R,R)(R,R)(R,R), (S,S)-
(S,S)(S,S), (R,R)(R,R)(S,S), and (S,S)(S,S)(R,R). The first
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
E
E
G
N
ACHTUNGTRENNUNG
and latter two form enantiomeric pairs that are diastereo-
meric to each other. It is interesting to note that 1 upon re-
action with excess enantiomerically pure (R,R)- or (S,S)-
trans-1,2-cyclohexanediamine for 20 h as in the reaction of
2rac provides a mixture with the intermediate condensation
product containing only two acetyl groups converted to keti-
mine units as the main component. A complete conversion
to 2RR and 2SS requires six days. This implies that the third
condensation is slower, if only one enantiomer of the di-
Results and Discussion
Synthesis and characterization: In contrast to the synthesis
of symmetrical salen-type ligands by one-pot condensation
reactions of one diamine with two identical aldehyde or
ketone derivatives, synthesis of unsymmetrical salen-type li-
gands is more challenging[72] and requires a so-called half
unit.[73–76] After several optimization steps, we synthesized
successfully two chiral half units, 2RR and 2SS, by reacting 1
with an excess of enantiomerically pure (R,R)- or (S,S)-
trans-1,2-cyclohexanediamine, respectively, for six days
(Scheme 1).[44] Condensation of three equivalents of 3,5-di-
tert-butylsalicylaldehyde with the half units 2RR and 2SS
yielded the corresponding enantiomerically pure chiral li-
gands H6chandRR and H6chandSS, respectively.[44] To synthe-
size the half unit 2rac, the above-stated synthetic protocol has
been applied. Specifically, triketone 1 has been reacted with
excess racemic trans-1,2-cyclohexanediamine for 20 h to
yield the half unit 2rac that upon further condensation with
AHCTUNGTREGaNNNU mine is used in comparison to a racemic mixture. A reason
for this might be attributed to some steric crowding for the
three pendent arms with the same chirality of the diamine
backbone. As a direct consequence, the bulk material of the
half unit 2rac and the ligand H6chandrac should not be a statis-
tical mixture of the four possible isomers, but should contain
the racemic (R,R)ACHTUNGTERNNU(NG R,R)ACHUTGTNRENNUG(S,S) and (S,S)ACHTUNEGRTG(NNUN S,S)CAHTNUGTREN(NNGU R,R) configura-
tions as the main products. The NMR spectra of 2rac and
H6chandrac provide a large set of multiplets of higher order.
However, the ketimine CH3 groups in the half unit 2rac give
rise to three sharp singlets. By taking into account that the
(R,R)ACHTUGNTRNE(UNG R,R)ACUTHNGTRENN(GUN R,R)/AHCNUTRTGEG(NNUN S,S)AHTCUNGTNER(UNNG S,S)AHCTGNRUETNNUG
CH3 signal and the (R,R)
G
R
G
N
ACHTUNGTRENNUNG
should provide two CH3 signals
in a 2:1 ratio shows that 2rac is
actually a mixture of two dia-
stereomers with (R,R)
(S,S)/(S,S)(S,S)(R,R) formed in
85 and (R,R)(R,R)(R,R)/(S,S)-
(S,S)(S,S) in 15% yield.
The reaction of one equiva-
ACHTUNGTRENNUNG(R,R)-
A
E
N
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
lent of the ligand (either
H6chandrac or H6chandRR) with
approximately three equivalents
of anhydrous FeCl3 in CH3CN
or in a CH3CN/MeOH solvent
mixture results in the separa-
tion of purple microcrystalline
trinuclear FeIII triplesalen com-
plexes. Recrystallization of the
microcrystalline solid from
a
CH2Cl2/CH3CN and a CH2Cl2/
CH3CN/MeOH solvent mixture
provided single crystals of com-
plex 3rac and complex 3RR, re-
spectively (vide infra), suited
for single-crystal X-ray diffrac-
tion.
The IR spectra of 2rac and 2RR
are almost identical and pro-
vide a sharp band at around
Scheme 1. Synthesis of chiral half units 2 and complexes 3.
1533 cmꢁ1 due to a n
ACTHNGUTERN(UNG C=N) vi-
10140
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10137 – 10149

Chiral Trinuclear FeIII Triplesalen Complexes
FULL PAPER
bration of the ketimine units. In the IR spectra of both li-
gands, H6chandrac and H6chandRR, a new band appears at
1628 cmꢁ1 due to a n
ACHTUNGTRENNUNG
ces. Upon complexation, the nACTHNUTRGNEUGN(C=N) vibration band appear-
ing in the ligand spectrum at 1628 shifts to 1626 cmꢁ1,
whereas the band at 1533 shifts to 1537 cmꢁ1. The strong
bands at 1479 and at 1254 cmꢁ1 are attributed to n
ACTHNUTRGNEU(GN C=C) vi-
ꢁ
brations of the aromatic rings and to phenolic nACTHNUTRGENUG(N C O) vibra-
tions, respectively.
Mass spectrometry has been found as an important tool
to characterize triplesalen half units, ligands, and corre-
sponding iron complexes. The ESI (positive mode) mass
spectra of the triplesalen half units and the ligands show
[M+H]+ as base peaks (100%). The MALDI-TOF-mass
spectrum of complex 3rac shows peaks at m/z: 1457.0 and
1421.9 that correspond to the composition [Fe3Cl3-
A
(chandrac)]+, respectively. The
2ACHTUNGTRENNUNG
Figure 1. Molecular structure of the ligand H6chandrac in crystals of
AHCTUNGTRENNUNG
H6chandrac·2CH3CN·MeOH in a ball and stick presentation. The enantio-
together with the peaks at m/z: 1457.1 and 1422.0.
mer shown has the (13R,18R)ACTHNUTRGENN(GU 23S,28S)ACHTUNGTRNE(NUGN 33S,38S) configuration. Solvent
ꢁ
molecules and hydrogen atoms have been omitted for clarity despite N
Molecular and crystal structures: The crystal structures of
the ligand H6chandrac·2CH3CN·MeOH and the complexes
ꢁ
H and O H hydrogen atoms.
[(FeCl)3ACHTUNGTRENNUNG 3ACHUTNGTRENNGUN
(chandrac)] (3rac) and [(FeCl) (chandRR)] (3RR) were
determined by single-crystal X-ray diffraction at 173 K.
First, the molecular structures of H6chandrac, 3rac, and 3RR
are described followed by a comparative analysis of bond
and structural parameters. Finally, the crystal structures of
a typical salen-like square-planar coordination environment.
The coordination of one Clꢁ ion per FeIII ion results in ap-
proximate square-pyramidal coordination environments (t
3
rac and 3RR are illustrated.
values:[77] rac: 0.18, 0.38, and 0.20 for Fe1, Fe2, and Fe3, re-
3
The triplesalen ligand H6chandrac·2CH3CN·MeOH crystal-
spectively; 3RR: 0.12, 0.26, and 0.22 for Fe1, Fe2, and Fe3,
respectively). In 3rac, the Fe1, Fe2, and Fe3 atoms are ex-
truded at 0.59, 0.61, and 0.55 ꢁ, respectively, out of the
lizes in the space group P21/n. The asymmetric unit of the
racemic ligand H6chandrac contains one enantiomer of the
molecule with the other enantiomer generated by a crystal-
lographic center of inversion and a mirror plane (the unit
cell contains four molecules, two of each enantiomer). The
molecular structure of one enantiomer of H6chandrac is given
in Figure 1 (thermal ellipsoid plot in Figure S1, Supporting
Information). The configurations of the trans-cyclohexane-
diimine backbone for this enantiomer are 13R,18R, 23S,28S,
and 33S,38S. It is interesting to note that the two pendent
arms of the same chirality of the cyclohexanediimine parts
are to the same side with regard to the central phlorogluci-
nol unit, whereas the third pendent arm of opposite chirality
points to the opposite side. This corroborates the argument
given in the synthesis and characterization section that the
third pendant arm of the same chirality would exhibit steric
crowding.
basal coordination plane consisting of the N2O2 donor atoms
6ꢁ
of the ligand (chandrac
)
in the direction of the apical chlo-
ride. In 3RR, the Fe1, Fe2, and Fe3 atoms are lying above the
best N2O2 plane at a respective distance of 0.58, 0.60, and
0.62 ꢁ towards the apical chloride atom.
In complex 3RR, all the trans-cyclohexanediimine back-
bones are in an R,R configuration and the chloride ions at-
tached to Fe1, Fe2, and Fe3 atoms are in a cis position to
each other. It is noticeable that the position of the Fe2 atom
is disordered by 10% to the Fe2A atom. In complex 3rac, the
trans-cyclohexanediimine moieties exhibit different configu-
rations. In the molecule shown in Figure 2a, the trans-13,18-
cyclohexanediimine ring with N11 and N12 atoms is in an
R,R configuration, whereas the two other atoms are in the
S,S configuration. This goes along with Cl1 bound to Fe1
(cyclohexadiamine ring in a R,R configuration), which
points in the opposite direction relative to Cl2 bound to Fe2
and Cl3 bound to Fe3 (cyclohexadiamine rings in a S,S con-
figuration). The absolute orientation of Cl1 in 3rac corre-
sponds to the orientations of all three chlorides in 3RR (all
The racemic complex [(FeCl)3ACTHNUTRGNEUNG
(chandrac)] (3rac) crystallizes
with one molecule in the asymmetric unit in the chiral space
group P61, that is, a spontaneous resolution of the enantio-
mers upon crystallization has occurred. The chiral complex
[(FeCl)3ACHTUNGTRENNUNG
(chandRR)] (3RR) crystallizes in the chiral space
group P212121.
cyclohexanediamine rings are in a R,R configuration).
The molecular structures of 3rac and 3RR are shown in Fig-
ure 2a and b, respectively. In both complexes, a six-fold de-
protonated triplesalen ligand coordinates three FeIII ions in
Hence, it is evident that the position of the chloride atom is
highly dependent on the configuration of the trans-cyclohex-
anediimine backbone.
Chem. Eur. J. 2010, 16, 10137 – 10149
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10141

T. Glaser et al.
Table 2. Selected interatomic distances [ꢁ] for H6chandrac, 3rac, and 3RR
.
H6chandrac
3rac
3RR
Fe1–O11
Fe1–O12
Fe2–O21
Fe2–O22
Fe3–O31
Fe3–O32
Fe1–N11
Fe1–N12
Fe2–N21
Fe2–N22
Fe3–N31
Fe3–N32
Fe1–Cl1
Fe2-Cl2
Fe3–Cl3
C1–O11
C101–O12
C3–O21
C201–O22
C5–O31
C301-O32
C11–N11
C19–N12
C21–N21
C29–N22
C31–N31
C39–N32
C2–C11
1.894(7)
1.876(8)
1.862(7)
1.860(10)
1.880(7)
1.878(2)
1.898(2)
1.885(2)
1.881(2)
1.894(2)
1.902(2)
2.132(3)
2.076(3)
2.142(3)
2.077(3)
2.169(2)
2.061(2)
2.217(1)
2.194(2)
2.192(1)
1.312(4)
1.318(4)
1.305(4)
1.319(4)
1.303(4)
1.315(4)
1.302(4)
1.279(5)
1.304(4)
1.282(4)
1.294(4)
1.271(4)
1.471(4)
1.474(4)
1.475(4)
1.425(6)
1.437(5)
1.451(5)
7.249(1)
7.056(1)
7.045(1)
1.890ACTHNUGRTNEUNG(8
2.127(9)
2.070(9)
2.113(11)
2.085(10)
2.142(8)
2.046(8)
2.205(5)
2.215(5)
2.212(5)
1.344(12)
1.337(14)
1.322(12)
1.358(15)
1.307(11)
1.322(12)
1.308(14)
1.324(13)
1.250(14)
1.266(16)
1.273(12)
1.278(13)
1.475(15)
1.472(14)
1.465(14)
1.443(16)
1.381(19)
1.426(14)
7.391(2)
7.177(2)
7.371(2)
1.277(5)
1.359(5)
1.259(5)
1.356(5)
1.264(5)
1.355(5)
1.327(5)
1.284(5)
1.307(5)
1.279(5)
1.304(5)
1.279(5)
1.429(5)
1.427(6)
1.428(5)
1.456(6)
1.462(6)
1.445(6)
C4–C21
C6–C31
C102–C19
C202-C29
C302–C39
Fe1···Fe2
Fe2···Fe3
Fe1···Fe3
planer form but a little bit twisted as evidenced by the larg-
est distance of a constituting carbon atom of 0.10 ꢁ from
the best plane. For the terminal benzene rings, the strongest
deviation is only 0.02 ꢁ. In the trinuclear complexes, the
ꢁ
central C C bond lengths are shortened to 1.41 and 1.42 ꢁ
for 3rac and 3RR, respectively, in comparison to H6chandrac
,
ꢁ
whereas the terminal benzene rings exhibit a mean C C
bond length of 1.40 ꢁ. Additionally, the deviation of the
central benzene ring from planarity is not pronounced in the
complexes (largest deviation from the best plane is 0.03 and
0.01 ꢁ for 3rac and 3RR, respectively).
Figure 2. Molecular structures of the trinuclear complexes a) 3rac and
b) 3RR; thermal ellipsoids are drawn at the 50% probability level; hydro-
gen atoms are omitted for clarity.
ꢁ
Selected interatomic distances are summarized in Table 2.
The longer C C bond lengths in the central ring of
Ph
rac
The mean Fe O , Fe N, and Fe Cl bond lengths for 3
H6chandrac coincide with shorter mean C O lengths (value
ꢁ
ꢁ
ꢁ
ꢁ
(3RR) are 1.88 (1.89), 2.10 (2.11), and 2.21 (2.20) ꢁ, respec-
tively. These are in good agreement with the other FeIII
chloride complexes of salen-type ligands.[78–87] On the other
hand, the bond lengths in the ligand show some interesting
for terminal rings in parenthesis) of 1.27 (1.36) ꢁ, longer C
ꢁ
N lengths of 1.31 (1.28) ꢁ, and shorter CAr C bond lengths
ꢁ
of 1.43 (1.45) ꢁ. These bond lengths are a clear indication
that in the ligand the central moiety is in the keto–enamine
form rather than in the conventionally assumed phenol–
imine form.[88,89] This bond-length-based assignment is corro-
borated by difference Fourier synthesis. The position of the
hydrogen atoms in the central moiety have been found to
features. In the molecular structure of the ligand H6chandrac
,
the average C–C distance of the central C6 ring is 1.45 ꢁ,
ꢁ
which is longer than the average C C bond length of the
terminal benzene rings (1.40 ꢁ) and considerably closer to
ꢁ
ꢁ
the typical values of a C C single bond length. Furthermore,
the central C6 ring (phloroglucinol unit) is no longer in a
localize on N11, N21, and N31 nitrogen atoms (N11 H11:
ꢁ
ꢁ
0.96, N21 H21: 0.92, N31 H31 1.05 ꢁ), rather than on O11,
10142
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10137 – 10149

Chiral Trinuclear FeIII Triplesalen Complexes
FULL PAPER
Table 3. Selected structural properties for the iron triplesalen complexes
O21, and O31 oxygen atoms. Contrarily, at the terminal phe-
nols the hydrogen atoms could be located at the phenol
oxygen atoms and not the imine nitrogen atoms. In conjunc-
tion with the structural parameters in bond distances it is
clearly established that the terminal donor sites in
H6chandrac are in the usual phenol–imine form and not in
the tautomeric keto–enamine form.
In the complexes, the acidic hydrogen atoms are absent
and the resulting forms, that is, the phenolate–imine and
keto–enamide, are no longer tautomers but rather resonance
structures of a delocalized system. From a comparison of
the relevant bond distances, the prevalence of the usually as-
sumed phenolate–imine resonance structure is indicated.
However, a partial loss of the central p system as evidenced
3rac and 3RR
.
3rac
3RR
0.84
0.71
1.20
d [ꢁ][a]
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
Fe1
Fe2
Fe3
ꢁ0.66
0.57
0.93
12.8
22.4
14.2
39.7
34.0
24.7
32.2
13.9
15.7
20.8
19.9
18.6
11.6
23.2
12.6
a [8][b]
b [8][c]
16.8
22.4
25.5
14.4
18.3
5.9
19.9
35.0
23.2
22.0
14.2
32.9
14.1
17.5
25.0
g [8][d]
fcentral [8]
ꢁ
by C C bond lengths of 1.41–1.42 ꢁ is the main result of
this structural analysis.
fterminal [8]
Trinuclear triplesalen complexes are known to exhibit var-
ious degrees of ligand folding.[32,35,37–39,42,44] A regular ligand
[a] d is the shortest distances of an FeIII ion from the best plane formed
by the six carbon atoms of the central benzene ring of the phloroglucinol
unit. A negative value corresponds to a displacement to the other side of
the central plane. [b] a is the angle between the best plane of 1) N2O2
and 2) the benzene of the phloroglucinol unit. [c] b is the angle between
the best plane of 1) N2O2 and 2) the terminal phenolate. [d] g is the angle
between the best plane of 1) the benzene of the phloroglucinol unit and
2) the benzene of the terminal phenolate.
folding results in an overall bowl-shaped molecular structure
6ꢁ
for the trinuclear CuII and NiII complexes of (talentBu ) . We
2
have applied several parameters for a quantitative descrip-
tion of the ligand folding in the study of the trinuclear
tripleACHTUNGTRENNUNG
salen complexes.[35,39] It turned out that the best param-
eters to quantitatively describe the ligand folding are the
bent angles fcentral and fterminal. The bent angle f (introduced
by Cavallo and Jacobsen)[90] is defined by f=1808ꢁ<(M-
X
NO-XR) (XNO: midpoint of adjacent N and O donor atoms,
XR: midpoint of the six-membered chelate ring containing
the N and O donor atoms). We identified the bent angles
fcentral and fterminal in trinuclear CuII and NiII triplesalen com-
plexes best suited to differentiate between a bending along
an idealized line through neighboring N and O ligands and
a line perpendicular to the former, resulting in a helical dis-
tortion.[35,39] The application to the heptanuclear complexes
[Mn6Cr]3+, [Mn6Fe]3+, and [Mn6Co]3+ also showed that a
strong folding occurred at the OPh–Nimine vector of the cen-
tral unit and only a minor folding appeared at the terminal
phenolates.[37,42,47] Common to the CuII, NiII, and MnIII ions
in these complexes is a relatively well-defined planar N2O2
plane of the salen-like ligand compartment. This is in con-
trast to the five-coordinate FeIII ions in 3rac and 3RR. The co-
ordination environment exhibits a more or less strong devia-
tion from square-pyramidal to triginal-bipyramidal (t values
vide supra) with the central phenolate and the terminal
imine donors as axial ligands.
Applying the usual parameters (Table 3) results in no
clear trend in the folding, which is also evident in the two
structural presentations provided in Figure 3. While the dif-
ferent cyclohehanediamine configurations in 3rac prohibit an
overall regular folding, the ligand folding in 3RR does also
not provide a bowl-shaped structure observed in the trinu-
6ꢁ
clear complexes of the ligand (talentBu ) . This is mainly evi-
2
dent by the relative bendings of the central phenolates and
the terminals phenolates, which are in opposite directions in
6ꢁ
RR, but in the same direction in (talentBu
)
complexes.
2
3
Figure 3. Molecular structures of a) 3rac and b) 3RR, which demonstrate
the unregular ligand folding. The tert-butyl groups were omitted for clari-
ty.
The chloride atoms Cl2 bound to Fe2 in 3rac form a weak
intermolecular interaction (3.426 ꢁ) to Fe3 of the next tri-
Chem. Eur. J. 2010, 16, 10137 – 10149
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10143

T. Glaser et al.
plesalen complex (Figure 4a). This chain formation occurs
along a C3 axis running along c leading to a helical structure.
These helices pack along c resulting in very large chiral
Figure 5. 57Fe Mçssbauer spectra of a) 3rac and b) 3RR measured at 80 K.
*
: Exp; c: Sim.
ly reported square-pyramidal FeN2O2Cl salen systems[79,91]
and contain the ferric high-spin Si =5/2 configuration for the
iron ions.
Electronic absorption spectrum: The electronic absorption
spectra (10000–40000 cmꢁ1) of the half units (2rac and 2RR),
the free ligands (H6chandrac and H6chandRR), and the corre-
sponding complexes (3rac and 3RR) in CH2Cl2 are shown in
Figure 6, whereas the spectral data are summarized in
Figure 4. a) Intermolecular Cl···Fe interactions in 3rac leading to helices
along c. b) Space-filling model of 3rac to visualize the large chiral channels
along c.
channels (Figure 4b). Though we cannot rule out some ill-
defined solvent molecule, no solvent molecule has been
found inside the channels. The diameter of these channels is
~14.5 ꢁ based on atom to atom distances and ~11 ꢁ consid-
ering van der Waals radii. These empty channels make up
33.6% of the volume of the crystals. In 3RR, there are no
specific intermolecular interactions. However, large empty
voids of 560 ꢁ3 are apparent.
Figure 6. Electronic absorption spectra of 2rac
,
2RR
,
H6chandrac
,
H6chandRR, 3rac, and 3RR measured at ambient temperatures in CH2Cl2.
Table 4. It should be noted that the spectra of the diastereo-
meric pairs, that is, 2rac and 2RR, H6chandrac and H6chandRR
,
and 3rac and 3RR, are not identical with the strongest differ-
ences observed in the spectra of the two ligands.
Mçssbauer spectroscopy: Mçssbauer spectra of 3rac and 3RR
at 80 K are shown in Figure 5. Both spectra exhibit a single
unsymmetrical quadrupole doublet with isomer shifts (d)
and quadrupole splittings (jDEQ j) of d=0.44 mmsꢁ1, jDEQ j
The spectra of the two triplesalen half units 2rac and 2RR
exhibit two strong absorption features at ~29000 and
~33000 cmꢁ1. These transitions have been assigned to p–p*
transitions involving the ketimine groups.[35] A strong fea-
ture between 35000 and 40000 cmꢁ1 for p–p* transitions as-
sociated with the phenolic chromophore is not observed in
=1.10 mmsꢁ1 for 3rac
,
and d=0.44 mmsꢁ1, jDEQ j=
1.13 mmsꢁ1 for 3RR. These parameters are close to previous-
10144
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10137 – 10149

Chiral Trinuclear FeIII Triplesalen Complexes
FULL PAPER
value for three noninteracting high-spin FeIII(d5) centers
(meff =10.25 mB, g=2.00). The meff value for 3rac·1.5H2O re-
mains roughly constant up to 40 K. Below 40 K, a pro-
nounced decrease in meff occurs resulting in meff =5.61 mB at
2 K. Complex 3RR shows almost a temperature-independent
value of meff =10.46 mB up to 15 K. Below 15 K, the meff value
decreases abruptly to meff =6.81 mB at 2 K. The close resem-
blance of the meff value between the temperature range 40–
290 K for 3rac·1.5H2O and 15–290 K for 3RR indicates that
both complexes are almost noninteracting trinuclear high-
spin FeIII systems.
Table 4. Spectral data of the triplesalen half units, ligands, and trinuclear
iron complexes.
n˜ [cmꢁ1] (e [mꢁ1 cmꢁ1])
2rac
28950
28600
29750
28750
18950
19250
N
ACHTUNGTRENNUNG
2RR
ACHTUNGTRENNUNG
H6chandrac
H6chandRR
3rac
ACHTUNGTRNE(NUNG 50000)
ACHTUNGTRENNUNG
G
ACHTUNGTRENNUNG
3RR
C
ACHTUNGTRENNUNG
the half units. Interestingly, by adding the terminal imine–
phenol units, that is, on going from 2X to H6chandX, a strong
absorption feature at ~38000 cmꢁ1 appears with some inten-
sity modification in the 25000–35000 cmꢁ1 region. This im-
plies that the terminal phenol-imine units absorb mainly at
higher energies (35000–40000 cmꢁ1), whereas the lower ab-
sorptions at 25000–35000 cmꢁ1 are mainly due to the central
keto–enamine unit, which provides a large delocalized p
system. Despite some changes in the region above
25000 cmꢁ1, the two FeIII complexes 3rac and 3RR exhibit a
new strong band at ~19000 cmꢁ1. This band is attributed to
The data were analyzed by full-matrix diagonalization of
the appropriate spin-Hamiltonian [Eq. (1)] for trinuclear
ferric high-spin complexes including isotropic HDvV ex-
change, zero-field splitting, and Zeeman interactions.[95]
^
H ¼ ꢁ2JðS1S2 þ S2S3 þ S3S1Þ
3
3
X
X
ð1Þ
1
3
þ
ðDðS2z,iꢁ SiðSi þ 1ÞÞÞ þ mB
ðgSiBÞ
i¼1
i¼1
phenolic oxygen(pp) to FeIII
tions.[81,92–94]
ACHTUGNRTEN(NUGN dp*) charge-transfer transi-
Magnetic moments were obtained from numerically gen-
erated derivatives of the eigenvalues of Equation (1) and
summed up over 16-field orientations along a 16-point Lebe-
dev grid to account for the powder distribution of the
sample.
Magnetic measurements: Variable-temperature magnetic
susceptibility measurements have been measured on pow-
dered samples of 3rac·1.5H2O and 3RR in the temperature
range of 2–290 K with an applied field of 1 T (Figure 7).
3rac·1.5H2O and 3RR show meff =10.10 and 10.46 mB at 290 K,
respectively. These values are close to the expected meff
The decrease of meff in the low-temperature range could
be either due to saturation effects (which are taken into ac-
count in our simulation), a weak anti-ferromagnetic intra-
molecular interaction between the high-spin FeIII(d5) centers,
zero-field splitting, and/or intermolecular antiferromagnetic
interactions. Simulations with no zero-field splitting (iso-
tropic limit), with no exchange interaction (uncoupled
limit), and by taking into account both contributions were
performed. Intermolecular anti-ferromagnetic interactions
were also considered.
By using the isotropic limit without zero-field splitting,
the experimental data are well reproduced by using g=1.99,
J=ꢁ0.14 cmꢁ1 for 3rac·1.5H2O (c in Figure 7a) and g=
2.04, J=ꢁ0.015 cmꢁ1 for 3RR (c in Figure 7b). In the un-
coupled limit the experimental results were reproduced by
using g=1.97, D=ꢁ8.2 cmꢁ1 for 3rac·1.5H2O (c in Fig-
ure 7a) and g=2.04, D=ꢁ2.5 cmꢁ1 for 3RR (c in Fig-
ure 7b). By considering only intermolecular interactions by
using a Curie–Weiss approach, the experimental data for
3RR were fitted with g=2.04 and q=ꢁ0.27 K (g in Fig-
ure 7b). Furthermore, by considering interactions between
the FeIII centers and the presence of zero-field splitting in
the system, the experimental data could be simulated by
using the following parameters: g=1.99, D=ꢁ2.0 cmꢁ1, J=
ꢁ0.12 cmꢁ1 for 3rac·1.5H2O (g in Figure 7a) and g=2.04,
D=ꢁ2.7 cmꢁ1, J=0.005 cmꢁ1 for 3RR (d in Figure 7b). It
is evident from the almost superimposed simulations in
Figure 7 that the system is highly over parameterized and
the identification of one unique parameter set is mathemati-
cally not justified. For this reason, we have only applied a
coupling scheme for a trinuclear unit of exact C3 symmetry,
Figure 7. Temperature-dependence of the effective magnetic moment,
meff, of a) 3rac·1.5H2O (c: g=1.99, J=ꢁ0.14 cmꢁ1, D=0; a: g=1.97,
J=0, D=ꢁ8.2 cmꢁ1
;
g: g=1.99, J=ꢁ0.12 cmꢁ1
,
D=ꢁ2.0 cmꢁ1
;
:
;
*
Exp) and b) 3RR measured at 1 T (c: g=2.04, J=q=0, D=ꢁ2.5 cmꢁ1
c: g=2.04, J=ꢁ0.0015 cmꢁ1, D=q=0; g: g=2.04, q=ꢁ0.27 K,
D=J=0; d: g=2.04, J=0.0015 cmꢁ1, D=ꢁ2.7 cmꢁ1
;
*
: Exp).
Chem. Eur. J. 2010, 16, 10137 – 10149
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10145

T. Glaser et al.
Table 5. Catalytic oxidation of sulfides by using 3rac [a]
.
which is a relative good approximation for 3RR, but less
good for 3rac. Reducing the symmetry would even lead to a
more pronounced over parameterization.
To focus further on the low-temperature magnetic behav-
ior of 3rac·1.5H2O and 3RR we have measured variable-tem-
perature variable-field (VTVH) magnetization data (1, 4,
and 7 T), which are strongly sensitive to zero-field splitting
effects. Fitting of the VTVH magnetization data for
3rac·1.5H2O (Figure 8a) indicates the existence of both weak
Sulfide
Sulfoxide
[%]
Sulfone
[%]
Selectivity
[%][b]
PhIO (a),
NaOCl (b)[c]
I
I
I
I
I
I
II
II
III
III
IV
IV
59
72
97
37
58
86
30
84
50
90
72
98
<1
<1
<1
63
42
5
>99
>99
>99
37
58
94
100
91
96
97
99
1.1 (a)
1.5 (a)
2.2 (a)
10 (b)
4 (b)
2 (b)
n.d.[d]
8
2
3
1
1.1 (a)
2.2 (a)
1.1 (a)
2.2 (a)
1.1 (a)
2.2 (a)
2
98
[a] 5 mol% catalyst, CH2Cl2 as the solvent, room temperature. [b] Selec-
tivity=[RSOR’]/[(RSOR’+RSO2R’)]. [c] Values of PhIO (a) and
NaOCl (b) correspond to the equivalents of the oxidant to the substrate.
[d] n.d.=not detected.
tivities (sulfoxide to sulfone) are good when using both 3rac
and 3RR complexes as the catalyst. The yield appears to
depend on the equivalents of the external oxidant PhIO
(Table 5). By using NaOCl as an external oxidant, the cata-
lytic oxidation reaction can also be performed. In this case,
the selectivity depends on the concentration of NaOCl.
When the concentration of NaOCl is in a 10-fold excess to
the concentration of thioanisole (Table 3), the selectivity is
only 37%. Upon lowering the equivalents of NaOCl to a 2-
fold excess relative to the equivalents of thioanisole, 94%
selective formation of sulfoxide over sulfone is obtained
with 91% conversion of the substrate.
We measured the enantiomeric excess (ee) for those cata-
lytic reactions for which the enantiomerically pure chiral
complex 3RR was used as the catalyst. The results are listed
in Table 6. Though, the enantiomeric excess is low for all
the substrates, it is worth noting that on addition of a bulky
Br group at the para position of the phenyl ring of thioani-
sole (substrate III), the ee increases from 13 to 26%. Simi-
larly, on addition of a bulky phenyl group to the methylene
position of the thioanisole (substrate II), 20% ee has been
Figure 8. Variable-temperature variable-field (VTVH) magnetization
measurements of a) 3rac·1.5H2O (g=1.99, J=ꢁ0.10 cmꢁ1, D=ꢁ2.75 cmꢁ1
;
RR
*
: Exp (1 T, 4 T, 7 T)) and b) 3
at 1 T, 4 T, and 7 T (g=2.04, J=
0.005 cmꢁ1, D=ꢁ2.0 cmꢁ1
;
: Exp (1 T, 4 T, 7 T)).
*
antiferromagnetic interactions (J=ꢁ0.1 cmꢁ1) and zero-field
splitting (D=ꢁ2.75 cmꢁ1) in the trinuclear high-spin FeIII
(d5, Si =5/2) system. The inspection of VTVH data for 3RR
(Figure 8b) shows a decrease in the magnetization value
with increasing mBH/kT at 4 T, which is more pronounced at
1 T. This decrease is most likely due to dipolar antiferro-
magnetic intermolecular interactions, which are depressed
with increasing magnetic field strength. The experimental
magnetization data obtained at 7 T, at which the intermolec-
ular interactions have the weakest effect, can be simulated
with g=2.04, D=ꢁ2.0ꢀ0.1 cmꢁ1, J=0.005ꢀ0.005 cmꢁ1.
In summary, the magnetic measurements reveal that in
both complexes the intramolecular interactions are only
very weak but the FeIII ions possess relatively strong mag-
netic anisotropies.
[a]
Table 6. Catalytic enantioselective oxidation of sulfides by using 3RR
.
Sulfide
Sulfoxide
[%]
Sulfone
[%]
Selectivity
[%][b]
ee [%]
(configuration)[c]
I
II
95
96
70
84
97
3
4
1
2
3
97
96
98
98
97
13 (R)
20 (R)
30 (R)
26 (R)
10 (R)
II[d]
III
IV
Catalytic reactivity: Oxidation of prochiral sulfides to the
corresponding sulfoxides: Complex 3rac and 3RR were used as
catalysts (5 mol%) to oxidize prochiral sulfides at room
temperature under the conditions stated in the Experimen-
tal Section. The results are summarized in Tables 4 and 5,
respectively. The yields of the sulfoxides as well as the selec-
[a] 5 mol% catalyst, 2.2 equivalents of PhIO, CH2Cl2 as the solvent, room
temperature (unless otherwise stated). [b] Selectivity=[RSOR’]/
[(RSOR’+RSO2R’)]. [c] ee values were measured by using a chiral OD-
H column. [d] 08C, 17 h.
10146
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10137 – 10149

Chiral Trinuclear FeIII Triplesalen Complexes
FULL PAPER
achieved. When the catalysis with substrate II is performed
at 08C, the ee value increases from 20 to 30%, but the yield
decreases from 96 to 70%. Interestingly, for all catalytic re-
actions using 3RR as the catalyst production of the R over S
isomer dominates.
cies. In this respect, the stabilization of the oxidized species
leading to a reduced activity and thus a higher selectivity is
a cooperative effect in the trinuclear catalysts 3RR and 3rac
.
The lower ee values for the trinuclear catalyst 3RR than
that of the mononuclear reference [FeClACTHNUTRGNEUNG(salen’)] is related
Due to a different solubility, we could not adapt the exact
catalytic protocol used for the application of [FeCl-
to the ligand folding observed for 3RR (Figure 3), which does
not produce a bowl-shaped structure. The structure of
[FeClACHTUGNTERN(UNG salen’)] has not been published yet. However, closely
A
properties of the trinuclear complex 3RR to its mononuclear
analogue, we performed parallel reactions with both cata-
lysts (Table 7) for the asymmetric sulfoxidation of benzyl
related complexes[84,86,98–100] exhibit only minor ligand fold-
ings. As is evident from Figure 3, the bending of the termi-
nal phenolates is to the opposite direction of the Fe–Cl unit.
Assuming a similar orientation in the active high-valent Fe–
O species, there is no strong steric hindrance for the attack
of a sulfide to the Fe–O unit. Furthermore, the accessibility
of the active Fe–O unit must be easier in the trinuclear cata-
lyst relative to the mononuclear analogue. While this result
is quite surprising, there are several possibilities in the tri-
plesalen ligand system to increase steric bulk.
Table 7. Catalytic oxidation of benzyl phenyl sulfide by using [(FeCl)3-
ACHTUNGTRENNUNG ACHTUNGTRENNUNG
(chandRR)] (3RR) and [FeCl(salen’)] complexes.[a]
Catalyst ([mol%]) PhIO
Sulfoxide Sulfone Selectivity ee [%]
A
[%]
[%][b]
(configuration)[c]
N
3A
96
4
96
20 (R)
A
E
2.2
78
9
22
78
100
63 (R)
n.m.
[e]
E
G
n.d.[e]
Conclusion
N
N
60
71
2
97
100
48 (R)
10 (R)
R
G
n.d.[e]
Triplesalen ligands bridge three salen coordination environ-
ments in a meta-phenylene arrangement by a common
phloroglucinol backbone. The hexa-anionic forms are tri-
ple(tetradentate) ligands and are capable of coordinating
three transition-metal ions. The three metal salen subunits
are electronically interacting through a strongly conjugated
central phloroglucinol unit, which we thought would be
useful for cooperative effects in catalysis.[35]
G
E
1.1
82
3
96
47 (R)
[a] All the reactions were carried out for 15 h, CH2Cl2 as the solvent, room
temperature (unless otherwise stated). Values of PhIO correspond to the equiv-
alent of PhIO to the concentration of the substrate. [b] Selectivity=[RSOR’]/-
ACHTUNGTRENNUNG([RSOR’]+[RSO2R’]). [c] ee values were obtained by measuring chiral HPLC
by using a chiral OD-H column. [d] Reactions were carried out for 1 h.
[e] n.d.=not detected; n.m.=not measured.
In this respect, we envisioned to incorporate the chiral
trans-1,2-cyclohexanediamine for the realization of the
chiral triplesalen ligand H6chand as the C3 symmetric trinu-
cleating extension of Jacobsenꢀs ligand H2salen’. Herein, we
have presented the synthesis and characterization of the two
diastereomeric ligands H6chandrac and H6chandRR in two
steps with the half units 2rac and 2RR as intermediates. The
phenyl sulfide. By using 5 mol% catalyst, the trinuclear
complex 3RR has a better performance with respect to yield
and selectivity, whereas the ee is tremendously reduced.
Going to only 1 mol% reduces the yield for 3RR from 96 to
9%. This implies that the reaction of 3RR is slower than that
of [FeCl
N
former is synthesized from the racemic (R,R)/
cyclohexanediamine mixture resulting in four stereoisomers.
The racemic (R,R)(R,R)(S,S)/(S,S)(S,S)(R,R) pair is ob-
tained in excess to the racemic (R,R)(R,R)(R,R)/(S,S)(S,S)-
(S,S) pair due to steric hindrance for the formation of the
third pendant arm of the same chirality, which are all on the
same side relative to the phloroglucinol backbone. The use
of the enantiomerically pure (R,R)-trans-1,2-cyclohexanedi-
ACHTUNGTREN(NUGN S,S)-trans-1,2-
G
E
G
R
ACHTUNGTRENNUNG
These results may be rationalized by the assumption, that
the high-valent, catalytically active Fe–O species is stabilized
in the trinuclear complex relative to the mononuclear com-
plex. This would explain not only the slower reaction of 3RR
but also its higher selectivity, so that the oxidized species is
less capable of oxygenating the sulfoxide product to sulfone.
The assumption of a stabilization of the active Fe–O spe-
cies in the trinuclear complex is supported by our studies on
the trinuclear NiII and CuII triplesalen complexes.[35,39]
G
G
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
Reaction of the half units with 3,5-di-tert-butylsalicylalde-
hyde provides the racemic ligand H6chandrac and the enan-
tiomerically pure ligand H6chandRR. Reaction of the ligands
with FeCl3 provided the trinuclear complexes [(FeIIICl)3-
These complexes are either ligand-centered oxidized at the
central unit[97] in the case of CuII or exhibit a valence-tauto-
merism between a ligand-centered oxidized form and a
(chandrac)] (3rac) and [(FeIIICl) (chandRR)] (3RR).
ACHUTNGTRENNUG CHUTNGTRENNUNG
3A
The molecular structure of H6chandrac shows the presence
metal-oxidized NiIII species in the case of NiII . Either form
is inherently mixed-valent. A comparison of the potential
for oxidation with those of the respective mononuclear ana-
logues provided a stabilization of the oxidized species in the
trinuclear complexes relative to the mononuclear salen spe-
of different tautomers: whereas the terminal phenols are in
the conventionally anticipated phenol–imine form, the cen-
tral unit is in the tautomeric keto–enamine form. Compari-
son of the structural parameters of the ligand and the com-
Chem. Eur. J. 2010, 16, 10137 – 10149
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10147

T. Glaser et al.
[5] M. C. Carreno, G.-H. Torres, M. Ribagorda, A. Urbano, Chem.
Commun. 2009, 6129–6144.
[6] K. K. Andersen, Tetrahedron Lett. 1962, 3, 93–95.
[7] H. B. Kagan, Ed.: I. Ojima, Asymmetric Synthesis, VCH, Wein-
heim, 1993, pp. 203–226.
[8] S. Colonna, N. Gaggero, L. Casella, G. Carrea, P. Pasta, Tetrahe-
dron: Asymmetry 1992, 3, 95–106.
[9] J.-M. Brunel, H. B. Kagan, Bull. Soc. Chim. Fr. 1996, 133, 1109–
1115.
[10] S. Colonna, A. Manfredi, M. Spadoni, L. Casella, M. Gullotti, J.
Chem. Soc. Perkin Trans. 1 1987, 71–73.
[11] F. Di Furia, G. Licini, G. Modena, R. Motterle, J. Org. Chem. 1996,
61, 5175–5177.
[12] H. B. Kagan, T. Luukas, M. Eds.: M. Beller, C. Bolm, Transition
Metals for Organic Synthesis, Wiley-VCH, Weinheim, 1998,
pp. 361–373.
[13] C. Bolm, K. Muniz, J. Hildebrand, Comprehensive Asymmetric Cat-
alysis, Springer, Berlin, 1999, pp. 697–713.
[14] J. Sun, C. Zhu, Z. Dai, M. Yang, Y. Pan, H. Hu, J. Org. Chem.
2004, 69, 8500–8503.
[15] A. H. Vetter, A. Berkessel, Tetrahedron Lett. 1998, 39, 1741–1744.
[16] C. Bolm, Coord. Chem. Rev. 2003, 237, 245–256.
[17] H. Fu, H. Kondo, Y. Ichikawa, G. C. Look, C. H. Wong, J. Org.
Chem. 2002, 67, 7265–7270.
[18] J. K. Whitesell, M. S. Wong, J. Org. Chem. 1994, 59, 597–601.
[19] K. Nakajima, M. Kojima, K. Toriumi, K. Saito, J. Fujita, Bull.
Chem. Soc. Jpn. 1989, 62-62, 760–767.
[20] J. Burnel, M.-H. B. Kagan, Bull. Chem. Soc. Fr. 1996, 133, 1109–
1115.
[21] C. Sasaki, K. Nakajima, M. Kojima, J. Fujita, Bull. Chem. Soc. Jpn.
1991, 64, 1318–1324.
[22] C. Bolm, O. A. G. Dabard, Synlett 1999, 360–362.
[23] G. Liu, D. A. Cogan, J. A. Ellman, J. Am. Chem. Soc. 1997, 119,
9913.
[24] C. Bolm, F. Bienewald, Angew. Chem. 1995, 107, 2883–2885;
Angew. Chem. Int. Ed. Engl. 1995, 34, 2640–2642.
[25] M. Palucki, P. Hanson, E. N. Jacobsen, Tetrahedron Lett. 1992, 33,
7111–7114.
[26] K. Noda, N. Hosoya, R. Irie, Y. Yamashita, T. Katsuki, Tetrahedron
1994, 50, 9609–9618.
plexes reveals that the terminal donors are well described
by the phenolate–imine resonance structure, whereas the
central donors have also some contributions from the keto–
enamide resonance structure. We are currently performing
detailed NMR spectroscopic studies on our phloroglucinol-
based ligand systems to obtain an experimental handle on
the weight of this resonance structure. The presence of this
resonance structure might be a reason for the relatively
weak ferromagnetic interactions by means of the spin-polar-
ization mechanism through the triplesalen-based systems.
This would provide us with a synthetic handle to rationally
optimize the strength of the ferromagnetic coupling.
The magnetic measurements reveal quite strong magnetic
anisotropies for the high-spin FeIII ions. As the high-spin
FeIII ion is usually inherently magnetically isotropic, there
have been reports on even stronger anisotropies for salen
FeIII complexes in the literature quantified by zero-field
splitting parameters jDj of 4–6 cmꢁ1,[101] 7.5,[102] 10.2,[103] and
7.2[103] In this respect, the strategy to incorporate salen-like
ligand environments to introduce a strong magnetic aniso-
tropy[32,37] seem to work.
Both complexes 3rac and 3RR show good catalytic perfor-
mance in the sulfoxidation of sulfides relative to mononu-
clear iron chloride salen systems. They provide high yields
and high selectivities favoring the formation of the sulfox-
ides relative to the corresponding sulfones. However, the
chiral catalyst 3RR is not as good in the enantioselective sul-
foxidation of prochiral sulfides providing lower ee values
relative to [FeClACHTUNGTRENNUNG(salen’)]. This might be related to strong
ligand foldings observed in 3RR, which open up access to the
catalytically active Fe–O species.
The high yields and selectivities on the other hand are re-
lated to a stabilization of the active high-valent Fe–O spe-
cies, which results in a slower reaction with less side reac-
tions and more selectivity due to the decreased driving force
for overoxidation. The stabilization of the intrinsically
mixed valence system occurs by a delocalization over the
whole central unit, which involves also some communication
with the other two iron centers. In this respect, the trinu-
clear system exhibits a cooperative effect relative to the
mononuclear system by reducing the reactivity of the active
species, which is not possible in the mononuclear system.
[27] J. Legros, C. Bolm, Angew. Chem. 2003, 115, 5645–5647; Angew.
Chem. Int. Ed. 2003, 42, 5487–5489.
[28] V. K. Sivasubramanian, M. Ganesan, S. Rajagopal, R. Ramaraj, J.
Org. Chem. 2002, 67, 1506–1514.
[29] W. Zhang, E. N. Jacobsen, J. Org. Chem. 1991, 56, 2296–2298.
[30] K. P. Bryliakov, E. P. Talsi, Angew. Chem. 2004, 116, 5340–5342.
[31] T. Glaser, M. Gerenkamp, R. Frçhlich, Angew. Chem. 2002, 114,
3984–3986; Angew. Chem. Int. Ed. 2002, 41, 3823–3825.
[32] T. Glaser, M. Heidemeier, T. Lꢂgger, Dalton Trans. 2003, 2381–
2383.
[33] T. Glaser, T. Lꢂgger, R. Frçhlich, Eur. J. Inorg. Chem. 2004, 394–
400.
[34] T. Glaser, M. Heidemeier, S. Grimme, E. Bill, Inorg. Chem. 2004,
43, 5192–5194.
[35] T. Glaser, M. Heidemeier, R. Frçhlich, P. Hildebrandt, E. Bothe,
E. Bill, Inorg. Chem. 2005, 44, 5467–5482.
[36] T. Glaser, M. Heidemeier, E. Krickemeyer, H. Bçgge, Z. Natur-
forsch. 2006, 61, 753–757.
[37] T. Glaser, M. Heidemeier, T. Weyhermꢂller, R.-D. Hoffmann, H.
Rupp, P. Mꢂller, Angew. Chem. 2006, 118, 6179–6183; Angew.
Chem. Int. Ed. 2006, 45, 6033–6037.
Acknowledgements
This work was supported by the DFG (FOR945 “Nanomagnets: from
Synthesis via Interactions with Surfaces to Function”), the Fonds der
Chemischen Industrie, and Bielefeld University. Mr. B. Sammet and
Prof. Dr. N. Sewald (Chemistry Department, Bielefeld University) are
thankfully acknowledged for assistance in the chiral HPLC measure-
ments.
[38] T. Glaser, M. Heidemeier, R. Frçhlich, C. R. Chim. 2007, 10, 71–
78.
[39] T. Glaser, M. Heidemeier, J. B. H. Strautmann, H. Bçgge, A.
Stammler, E. Krickemeyer, R. Huenerbein, S. Grimme, E. Bothe,
E. Bill, Chem. Eur. J. 2007, 13, 9191–9206.
[1] I. Fernꢆndez, N. Khiar, Chem. Rev. 2003, 103, 3651–3705.
[2] P. Renaud, M. Gerster, Angew. Chem. 1998, 110, 2704–2722;
Angew. Chem. Int. Ed. 1998, 37, 2562–2579.
[40] H. Theil, C.-G. Freiherr von Richthofen, A. Stammler, H. Bçgge, T.
Glaser, Inorg. Chim. Acta 2008, 361, 916–924.
[3] H. Pellissier, Tetrahedron 2006, 62, 5559–5601.
[4] A. Korte, J. Legros, C. Bolm, Synlett 2004, 2397–2399.
10148
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 10137 – 10149

Chiral Trinuclear FeIII Triplesalen Complexes
FULL PAPER
[41] T. Glaser, H. Theil, M. Heidemeier, C. R. Chim. 2008, 11, 1121–
1136.
[42] T. Glaser, M. Heidemeier, E. Krickemeyer, H. Bçgge, A. Stamm-
ler, R. Frçhlich, E. Bill, J. Schnack, Inorg. Chem. 2009, 48, 607–
620.
[43] C.-G. Freiherr von Richthofen, C.-G. A. Stammler, H. Bçgge,
M. W. DeGroot, J. R. Long, T. Glaser, Inorg. Chem. 2009, 48,
10165–10176.
[73] A. W. Kleij, Eur. J. Inorg. Chem. 2009, 193–205.
[74] J. Lopez, S. Liang, X. R. Bu, Tetrahedron Lett. 1998, 39, 4199–
4202.
[75] K. B. M. Janssen, I. Laquierra, W. Dehaen, R. F. Parton, I. F. J.
Vankelecom, P. A. Jacobs, Tetrahedron: Asymmetry 1997, 8, 3481–
3487.
[76] J.-P. Costes, F. Dahan, A. Dupuis, J.-P. Laurant, Dalton Trans. 1998,
1307–1314.
[44] C. Mukherjee, A. Stammler, H. Bçgge, T. Glaser, Inorg. Chem.
2009, 48, 9476–9484.
[77] A. W. Addison, T. N. Rao, J. Reedijk, J. van Rijn, G. C. Verschoor,
Dalton Trans. 1984, 1349–1356.
[45] H. Theil, R. Frçhlich, T. Glaser, Z. Naturforsch. B 2009, 64, 1633–
1638.
[46] T. Glaser, M. Heidemeier, H. Theil, A. Stammler, H. Bçgge, J.
Schnack, Dalton Trans. 2010, 39, 192–199.
[47] E. Krickemeyer, V. Hçke, A. Stammler, H. Bçgge, J. Schnack, T.
Glaser, Z. Naturforsch. B 2010, 65, 295–303.
[78] J. E. Davies, B. M. Gatehouse, Chem. Commun. 1970, 1166.
[79] H.-L. Shyu, H.-H. Wei, G.-H. Lee, Y. Wang, Dalton Trans. 2000,
915–918.
[80] M. Gerloch, F. E. Mabbs, J. Chem. Soc. A 1967, 1900–1908.
[81] M. Lubben, A. Meetsma, F. v. Bolhuis, B. L. Feringa, Inorg. Chim.
Acta 1994, 215, 123–129.
[48] H. C. Longuet-Higgins, J. Chem. Phys. 1950, 18, 265–274.
[49] H. Iwamura, Adv. Phys. Org. Chem. 1991, 26, 179–253.
[50] K. Yoshizawa, R. Hoffmann, J. Am. Chem. Soc. 1995, 117, 6921–
6926.
[51] D. A. Dougherty, Acc. Chem. Res. 1991, 24, 88–94.
[52] A. A. Ovchinnikov, Theor. Chim. Acta 1978, 47, 297–304.
[53] S.-i. Mitsubori, T. Ishida, T. Nogami, H. Iwamura, Chem. Lett.
1994, 285–288.
[82] H. Fujii, Y. Funahashi, Angew. Chem. 2002, 114, 3790–3793;
Angew. Chem. Int. Ed. 2002, 41, 3638–3641.
[83] M. Gerloch, F. E. Mabbs, J. Chem. Soc. A 1967, 1598–1608.
[84] Z. Chu, W. Huang, L. Wang, S. Gou, Polyhedron 2008, 27, 1079–
1092.
[85] K. Oyaizu, E. Tsuchida, Inorg. Chim. Acta 2003, 355, 414–419.
[86] T. Kurahashi, K. Oda, M. Sugimoto, T. Ogura, H. Fujii, Inorg.
Chem. 2006, 45, 7709–7721.
[54] T. Ishida, S.-i. Mitsubori, T. Nogami, N. Takeda, M. Ishikawa, H.
Iwamura, Inorg. Chem. 2001, 40, 7059–7064.
[87] M. Tomita, N. Matsumoto, H. Akagi, H. Okawa, S. Kida, Dalton
Trans. 1989, 179–184.
[55] F. Lloret, G. De Munno, M. Julve, J. Cano, R. Ruiz, A. Caneschi,
Angew. Chem. 1998, 110, 143–145; Angew. Chem. Int. Ed. 1998,
37, 135–138.
[56] L. C. Francesconi, D. R. Corbin, D. N. Hendrickson, G. D. Stucky,
Inorg. Chem. 1979, 18, 3074–3080.
[57] H. Oshio, Chem. Commun. 1991, 240–241.
[58] J. A. McCleverty, M. D. Ward, Acc. Chem. Res. 1998, 31, 842–851.
[59] A. M. W. Cargill Thompson, D. Gatteschi, J. A. McCleverty, J. A.
Navas, E. Rentschler, M. D. Ward, Inorg. Chem. 1996, 35, 2701–
2703.
[60] V. A. Ung, A. Thompson, D. A. Bardwell, D. Gatteschi, J. C. Jeff-
ery, J. A. McCleverty, F. Totti, M. D. Ward, Inorg. Chem. 1997, 36,
3447–3454.
[61] I. Fernꢆndez, R. Ruiz, J. Faus, M. Julve, F. Lloret, J. Cano, X. Ot-
tenwaelder, Y. Journaux, C. Munoz, Angew. Chem. 2001, 113,
3129–3132; Angew. Chem. Int. Ed. 2001, 40, 3039–3042.
[62] X. Ottenwaelder, J. Cano, Y. Journaux, E. Riviere, C. Brennan, M.
Nierlich, R. Ruiz-Garcia, Angew. Chem. 2004, 116, 868–870;
Angew. Chem. Int. Ed. 2004, 43, 850–852.
[63] K. B. Hansen, J. L. Leighton, E. N. Jacobsen, J. Am. Chem. Soc.
1996, 118, 10924–10925.
[64] D. J. Darensbourg, J. C. Yarbrough, J. Am. Chem. Soc. 2002, 124,
6335–6342.
[65] D. J. Darensbourg, R. M. Mackiewicz, A. L. Phelps, D. R. Billo-
deaux, Acc. Chem. Res. 2004, 37, 836–844.
[66] E. N. Jacobsen, Acc. Chem. Res. 2000, 33, 421–431.
[67] H. Friese, Chem. Ber. 1931, 64, 2109–2112.
[68] J. F. Larrow, E. N. Jacobsen, Y. Gao, Y. Hong, X. Nie, C. M. Zepp,
J. Org. Chem. 1994, 59, 1939–1942.
[69] SADBS, G. M. Sheldrick, University of Gçttingen, Gçttingen,
2003.
[70] G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[71] PLATON/SQUEEZE, A. L. Spek, Bijvoet Center for Biomolecu-
lar Research, Utrecht University, Utrecht.
[88] M. Sauer, C. Yeung, J. H. Chong, B. O. Patrick, M. J. MacLachlan,
J. Org. Chem. 2006, 71, 775–788.
[89] J. H. Chong, M. Sauer, B. O. Patrick, M. J. MacLachlan, Org. Lett.
2003, 5, 3823–3826.
[90] L. Cavallo, H. Jacobsen, Eur. J. Inorg. Chem. 2003, 892–902.
[91] K. J. Berry, P. E. Clark, K. S. Murray, C. L. Raston, A. H. White,
Inorg. Chem. 1983, 22, 3928–3934.
[92] B. P. Gaber, V. Miskowski, T. G. Spiro, J. Am. Chem. Soc. 1974, 96,
6868–6873.
[93] J. B. H. Strautmann, S. DeBeer George, E. Bothe, E. Bill, T. Wey-
hermꢂller, A. Stammler, H. Bçgge, T. Glaser, Inorg. Chem. 2008,
47, 6804–6824.
[94] C. Flassbeck, K. Wieghardt, Z. Anorg. Allg. Chem. 1992, 608, 60–
68.
[95] The program package JulX was used for spin-Hamiltonian simula-
tions and fittings of the data by a full-matrix diagonalization ap-
proach (E. Bill, unpublished results).
[96] K. P. Bryliakov, E. P. Talsi, Chem. Eur. J. 2007, 13, 8045–8050.
[97] S. Walleck, H. Theil, M. Heidemeier, H. Heinze-Brꢂchner, A.
Stammler, H. Bçgge, T. Glaser, unpublished results.
[98] H. Fujii, Y. Funahashi, Angew. Chem. 2002, 114, 3790–3793;
Angew. Chem. Int. Ed. 2002, 41, 3638–3641.
[99] M. Lubben, A. Meetsma, F. von Bolhuis, B. L. Feringa, R. Hage,
Inorg. Chim. Acta 1994, 215, 123–129.
[100] A. V. Wiznycia, J. Desper, C. J. Levy, Can. J. Chem. 2009, 87, 224–
231.
[101] B. J. Kennedy, G. Brain, E. Horn, K. S. Murray, M. R. Snow, Inorg.
Chem. 1985, 24, 1647–1653.
[102] B. J. Kennedy, K. S. Murray, M. R. Snow, Inorg. Chim. Acta 1987,
132, 153–155.
[103] H.-L. Shyu, H.-H. Wei, G.-H. Lee, Y. Wang, J. Chem. Soc. Dalton
Trans. 2000, 915–918.
[104] H. D. Flack, Acta Cryst. 1983, A39, 876–881.
[72] R. Hernꢆndez-Molina, A. Mederos in Comprehensive Coordination
Chemistry II, Vol. 1 (Eds.: J. A. McCleverty, T. J. Meyer), Elsevier,
Oxford, 2004, pp. 411–446.
Received: April 12, 2010
Published online: July 14, 2010
Chem. Eur. J. 2010, 16, 10137 – 10149
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10149