Synthesis and characterization of manganese triple-decker complexes

Article abstract of DOI:10.1039/c5dt01750e

The use of the highly sterically demanding Cp^BIG ligand (Cp^BIG = C₅(4-nBuC₆H₄)₅) and white phosphorus (P₄) enables the synthesis of new P-rich derivatives of the rare P_n ligand complexes of manganese. The obtained complexes, [{Cp^BIGMn}₂(μ,η^5:5-P₅)] (2) and [{Cp^BIGMn}₂(μ,η^2:2-P₂)₂] (3), exhibit the highest number of P atoms in this class of manganese compounds identified by X-ray structure analyses. The EPR spectrum of the 29 VE triple-decker complex 2 shows one unpaired electron coupling with two 5/2 spin Mn nuclei.

Full text of DOI:10.1039/c5dt01750e

Dalton
Transactions
View Article Online
View Journal
PAPER
Synthesis and characterization of manganese
triple-decker complexes†‡
Cite this: DOI: 10.1039/c5dt01750e
Sebastian Heinl, Gábor Balázs, Michael Bodensteiner and Manfred Scheer*
The use of the highly sterically demanding CpBIG ligand (CpBIG = C5(4-nBuC6H4)5) and white phosphorus
(P4) enables the synthesis of new P-rich derivatives of the rare Pn ligand complexes of manganese. The
obtained complexes, [{CpBIGMn}2(µ,η5:5-P5)] (2) and [{CpBIGMn}2(µ,η2:2-P2)2] (3), exhibit the highest
number of P atoms in this class of manganese compounds identified by X-ray structure analyses. The EPR
spectrum of the 29 VE triple-decker complex 2 shows one unpaired electron coupling with two 5/2 spin
Mn nuclei.
Received 8th May 2015,
Accepted 2nd June 2015
DOI: 10.1039/c5dt01750e
www.rsc.org/dalton
Introduction
Since Ginsberg and Lindsell1 first reported on the synthesis of
complexes with substituent-free phosphorus ligands, numer-
ous so-called Pn ligand complexes have been prepared and
described in the literature.2 In most cases, the tetrahedral
molecule P4 is the starting material of choice.3 Up to now, Pn
ligand complexes have been known for almost all transition
metals4 and a large number of main group elements.5
However, there are some unexplored spots within the d-block
elements, where no or only very few compounds with substitu-
ent-free Pn ligands are known. The 7th group might be seen as
a border between the early and the late transition metals,
where in both cases a certain number of substances with Pn
units are known. The absence of technetium derivatives can
most likely be explained by their artificial and radioactive
nature. However, while for rhenium several compounds are
known, for manganese only very few examples exist.4 To the
best of our knowledge, the only crystallographically character-
ized Pn ligand complex of manganese is [{CpMo(CO)2}2{Mn
(CO)4}(µ3-P)] (A) with a P1 unit, reported by Mays et al.
(Scheme 1).6 Huttner et al. also reported on Mn containing
dimetallaphosphacumulenes B containing P1 units.7 The only
Mn complex with a Pn moiety with n > 1 was described by
Baudler et al. in 1991 as [(CO)3Mn(η5-P5)] (C).8 Due to the
limited characterization of C, its existence is questionable (cf.
Scheme 1 Reported Pn ligand complexes of manganese.
ESI‡). There are also several other examples of Pn ligand com-
plexes with direct Mn–P bonds. However, they are based on the
coordination of the phosphorus lone pair towards 16 valence
electron (VE) {CpMn(CO)2} fragments.4
The lack of manganese complexes with substituent-free Pn
ligands prompted us to investigate this class of compounds.
Herein, we report on the synthesis of the precursor complex
[CpBIGMn(cht)] (1; CpBIG = C5(4-nBuC6H4)5; cht = cyclohepta-
triene) and its reaction with P4 forming the first triple-decker
complexes of manganese, [{CpBIGMn}2(µ,η5:5-P5)] (2) and
[{CpBIGMn}2(µ,η2:2-P2)2] (3). The compounds were character-
ized by NMR and EPR spectroscopy, mass spectrometry, single
crystal X-ray diffraction as well as by DFT calculations.
Results and discussion
The UV irradiation of the sterically crowded cymantrene
derivative [CpBIGMn(CO)3]9 in toluene with an excess of cyclo-
heptatriene (cht) results in the substitution of all three carbo-
nyl groups by one cht ligand (eqn (1)). After filtration over
Celite, [CpBIGMn(cht)] (1) is obtained pure as a brown powder
in 90% yield. Interestingly, if silica gel is used instead of
Celite, decomposition is observed. As expected, complex 1
exhibits a sandwich structure with a η6 bound cht ligand and
the methylene group pointing away from the {CpBIGMn} frag-
University of Regensburg, Universitätsstr. 31, 93040 Regensburg, Germany.
E-mail: manfred.scheer@chemie.uni-regensburg.de; Fax: +49 941 943 4439;
Tel: +49 941 943 4440
†The presented results are part of the doctoral thesis of Sebastian Heinl. Heinl,
S. PhD thesis, University of Regensburg (Regensburg, Germany), 2014.
‡Electronic supplementary information (ESI) available: Details of the X-ray struc-
ture analysis, full crystallographic data, and NMR spectra. CCDC 1048598 and
1048599. For ESI and crystallographic data in CIF or other electronic format see
DOI: 10.1039/c5dt01750e
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
ment (Fig. 1). This methylene group is disordered over two
positions (occupancies of 78% and 22%). The structural para-
meters compare rather well with the only other structurally
known cht-Mn complex [(C5H4Me)Mn(cht′)] (cht′ = 1-phenyl-
cycloheptatriene).10 In both cases no alternating bond lengths
in the triene moiety are observed, which were also found for
other η6-cht complexes.11 In 1 this can be explained by the dis-
order discussed above.
The precursor complex 1 reacts with P4 in toluene in a co-
thermolysis reaction under elimination of the cht ligand. Chro-
matographic workup affords a single brown fraction of a
mixture of [{CpBIGMn}2(µ,η5:5-P5)] (2) and [{CpBIGMn}2(µ,η2:2
-
P2)2] (3), which could not be separated from each other even by
thin-layer chromatography (eqn (1)). Single crystals suitable for
X-ray diffraction were obtained from a concentrated toluene
solution as the solvate with six toluene molecules per formula
unit. Both compounds co-crystallize at the same crystallo-
graphic position. Therefore, effectively a triple-decker complex
with a disordered middle-deck is observed. According to the
occupancies of the cyclo-P5 ligand (70%) and that of the two P2
ligands (30%), a 7 : 3 mixture of 2 and 3 is obtained (Fig. 2).
The P1 and P2 atoms belong to both complexes. This compo-
sition is also confirmed by the elemental analysis. The P–P
bond lengths in 2 are with an average value of 2.20 Å, in good
agreement with other known triple-decker complexes contain-
ing a cyclo-P5 middle-deck (2.15 Å–2.28 Å).12
Fig. 2 Molecular structures of (a) 2 and (b) 3 in the crystal. For clarity C
atoms are shown in the wireframe model and H atoms and solvent
molecules are omitted. Ellipsoids are drawn at the 50% probability level.
(c) Schematic illustration of the disorder of 2 and 3 in the crystal (left:
side view; right: top view). Different parts are colored differently; hashed
globes belong to both parts. Selected bond lengths [Å] and angles [°] of
2: Mn1–Mn1’ 2.7953(9), P1–P2 2.150(2), P2–P3 2.291(2), P3–P4 2.157(2),
P4–P5 2.262(2), P1–P5 2.114(2), Mn1–P1 2.3295(7), Mn1–P2 2.3276(8),
Mn1–P3 2.3283(10), Mn1–P4 2.3407(12), Mn1–P5 2.3369, angle
between P5 and Cp 4.40(6), angle between CpBIG ligands 8.80(11).
Selected bond lengths [Å] and angles [°] of 3: Mn1–Mn1’ 2.7953(9), P1–
P2 2.150(2), P6–P7 2.123(8), P1⋯P7 3.243(4), P2⋯P6 2.843(4), Mn1–P1
2.3295(7), Mn1–P2 2.3276(8), Mn1–P6 2.319(2), Mn1–P7 2.342(3), angle
between (P2)2 and Cp 4.40(6), angle between CpBIG ligands 8.80(11),
angle between P2 units 10.75(15).
bond lengths P1–P2 and P6–P7 are in between a single and a
double bond with 2.150(2) Å and 2.123(8) Å and roughly 0.1 Å
longer than in the previously described cobalt and iron
examples. The CpBIG ligands in 2 and 3 are not co-planar but
tilted against each other by 8.80(11)° and by 4.40(6)° against
the phosphorus middle-decks. In complex 2, the CpBIG moi-
eties and the P5 ring are almost eclipsed (∼10°), which results
in a staggered conformation of the P5 ring and the downward
orientated Ph groups. This is also the case for the pentapho-
sphaferrocene [CpBIGFe(η5-P5)].15 A further interesting feature
of 2 and 3 is the Mn1–Mn1′ distance of 2.7953(9) Å, which is
in a usual range for Mn–Mn single bonds (CCDC search:
2.5–3.1 Å).
Binuclear compounds with two bridging P2 units and Cp
ligands are rare, only [(CpRCo)2(µ,η2:2-P2)2] (CpR = Cp′′′,13
Cp*14) and [(Cp*Fe)2(µ,η2:2-P2)2]14 are known. In complex 3 the
In the 31P{1H} NMR spectrum in C6D6 of dissolved crystals,
which represent a mixture of 2 and 3, only one broadened
singlet is observed at δ = 396.9 ppm (ω1/2 = 467 Hz). Therefore,
the signal of one of the two compounds is not observed, or 2
and 3 have incidentally the same chemical shifts. Since 2 is
calculated to have a dublet spin state, it is very likely that the
observed signal corresponds only to 3. By cooling a solution in
CD2Cl2, in the NMR spectrum, the signal steadily sharpens,
which is accompanied by an intensity decrease. At 213 K the
signal is not observed anymore. Heating the solution in
toluene-d8 to temperatures up to 373 K results in further
broadening of the signal. The 1H NMR spectrum shows several
superimposed signals, which cannot further be assigned to a
Fig. 1 Molecular structure of 1 in the crystal. Thermal ellipsoids are
drawn at the 50% probability level. Because of disorder, only the main
part is depicted. H atoms on the CpBIG ligand are omitted for clarity and
it is drawn in the ‘wire-or-stick’ model. Selected bond lengths [Å] and
angles [°]: Mn1⋯C56/C56’ 2.749(3)/2.789(11), Mn1–C57/C57’ 2.173(3)/
2.220(12), Mn1–C58 2.109(2), Mn1–C59 2.110(2), Mn1–C60 2.114(2),
Mn1–C61 2.084(2), Mn1–C62 2.143(2), C56/C56’-chtPlane 115.7(3)/113.8
(8).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
particular CpBIG ligand of the products. Cooling the sample to
193 K does not result in the splitting of the signal (see the
ESI‡ for spectra). The NMR behavior may be explained by the
odd number of valence electrons in complex 2, making the
sample paramagnetic.
EPR measurements at room temperature in toluene do not
show an EPR signal, as would be expected for the 29 VE triple-
decker complex 2. However, the spectrum at 77 K shows a well
resolved anisotropic multiplet (Fig. 3). The splitting can be
explained by coupling of the electron with two Mn centers
with a spin of 5/2 each. The giso value was identified to be
1.9228 (simulation). This is slightly smaller than the value of
1.9571 we obtained by DFT calculations.16 The absence of an
EPR signal at room temperature can be explained with a rapid
relaxation process.
The LIFDI mass spectrum of a solution of the crystals exhi-
bits a base peak at m/z = 1716.4 corresponding to 2. The peak
for 3 (m/z = 1685.6) can be detected with a relative intensity
ca. 25%, which is in good agreement with the product distri-
bution obtained by the X-ray analysis.
Fig. 4 Selected α-spin molecular orbitals of [{(C5Ph5)Mn}2(μ,η5:5-P5)]
(2’), calculated at the BP86/def2-SVP level of theory.
the two manganese atoms. Since the in-phase combination is
doubly occupied, a Mn–Mn bond order of 0.5 results. This is
also confirmed by the calculated Wiberg bond index of 0.6.
The localized molecular orbital representing the Mn–Mn bond
is depicted in Fig. 5.
The geometry of the model compound [{(C5Ph5)Mn}2(μ,η2:2
-
P2)2] (3′) has been optimized using the same functional and
basis set as for 2′. The ground state of 3′ is calculated to be a
singlet. This explains the absence of a signal of 3 in the EPR
spectrum. The optimized geometry shows a short Mn–Mn dis-
tance of 2.597 Å which is shorter than the corresponding dis-
tance in 2′ (Mn–Mn 2.878 Å) as well as the experimentally
determined distance (2.7953(9) Å). This is probably a result of
the overestimation of the Mn–Mn interaction in 3′. The pres-
ence of a Mn–Mn bond is clearly indicated by the localized
molecular orbitals (Fig. 6). Furthermore, the Wiberg bond
index for the Mn–Mn bond is high (1.18) and indicates a very
strong bond. Similarly high bond orders were obtained for the
Mn–P (0.97) and P–P (0.84 and 0.86) bonds. The Natural Popu-
lation Analysis shows a moderate charge separation. The two
Mn atoms carry a partial positive charge of 0.45 each, while
the negative partial charge is delocalized over the Cp ligands.
In summary, we reported on the synthesis of two novel
triple-decker complexes [{CpBIGMn}2(µ,η5:5-P5)] (2) and
[{CpBIGMn}2(µ,η2:2-P2)2] (3), the so far P-richest manganese
compounds characterized by X-ray crystallography. They rep-
resent two new derivatives of the rare class of Pn ligand com-
plexes of manganese. The compounds were characterized by
To gain more insight into the electronic structure of 2, DFT
calculations have been performed on the model compound
[{(C5Ph5)Mn}2(μ,η5:5-P5)] (2′).16 For comparison with other
triple-decker complexes with inorganic middle-decks see ref.
17. The geometric parameters of the optimized structure of 2′
in the doublet spin state are in agreement with the experi-
mental values determined by X-ray diffraction. The calculated
Mn–Mn distance in 2′ is 2.878 Å, which is slightly longer than
the experimentally determined value of 2.7953(9) Å for 2. Inter-
estingly, the optimized geometry of 2′ in the excited, quartet
spin state differs considerably from the experimental geome-
try, and is 67.31 kJ mol−1 higher in energy. In the quartet spin
state the P5 middle deck adopts an envelope conformation
with a μ,η4:3 coordination mode. The Mn–Mn distance is
3.164 Å, considerably longer than the corresponding distance
in the doublet spin state. All these results suggest that the
ground spin state of 2′ is a doublet, which could be definitely
proven by the simulation of the EPR spectrum. According to
the calculations, the unpaired electron is equally localized on
both manganese atoms. The single occupied molecular orbital
(SOMO) is the highest occupied molecular orbital (Fig. 4). It
represents an out-of-phase combination of the dz2 orbitals of
Fig. 5 Localized molecular orbital representing the Mn–Mn bond in
[{(C5Ph5)Mn}2(μ,η5:5-P5)] (2’). Calculated at the BP86/def2-SVP level of
theory.
Fig. 3 EPR spectrum of a mixture of 2 and 3 in toluene at 77 K.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
also the ESI.‡ 13C{1H} NMR (CD2Cl2): δ [ppm] = 14.1 (nBu),
22.2 (CH2-cht), 22.7 (nBu), 33.5 (nBu), 35.6 (nBu), 91.0 (CH-
cht), 91.4 (CH-cht), 92.5 (CH-cht), 132.7 (Ph), 134.1 (Ph), 140.7
(Ph); one Ph signal is obscured by the C6D6 signal.
Synthesis of [{CpBIGMn}2(µ,η5:5-P5)] (2) and [{CpBIGMn}2-
(µ,η2:2-P2)2] (3)
A solution of [CpBIGMn(cht)] (400 mg, 0.46 mmol) and P4
(284 mg, 2.3 mmol) in 100 mL toluene is refluxed for 16 h.
The brown reaction mixture is dried in vacuum. Column chro-
matographic work-up (25 × 3 cm, hexane/toluene 2 : 1, silica)
of the residue affords a brown fraction of a mixture of 2 and 3.
A co-crystallized product mixture is obtained from a concen-
trated toluene solution (310 mg, yield 79%).
Fig. 6 Selected localized molecular orbitals representing the Mn–Mn,
Mn–P and P–P bonds in [{(C5Ph5)Mn}2(μ,η2:2-P2)2] (3’). Calculated at the
BP86/def2-SVP level of theory.
single crystal X-ray diffraction, mass spectrometry, VT NMR
and EPR spectroscopy and elemental analysis. DFT calcu-
lations confirmed a doublet ground state for 2 with a Mn–Mn
bond through the cyclo-P5 ligand with a calculated Wiberg
bond index of 0.6.
The mixture of 2 and 3: [70% 2/30% 3] calc.: C, 77.37; H,
7.67. Found: C, 77.76; H, 7.64. LIFDI-MS (toluene): m/z (%) =
1716.4 (100%, [2]+), 1686.6 (25%, [3]+), 726.3 (60%, [CpBIG]+).
1H NMR (C6D6): δ [ppm] = 0.73 (s-br, 30H, nBu), 1.09 (s-br,
n
n
n
20H, Bu), 1.29 (s-br, 10H, Bu), 1.41 (s, 10H, Bu), 2.18 (s-br,
n
10H, Bu), 6.2 (very broad, ω1/2 = 236 Hz, 5H), 6.67 (s-br, 10H,
Ph), 7.31 (s-br, 10H). 31P{1H} NMR (C6D6): δ [ppm] = 396.9 (s-
br, ω1/2 = 467 Hz). 13C{1H} NMR (CD2Cl2): δ [ppm] = 14.1
(CH3), 22.8 (CH2), 33.4 (CH2), 35.8 (CH2), 118.5 (Cp), 127.9
(Ph), 131.3 (Ph), 140.5 (Ph), 141.6 (Ph).
Experimental details
General remarks
All experiments were carried out under an atmosphere of dry
argon or nitrogen using glovebox and Schlenk techniques. The
solvents toluene, hexane and C6D6 were degassed and refluxed
over potassium prior to use. The same procedure was applied
for CD2Cl2 by using CaH2 as a drying reagent. Cht was used as
obtained from ABCR GmbH and [CpBIGMn(CO)3] was prepared
according to the literature procedure.9 The NMR spectra were
measured on Bruker Avance 300, 400, or 600 spectrometer.
LIFDI-MS spectra were measured on Finnigan MAT 95 (1) or a
JOEL JMS-700 (2, 3) mass spectrometer. The elemental ana-
lyses were determined on Vario EL III apparatus. The IR
spectra were measured on a VARIAN FTS-800 FT-IR spectro-
meter. For irradiation a TQ150-Z1 Hg lamp from Heraeus was
used. The X-band EPR measurements were carried out with a
MiniScope MS400 device with a frequency of 9.5 GHz and a
rectangular resonator TE102 from Magnettech GmbH.
Acknowledgements
This work was supported by the Deutsche Forschungs-
gemeinschaft. S. Heinl is grateful to the Fonds der Che-
mischen Industrie for a PhD fellowship. Prof. Dr Hans-Jörg
Himmel and Hendrik Herrmann (University of Heidelberg,
Germany) are gratefully thanked for recording the LIFDI mass
spectrum of 2/3.
Notes and references
1 A. P. Ginsberg and W. E. Lindsell, J. Am. Chem. Soc., 1971,
93, 2082–2084.
2 (a) O. J. Scherer, Angew. Chem., Int. Ed. Engl., 1990, 102,
1137–1155; (b) V. M. Di, P. Stoppioni and M. Peruzzini,
Polyhedron, 1987, 6, 351–382.
Synthesis of [CpBIGMn(cht)] (1)
A solution of [CpBIGMn(CO)3] (1.0 g, 1.16 mmol) in 50 mL
toluene is irradiated together with cht (0.32 g, 0.36 mL,
3.46 mmol) for 30 min. The brown reaction mixture is filtered
through Celite and carefully dried in a vacuum. Compound 1
is obtained as a brown powder (914 mg, 90%).
3 B. M. Cossairt, N. A. Piro and C. C. Cummins, Chem. Rev.,
2010, 110, 4164–4177.
4 (a) M. Caporali, L. Gonsalvi, A. Rossin and M. Peruzzini,
Chem. Rev., 2010, 110, 4178–4235. For a recent example cf.:
(b) S. Heinl, E. Peresypkina, A. Y. Timoshkin, P. Mastrorilli,
V. Gallo and M. Scheer, Angew. Chem., Int. Ed., 2013, 52,
10887–10891.
5 (a) M. Scheer, G. Balazs and A. Seitz, Chem. Rev., 2010, 110,
4236–4256; (b) N. A. Giffin and J. D. Masuda, Coord. Chem.
Rev., 2011, 255, 1342–1359.
1: [C62H73Mn] calc.: C, 85.28; H, 8.43. Found: C, 84.90; H,
8.37. LIFDI-MS (toluene): m/z (%) = 872.7 (100%, [M]+), 818.9
(30%, [CpBIGH + cht]+, Diels–Alder product). IR (CH2Cl2,
1
cm−1): νCO 1917 (s), 1858 (s). H NMR (C6D6): δ [ppm] = 0.10
3
3
(d, JHH = 12.1 Hz, 1H, CH2-cht), 0.82 (t, JHH = 6.8 Hz, 30H,
nBu), 1.22 (m, 20H, nBu), 1.47 (m, 20H, Bu), 2.42 (t, 3JHH = 7.3
n
3
Hz, 20H, nBu), 2.79 (q, JHH = 10.0 Hz, 1H, CH2-cht), 3.06 (s,
2H, CH-cht), 5.03 (s, 2H, CH-cht), 5.97 (s, 2H, CH-cht), 6.88 (d,
6 A. J. Bridgeman, M. J. Mays and A. D. Woods, Organometal-
lics, 2001, 20, 2932–2935.
3
3JHH = 7.0 Hz, 20H, Ph), 7.24 (d, JHH = 7.0 Hz, 20H, Ph); see
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
7 A. Strube, J. Heuser, G. Huttner and H. Lang, J. Organomet.
Chem., 1988, 356, C9–C11.
8 M. Baudler and T. Etzbach, Angew. Chem., Int. Ed. Engl.,
1991, 30, 580–582.
9 S. Heinl, E. V. Peresypkina, A. Y. Timoshkin, P. Mastrorilli,
V. Gallo and M. Scheer, Angew. Chem., Int. Ed., 2013, 52,
10887–10891.
10 J. A. D. Jeffreys and J. MacFie, J. Chem. Soc., Dalton Trans.,
1978, 144–147.
364; (b) A. R. Kudinov, D. A. Loginov, Z. A. Starikova,
P. V. Petrovskii, M. Corsini and P. Zanello, Eur. J. Inorg.
Chem., 2002, 2002, 3018–3027; (c) O. J. Scherer, T. Brück
and G. Wolmershäuser, Chem. Ber., 1989, 122, 2049–2054;
(d) L. Y. Goh, R. C. S. Wong, C. K. Chu and T. W. Hambley,
J. Chem. Soc., Dalton Trans., 1990, 977–982.
13 (a) O. J. Scherer, G. Berg and G. Wolmershäuser, Chem.
Ber., 1995, 128, 635–639; (b) F. Dielmann, M. Sierka,
A. V. Virovets and M. Scheer, Angew. Chem., Int. Ed., 2010,
49, 6860–6864.
11 (a) M. L. H. Green and N. M. Walker, J. Chem. Soc., Chem.
Commun., 1989, 850–852; (b) C. Jiabi, Y. Jienguo, L. Guixin, 14 M. E. Barr and L. F. Dahl, Organometallics, 1991, 10, 3991–
X. Weihua, S. Meicheng, Z. Zeying and T. Youqi, J. Organo- 3996.
met. Chem., 1980, 329, 69–83; (c) B. Klingert and G. Rihs, 15 S. Heinl, G. Balázs and M. Scheer, Phosphorus, Sulfur
J. Chem. Soc., Dalton Trans., 1991, 2749–2760; (d) T. Meyer- Silicon Relat. Elem., 2014, 189, 924–932.
Friedrichsen, H. Wong, M. H. Prosenc and J. Heck, 16 For details on DFT calculations and EPR simulation see the
Eur. J. Inorg. Chem., 2003, 2003, 936–946. ESI.‡
12 (a) O. J. Scherer, J. Schwalb, G. Wolmershäuser, W. Kaim 17 W. Tremel, R. Hoffmann and M. Kertesz, J. Am. Chem. Soc.,
and R. Gross, Angew. Chem., Int. Ed. Engl., 1986, 25, 363–
1989, 111, 2030–2039.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.