Organometallic cluster complexes with face-capping arene ligands, 91. Hydrogenation of the side-chain in [{(C5H5)Co}3(μ3-alkenylbenzene)]: Synthesis and structure of the cluster complexes [{(C5H5)Co}3(μ3-η 2:η2:η2-C6H5-CHR1-CH2R2)]2

Article abstract of DOI:10.1016/S0022-328X(98)00594-4

Catalytic hydrogenation (1 bar H₂, 20°C, Pd/C catalyst) of the C=C double bond in the side-chain of the facial alkenylbenzene ligands of the cluster complexes [{(C₅H₅)Co}₃{μ₃-C ₆H₅(CR¹)(CHR²)}] 2b-c gave the derivatives [{(C₅H₅)Co}₃{μ₃-C ₆H₅(CHR¹)(CH₂R²)}] 3b-d in high yield (b, R¹=CH₃, R²=H; c, R¹=H, R²=Ph; d, R¹=Ph, R²=H). The X-ray crystal and molecular structures of 3c and 3d were determined. In both derivatives a facial μ₃-η²:η²:η² coordination of one phenyl ring to the tricobalt cluster was found. Only one of two possible diastereomeric conformers is present in the crystals of 3d. The expanded μ₃-phenyl rings show a small but statistically significant Kekule-type (trigonal) distortion. In solution hindered rotation of the μ₃-arenes on top of the [(C₅H₅)Co]₃ clusters is observed by NMR spectroscopy.

Full text of DOI:10.1016/S0022-328X(98)00594-4

Journal of Organometallic Chemistry 573 (1999) 22–29
Organometallic cluster complexes with face-capping arene ligands, 91.
Hydrogenation of the side-chain in [{(C5H5)Co}3(v3-alkenylbenzene)]:
synthesis and structure of the cluster complexes
[{(C5H5)Co}3(v3-p2:p2:p2-C6H5–CHR1–CH2R2)]2
Hubert Wadepohl *, Klaus Bu¨chner, Michael Herrmann, Hans Pritzkow
Anorganisch-chemisches Institut der Uni6ersita¨t, Im Neuenheimer Feld 270, D-69120 Heidelberg, Germany
Received 24 January 1998
Abstract
Catalytic hydrogenation (1 bar H2, 20°C, Pd/C catalyst) of the CꢀC double bond in the side-chain of the facial alkenylbenzene
ligands of the cluster complexes [{(C5H5)Co}3{v3-C6H5(CR1)(CHR2)}] 2b–c gave the derivatives [{(C5H5)Co}3{v3-
C6H5(CHR1)(CH2R2)}] 3b–d in high yield (b, R1=CH3, R2=H; c, R1=H, R2=Ph; d, R1=Ph, R2=H). The X-ray crystal and
molecular structures of 3c and 3d were determined. In both derivatives a facial v3-p2:p2:p2 coordination of one phenyl ring to the
tricobalt cluster was found. Only one of two possible diastereomeric conformers is present in the crystals of 3d. The expanded
v3-phenyl rings show a small but statistically significant Kekule´-type (trigonal) distortion. In solution hindered rotation of the
v3-arenes on top of the [(C5H5)Co]3 clusters is observed by NMR spectroscopy. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: Arene Clusters; Cyclopentadienyl; Cobalt
1. Introduction
zene [4]. Not even traces of arene cluster complexes
were formed under the same reaction conditions with
benzene or alkylbenzenes as the aromatic substrates.
Obviously, the unsaturated side chain plays a crucial
role during the assemblage of the trimetal cluster [5].
Indeed, a considerable number of cluster complexes 2
have been prepared in like manner from 1 and a variety
of 1-alkenylbenzenes [6,7].
A number of features of the solid state structures,
dynamic behaviour and reactivity of 2 appear to be
influenced by the unsaturated substituent. Crystallo-
graphic studies on a number of derivatives 2 showed a
small Kekule´-type distortion of the v3-arene rings [2,8].
Electronic conjugation between the ring and side-chain
may have a non-negligible effect on the alternating
longer and shorter endocyclic C–C bonds. The steric
bulk of the substituent, which is of course a function of
the hybridisation (connectivity) of the h-C, is expected
to influence the dynamic behaviour of the arene cluster.
The organometallic cluster complexes [{(C5H5)
M}3(v3-p2:p2:p2-arene)] (M=Co, Rh) are kinetically
very stable molecules [2]. It has recently been shown
that the rhodium derivatives can be synthesised by
stepwise addition of (C5H5)Rh fragments (photogener-
ated from [(C5H5)Rh(C2H4)2]) directly to the six-mem-
bered ring of the arene (typically benzene or
hexamethylbenzene) [3]. Such a straightforward syn-
thetic route is impracticable for the tricobalt analogues.
Some time ago, we reported the facile formation of
[{(C5H5)Co}3(v3-p2:p2:p2-i-methylstyrene)] 2a from
[(C5H5)Co(C2H4)2] 1 and i-methylstyrene or allylben-
* Corresponding author. Fax.: +49 6221 544197; e-mail:
bu9@ix.urz.uni-heidelberg.de
1 For part 8 see ref. [1].
2 Dedicated to Professor Brian Johnson on the occasion of his 60th
birthday.
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII S0022-328X(98)00594-4

H. Wadepohl et al. / Journal of Organometallic Chemistry 573 (1999) 22–29
23
Table 1
1H-NMR spectroscopic data (l in C6D6) for complexes 3b–d
Complex
C5H5
v3-Phenyl
Side chain
3ba
3cb
3da
4.72 (s, 15H)
4.68 (s, 15H)
4.72 (s, 15H)
4.20 (m, 2H), 4.40 (m, 3H)
4.23 (m, 5H)
4.21 (m, 3H, m-H, p-H)
4.53 (d, 1H, o-H),
4.61 (d, 1H, o-H)
1.09 (d, 6H, CH3), 2.15 (sept, 1H, CH)
2.24 (t, 2H, CH2), 2.98 (t, 2H, CH2), 7.1–7.3 (m, Ph)c
1.25 (d, 3H, CH3), 3.46 (q, 1H, H–Ch),
7.1–7.2 (m, p-H)c, 7.30 (t, 2H, m-H),
7.47 (d, 2H, o-H)
a 333 K. b 313 K. c Overlap with solvent resonances.
Finally, an olefinic side-chain offers additional func-
tionality, which may mask the reactivity of the facial
arene ligand itself [1]. These problems prompted us to
look into the synthesis of v3-arene tricobalt derivatives
without a CꢀC double bond in the h-position to the
face-capping arene ring.
In this paper we report on the catalytic hydrogena-
tion of the unsaturated side-chain in some cluster com-
plexes 2 to give the target molecules 3 which are not
accessible via the direct route utilising (C5H5)Co frag-
ments and the arene ligand. Catalytic hydrogenation of
[{(C5H5)Co}3(v3-p2:p2:p2-i-methylstyrene)] 2a to give
the n-propyl derivative 3a has already been briefly
described in an earlier paper ([6]a). However, since due
to the poor quality of the crystalline material an X-ray
crystal structure analysis could not be carried out, no
details of the structure of this complex are known.
2.2. Spectroscopic in6estigations
The proton and carbon resonances of the v3-phenyl
rings in 3b–d are shifted to high field with respect to
those of the uncoordinated phenyl groups (Tables 1
and 2). Hence, two sets of four C resonances each (3
CH, 1 C) are observed for the 1,2-diphenylethane
(dibenzyl) ligand in 3c. The isopropylbenzene ligand
in 3b also gives four signals for the facial phenyl ring.
In contrast, six 13C resonances can be assigned to the
cluster coordinated phenyl ring in the 1,1-
diphenylethane derivative 3d, in addition to the four
signals at lower field, which are due to the uncoordi-
nated phenyl ring of this ligand. The methyl hydro-
gens of the isopropyl substituent in 3b appear as a
broad resonance at room temperature (r.t.), accompa-
nied by a sharp septet due to the methyne proton.
On cooling to 235 K, the broad feature splits into
two sharp doublets, without any noticeable change of
the septet resonance. At 333 K, a sharp doublet/sep-
tet pattern is observed. At ambient temperature, the
1H-NMR spectrum of 3c shows two separate pseudo-
triplets (Dl:0.7) for the two methylene groups of
the 1,2-diphenylethane ligand.
2. Results
2.1. Syntheses
Toluene solutions of the complexes 2b (R1=CH3,
R2=H), 2c (R1=H, R2=Ph) and 2d (R1=Ph, R2=H)
were treated with dihydrogen (1 bar) in the presence of
a Pd/charcoal catalyst. After reaction times between
several hours (2b, 2d) and more than a week (2c) the
educts had been quantitatively converted into the com-
plexes 3b (R1=CH3, R2=H), 3c (R1=H, R2=Ph) and
3d (R1=Ph, R2=H), respectively. The products were
isolated as dark brown crystalline solids in 50–95% yield.
The three cyclopentadienyl ligands in 3b–d give rise
1
to only one slightly broadened singlet both in the H-
(l:4.7) and 13C-NMR spectra (l:82) at r.t. These
resonances sharpen when the samples are warmed to
330 K.
2.3. Crystal structure analyses
Single crystal X-ray structure determinations
were carried out with 3c and 3d. Both com-
plexes crystallise monoclinic in non-centrosymmetric
space groups (3c P21, 3d Cc). Important bond
lengths and angles are given in Table 3. Views of
the molecules are presented in Figs. 1 and 2. Only
one of the two possible diastereomers is present
as a racemic mixture in the crystals of 3d (Fig. 1).
Complex 3c crystallised enantiomerically pure
(Fig. 2). The crystal packing of 3c and 3d has been
examined in detail; the results are described elsewhere
[9].

24
H. Wadepohl et al. / Journal of Organometallic Chemistry 573 (1999) 22–29
Table 2
13C-NMR spectroscopic dataa (l in C6D6) for complexes 3b–d
Complex
C5H5
v3-Phenyl
Ch
Ci
Phenyl
CH
CH
Cipso
Cipso
3bb
3cd
3df
82.5
82.5
82.6
39.8c, 40.8c, 41.9c
63.5
60.7
65.8
38.6c
39.7e
39.0c
24.0
45.2e
21.1
39.3, 40.5, 44.6
126.1, 128.6, 129.0
126.2, 127.4, 128.6
142.8
149.2
39.5c, 41.0c, 41.2c, 45.0, 49.3
a Assignment was supported by determination of signal multiplicities using the DEPT and J-modulated spin–echo techniques. b 333 K. c v3-Phenyl
CH or ChH. d 313 K. e Ch or Ci. f 298 K.
In both complexes, the diphenylethane ligand is
bound to a [(C5H5)Co]3 cluster in the v3-p2:p2:p2 coor-
dination mode via one of the phenyl rings. The v3-
phenyl rings are essentially planar (root mean square
deviation of the C atoms from the best plane: 0.015 (3c)
and 0.004 (3d)) and are expanded considerably com-
pared to the uncoordinated phenyl groups. A staggered
conformation is adopted by the ca. coplanar Co3 and
C6 rings. There is a slight alternation of the endocyclic
C–C bond lengths within the v3-phenyl rings, which is
however not significant in all the cases (the differences
in length, Dd(CC), of adjacent bonds are one to nine
times larger than the esds of the individual bond dis-
tances). In all cases, the bonds ‘on top’ of a metal atom
are nominally shorter than the two adjacent bonds,
which are ‘in between’ the metals.
(3d), corresponding to bond angles of 24 and 20°,
respectively), but is also noticeable with the H atoms,
˚
which are less accurately localised (0.2–0.4 A off the
C6-plane).
3. Discussion
3.1. Syntheses and structures
In previous work, we noted that 2a was rather inert
to a number of conventional addition reagents aimed at
the olefinic functionality [5]. In contrast to this be-
haviour, hydrogenation of the CꢀC double bonds in the
side chains of 2 occurs under conditions which are
similar to those needed for the free ligands. The ob-
served considerable difference of the rates of this reac-
The substituents on the v3-phenyl rings are displaced
from the C6-plane in the direction away from the
Co3-cluster. This distortion is most obvious with the
h-C atoms of the side chains (0.623(7) (3c) and 0.544(7)
Table 3
˚
Selected bond lengths (A) and angles (°) for complexes 3c and 3d
3c
3d
Co(1)–Co(2)
Co(1)–Co(3)
Co(2)–Co(3)
Co(1)–C(1)
Co(1)–C(2)
Co(2)–C(3)
Co(2)–C(4)
Co(3)–C(5)
Co(3)–C(6)
C(1)–C(2)
2.5319(14)
2.4905(14)
2.4970(14)
2.041(5)
2.027(5)
2.038(6)
2.018(6)
2.037(5)
2.028(6)
1.406(7)
1.449(8)
1.438(8)
1.412(8)
1.446(9)
1.427(8)
1.511(8)
1.547(8)
2.5014(11)
2.4895(12)
2.5218(11)
2.069(4)
2.011(5)
2.021(5)
2.022(5)
2.023(5)
2.021(5)
1.416(7)
1.443(6)
1.437(7)
1.412(7)
1.434(8)
1.405(7)
1.524(6)
1.542(7)
1.526(6)
C(1)–C(6)
C(2)–C(3)
C(3)–C(4)
C(4)–C(5)
C(5)–C(6)
C(1)–C(13)
C(13)–C(14)
C(13)–C(7)
C(14)–C(7)
1.507(8)
Fig. 1. Molecular structure of 3c. Displacement ellipsoids are at the
30% level.

H. Wadepohl et al. / Journal of Organometallic Chemistry 573 (1999) 22–29
25
sponding to a v3-p2:p2:p2 coordination, is typical for
arene ligands capping a triangular face of a metal
cluster [7,11]. Minor deviations from the ideal staggered
geometry, which occur in some derivatives with bulky
substituents, can be explained with steric repulsion
between the substituents on the v3-arene and the other
ligands on the metal cluster. For example, in the
complex [(v3-H){(C5H5)Co}3(v3-p2:p2:p2-1,1-diphenyle-
thane)]+ [3d+H]+, the protonation product of 3d, the
v3-arene is rotated by a few degrees within the plane
parallel to the Co3-triangle, away from the ideal stag-
gered conformation [1]. From the behaviour in solu-
tion, where full rotation of the v3-arenes is generally
observed [2,8], it is obvious that the intrinsic rotational
barrier is relatively small. In 3d, steric repulsions be-
tween the substituents on the h-carbon atom C13 and
the cyclopentadienyl rings are minimised in the confor-
mation which is actually found in the crystal (Fig. 2).
Here, the methyl group (C14)H3 nicely fits in between
the cyclopentadienyl ligands on cobalt atoms Co1 and
Co3, even with a nearly perfect staggered arrangement
of the Co3 and C6 rings. However, in the complex
[3d+H]+ steric strain is imposed on the [(C5H5)Co]3
(v3-phenyl) unit by the face capping hydrido ligand,
which pushes the C5H5 groups apart, into the direction
of the v3-phenyl, and causes the observed deviation
from the ideal geometry. With the h-carbon atom C13
being secondary rather than tertiary, steric congestion
in 3c is even less severe than in 3d.
A collection of data relating to the endocyclic C–C
bond lengths in the two different types of phenyl
groups in 3c and 3d is compiled in Table 4. Relevant
data for 2d and [3d+H]+ are also included for com-
parison. An expansion of the metal coordinated arene
rings compared to the free phenyl groups is obvious.
This lengthening on coordination of the C–C bonds is
fairly independent of the particular complex in Table 4
and even of the type of metal cluster (similar effects are
found in the second-row metal (Rh) cyclopentadienyl
and second- (Ru) or third-row metal (Os) carbonyl
cluster complexes [2,11]).
Fig. 2. Molecular structure of 3d. Displacement ellipsoids are at the
30% level.
tion for the two isomeric substrates 2c and 2d (and also
between 2b and 2c) also parallels the behaviour of the
free ligands. The commonly observed less facile hydro-
genation of internal compared to terminal alkenes
merely is an example of deactivation of the C–C y-
bond with increasing degree of substitution [10]. This
effect should be even more pronounced with the very
bulky [(C5H5)Co]3(v3-phenyl) substituent.
Interestingly, hydrogenation of the side chains in
some derivatives of 2 was observed on several other
occasions where it had not initially been expected. The
side-chain hydrogenated alkylbenzene complexes 3a
and 3d were the only isolatable products of the reac-
tions of the alkenylbenzene derivatives 2a and 2d, re-
spectively, with dimethylphenylsilane in the presence of
AIBN or H2PtCl6 [5]. Apparently, hydrogenation is
preferred to hydrosilation in these systems. Likewise, 3d
was observed along with free 1,1-diphenylethane and 2d
in a 1:1:1 ratio when ca. equimolar mixtures of 2d and
CF3COOH were allowed to stand for 24 h [1]. The
cluster stabilised carbonium ion [{(C5H5)Co}3(v3-
C6H5CMe2)]+, obtained by protonation of the
methylene group in 2b, also decomposed within weeks
at −20°C to give a 3:1 mixture of 2b and 3b [1]. As an
explanation for the latter two observations, reaction
with hydrogen ‘in statu nascendi’, generated from oxi-
dation [1] of some cobalt containing species by the acid,
can be envisaged.
The presence and significance of a true Kekule´-type
(trigonal) distortion of the v3-C6 rings is much more
difficult to assess. In our theoretical study of the hypo-
thetical [{(C5H5)Co}3(v3-C6H6)] we have pointed out
that such a distortion is the consequence of a reduction
of symmetry of the C6-ligand (from 6-fold in free
benzene to 3-fold in the v3-coordination site), which
triggers a mixing of benzene HOMO and LUMO, in
phase with the metal orbitals [12]. As a consequence, a
stronger interaction of the arene with the metal cluster
is expected to induce a stronger distortion of the C6-lig-
and. This explains the general trend that arenes bonded
to metal clusters composed of second- or third-row
transition metals tend to show a more pronounced
alternation of endocyclic bond lengths than that which
The [(C5H5)Co]3(v3-phenyl) cores of the cluster com-
plexes 3c and 3d have very similar geometries. The
staggered arrangement of the Co3 and C6 rings, corre-

26
H. Wadepohl et al. / Journal of Organometallic Chemistry 573 (1999) 22–29
ent in length by as much as seven times the esd. In
principle, in a benzene derivative any substituent is
expected to have some influence on the geometry of the
benzene ring to which it is attached. This problem has
been addressed by systematic studies [13]. The effects of
alkyl substituents on endocyclic bond lengths were
found to be considerably less pronounced and more
prone to errors than those on bond angles. In any case
they should be comparable for both types of phenyl
rings (v3 cluster coordinated and free), and vary in a
systematic way around the C6 ring [13].
In the structures of 3c and 3d there is no obvious
regularity in the observed pattern of bond lengths
within the free phenyl groups. This contrasts the more
regular pattern of alternating longer and shorter bonds
in the cluster coordinated phenyl rings. The geometry
of the free phenyl rings, as obtained from the refine-
ment of our X-ray diffraction data at ambient tempera-
ture, is more likely to be influenced by systematic errors
than that of the metal coordinated rings. The former
are less restricted to undergo librational motion in the
crystal than the latter, which are tightly bound to the
metal cluster frame. Indeed, inspection of the Hirshfeld
DA,B values [14] indicates considerable motion of the
free phenyl groups relative to the rest of the molecules
[15]. In effect, the above arguments attach a higher
significance to the observed distortion of the face-cap-
ping arene rings than to that of the phenyl substituents
in the side-chains. A similar case can be put forward for
complex 2d.
is found in the first-row (cobalt) cluster complexes. The
bending of the substituents out of the plane of the facial
arene ligand away from the metal cluster can also be
related to an optimisation of overlap between the metal
cluster and arene y-orbitals [12].
The presence of a second, non-coordinated phenyl
ring in 3c and 3d offers the possibility to further assess
the significance of the Kekule´-type distortion observed
in the facial phenyl ligands. In the previously published
structure of 2d, complications caused by the conjuga-
tion of the two phenyl rings with the CꢀC double bond
in the side chain could not be ruled out completely [5].
This problem is not present with the complexes of type
3. Inspection of Table 4 reveals that those three C–C
bonds within the v3-arene ligands which come to lie on
top of the cobalt atoms are consistently shorter than
the other set of three bonds, which fall in between the
metals. The difference is however small and not even
significant for some individual pairs of adjacent bonds
(Table 3).
As can be seen from Tables 3 and 4, there are some
pairs of C–C bonds within the free phenyl groups in
2d, 3c, 3d and [3d+H]+ which are significantly differ-
Table 4
˚
Selected distances (A) relating to the geometry of the phenyl rings in
complexes 3c, 3d, [3c+H]+ and 2d
3c
3d
[3d+H]+a 2db
v3-p2:p2:p2-C6H5R
The length of the central bond C13–C14 of the
1,2-diphenylethane ligand in 3c compares well with that
of the C13–C14 bond of the 1,1-isomer in 3d (Table 3)
and with results [16] obtained by ab initio MO-calcula-
tions on dibenzyl. An apparent discrepancy with the
values reported for free 1,2-diphenylethane (1.48–1.517
dmean(CC,
1.419(8)
1.411(4)
0.011
1.411(8)
0.020
1.418(6)
short)c
D
max[d(CC,
0.021
0.014
short)]d
dmean(CC,
1.440(9)
0.021
1.438(4)
0.010
1.442(9)
0.021
1.449(3)
0.006
long)e
D
D
D
D
max[d(CC,
˚
A) results from experimental artifacts, due to a particu-
long)]f
lar torsional vibration of the molecules in crystalline
dibenzyl [16,17].
mean[d(CC,
short–long)]g
min[d(CC,
0.021(10)
0.009
0.027(6)
0.020
0.031(8)
0.022
0.031(6)
0.021
short–long)]h
max[d(CC,
short–long)]i
3.2. Dynamic stereochemistry
0.041
0.039
0.043
0.041
The NMR spectroscopic data of 3b and c compare
well with those of other derivatives. In particular, the
proton and carbon chemical shifts of the facial v3-
phenyl groups are similar in 2 and 3. As in the free
ligands, there is an h-effect [18] of the substituents on
the chemical shifts of the ipso C atoms. These resonate
at a lower field than the CH resonances. The magnitude
of this shift (Dl=10–20 ppm) is independent of the
hybridisation (sp2 in 2, sp3 in 3) of the h-C.
In the solid state, the three cyclopentadienyl cobalt
units which form the cluster backbone of our complexes
are rendered inequivalent by the presence of the sub-
stituent on the facial arene ligands. The presence of
only one singlet for all the protons (and carbons,
C6H5R
dmean(CC)j
1.377(11)
0.014
0.014(8)
1.382(5)
0.029
0.005(2)
1.375(16)
0.054
0.020(13)
1.387(18)
0.057
0.021(14)
D
D
max[d(CC)]k
mean[d(CC)]l
a Ref. [1]. b Ref. ([6]a). c Mean of the lengths of the three C–C bonds
‘on top’ of the metal atoms with standard deviation in parentheses.
d Largest difference of the lengths of any two in the set of three C–C
bonds ‘on top’ of the metal atoms. e Same as but for the bonds ‘in
c
between’ the metal atoms. f Same as but for the bonds ‘in between’
d
the metal atoms. g Mean difference of the lengths of adjacent C–C
bonds with standard deviation in parentheses. h smallest difference
between the lengths of any two adjacent C–C bonds. i Same as but
h
largest difference. j Same as and but for all six C–C bonds. k Same
c
d
d
f
g
as and but for any two C–C bonds. l Same as but for all C–C
bonds.

H. Wadepohl et al. / Journal of Organometallic Chemistry 573 (1999) 22–29
27
Scheme 1.
Scheme 2.
respectively) of the three cyclopentadienyl ligands in the
1H- and 13C-NMR spectra is indicative of a dynamic
process which equilibrates the three cyclopentadienyl
ligands in solution [19]. A rotation of the v3-arene in the
plane parallel to the Co3 triangle is consistent with all
the experimental data, as will be illustrated in detail
below.
4. Conclusions
(1) Starting with the free arenes, the tricobalt cluster
complexes of the type [{(C5H5)Co}3(v3-arene)] can only
be synthesised with alkenylbenzenes as the facial lig-
ands. We have shown in this paper that the CꢀC double
bond in the side chain is by no means required for such
complexes to be stable once they have been formed.
This supports our view of a crucial involvement of the
alkenyl substituent in intermediate steps during the
assembly of the metal cluster [5].
(2) The structure of the [(C5H5)Co]3(v3-arene) cluster
core is very similar in the complexes 2 (with alkenylben-
zenes as facial ligands) and 3 (with alkylbenzenes as
facial ligands). A detailed comparison of the geometries
of the free and Co3 face-capping phenyl groups attaches
statistical significance to the small Kekule´-type distor-
tion of the latter ligands. The influence of the unsatura-
tion within the side chain on the structure of the
v3-phenyl ring is considered negligible.
Any derivative [{(C5H5)Co}3(v3-C6H5R)] with
a
monosubstituted facial arene ligand is intrinsically chi-
ral in the staggered (p2:p2:p2) conformation. This does
not only lead to inequivalence of the three (C5H5)Co
groups and of all five CH units of the v3-phenyl nucleus,
but also makes the two methyl groups of the prochiral
isopropyl substituent in 3b diastereotopic. The latter is
quite obvious in the low temperature proton spectra of
this complex. A 60° rotation of the arene (A“A%, e.g.
process I in Scheme 1) causes inversion of chirality.
Such a movement pairwise equilibrates the sets ortho
and meta CH-groups of the v3-phenyl ligand as well as
the methyl resonances in 3b [20], but it is not sufficient
to average the chemical shifts of all three (C5H5)Co
groups (only two of these are interchanged). However,
starting from A, A% can be reached by a second process
(II in Scheme 1), which has the same effect on the
v3-phenyl ring and on the methyl substituents but inter-
changes a different set of two out of the three (C5H5)Co
groups. Thus, any combination of the dynamic pro-
cesses I and II will reproduce the experimental spectra
at the high temperature limit, provided that both are
fast on the NMR time scale. In effect, this corresponds
to a full (360°) rotation of the facial arene ligand.
The stereochemical situation is somewhat different in
3d, due to the asymmetry of the h-carbon atom C13. In
this complex, the two staggered conformations B and C
(Scheme 2) are diastereomers, each of which may be a
pair of enantiomers. A fast rotation of the v3-arene
interconverts the diastereomers and equilibrates the
three (C5H5)Co groups. It does, however, not lead to
equivalence of the two ortho and the two meta CH-units
of the v3-phenyl ligand. In the proton spectra, the two
pseudo-doublets for the ortho protons of the v3-phenyl
group are clearly separated (Table 1). Furthermore, all
five carbon resonances (4 CH, 1 C) for the v3-phenyl
ligand are resolved even at the high temperature limit, in
complete agreement with the above arguments.
(3) Hydrogenation of the side-chains in 2 may be used
to synthesise derivatives which are particularly suited to
test and define the interesting dynamic stereochemistry
of facial arene coordination.
5. Experimental section
5.1. General procedures
All operations were carried out under an atmosphere
of purified nitrogen or argon (BASF R3-11 catalyst)
using Schlenk techniques. Solvents were dried by con-
ventional methods. Alumina used as a stationary phase
for column chromatography was first heated to 180–
200°C under vacuum for several days, then treated with
5% deoxygenated water and stored under nitrogen. The
cluster complexes 2b–d were prepared by published
methods ([6]a). NMR spectra were obtained on a
1
Bruker AC 200 instrument (200.1 MHz for H, 50.3
MHz for 13C). H and 13C chemical shifts are reported
1

28
H. Wadepohl et al. / Journal of Organometallic Chemistry 573 (1999) 22–29
versus TMS and were determined by reference to internal
TMS or residual solvent peaks. To assign the carbon
resonances, their multiplicities were determined using the
DEPT or J-modulated spin–echo techniques. Mass
spectra were measured in the electron impact ionisation
mode (EI) at 70 eV on Finnegan MAT 8230 and MAT
CH7 spectrometers. Elemental analyses were performed
locally by the microanalytical laboratory of the or-
ganisch-chemisches Institut der Universita¨t Heidelberg
and by Mikroanalytisches Labor Beller, Go¨ttingen.
M+), 371 (11, [(CpCo)3–H]+), 370 (64, [(CpCo)3–
2H]+), 365 (26), 247 (12, [(CpCo)2–H]+), 189 (100,
[Cp2Co]+), 182 (28, L+), 168 (11), 167 (77, [L–CH3]+),
165 (16), 124 (28, [CpCo]+), 59 (18, Co+); (L=1,1-
diphenylethane). Anal. Calc. (found): C, 62.83 (62.16);
H, 5.27 (5.07).
5.3. Crystal structure determinations
Single crystals were grown from toluene/petroleum
ether solutions at 5 to −20°C. Intensity data were
collected on a Siemens Stoe AED2 four circle diffrac-
tometer at ambient temperature and corrected for
Lorentz, polarisation and absorption effects (Table 5).
The structures were solved by direct methods, and refined
by full-matrix least-squares based on F2 using all mea-
sured unique reflections. All non-H atoms were given
anisotropic displacement parameters. The H atoms on
the v3-phenyl rings and on C13 (complex 3d only) were
located in difference Fourier syntheses and refined with
isotropic displacement parameters. All other H atoms
5.2. General procedure for hydrogenations
In a 500 ml flask a ca. 1 mmol sample of 2 is dissolved
in 50–150 ml of toluene. After addition of 30–50 mg Pd
on charcoal (10% Pd) the mixture is frozen with liquid
nitrogen. The flask is evacuated, filled with hydrogen and
warmed to r.t. The mixture is mechanically agitated until
hydrogenation is complete (checked by 1H-NMR of
samples taken periodically). The catalyst is then removed
by filtration (G4 frit). The filtrate is concentrated to a
small volume and chromatographed on alumina (2.5×
20 cm, eluent toluene/petroleum ether 1:1). The product
3 is crystallised from the eluate at −20°C.
Table 5
Details of the crystal structure determinations of complexes 3c and 3d
5.2.1. [{(C5H5)Co}3(v3-p2:p2:p2-iso-propylbenzene)]
(3b)
3c
3d
Formula
Mol. wt.
C29H29Co3
554.35
0.12 0.40 0.65
Monoclinic
P21
8.853(4)
9.460(5)
13.858(7)
90.56(3)
1160.5(10)
2
C29H29Co3
554.35
0.10 0.20 0.75
Monoclinic
Cc
17.358(5)
14.766(4)
9.296(3)
105.77(3)
2293.0(12)
4
Using 1.48 g (3.02 mmol) of 2b, a 750 mg yield (51%)
of 3b was obtained as dark brown microcrystals (reaction
time 24 h). M.p. 155°C. MS (EI) m/z (relative intensity):
492 (23, M+), 372 (15, [(CpCo)3]+), 371 (18, [(CpCo)3–
H]+), 370 (100, [(CpCo)3–2H]+), 311 (13), 310 (18), 247
(17, [(CpCo)2–H]+), 189 (43, [Cp2Co]+), 149 (17), 124
(13, [CpCo]+), 120 (17, L+), 119 (42, [L–H]+), 105 (45,
[L–CH3]+), 91 (26, [PhCH2]+), 77 (12, Ph+), 71 (20),
59 (6, Co+), 57 (37), 43 (40), 41 (32); (L=iso-propylben-
zene). Anal. Calc. (found): C, 58.54 (58.30); H, 5.53
(5.45).
Crystal size (mm)
Crystal system
Space group
˚
a (A)
˚
b (A)
˚
c (A)
i (°)
V (A )
3
˚
Z
Dcalc. (g cm−3
F(000)
)
1.586
568
2.13
1.606
1136
2.16
v(Mo–Kh) (mm−1
)
˚
X-radiation, u (A)
Mo–Kh, graphite monochromated,
0.71069
ambient
5.2.2. [{(C5H5)Co}3(v3-p2:p2:p2-1,2-diphenylethane)]
(3c)
Data collection tempera-
ture (K)
2qmax (°)
Using 660 mg (1.19 mmol) of 2c, a 560 mg yield (85%)
of 3c was obtained as black crystals (reaction time 10
days). M.p. 156°C. MS (EI) m/z (relative intensity): 554
(14, M+), 370 (45, [(CpCo)3–2H]+), 365 (12), 247 (20,
[(CpCo)2–H]+), 241 (47), 189 (97, [Cp2Co]+), 182 (8,
L+), 150 (17), 124 (47, [CpCo]+), 91 (100, [PhCH2]+),
65 (12, Cp+), 59 (24, Co+), 39 (18, [C3H3]+); (L=1,2-
diphenylethane). Anal. Calc. (found): C, 62.83 (62.76);
H, 5.27 (5.24).
62
60
hkl-range
−12/12, 0/13, 0/ −24/23, 0/20, 0/
20
3902
3902
2170
Empirical, „-
scans
345
0.945
13
3333
3333
2053
Empirical, „-
scans
334
Reflections measured
Unique
Observed [I]2|(I)]
Absorption correction
Parameters refined
GOF
0.927
R (obs. reflections only)
0.038
0.090
0.038, 0
[Max(F2o, 0)+
2F2c]/3
0.030
0.069
0.029, 0
[Max(F2o, 0)+
2F2c]/3
wR2 (all reflections)a
5.2.3. [{(C5H5)Co}3(v3-p2:p2:p2-1,1-diphenylethane)]
(3d)
Using 590 mg (1.07 mmol) of 2d, a 550 mg yield (93%)
of 3d was obtained as black crystals (reaction time 16 h).
M.p. 151°C. MS (EI) m/z (relative intensity): 554 (21,
A, B
P
a w=1/[|2(F2o)+(AP)2+BP].

H. Wadepohl et al. / Journal of Organometallic Chemistry 573 (1999) 22–29
29
Organometallics 10 (1991) 861. (b) H. Wadepohl, T. Borchert,
K. Bu¨chner, H. Pritzkow, Chem. Ber. 126 (1993) 1615. (c) H.
Wadepohl, T. Borchert, K. Bu¨chner, M. Herrmann, F.-J. Paffen,
H. Pritzkow, Organometallics 14 (1995) 3817. (d) H. Wadepohl,
T. Borchert, H. Pritzkow, J. Organomet. Chem. 516 (1996) 187.
[7] (a) H. Wadepohl, Angew. Chem. 104 (1992) 253. (b) H. Wade-
pohl, Comments Inorg. Chem. 15 (1994) 369. (c) H. Wadepohl,
S. Gebert, Coord. Chem. Rev. 143 (1995) 535.
[8] H. Wadepohl, The synergy between dynamics and reactivity at
clusters and surfaces, in: L.J. Farrugia (Ed.), Mathematical and
Physical Sciences, vol. 465, Nato ASI Series C, Kluwer, Dor-
drecht, 1995, p. 175.
[9] D. Braga, F. Grepioni, E. Parisini, H. Wadepohl, S. Gebert, J.
Chem. Soc. Dalton Trans. (1995) 1089.
[10] F. Zymalkowski, in: H. Kropf (Ed.), Methoden der organischen
Chemie (Houben-Weyl), vol. IV/1c, Part 1, 4th ed., Thieme
Verlag, Stuttgart, 1980.
[11] D. Braga, P.J. Dyson, F. Grepioni, B.F.G. Johnson, Chem. Rev.
94 (1994) 1585.
[12] (a) H. Wadepohl, L. Zhu, J. Organomet. Chem. 376 (1989) 115.
(b) L. Zhu, H. Wadepohl, N. Kostic, Wuji Huaxue Xuebao 8
(1992) 1.
[13] A. Domenicano, in: A. Domenicano, I. Hargittai (Eds.), Accu-
rate Molecular Structures, IUCr Monographs on Crystallogra-
phy 1, International Union of Crystallography, Oxford
University Press, Oxford, 1992, Ch. 18.
[14] (a) F.L. Hirshfeld, Acta Crystallogr. A32 (1976) 239. (b) R.E.
Rosenfield Jr., K.N. Trueblood, J. Dunitz, Acta Crystallogr. A34
(1978) 828.
were inserted in calculated positions (see Section 6).
The absolute structure could be determined for both
crystals by means of the Flack x-parameter, which was
close to zero (0.011(27) for 3c, −0.021(19) for 3d). The
calculations were performed using the programs
SHELXS-86 and SHELXL-97 [21]. Graphical represen-
tations were drawn with ORTEP-II [22].
6. Supplementary material available
Crystallographic data (excluding structure factors)
for the structures reported in this paper have been
deposited with the Cambridge Crystallographic Data
Centre as supplementary publication no. CCDC-
100813. Copies of the data can be obtained free of
charge on application to The Director, CCDC, 12
Union Road, Cambridge CB2 1EZ, UK (Fax: +44
1223 336033; e-mail: teched@chemcrys.cam.ac.uk).
Acknowledgements
This work was supported by the Deutsche
Forschungsgemeinschaft and the Fonds der Chemi-
schen Industrie. H. Wadepohl gratefully acknowledges
the award of a Heisenberg Fellowship.
[15] A quantitative treatment of these effects, which is possible in
principle and ultimately leads to corrections of the individual
bond lengths, was not considered sensible because of the limited
accuracy and q-range of our room temperature diffraction data.
[16] B. Kahr, C.A. Mitchell, J.M. Chance, R.V. Clark, P. Gantzel,
K.K. Baldridge, J.S. Siegel, J. Am. Chem. Soc. 117 (1995) 4479.
[17] J. Harada, K. Ogawa, S. Tomoda, J. Am. Chem. Soc. 117 (1995)
4476.
References
[18] H.O. Kalinowski, S. Berger, S. Braun, 13C-NMR-Spektroskopie,
Thieme Verlag, Stuttgart, 1984, Ch. 3, Part 4.10.
[19] Rapid rotation of the C5H5 ligands around their bonding axis to
the cobalt is implied as usual.
[20] The same is true for the diastereotopic methylene protons in 3c.
[21] (a) G.M. Sheldrick, SHELXS-86, Acta Crystallogr. A46 (1990)
467. (b) G.M. Sheldrick, SHELXL-97, Universita¨t Go¨ttingen,
1997.
[22] C.K. Johnson, ORTEP-II, Report ORNL-5138, Oak Ridge Na-
tional Laboratory, TN, USA.
[1] Part 8: H. Wadepohl, M.J. Calhorda, M. Herrmann, C. Jost,
P.E.M. Lopes, H. Pritzkow, Organometallics 15 (1996) 5622.
[2] H. Wadepohl, Angew. Chem. 104 (1992) 253.
[3] J. Mu¨ller, P. Escarpa Gaede, Ke Qiao, Angew. Chem. 105 (1993)
1809.
[4] H. Wadepohl, K. Bu¨chner, H. Pritzkow, Angew. Chem. 99
(1987) 1294.
[5] H. Wadepohl, K. Bu¨chner, H. Pritzkow, Organometallics 8
(1989) 2745.
[6] (a) H. Wadepohl, K. Bu¨chner, M. Herrmann, H. Pritzkow,
.