Mechanistic insights into an unprecedented C-C bond activation on a Rh/Ga bimetallic complex: A combined experimental/computational approach

Article abstract of DOI:10.1021/ja055298d

The unusual rearrangement of [RhCp*(GaCp*)(CH₃) ₂] (1c) to [RhCp*(C₅Me₄Ga(CH ₃)₃)] (2) is presented and its mechanism is discussed in detail. ¹³C MAS NMR spectroscopy revealed that the title reaction proceeds cleanly not only in solution but also in solid state, which supports a unimolecular reaction pathway. On the basis of ¹H, ¹³C, and ROESY NMR spectroscopy as well as isolation and structural elucidation of the hydrolysis product, the compound [RhCp*(endo-η⁴-C ₅Me₅GaMe₂)] (3a) was identified as a crucial reaction intermediate. DFT calculations on the B3LYP level of theory support this assignment and suggest a concerted C-C bond activation mechanism that topologically takes place at the gallium center. Furthermore, two fluxional processes of the reaction intermediate 3a were studied experimentally as well as by computational methods. First, a mechanism takes place similar to a ring-slipping process that exchanges a GaMe₂ group between adjacent ring carbon atoms within the same Cp* ring. This process proceeds at a rate comparable to the NMR time scale and indeed is calculated to be energetically very favorable. Second, a unimolecular exchange process of the GaMe₂ group between the two Cp* rings of 3a could be experimentally proven by the introduction of phenyl substituents as a label into the Cp* ligands at both sites, the rhodium as well as the gallium center. A series of experiments including deuteration studies and competition reactions was performed to substantiate the suggested mechanism being in accordance with DFT calculations on possible transition states.

Full text of DOI:10.1021/ja055298d

Published on Web 11/10/2005
Mechanistic Insights into an Unprecedented C-C Bond
Activation on a Rh/Ga Bimetallic Complex: A Combined
Experimental/Computational Approach
Thomas Cadenbach, Christian Gemel, Rochus Schmid, and Roland A. Fischer*
Contribution† from the Anorganische Chemie II, Organometallics & Materials, Ruhr-UniVersita¨t
Bochum, D-44780 Bochum, Germany
Received August 13, 2005; E-mail: roland.fischer@ruhr-uni-bochum.de
Abstract: The unusual rearrangement of [RhCp*(GaCp*)(CH3)2] (1c) to [RhCp*(C5Me4Ga(CH3)3)] (2) is
presented and its mechanism is discussed in detail. 13C MAS NMR spectroscopy revealed that the title
reaction proceeds cleanly not only in solution but also in solid state, which supports a unimolecular reaction
pathway. On the basis of 1H, 13C, and ROESY NMR spectroscopy as well as isolation and structural
elucidation of the hydrolysis product, the compound [RhCp*(endo-η4-C5Me5GaMe2)] (3a) was identified as
a crucial reaction intermediate. DFT calculations on the B3LYP level of theory support this assignment
and suggest a concerted C-C bond activation mechanism that topologically takes place at the gallium
center. Furthermore, two fluxional processes of the reaction intermediate 3a were studied experimentally
as well as by computational methods. First, a mechanism takes place similar to a ring-slipping process
that exchanges a GaMe2 group between adjacent ring carbon atoms within the same Cp* ring. This process
proceeds at a rate comparable to the NMR time scale and indeed is calculated to be energetically very
favorable. Second, a unimolecular exchange process of the GaMe2 group between the two Cp* rings of 3a
could be experimentally proven by the introduction of phenyl substituents as a label into the Cp* ligands
at both sites, the rhodium as well as the gallium center. A series of experiments including deuteration
studies and competition reactions was performed to substantiate the suggested mechanism being in
accordance with DFT calculations on possible transition states.
Introduction
been extensively studied aiming at the understanding of
mechanistic aspects of catalytic transformations involving
The coordination chemistry of carbenoid ligands of the
heavier main group elements at transition metal centers has
continued to be a major driving force in the development of
fundamental organometallic chemistry since the early days of
Fischer’s discovery of the transition metal-carbon multiple
bonds.1 However, with the now very famous exception of the
so-called N-heterocyclic carbenes, NHCs,2-4 almost none of the
various subclasses of transition metal carbenoid ligand com-
plexes gained importance beyond the detailed understanding of
their structure and bonding situations. Strictly speaking, serious
examples for a potential use of carbenoid species of heavier
main group elements as controlling ligands in transition-metal-
mediated organic synthesis, as is the case for CO, phosphines,
or NHCs, aresnot surprisinglysvirtually nonexistent. Namely,
the coordination chemistry of silylenes5-7 and borylenes8-11 has
silylations and borylations mediated by transition-metal centers,
but not in the view of tuning the reactivity of the transition-
metal center itself. On the other hand, the recent report on the
first triple bond between lead and a transition metal center, e.g.
trans-[Br(PMe3)4MotPb(2,6-C6H3R′2)] [R′ ) (triisopropylphe-
nyl)phenyl] may be regarded as a typical case of research
primarily focusing on static aspects of structure and bonding
only.12 The same is largely true for the chemistry of ECp* (E
) Al, Ga, In) and related low-valent group 13 metal compounds
ER (R ) C(SiMe3)3, bulky bis-imidinates, etc.) being formally
isolobal to CO and/or cationic alkylidine fragments CR+, the
chemistry of which developed into its own field in the past
decade.13 For instance, we described novel homoleptic clusters
[Ma(ECp*)b], adopting solid-state structures similar to the
classical carbonyl clusters [Ma(CO)b] and thus going well
beyond the coordination chemistry of other heavier carbenoid
complexes.14-17 In solution, the clusters [Ma(ECp*)b] exhibit a
rich fluxional behavior following inter- as well as intramolecular
† Organo group 13 complexes of transition metals XXXIX; for XXXVIII,
see ref 17.
(1) Tokitoh, N.; Okazaki, R. Coord. Chem. ReV. 2000, 210, 251.
(2) Crabtree, R. H.; Peris, E. Coord. Chem. ReV. 2004, 248, 2239.
(3) Gade, L. H.; Stephane, B.-L.; Vincent, C. Chem. Soc. ReV. 2004, 33, 619.
(4) Hillier, A. C.; Grasa, G. A.; Viciu, M. S.; Lee, H. M.; Yang, C.; Nolan,
S. P. J. Organomet. Chem. 2002, 653, 69.
(5) Brunner, H. Angew. Chem., Int. Ed. 2004, 43, 2749.
(6) Milstein, D.; Goikhman, R. Chem. Eur. J. 2005, 11, 2983.
(7) Tilley, T. D.; Glaser, P. B. J. Am. Chem. Soc 2003, 125, 13640.
(8) Braunschweig, H. Angew. Chem., Int. Ed. Engl. 1998, 37, 1786.
(9) Braunschweig, H.; Colling, M. Coord. Chem. ReV. 2001, 223, 1.
(10) Braunschweig, H.; Colling, M. Eur. J. Inorg. Chem. 2003, 393.
(11) Braunschweig, H. AdV. Organomet. Chem. 2004, 51, 163.
(12) Filippou, A. C.; Rohde, H.; Schnakenburg, G. Angew. Chem. 2004, 116,
2293.
(13) Uhl, W.; Pohlmann, M.; Wartchow, R. Angew. Chem. 1998, 110, 1007.
(14) Steinke, T.; Gemel, C.; Winter, M.; Fischer, R. A. Angew. Chem. 2002,
114, 4955.
(15) Gemel, C.; Steinke, T.; Weiss, D.; Cokoja, M.; Winter, M.; Fischer, R. A.
Organometallics 2003, 22, 2705.
9
17068
J. AM. CHEM. SOC. 2005, 127, 17068-17078
10.1021/ja055298d CCC: $30.25 © 2005 American Chemical Society

C−C Bond Activation on a Rh/Ga Bimetallic Complex
A R T I C L E S
Scheme 1. Reaction of [RhCp*(GaCp*)(CH3)2] (1c) Giving
[RhCp*(C5Me4E(CH3)3)] (2)
mechanisms of ligand exchange. In part, this exclusivity of ECp*
ligands compared to other heavier main group carbenoids is
based on their distinctive electronic characteristics. They
combine exceptionally potent σ-donor and rather weak π-ac-
ceptor properties with the soft coordination behavior of Cp*,
allowing haptotropic shifts and thus a delicate mediation of the
steric and electronic situation at the group 13 center.18-20 Only
quite recently, coordinatively unsaturated intermediates such as
[Ni(AlCp*)n] (n < 4)21 and [Fe(AlCp*)m] (m < 5)22 were found
to activate C-H bonds under very mild conditions, i.e. aromatic
C-H bonds by an intermolecular reaction in the former case
and aliphatic C-H bonds by an intramolecular reaction in the
latter case. On coordination of ECp*, the electron density of
the transition metal is considerably increased, leaving at the same
time an electrophilic group 13 metal center. Both the resulting
nucleophilic, oxidizable, transition metal as well as the elec-
trophilic main group metal represent the principal requirements
for bond activation reactions. These findings point to the quite
unique potential of these rather exotic species ECp* (or ER in
general) to be considered as novel controlling ligands in
organometallic chemistry.
In the course of these investigations we recently discovered
an unprecedented C-C bond activation in the reaction of
[RhCp*(L)(CH3)2] [L ) DMSO (1a), pyridine (1b)] with GaCp*
or AlCp*, giving the zwitterionic species [RhCp*(C5Me4E-
(CH3)3)] (2) (Scheme 1).23 The stability of the 18 VE rhode-
nocenium structure together with the oxidation of Ga(I) to
Ga(III) clearly represents a strong driving force. This complex
rearrangement for E ) Ga was shown to proceed via the
substitution product [RhCp*(GaCp*)(CH3)2] (1c), which could
be isolated and characterized. Complex 1c was observed to be
very labile also in the solid state, showing a color change from
deep orange to pale brown within a few minutes at 60 °C. The
solution NMR spectra of this pale residue were identical with
the spectra of the final reaction product 2, which was taken as
the first indication that the reaction from 1c to 2 also proceeds
cleanly in the solid state.
Figure 1. Representative cutout (85-130 ppm) of VT 13C-MAS NMR
spectra of the solid-state reaction of 1c giving 2.
The moderate reaction rates matching the NMR time scale
combined with the synthetic accessibility of potential intermedi-
ates and the rather unusual possibility to directly compare the
reaction in the solid state with the solution prompted us to look
in detail into the mechanism of the title reaction in order to
learn more about the reactiVe aspects of carbenoid group 13
ligands attached to a transition-metal center. Our investigations
are based on a thorough experimental analysis of a crucial
intermediate by means of 1D and 2D NMR spectroscopic
techniques, isotope labeling studies, cross-mixing experiments,
as well as theoretical modeling of the whole reaction sequence
searching for possible transition states, intermediates, and
alternative mechanisms on the B3LYP24,25/LanL2DZ26 level of
theory.27
Results and Discussion
Reaction in the Solid State. To verify that the reaction of
1c to 2 does indeed proceed not only in solution but also in the
solid state, a series of 13C MAS NMR spectra of 1c was
recorded. Figure 1 illustrates a representative cutout of spectra
obtained as a subject to temperature and time (a figure
containing the full spectra is available in the Supporting
Information). Evidently, the formation of 2 proceeds cleanly
also in the solid state, however, without the spectroscopic
appearance of intermediates.
The most important consequence from this observation is the
fact that at least one intermolecular pathway connecting 1c with
2 must be energetically available, not ruling out competing
alternatives in solution. Foremost, evidence of such an intramo-
lecular pathway validates efforts to support the experiments by
a computational approach, as likely mechanistic alternatives are
more strongly limited and thus much easier to model than would
be in the case of distinctly intermolecular mechanisms.
Reaction in Solution. Monitoring the reaction from 1c to 2
by 1H NMR shows a strong dependence of the resulting spectra
on the solvent used. We found, that in the presence of donors,
such as pyridine or THF, an intermediate (3) appears, which
subsequently disappears in due course of the reaction, finally
leading to a clean 1H NMR spectrum of 2 under all conditions.
(16) Steinke, T.; Gemel, C.; Winter, M.; Fischer, R. A. Chem. Eur. J. 2005,
11, 1636.
(17) Buchin, B.; Gemel, C.; Cadenbach, T.; Fischer, R. A. Inorg. Chem. 2005,
44, submitted.
(18) Steinke, T.; Gemel, C.; Cokoja, M.; Fischer, R. A. J. Chem. Soc., Dalton
Trans. 2005, 55.
(19) Cokoja, M.; Gemel, C.; Steinke, T.; Schroeder, F.; Fischer, R. A. J. Chem.
Soc., Dalton Trans. 2005, 44.
(20) Steinke, T.; Gemel, C.; Cokoja, M.; Winter, M.; Fischer, R. A. Chem.
Commun. 2003, 1066.
(24) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(21) Steinke, T.; Gemel, C.; Cokoja, M.; Winter, M.; Fischer, R. A. Angew.
Chem. 2004, 116, 2349.
(25) Parr, R. G.; Yang, W.; Lee, C. Phys. ReV. B 1988, 37, 785.
(26) Dunning, T. H.; Hay, P. J. Modern Theoretical Chemistry, Plenum: New
York, 1976; Vol. 3.
(22) Steinke, T.; Cokoja, M.; Gemel, C.; Kempter, A.; Krapp, A.; Frenking,
G.; Zenneck, U.; Fischer, R. A. Angew. Chem. 2005, 117, 3003.
(23) Cadenbach, T.; Gemel, C.; Schmid, R.; Block, S.; Fischer, R. A. J. Chem.
Soc., Dalton Trans. 2004, 3171.
(27) All energies in this paper are given as relative free energies ∆G° for the
standard state. For comparison the zero-point energy corrected EDFT values
are given in the Figures 3, 4, and 6 in brackets.
9
J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005 17069

A R T I C L E S
Cadenbach et al.
The fact that C1 and as well C6 show a rather large Rh-C
coupling (JRhC ) 6.85 and 4.53 Hz, respectively) cannot be
explained at this point. An agostic interaction of the C-CH3
bond and the rhodium center, for example (which could be the
case for structure 3b), is ruled out due to the formal 18 VE
count of the transition metal. However, the question whether
3a or 3b corresponds to the actually spectroscopically observed
species cannot be decided solely on the basis of the 13C NMR
spectrum.
Mechanistic Hypotheses for the Reaction of 1 to 2. In
principle, three different pathways for the reaction of 1c to 2
are feasible with either the endo-complex 3a or the exo-complex
3b as a reaction intermediate (Scheme 2). In all three mecha-
nisms the first step is a transfer of a single methyl group from
the rhodium to the gallium center, giving the spectroscopically
nondetectable intermediate “Rh(II)-Ga(II)” species 4. The
transfer of the second methyl group from the rhodium to the
gallium under coordination of the diene moiety of the Cp*GaMe2
formed yields either 3a (pathway A) or 3b (pathways B and C)
as the consecutive reaction intermediates.
Figure 2. Possible intermediates 3a and 3b.
1
The H NMR spectrum of 3 consists of three singlets in an
integral ratio of 15:15:6, representing two chemically non-
equivalent Cp* rings and two chemically equivalent CH3 groups.
Noticeable, one of the Cp* signals is considerably broadened,
suggesting that some kind of fluxional behavior is involved.
Indeed, at -40 °C this broad signal splits into three sharp
singlets in an integral ratio of 6:6:3. The fluxional process
becomes slower with increasing donor strength of the solvent
[w1/2(C6D6/1 equiv py) ) 9.23 Hz; w1/2(py) ) 34.6 Hz], and at
the same time the stability of 3 increases. In neat pyridine, for
example, 3 is stable at room temperature for at least 5 days.
Only heating the solution above 40 °C for 1 h allows detection
of 2.
The fact that the signal for the two CH3 groups initially bound
to the rhodium atom does not show a Rh-CH3 coupling clearly
suggests that both methyl groups have migrated from the
rhodium to the gallium center. Additionally, the presence of
two signals each representing 15 protons points to the fact that
both Cp* ligands are still intact, i.e. no C-C activation has
taken place at this point of the reaction. Further insight into the
molecular structure of 3 is gained by analysis of the low-
temperature 13C NMR spectrum: At -100 °C the 13C NMR
spectrum in THF shows five doublets at 90.1 (JRhC ) 5.48 Hz),
In pathway A, the actual C-C bond rupture takes place at
the gallium center of 3a via a direct transfer of the exo-oriented
methyl group (C6) to the endo-oriented GaMe2 moiety. Pathway
B describes a similar mechanism, with the exception that the
locations of the migrating CH3 group and the GaMe2 moiety
are now swapped, i.e. the endo-oriented CH3 (C6) group in 3b
is directly transferred to the exo-oriented GaMe2 moiety via a
concerted mechanism. Pathway C takes into account that the
endo-oriented CH3 group in 3b can first migrate to the rhodium
center (i.e. a “classic” C-C bond activation at a Rh(I)
center),28,29 giving the half-sandwich Rh(III) complex 5, bearing
a methyl group and a σ-bound (C5Me4GaMe2) ligand. Subse-
quent transfer of the methyl group to the gallium and coordina-
tion of the resulting (C5Me4GaMe3) unit to the rhodium center
finally lead to the formation of 2. The question if 3a or 3b is
the experimentally observed intermediate must be addressed
experimentally. However, to compare all three pathways with
respect to the energies of their intermediates and transition states,
also DFT calculations were performed.
83.8 (JRhC ) 9.10 Hz), 62.3 (JRhC ) 13.17 Hz), 56.9 (JRhC
)
6.85 Hz), and 29.3 ppm (JRhC ) 4.53 Hz), as well as four
singlets at 11.5, 9.7, 8.3, and -3.9 ppm. The doublet at 90.1
ppm integrating to approximately five carbons can unequivocally
be assigned to the aromatic carbon atoms of the RhCp* moiety.
The presence of two more doublets in the olefinic region, each
integrating to two carbon atoms, indicates a structure where
direct Rh-C bonds are established also to the second Cp* ring.
η4-Coordination of the diene moiety of (η1-Cp*)GaMe2 to the
RhCp* fragment forming an 18 VE Rh(I) complex is a
reasonable suggestion for the molecular structure of 3 and
consistent with the NMR spectra. The signals at 83.8 and 62.3
ppm can be assigned to the internal and the terminal carbon
atoms of the diene moiety, respectively. The four singlets in
the aliphatic region of the 13C NMR spectrum subsequently
belong to the methyl groups of the Rh(C5Me5) moiety (5C, 8.3
ppm), the C5Me4(Me)GaMe2 unit (2C, 11.5 ppm and 2C, 9.7
ppm), and the GaMe2 group (2C, -3.9 ppm). Consequently,
the sp3 carbon of the Cp*GaMe2 fragment (C1) and its adjacent
methyl group (C6) give rise to the remaining two doublets at
56.9 and 29.3 ppm.
Characterization of the Key Intermediate 3a/b. Several
attempts to crystallize the intermediate 3a/b failed due to its
high solubility under the conditions of formation and stabiliza-
tion. We followed two major strategies for an adequate
stereochemical analysis of 3a/b: (i) a reasonable chemical
modification of the intermediate in order to sufficiently suppress
the bond activation step and thus to obtain an X-ray single-
crystal structure of the modified intermediate and (ii) 2D
ROESY NMR spectroscopy.
In an attempt to chemically modify 3a/b, a freshly prepared
solution of this intermediate in THF was treated with an
equimolar amount of salicylic acid with the intent to hydrolyze
the sterically accessible Ga-CH3 groups under formation of a
chelating gallium salicylate complex. However, monitoring this
1
reaction by H NMR spectroscopy shows that both Ga-CH3
bonds remain intact and instead the hydrolysis of the Cp*-Ga
bond takes place, forming [Cp*Rh(C5Me5H)] (6) and [(salicyl-
ate)GaMe2] as the reaction products. 6 is also obtained by
Two stereoisomers of this intermediate are possible: (i) a
structure where the GaMe2 unit is located in an endo-position
with respect to the RhCp* moiety (3a) and (ii) a structure with
an exo-orientation of the GaMe2 unit (3b). Figure 2 illustrates
the calculated optimized structures fore these two isomers.
(28) Crabtree, R. H.; Dion, R. P.; Gibboni, D. J.; McGrath, R. B.; Holt, M. S.
J. Am. Chem. Soc. 1986, 108, 7222.
(29) Crabtree, R. H.; Dion, R. P. Chem. Commun. 1984, 1260.
9
17070 J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005

C−C Bond Activation on a Rh/Ga Bimetallic Complex
A R T I C L E S
Scheme 2. Mechanistic Working Hypotheses for the Intramolecular Rearrangement of 1c into 2
hydrolysis with an excess of acetic acid, whereas HCl/Et2O or
H2O lead to complete decomposition of the material. 6 can be
obtained analytically pure in high yields by sublimation of the
crude product in vacuo as well-shaped yellow single crystals.
An X-ray crystal structure of 6 (Figure 5) reveals the methyl
group C6 to be located in the exo-position with respect to the
RhCp* fragment. It should be mentioned that the respective
endo-isomer of this compound has been electrochemically
To collect further experimental support for the endo-structure
3a as the key intermediate of the title reaction, a 2D ROESY
spectrum was recorded. The minimum H-H distance of the
RhCp* and the GaMe2 fragments in the minimized structure of
3a is 2.38 Å, and therefore, a cross-peak between these two
signals is expected in the case of an endo-orientation of the
GaMe2 fragment. In the case of 3b, no such signal should be
observed. Indeed, the appearance of a strong ROESY cross-
peak for the RhCp* and the GaMe2 signals unequivocally points
to 3a.
DFT Calculations for Pathways A, B, and C. Figure 3
depicts the calculated pathways for the formation of 3a and 3b
from 1c. The formation of the “Rh(II)-Ga(II)” intermediate 4
resulting from 1c via migration of one CH3 group from the
rhodium to the gallium center is calculated to be endergonic
by 7.9 kcal/mol. This cost in energy is partially a result of a
change in the coordination mode of the GaCp* ring, shifting
from η5 in 1c to η1 in 4. The energy of the transition state of
this process (TS1) is 11.3 kcal/mol above that of 1c and also
points to the loss of aromatization energy of the GaCp* ring.
The next step in the reaction is the migration of the second
Rh-CH3 group to the gallium center under simultaneous η4-
coordination of the diene moiety of the [(η1-C5Me5)(GaMe2)]
fragment to the rhodium atom. In the case of 3a, i.e. an endo-
arrangement of the GaMe2 group with respect to the RhCp*
fragment, a transition state (TS2a) could be located, being only
2.8 kcal/mol above 4 (which is 10.7 kcal/mol above 1c).
In contrast to the formation of 3a, no transition state could
be located for the direct conversion of 4 to 3b (TS2b). It is
questionable if such a direct conversion is possible at all, since
a 180° rotation of the [C5Me5(GaMe2)] moiety perpendicular
1
prepared and analyzed by H NMR spectroscopy.30
Since the mechanism of the hydrolysis reaction and particu-
larly its stereochemical consequences are unclear, the exo-
orientation of the methyl group in 6 is not necessarily an
indication for an endo-orientation of the GaMe2 group in 3.
However, the 13C NMR spectrum of 6 reveals a strong
spectroscopic similarity to 3a/b (Table 1).
Above all, C1 as well as the C6 exhibit large Rh-C coupling
constants in both complexes. At this point the origin of this
unusually strong 2JRhC and 3JRhC coupling is uncertain; probably
the nearly ideal zigzag geometry in the bonding arrangement
of the rhodium atom, the exo-methyl group, and the two linking
C5 ring carbon atoms is responsible for the strong magnetic
communication between the nuclei. Thus, the C1-C6 axis in
the solid-state structure of 6 is almost perfectly perpendicular
to the diene plane of the C5Me5H ligand (87.5°) and hence
almost perfectly parallel to the Rh-dienecentroid axis. Noteworthy,
the endo-oriented proton in 6 does not show a detectable JRhH
coupling. However, the strong similarity of the 13C NMR spectra
of 3a/b and 6 is an indication of, yet not convincing evidence
for the presence of 3a.
(30) Gusev, O. V.; Denisovich, L. I.; Peterleitner, M. G.; Rubezhov, A. Z.;
Ustynyuk, N. A.; Maitlis, P. M. J. Organomet. Chem. 1993, 452, 219.
9
J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005 17071

A R T I C L E S
Cadenbach et al.
Figure 3. Calculated pathway for the formation of 3a and 3b from 1c.
to its bond axis to the rhodium center is needed, a movement
that seems to be impossible without prior dissociation of the
whole fragment. Such a dissociative process, however, is
expected to be energetically very unfavorable and does, above
all, not cope with a unimolecular rearrangement sequence for
the conversion of 1c to 2. Even so, a unimolecular conversion
from 4 to 3b can certainly not be ruled out solely on the basis
of its transition state not being located computationally.
Thermodynamically 3b is more favorable than 3a by 2.3 kcal/
mol, which is apparently a consequence of the steric interference
of the GaMe2 group with the RhCp* fragment in 3a (smallest
H-H distance ) 2.38 Å). When comparing the energies of the
pyridine adducts of 3a and 3b, this difference becomes even
more evident, now being 7.5 kcal/mol. The stronger interaction
of the pyridine with the acidic gallium in 3b is well-reflected
by its significantly smaller Ga-N bond length (Ga-N ) 2.13
Å in 3b vs 2.19 Å in 3a).
Figure 4 illustrates pathways A, B, and C for the C-C
activation process. The final reaction product 2 is calculated to
be more stable than the starting complex 1c by 12.4 kcal/mol.
It should be noted here that the geometrical features of the
optimized structures of both 1c and 2 are in good agreement
with the X-ray crystal structures (see Supporting Information
for further information).
the solid state. The reaction is a concerted migration of the CH3
group (C6), with TS3A exhibiting a classical five-coordinate
carbon “halfway” between the C5 ring and the gallium center.
On a closer look at the bond lengths of TS3A and comparison
with the respective structural features of 1c and 2, it becomes
apparent that the migration of the gallium atom toward the
methyl group proceeds concomitantly with an elongation of the
C-C bond. This movement of the gallium in the direction of
the C5 ring plane of the product allows an interaction of C1
with the rhodium center, i.e. a partial aromatization of the C5
ring already in the stage of TS3A, which is probably responsible
for the low activation barrier.
The transition state TS3B for pathway B, starting out from
the exo-isomer 3b, is energetically by far more unfavorable than
TS3A, being 33.2 kcal/mol above 3b. The geometrical reaction
characteristics for pathways A and B are in general very similar,
where TS3B is also representing a “classical” transition state
for a concerted bond activation process. However, as the
positions of the GaMe2 group and the CH3 group (C6) are now
swapped with respect to pathway A, an interaction of C1 with
the rhodium atom in TS3B is sterically impossible. This missing
interaction, which is also reflected by the higher distortion of
the C5 ring from planarity in TS3B, evidently accounts for its
relatively high energy.
The transition state TS3A for the rearrangement of 3a to 2
could be located and is energetically 17.4 kcal/mol above 3a.
This is a remarkably small activation barrier for a C-C bond
rupture and thus is in good agreement with the experimentally
observed fact that the reaction of 1c to 2 is taking place under
exceptionally mild reaction conditions in solution as well as in
For the third pathway (pathway C) the intermediate 5 as well
as the respective transition state TS3C could be optimized on
the potential energy surface. 5 is energetically less favorable
than 3b by 19.3 kcal/mol. TS3C is unexpectedly high in energy,
being 56.9 kcal/mol above 3b. This is a surprising result, since
C-C activation barriers on Rh(I) centers are typically consider-
9
17072 J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005

C−C Bond Activation on a Rh/Ga Bimetallic Complex
A R T I C L E S
Figure 4. Calculated pathways A, B and C for the C-C bond-breaking step.
suggest that the methyl group migration from the rhodium to
the gallium atom should not be the limiting step for pathway
C.
Several important conclusions can be gained from these
computations: 3a as well as 3b clearly represent likely
intermediates in the reaction of 1c to 2. Isomer 3b is thermo-
dynamically more stable than 3a, particularly when stabilized
by pyridine, which is present in the actual reaction system. Yet,
no direct path from 1c to 3b could be found. In addition, the
barrier for the C-C activation is very low for 3a but consider-
ably higher for both pathways involving 3b. This is in very
good agreement with the experimental facts suggesting 3a as
the important reaction intermediate (vide supra).
Fluxional Processes of 3a. On the basis of the structure of
3a, two possible fluxional processes become evident (Scheme
3): First, a ring-slipping process of the Ga(C5Me5)Me2 ligand,
best described as a “spinning” movement of the GaMe2 fragment
around the Rh-Cp*centroid axis (fluctuation A). This movement
is presumably responsible for the experimentally observed
coalescence of the Cp* methyl groups. Second, it is conceivable
that a Cp* ring-exchange process takes place, i.e. a movement
where the direct Ga-C bond to one Cp* ring is broken and
subsequently a new Ga-C bond to the second Cp* ring is
formed (fluctuation B). No evidence for this process is gained
Figure 5. Structure of the hydrolysis product 6 of 3a in the solid state.
Table 1. Comparison of Chemical Shifts and Coupling Constants
JRhC of 3a/b and 6
3a/b
6
δ(C1), ppm
δ(C6), ppm
2JRhC(C1), Hz
3JRhC(C6), Hz
56.9
29.3
6.85
4.53
61.5
24.2
5.34
3.47
ably lower.31 The transition state for the final methyl group
migration in pathway C (TS3C′) could not be located despite
several attempts. However, the low energies of TS1 and TS2a
(31) Rybtchinski, B.; Milstein, D. Angew. Chem. 1999, 111, 918.
9
J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005 17073

A R T I C L E S
Cadenbach et al.
Scheme 3. Possible Fluxional Processes of 3a
The introduction of one phenyl group instead of one methyl
group in one of the two Cp* rings in 1c allows us to chemically
and spectroscopically distinguish between the two Cp* rings.
This results in either [Rh(C5Me4Ph)(GaCp*)Me2] (9), if the
phenyl group is attached to the RhCp* fragment, or [Rh(Cp*)-
(GaC5Me4Ph)Me2] (10), if the phenyl group is introduced into
the GaCp* moiety. 9 could be isolated and spectroscopically
characterized, whereas 10 can be prepared only in situ due to
1
its high reactivity. According to H NMR, starting from both
compounds, a mixture of two products is formed: [Rh(C5Me4-
Ph)(C5Me4PhGaMe3)] (11a), a complex where the GaMe3 and
the phenyl group are attached to two different rings, and
[RhCp*(C5Me3PhGaMe3)] (11b), where both groups are at-
tached to the same ring. Interestingly, the molar ratio of these
two isomers is identical, being 80 (11a) to 20 (11b) for both
reactions!
The aliphatic region of the 1H NMR spectrum of 11a consists
of five singlets at 1.79 (6H), 1.49 (6H), 1.43 (6H), 1.02 (6H),
and 0.20 ppm (9H). The aliphatic part of the 1H NMR spectrum
of the minor isomer 11b also consists of five peaks, however,
showing different integral ratios. The signal at 1.35 ppm (15H)
can be easily assigned to the RhCp* fragment and the signal at
0.22 ppm to the GaMe3 group. The ring methyl groups of the
[C5Me3Ph(GaMe3)] moiety give rise to three signals at 1.98,
1.933, and 1.932 ppm, all integrating to three protons. On the
basis of these spectral data it cannot be unequivocally decided
if either the ortho- or meta-isomer of 11b, respectively, is
formed. The assigned chemical shifts for 11a and 11b were
cross-checked by isolation and spectroscopic characterization
of [Rh(C5Me4Ph)(C5Me3PhGaMe3)] (12), which can be obtained
by reaction of [Rh(C5Me4Ph)(pyridine)Me2] (12) and Ga-
(C5Me4Ph). Single crystals of the major isomer 11a were
obtained by crystallization from THF/Et2O. The solid-state
structure does not show any unexpected geometric features, and
a representation of the molecular structure can be obtained from
the Supporting Information.
1
from the H or 13C NMR spectra of 3a, which either means
that it is slower than the NMR time scale or nonexistent at all.
However, experimental evidence of a unimolecular mechanism
for fluctuation B would be a strong proof for 3a as the reaction
intermediate. Thus, both processes were studied experimentally
by spectroscopic and synthetic techniques, as well as by
computational modeling of the reactions’ pathways.
The activation energy for fluctuation A could be experimen-
tally determined by low-temperature NMR spectroscopy. H
1
NMR spectra were measured at -33, -25, -14, -3, and 13
°C and the line widths of the resulting spectra compared to fitted
spectra resulting from simulations. By a thus obtained Eyring
plot (R2 ) 0.98) the activation energy is calculated to be 22.6
kcal/mol. This value is considerably higher than the energy of
the respective calculated transition state (TS4). However, it must
be taken into account that the rate was determined in a solution
containing a comparably high concentration of pyridine, which
stabilizes 3a/b under the conditions of our experiments (vide
supra). Thus, the actual concentration of 3a in solution is
certainly decreased in a dissociation/association equilibrium with
the pyridine complex, which itself is not expected to undergo a
similar fluxional process. The fact that both high concentrations
of pyridine and very low temperatures retard the fluxional
processes substantiates this assumption.
Thus, it seemed impossible to gain spectroscopic evidence
for the slower process, i.e. fluctuation B. Broadening of the 1H
NMR line width of the RhCp* signal never observed, no matter
what conditions we tried. For that reason, no further NMR
studies such as ECSY or DANTE experiments were performed;
instead a labeling approach for the elucidation of process B was
given preference.
The fact that the same ratio of 11a and 11b are obtained by
starting from either 9 or 10 can be regarded as experimental
evidence for fluctuation B, however, only if a unimolecular
mechanism is assumed. An experimental indication for this can
be gained from the solid-state reaction of 9. Thus, a crystalline
sample of 9 was heated to 60 °C for 1 min, leading to a color
1
change from orange red to pale brown. Indeed, the H NMR
spectrum in C6D6 of the solid obtained consists of exactly the
same signals in the same ratios as the 1H NMR spectra obtained
from the solution reactions of 9 and 10.
The calculated energetic minima and transition states respon-
sible for both fluxional processes in 3a are depicted in Figure
Scheme 4. Reaction of Complexes of the Type 1c Labeled with One Phenyl Group
9
17074 J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005

C−C Bond Activation on a Rh/Ga Bimetallic Complex
A R T I C L E S
Figure 6. Calculated minima and transition states for the fluctional processes of 3a.
6. Complexes 13 and 14 both represent isomers of 3a, in which
no strong Ga-Cp* bond exists, but rather a direct Rh-GaMe2
bond is formed. Both are energetic minima that are geometrically
very similar to each other and basically differ in the exact
position of the GaMe2 group as well as the position of one of
the two Cp* rings. Species 14 represents a C2V symmetric
molecule; i.e. the rhodium, the gallium and the two carbons
bound to the gallium are exactly coplanar, constructing a mirror
plane for the two Cp* rings. Thus, it can be considered as the
key intermediate for fluctuation B. In 13 this C2V symmetry is
broken by a slight movement of the GaMe2 fragment toward
one Cp* ring [Ga-C1 ) 2.63 Å (7), 3.03 Å (8)]. 13 is lower
in energy than 14 by 1.4 kcal/mol, which itself is more stable
in energy than 3a by 1.8 kcal/mol.
No direct pathway from 3a to 14 was found, instead two
transition states could be localized between 3a and 13 (TS4)
and between 13 and 14 (TS5), respectively. TS4 is energetically
extremely low (0.6 kcal/mol above 3a). In contrast, the transition
state for the reaction of 13 to 14 (TS5) is calculated to be
astonishingly high, being 7.9 kcal/mol above 13, although 13
and 14 are structurally very similar. However, this is well
consistent with the fact that fluctuation A is spectroscopically
observed, while fluctuation B is not.
a favorable unimolecular reaction pathway. However, the solid-
state reaction does not exclude competing bimolecular mech-
anisms working in solution! Indeed, a simple cross-mixing H
1
NMR experiment proves the existence of such a mechanism of
exchanging CH3 groups between two complexes. Thus, a
solution of the selectively deuterated complex [RhCp*(GaCp*)-
(CD3)2] (1cD6) and the nondeuterated phenyl-marked complex
[Rh(C5Me4Ph)(GaCp*)Me2] (9) (1:1 ratio) in C6D6 was heated
to 60 °C for 45 min. As expected, the product mixture consists
of the three compounds 2, 11a, and 11b. As evident from the
integral ratios of the three GaMe3 signals at 0.17 (2), 0.20 (11a),
and 0.22 ppm (11b), a statistic distribution of the CD3 groups
to all three products occurs. The same result is obtained if two
solutions of the respective intermediates analogous to 3a are
prepared independently and subsequently mixed and heated to
induce the activation reaction. This means that either methyl
groups are exchanged between two gallium centers of 3a (e.g.
via Ga-CH3-Ga bridges) or GaMe2 groups are exchanged
between two complexes of 3a. An exchange of whole (C5Me4-
GaMe2) units should be energetically the most unfavorable, since
dissociation of the strongly bound diene moiety from the RhI
center is necessary. Furthermore, in the reactions of 9 and 10,
no traces of either 2 or 12 were observed, which should be
expected if the whole diene fragments are exchanged between
complexes. However, experimentally it is hard to finally prove
whether such a diene exchange reaction takes place at all.
Preliminary DFT calculations do not exclude the dissociation
of [GaMe2]+ ions (stabilized by pyridine) from 3a under
All experimental as well as computational results presented
so far suggest 3a as the reaction intermediate and thus pathway
A (Scheme 2) as the reaction mechanism. This is strongly based
on the fact that the solid-state reactions of 1c to 2 (or from 9 to
11a/b, respectively) can be taken as experimental evidence for
9
J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005 17075

A R T I C L E S
Cadenbach et al.
formation of the anion [Rh(Cp*)2]-, although the energy for
this process is calculated to be rather high (26.3 kcal/mol;
calculated in THF as the solvent, with the PCM method).32-35
At this point, the nature of the exchange process in solution
cannot be further elucidated; more detailed DFT calculations
are needed and are the subject of current research.
both metal centers effectively cooperating in a redox cascade
creating intermediates with exceptionally reactive metal centers
in close proximity to each other and opening a number of low
activation energy pathways. The question of how to capitalize
on this rather fascinating and complex situation in catalytic or
stoichiometric processes has still to be addressed. Admittedly,
the group-13 carbenoid and metalloid ligands of the general
type ER are still quite exotic. However, GaCp* and its congeners
may represent model cases of strong donor ligands not being
innocent spectators such as phosphanes or NHCs but taking
active part in specific reactions, thus representing an intermediate
state between ancillary ligands and cooperative metal centers.
Apart from this more traditional point of view, such complexes
may also be considered as molecular models for the surface
reactivity of intermetallic materials such as transition-metal
aluminides, gallides, and indides, being potentially relevant for
heterogeneous catalysis. Since it is known that AlCp* can
stabilize molecular clusters such as [Al38(AlCp*)12]36 or [Ni8-
(AlCp*)6],37 the mechanistic study presented herein may at least
stimulate thinking about the reactivity of such metal-rich
compounds or related nanoparticles.
Conclusions
The thermal rearrangement of [Cp*Rh(GaCp*)(CH3)2] (1c)
proceeds under very mild condition in solution as well as in
the solid state, giving the zwitterionic species [RhCp*(C5Me4)-
(GaMe3)] (2) as the final product. As an intermediate, the Rh(I)
species [RhCp*(η4-C5Me5GaMe2] (3a) could be spectroscopi-
cally identified, with 2D ROESY spectroscopy pointing to an
endo-arrangement of the GaMe2 moiety with respect to the
RhCp* fragment. DFT calculations strongly support these
experimental results, indicating that 3a indeed represents an
energetic minimum at reasonable energy. The C-C activation
step from 3a giving 2 is calculated to have a very low activation
barrier (17.4 kcal/mol), which is well-reflected by the mild
experimental conditions for the overall reaction. In contrast,
alternative mechanisms involving the exo-isomer 3b were
calculated to have considerably higher activation barriers. The
crucial reaction intermediate 3a is shown to undergo two
fluxional processes: (i) migration of the GaMe2 group from
one Cp* ring carbon atom to an adjacent one (fluctuation A)
and (ii) the migration of the GaMe2 group from one Cp* ligand
to the second one (fluctuation B). Both processes could be
calculated to proceed via the same intermediates. In solution,
both intramolecular as well as intermolecular processes seem
to be activated as evidenced from labeling studies and cross-
mixing experiments. The intermolecular mechanisms could be
neither experimentally proven nor computationally understood,
which is basically a result of their expected complexity.
What is finally achieved by this mechanistic study? As stated
above, the major part of the transition-metal chemistry of low-
valent main-group metal ligands has been focused on structural
and theoretical issues. Aspects such as the abilities of carbenoid
metalloid ligands to tune the reactivity of transition-metal centers
have not been addressed before in detail. The title reaction
provided us with a model for a detailed study of the cooperative
effects originating from the proximity or coordination of a
potentially very electrophilic gallium center to a very nucleo-
philic d8 rhodium(I) center (as in 3a) or to a more electrophilic
d6 rhodium(III) center (as in 1c) and its consequences for typical
organometallic reactions: migrations of alkyl groups to coor-
dinated electrophilic ligands and C-C bond splitting. The
carbenoid character as well as the electrophilicity of the GaCp*
ligand in the starting complex provokes a reduction of the
Rh(III) to a reactive Rh(I) center in the course of the alkyl
migration, at the same time creating a very Lewis acidic [(η1-
C5Me5)Ga(Me)2] moiety. This oxidized Ga(III) center represents
the actual reaction site for a C-C bond splitting process with
a surprisingly small activation barrier, the electrons for this
process evidently coming from the nucleophilic Rh(I) center.
Altogether, the course of the reaction is clearly determined by
Experimental Section
General Remarks. All manipulations were carried out in an
atmosphere of purified argon using standard Schlenk and glovebox
techniques. Hexane, toluene, THF, and Et2O were dried using an
mBraun solvent purification system, all other solvents were dried by
distillation over standard drying agents. The final H2O content in all
solvents used was checked by Karl Fischer titration and did not exceed
5 ppm. [(RhCp*Cl2)2],38 GaCp*,39 C5Me4PhH,40 Ga(C5Me4Ph),41 Zn-
(CD3)2,42 and [Cp*Rh(DMSO)(CH3)2] (1a)23 were prepared according
to literature methods. Elemental analyses were performed by the
Microanalytical Laboratory of the University of Essen. NMR spectra
were recorded on a Bruker Avance DPX-250 spectrometer (1H, 250.1
MHz; 13C, 62.9 MHz) at 298 K unless otherwise stated. Chemical shifts
are given relative to TMS and were referenced to the solvent resonances
as internal standards.
The crystal structures of 6 and 13a were measured on a Oxford
Excalibur 2 diffractometer using Mo KR radiation (λ ) 0.710 73 Å).
The structures were solved by direct methods using SHELXS-97 and
refined against F2 on all data by full-matrix least-squares with SHELXL-
97 (SHELX-97 program package, Sheldrick, Universita¨t Go¨ttingen
1997).
CCDC-288415 (6) and CCDC-288416 (13a) contain the supple-
mentary crystallographic data for this paper. These data can be obtained
free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from
the Cambridge Crystallographic Data Center, 12 Union Road, Cam-
bridge CB21EZ, UK; fax: (+44)1223-336-033; deposit@ccdc.cam.uk).
Computational Details. All calculations were performed with the
Gaussian98 (rev. A11) program package.43 DFT calculations were
carried out using the hybrid exchange-correlation functional B3LYP24,25
together with the Los Alamos National Laboratory double-ú LanL2DZ
basis set.26 All structures were fully optimized without symmetry
(36) Vollet, J.; Hartig, J. R.; Schno¨ckel, H. Angew. Chem. 2004, 116, 3248.
(37) Steinke, T.; Gemel, C.; Fischer, R. A. Angew. Chem. 2005, 117, to be
submitted.
(38) Wang, J. W.; Moseley, K.; Maitlis, P. M. J. Am. Chem. Soc. 1969, 91,
5970.
(39) Jutzi, P.; Schebaum, L. O. J. Organomet. Chem. 2002, 654, 176.
(40) Bjoergvinsson, M.; Halldorsson, S.; Arnason, I.; Magull, J.; Fenske, D. J.
Organomet. Chem. 1997, 544, 207.
(32) Barone, V.; Scalmani, G.; Rega, N.; Cossi, M. J. Chem. Phys. 2001, 114,
(41) Buchin, B.; Steinke, T.; Gemel, C.; Cadenbach, T.; Fischer, R. A. Z. Anorg.
Allg. Chem. 2005, in press.
5691.
(33) Barone, V.; Rega, N.; Scalmani, G.; Cossi, M. J. Chem. Phys. 2002, 117,
43.
(42) Petrier, C.; Souza Barbosa, J. C. d.; Dupuy, C.; Luche, J. L. J. Org. Chem.
1985, 50, 5761.
(34) Tomasi, J.; Cance´s, E.; Mennucci, B. J. Phys. Chem. B 1997, 101, 10506.
(35) Tomasi, J.; Mennucci, B.; Cance´s, E. J. Chem. Phys. 1997, 107, 3032.
(43) Frisch, M. J.; et al. Gaussian 98, Revision A.11; Gaussian, Inc.: Pittsburgh,
PA, 2001.
9
17076 J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005

C−C Bond Activation on a Rh/Ga Bimetallic Complex
A R T I C L E S
[RhCp*(pyridine)(CD3)2] (1bD6). A suspension of CD3I (1.450 g,
10 mmol), ZnBr2 (1.115 g, 5 mmol), and Li wire (0.140 mg, 20 mmol)
in a mixture of toluene (20 mL) and THF (3 mL) was sonicated at 0
°C for 5 h. The black suspension was filtered at -80 °C to a freshly
prepared solution of [RhCp*(pyridine)Cl2] (1.553 g, 4 mmol) in THF.
The mixture was warmed to room temperature and stirred for 30 min,
whereupon an orange solution formed. After hydrolysis (3.6 mL of
H2O, 0.2 mol) all volatiles were removed in vacuo, and the residue
was extracted with toluene (3 × 10 mL doped with pyridine) at 0 °C.
After removal of the solvent in vacuo, the residue was then extracted
with n-hexane (3 × 15 mL) at 0 °C. The solvent volume was reduced
to ca. 10 mL and cooled at -40 °C for 16 h to give the product as
orange crystals. The crystals were isolated by means of canullation,
washed with a small amount of cold n-hexane, and dried in vacuo.
Yield: 1.11 g of orange crystals (80%). 1H NMR δH (C6D6, 250 MHz,
25 °C): 8.19 (d, 2H), 6.68 (t, 1H), 6.27 (t, 2H), 1.56 (s, 15H); 13C
NMR δC (298 K, 62.9 MHz, C6D6): 154.3 (s), 134.6 (s), 124.1 (s),
92.1 (d, JRhC ) 4.7 Hz), 8.8 (s), 2.6 (br). Anal. Calcd for (C17H20D6-
NRh): C, 57.79; H, 9.13; N, 3.96. Found: C, 57.54; H, 8.10; N, 3.88.
[RhCp*(GaCp*)(CD3)2] (1cD6). 1bD6 (0.300 g, 0.864 mmol) was
dissolved in toluene (7 mL) and GaCp* (0.212 g, 1.037 mmol) was
added at 0 °C. Immediately, the color changed from orange to red.
After stirring the solution for a further 10 min at 0 °C, the solvent was
removed in vacuo and the residue fast-extracted with n-hexane (3 × 7
mL) at 0 °C. The solvent volume was reduced to ca. 5 mL and cooled
to -40 °C for 16 h to give the product as light red crystals. The crystals
were isolated by means of canullation, washed with a small amount of
cold n-hexane, and dried in vacuo. Yield: 0.270 g of light red crystals
(66%). 1H NMR δH (173 K, 250.1 MHz, THF-d8): 2.07 (s, 15H), 1.70
(s, 15H). 13C NMR δC (173 K, 62.9 MHz, THF-d8): 114.1 (s), 95.1
(d, JRhC ) 4.3 Hz), 10.5 (s), 9.97 (s), -14.7 (br). Anal. Calcd for
(C22H30D6GaRh): C, 55.14; H, 8.83. Found: C, 55.62; H, 7.92.
[RhCp*(C5Me4GaMe3)] (2). Method A. [Cp*Rh(DMSO)(CH3)2]
(1a) (0.150 g, 0.433 mmol) was dissolved in toluene (8 mL), and GaCp*
(0.102 g, 0.498 mmol) was added. The solution was warmed to 60 °C
and stirred for 1 h, whereupon a white precipitate was formed. After
removal of all volatiles in vacuo, the residue was washed with hexane
(3 × 5 mL) and dried in vacuo. The precipitate was redissolved in
toluene (ca. 4 mL) and crystallized by slow evaporation of the solvent.
Table 2. Basis Set Comparison for the Calculation of Two Key
Steps of the Reaction Sequence
6-311G**(C, H, Ga)/
Stuttgart RSC 1997
LanL2DZ
ECP (Rh)
3a
0
0
3b
+0.5
+18.9
+57.9
+21.7
-7.7
+1.0
+21.3
+64.7
+29.8
-7.7
TS3A
TS3C
5
2
constraints. Vibrational frequencies were calculated for all stationary
points to ensure that local minima were located and to confirm that
transition states had only one imaginary frequency. These vibrations
were visually inspected using the MOLDEN program.44 Transition states
were located by means of the QST2 or QST3 approach, with an initial
guess for the transition state gained from manipulation of the geometries
of either the educts or the products, also using the MOLDEN software.
To verify that the overall qualitative mechanistic picture is not
affected by the choice of the comparably small LanL2DZ basis set,
single point calculations for two key steps (3a/TS1A/2 and 3a/TS3C/
5) were performed using a significantly larger basis set (6-311G** 45,46
for C, H, Ga and Stuttgart RSC 1997 ECP47 for Rh). The results (Table
2) indicate that the qualitative trends in energies and activation barriers
are reasonably well reproduced by the LanL2DZ basis set.
Syntheses. [RhCp*(pyridine)Me2] (1b). [{Cp*RhCl2}2] (1.03 g,
1.66 mmol) in 15 mL of THF was treated with pyridine (2.7 mL, 33.4
mmol) at room temperature. After stirring for 30 min, ZnMe2 (2.4 mL
of a 2.0 M solution in toluene, Aldrich; 4.80 mmol) was added at -80
°C to the orange suspension. The mixture was warmed to room
temperature and stirred for 30 min, whereupon an orange solution
formed. After hydrolysis by addition of H2O (0.35 mL, 19.2 mmol),
all volatiles were removed in vacuo, and the residue was extracted with
toluene (3 × 5 mL doped with pyridine) at 0 °C. After removal of the
solvent in vacuo, the residue was then extracted with n-hexane (3 × 5
mL) at 0 °C. The solvent volume was reduced to ca. 8 mL and cooled
to -40 °C for 16 h to give the product as pale orange crystals. The
crystals were isolated by means of canullation, washed with a small
amount of cold n-hexane, and dried in vacuo. Yield: 0.89 g of orange
crystals (77%). 1H NMR δH (298 K, 250.1 MHz, C6D6): 8.21 (d, 2H),
6.69 (t, 1H), 6.29 (t, 2H), 1.55 (s, 15H), 0.75 (d, 6H, 3JRhH ) 2.5 Hz);
13C NMR δC (298 K, 62.9 MHz, C6D6): 154.5 (s), 134.7 (s), 124.3
(s), 92.3 (d, JRhC ) 4.8 Hz), 9.0 (s), 3.6 (d, JRhC ) 31.9 Hz). Anal.
Calcd for (C17H26NRh): C, 58.79; H, 7.55; N, 4.03. Found: C, 58.01;
H, 7.46; N, 4.0.
1
Yield: 0.141 g of white crystals (69%). H NMR δH (298 K, 250.1
MHz, C6D6): 1.77 (s, 6H), 1.34 (s, 15H), 1.06 (s, 6H), 0.15 (s, 9H);
13C NMR δC (298 K, 62.9 MHz, C6D6): 117.8 (br), 106.8 (d, JRhC
)
8.5 Hz), 98.5 (d, JRhC ) 7.9 Hz), 96.5 (d, JRhC ) 7.3 Hz), 12.2 (s), 8.9
(s), 8.5 (s), 1.6 (s), -2.7 (s). Anal. Calcd (C22H36GaRh): C, 55.85; H,
7.67. Found: C, 55.14; H, 7.29.
Method B (Solid State). A crystalline sample of 1c was heated to
60 °C for 2-3 min, whereupon a color change from red to pale brown
was observed. The thus obtained product was analyzed by means of
solution 1H and 13C NMR as well as 13C MAS NMR. All spectroscopic
data were identical with those of the product obtained by method A.
[RhCp*(C5Me5GaMe2)] (3a). [Cp*Rh(GaCp*)(CH3)2] (0.043 g,
0.0906 mmol) was dissolved in pyridine-d5 (1 mL) and stirred for 45
min at 25 °C. The thus obtained solution of 3a was directly used for
analysis without further workup. 1H NMR δH (THF-d8, 250 MHz, -100
°C): 1.77 (s, 15H), 1.76 (s, 6H), 0.50 (s, 6H), -0.04 (s, 3H), -0.16
[RhCp*(GaCp*)Me2] (1c). (RhCp*(pyridine)(CH3)2] (0.282 g, 0.812
mmol) was dissolved in toluene (8 mL), and GaCp* (0.183 g, 0.893
mmol) was added at 0 °C. Immediately, the color changed from orange
to red. After stirring the solution for a further 10 min at 0 °C, the solvent
was removed in vacuo and the residue extracted quickly with n-hexane
(3 × 5 mL) at 0 °C. The solvent volume was reduced to ca. 5 mL and
cooled to -40 °C for 16 h to give the product as light red crystals.
The crystals were isolated by means of canullation, washed with a small
amount of cold n-hexane, and dried in vacuo. Yield: 0.197 g of light
(s, 6H); 13C NMR δC (THF-d8, 62.9 MHz, -100 °C): 90.9 (d, JRhC
)
1
red crystals (51%). H NMR δH (298 K, 250.1 MHz, C6D6): 1.89 (s,
5.48 Hz), 83.8 (d, JRhC ) 9.10 Hz), 62.3 (d, JRhC ) 13.17 Hz), 56.9 (d,
JRhC ) 6.85 Hz), 29.3 (d, JRhC ) 4.53 Hz), 11.5 (s), 9.7 (s), 8.3 (s),
-3.9 (s).
[RhCp*(C5Me5H)] (6). [Cp*Rh(GaCp*)(CH3)2] (0.700 g, 1.479
mmol) was dissolved in pyridine (10 mL) and stirred at 25 °C. After
45 min, acetic acid was added (0.266 g 4.437 mmol) and the reaction
mixture was stirred for a further 5 min, whereupon a yellow solution
and a white precipitate was formed. After removal of all volatiles in
vacuo, the residue was extracted with hexane (3 × 5 mL). The solvent
was removed in vacuo and the yellow residue sublimed at 50 °C in
3
15H), 1.77 (s, 15H), 0.50 (d, 6H, JRhH ) 2.7 Hz); 13C NMR δC (298
K, 62.9 MHz, C6D6): 113.8 (s), 95.1 (d, JRhC ) 4.5 Hz), 10.4 (s), 9.7
(s), -14.3 (d, JRhC ) 26.8 Hz). Anal. Calcd for (C22H36GaRh): C, 55.85;
H, 7.67. Found: C, 55.29; H, 7.58.
(44) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Des. 2000, 14, 123-
134.
(45) Pople, J. A. J. Chem. Phys. 1980, 72, 650.
(46) Curtiss, L. A.; McGrath, M. P.; Blaudeau, J.-P.; Davis, N. E.; Binning, R.
C., Jr.; Radom, L. J. Chem. Phys. 1995, 103, 6104.
(47) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys. 1987, 86, 866.
9
J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005 17077

A R T I C L E S
Cadenbach et al.
vacuo. Yield: 0.305 g (55%). 1H NMR δH (298 K, 250.1 MHz, C6D6):
2.53 (q, 1H, JCH ) 6.08 Hz), 1.75 (s, 6H), 1.74 (s, 15H), 1.21 (s, 6H),
0.56 (d, 3H, JCH ) 6.06 Hz); 13C NMR δC (298 K, 62.9 MHz, C6D6):
precipitate was formed. After removal of all volatiles in vacuo, the
residue was washed with hexane (3 × 7 mL) and dried in vacuo. The
residue was redissolved in THF and precipitated by slow diffusion of
Et2O to give the product as colorless crystals. Yield: 0.392 g (75%).
1H NMR δH (C6D6, 250 MHz, 25 °C) [Rh(C5Me4Ph)(C5Me4GaMe3)],
11a: 7.30 (m, 5H) 1.79 (s, 6H), 1.49 (s, 6H), 1.43 (s, 6H), 1.02 (s,
6H), 0.20 (s, 9H). 1H NMR δH (C6D6, 250 MHz, 25 °C) [RhCp*-
(C5Me3PhGaMe3)], 11b: 7.30 (m, 5H) 1.98 (s, 3H), 1.93 (s, 3H), 1.93
(s, 3H), 1.35 (s, 15H), 0.22 (s, 9H); 13C NMR δC (298 K, 62.9 MHz,
95.1 (d, JRhC ) 5.64 Hz), 88.4 (d, JRhC ) 9.58 Hz), 61.5 (d, JRhC
)
5.34 Hz), 59.8 (d, JRhC ) 13.06 Hz), 24.2 (d, JRhC ) 3.47 Hz), 15.1
(s),13.5 (s), 12.5 (s), 4.1 (s). Anal. Calcd for (C20H31Rh): C, 64.17; H,
8.35. Found: C, 65.15; H, 8.15.
[Rh(C5Me4Ph)Cl2]2 (7). A mixture of RhCl3‚3H2O (5. g, 0.019 mol)
and C5Me4HPh (3.76 g, 0.019 mol) in 40 mL of methanol was refluxed
under nitrogen with stirring for 48 h. After the reaction mixture was
cooled, the shiny red crystals were isolated by means of canullation,
washed with a small amount of methanol, and dried in vacuo. Yield:
5.01 g (71%). 1H NMR δH (THF-d8, 250 MHz, 25 °C): 7.73 (m, 4H),
7.35 (m, 6H), 1.66 (s, 12H), 1.63 (s, 12H); 13C NMR δC (C6D6, 62.9
MHz, 25 °C): 131.5 (s), 130.2 (s), 129.3 (s), 129.1 (s), 100.9 (d, JRhC
) 8.4 Hz), 93.8, (d, JRhC ) 7.5 Hz), 90.4 (d, JRhC ) 9.7 Hz), 10.7 (s),
9.6 (s). Anal. Calcd for (C30H34Cl4Rh2): C, 48.55, H, 4.62. Found: C,
48.40; H, 4.26.
C6D6) 130.8 (s), 130.6 (s), 107.3 (d, JRhC ) 8.3 Hz), 99.7 (d, JRhC
)
7.4 Hz), 99.1 (d, JRhC ) 7.6 Hz), 98.3 (d, JRhC ) 7.2 Hz), 97.3 (d, JRhC
) 7.3 Hz), 96.0 (d, JRhC ) 7.1 Hz), 12.3 (s), 9.8 (s), 8.9 (s), 8.4 (s),
-3.0 (s). Anal. Calcd for (C27H38GaRh): C, 60.59; H, 7.16. Found:
C, 59.31; H, 6.74.
Method B. Compound 1b (0.282 g, 0.812 mmol) was dissolved in
toluene (8 mL) and [Ga(C5Me4Ph)] (0.325 g, 1.218 mmol) was added.
The solution was warmed to 60 °C and stirred for 1 h, whereupon a
white precipitate was formed. After removal of all volatiles in vacuo,
the residue was washed with hexane (3 × 5 mL) and dried in vacuo.
Yield: 0.304 g (70%). All spectroscopic data were identical with those
of the product obtained by method A.
[Rh(C5Me4Ph)(pyridine)Me2] (8). Compound 9 (1.0 g, 1.35 mmol)
in 15 mL of THF was treated with pyridine (2.7 mL, 33.4 mmol) at
room temperature. After stirring for 30 min, ZnMe2 (1.96 mL of a 2.0
M solution in toluene, Aldrich; 3.92 mmol) was added at -80 °C to
the orange suspension. The mixture was warmed to room temperature
and stirred for 30 min, whereupon a pale orange solution formed. After
hydrolysis (0.35 mL H2O, 19.6 mmol) all volatiles were removed in
vacuo, and the residue was extracted with toluene (3 × 8 mL) at 0 °C.
After removal of the solvent in vacuo, the residue was then extracted
with n-hexane (3 × 10 mL) at 0 °C. The solvent volume was reduced
to ca. 7 mL and cooled to -40 °C for 16 h to give the product as
yellow crystals. The crystals were isolated by means of canullation,
washed with a small amount of cold n-hexane, and dried in vacuo.
Method C (Solid State). A crystalline sample of 9 was heated to
60 °C for 2-3 min, whereupon a color change from red to pale brown
was observed. The thus obtained product was analyzed by means of
solution 1H and 13C NMR as well as 13C MAS NMR. All spectroscopic
data were identical with those of the product obtained by methods A
and B.
[Rh(C5Me4Ph)(C5Me3PhGaMe3)] (12). Compound 8 (0.300 g,
0.733 mmol) was dissolved in toluene (8 mL), and [Ga(C5Me4Ph)]
(0.254 g, 0.953 mmol) was added. The solution was warmed to 60 °C
and stirred for 45 min, whereupon a pale brown precipitate formed.
After removal of all volatiles in vacuo, the residue was washed with
1
Yield: 0.88 g (80%). H NMR δH (C6D6, 250 MHz, 25 °C): 8.18 (d,
2H), 7.18 (m, 5H), 6.59 (t, 1H), 6.18 (t, 2H), 1.67 (s, 15H), 1.59 (s,
15H), 0.88 (d, 6H, 3JRhH ) 2.5 Hz); 13C NMR δC (C6D6, 62.9 MHz, 25
°C): 154.3 (s), 136.6 (s), 134.7 (s), 130.6 (s), 128.2 (s), 125.8 (s),
124.1 (s), 98.8 (d, JRhC ) 3.8 Hz), 98.5 (d, JRhC ) 4.1 Hz), 88.9 (d,
JRhC ) 5.5 Hz), 9.9 (s), 8.7 (s), 4.6 (d, JRhC ) 31.5 Hz). Anal. Calcd
for (C22H28NRh): C, 64.55; H, 6.89; N, 3.42. Found: C, 64.28; H,
6.79; N, 3.36.
1
hexane (3 × 8 mL) and dried in vacuo. Yield: 0.333 g (76%). H
NMR δH (298 K, 250.1 MHz, C6D6): 7.10 (m, 10H), 2.04 (s, 3H),
1.85 (s, 3H), 1.63 (s, 3H), 1.61 (s, 3H), 1.46 (s, 6H), 1.26 (s, 3H), 0.21
(s, 9H); 13C NMR δC (C6D6, 62.9 MHz, 25 °C): 131.4 (s), 130.9 (s),
130.7 (s), 130.6 (s), 128.8 (s), 128.5 (s), 128.0 (s), 119.7 (d, JRhC
)
6.7 Hz), 107.9 (d, JRhC ) 8.2 Hz), 107.5 (d, JRhC ) 8.1 Hz), 107.3 (d,
JRhC ) 8.0 Hz), 104.6 (d, JRhC ) 7.4 Hz), 99.2 (d, JRhC ) 7.5 Hz), 99.1
(d, JRhC ) 7.4 Hz), 98.1 (d, JRhC ) 7.3 Hz), 97.4 (d, JRhC ) 7.1 Hz),
96.4 (d, JRhC ) 7.2 Hz), 13.9 (s), 12.6 (s), 10.4 (s), 10.1 (s), 9.94 (s),
9.90 (s), 9.2 (s), -2.5 (s). Anal. Calcd for (C32H40GaRh): C, 64.35;
H, 6.75. Found: C, 63.57; H, 7.23.
[Rh(C5Me4Ph)(GaCp*)Me2] (9). Compound 8 (0.5 g, 0.934 mmol)
was dissolved in toluene (8 mL) and GaCp* (0.305 g, 1.50 mmol) was
added at 25 °C. Immediately, the color changed from yellow to red.
After stirring the solution for further 3 min at 25 °C, the solvent was
removed in vacuo at 0 °C and the residue fast-extracted with n-hexane
(3 × 7 mL) at 0 °C. The solvent volume was reduced to ca. 5 mL and
cooled to -40 °C for 16 h to give the product as red crystals. The
crystals were isolated by means of canullation, washed with a small
amount of cold n-hexane, and dried in vacuo. Yield: 0.464 g (71%).
1H NMR δH (C6D6, 250 MHz, 25 °C): 7.51 (m, 2H), 7.17 (m, 3H),
Acknowledgment. We thank Gregor Barchan and Martin
Gartmann for performing the 2D NMR experiments as well as
Jochen Hauswald for his help in recording 13C-MAS and VT-
NMR spectra.
3
1.91 (s, 6H), 1.82 (s, 15H), 1.61 (s, 6H), 0.61 (d, 6H, JRhH ) 2.63
Hz); 13C NMR δC (THF-d8, 62.9 MHz, -100 °C): 137.6 (s), 132.8
(s), 128.4 (s), 127.1 (s), 113.9 (s), 102.1 (d, JRhC ) 2.6 Hz), 99.6 (d,
JRhC ) 1.8 Hz), 92.4 (d, JRhC ) 4.6 Hz), 12.1 (s), 9.8 (s), 8.1 (s), -13.4
(d, JRhC ) 26.8 Hz). Anal. Calcd for (C27H38GaRh): C, 60.59; H, 7.16.
Found: C, 58.96; H, 6.45.
Supporting Information Available: Copies of NMR spectra
(solution and solid state), crystal structure of 11a, Eyring plot
for fluctuation A, and comparison of computed and measured
bond lengths and angles of 1c and 2, as well as Cartesian
coordinates and energies of all structures (minima and TS)
presented in this paper (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
[Rh(C5Me4Ph)(C5Me4GaMe3)] (11a) and [RhCp*(C5Me3PhGaMe3)]
(11b). Method A. Compound 8 (0.400 g, 0.977 mmol) was dissolved
in toluene (8 mL), and GaCp* (0.260 g, 1.270 mmol) was added. The
solution was warmed to 60 °C and stirred for 1 h, whereupon a white
JA055298D
9
17078 J. AM. CHEM. SOC. VOL. 127, NO. 48, 2005