Carbon monoxide activation via O-bound CO using decamethylscandocinium- hydridoborate ion pairs

Article abstract of DOI:10.1021/ja300591v

Ion pairs [Cp*₂Sc]⁺[HB(p-C₆F ₄R)₃]^- (R = F, 1-F; R = H, 1-H) were prepared and shown to be unreactive toward D₂ and α-olefins, leading to the conclusion that no back-transfer of hydride from boron to scandium occurs. Nevertheless, reaction with CO is observed to yield two products, both ion pairs of the [Cp*₂Sc]⁺ cation with formylborate (2-R) and borataepoxide (3-R) counteranions. DFT calculations show that these products arise from the carbonyl adduct of the [Cp*₂Sc]⁺ in which the CO is bonded to scandium through the oxygen atom, not the carbon atom. The formylborate 2-R is formed in a two-step process initiated by an abstraction of the hydride by the carbon end of an O-bound CO, which forms an η²-formyl intermediate that adds, in a second step, the borane at the carbon. The borataepoxide 3-R is suggested to result from an isomerization of 2-R. This unprecedented reaction represents a new way to activate CO via a reaction channel emanating from the ephemeral isocarbonyl isomer of the CO adduct.

Full text of DOI:10.1021/ja300591v

Article
pubs.acs.org/JACS
Carbon Monoxide Activation via O-Bound CO Using
Decamethylscandocinium−Hydridoborate Ion Pairs
Andreas Berkefeld,† Warren E. Piers,*,† Masood Parvez,† Ludovic Castro,§ Laurent Maron,*,§
and Odile Eisenstein*,¶
†Department of Chemistry, University of Calgary, 2500 University Drive NW, Calgary, Alberta T2N 1N4, Canada
§LPCNO, Universite
́
de Toulouse, INSA, UPS, LPCNO, 135 avenue de Rangueil, F- 31077 Toulouse, France, and CNRS, LPCNO,
F-31077 Toulouse, France
¶Institut Charles Gerhardt, Universite
́
Montpellier 2, CNRS 5253, cc 1501, Place E. Bataillon, F-34095 Montpellier, France
S
* Supporting Information
ABSTRACT: Ion pairs [Cp*2Sc]+[HB(p-C6F4R)3]− (R = F, 1-F; R = H,
1-H) were prepared and shown to be unreactive toward D2 and α-olefins,
leading to the conclusion that no back-transfer of hydride from boron to
scandium occurs. Nevertheless, reaction with CO is observed to yield two
products, both ion pairs of the [Cp*2Sc]+ cation with formylborate (2-R)
and borataepoxide (3-R) counteranions. DFT calculations show that these
products arise from the carbonyl adduct of the [Cp*2Sc]+ in which the CO
is bonded to scandium through the oxygen atom, not the carbon atom. The formylborate 2-R is formed in a two-step process
initiated by an abstraction of the hydride by the carbon end of an O-bound CO, which forms an η2-formyl intermediate that adds,
in a second step, the borane at the carbon. The borataepoxide 3-R is suggested to result from an isomerization of 2-R. This
unprecedented reaction represents a new way to activate CO via a reaction channel emanating from the ephemeral isocarbonyl
isomer of the CO adduct.
A third (and much more rare) mode of coordination for the
CO ligand occurs when the lone pair on the oxygen atom
engages in bonding to the metal, the O-bound isomer (Figure
1c).
1. INTRODUCTION
Carbon monoxide is one of the most ubiquitous ligands in
transition metal chemistry. Classical carbonyl complexes
possess d electron configurations of n ≥ 2 since π back-
donation from filled metal d orbitals to the empty π
antibonding (π*) orbitals of the CO ligand is a major
component of the M−CO bond (Figure 1a). Indeed, the
metal−carbonyl bond is a textbook example1 of the
fundamental concept of synergic bonding involving ligand-to-
metal σ donation and metal-to-ligand π back-donation in
transition metal chemistry. Since the π* accepting orbitals and
the σ symmetric highest occupied molecular orbital (HOMO)
of CO are more localized on the carbon atom, the vast majority
of carbonyl compounds contain C-bound CO ligands. Even
when the π* component of the M−CO bond is diminished or
eliminated, as in the so-called “non-classical” carbonyl
complexes (Figure 1b), studied in much detail in the mid
1990s,2 the carbonyl ligand binds to the metal through the
orbitals associated with the carbon atom. In such compounds,
the metal either possesses a d0 electron configuration3,4 (and
therefore cannot engage in π bonding) or is a positively
charged, highly electrophilic ion whose d electrons are too low
in energy to interact strongly with the CO π* orbitals.5,6
Classical and nonclassical C-bonded CO complexes are
distinguished by their C−O stretching frequencies in the
infrared (IR) spectrum, with the former exhibiting values below
and the latter above that of free CO (2143 cm−1).
Very few examples of authenticated monomeric O-bound
complexes exist, and those that have been observed or
postulated7 occur only in highly electropositive metal fragments
that are not capable of π-bonding, in other words, nonclassical
systems. For example, Andersen et al.8 have prepared mono
and dicarbonyl adducts of decamethylytterbocene, Cp*2Yb
(Cp* is the pentamethylcyclopentadienide anion), whose IR
spectroscopic stretching frequencies are lower than that of free
CO, contrary to expectations for this metal fragment which
cannot engage in π back-donation to the CO ligand(s). It has
been shown computationally9 that these observations are best
explained by invoking O-bound structures, since O-bound
ligands are expected to exhibit stretching frequencies below
2143 cm−1 despite the absence of π back-bonding. This system
remains, to the best of our knowledge, the only bona fide
example of an O-bound carbonyl complex that is stable under
ambient conditions.
Metal complexes of carbon monoxide are significant not only
for their role in the development of bonding models and
theories but also because they are crucial intermediates in
catalytic reactions that involve CO, an important C1 synthon.
Received: January 18, 2012
Published: June 6, 2012
© 2012 American Chemical Society
10843
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
transition metal ion, may transform it through transfer of a
reactive anion to the substrate, for example a hydride or alkyl.
Such systems may be thought of as “ionic FLPs”, in contrast to
the standard FLP systems that utilize neutral Lewis acids and
bases to activate small molecules. Herein we report the
development of an ionic FLP ensemble based on the
decamethylscandocinium cation partnered with a weakly
coordinating perfluoroarylhydrido borate anion and its use to
activate carbon monoxide. The products formed strongly
suggest that the CO is activated through its O-bound isomer,
resulting in a new type of reactivity for coordinated CO. This
proposal is corroborated by DFT computations.
2. RESULTS AND DISCUSSION
Decamethylscandocinium cations have been prepared by
Hessen et al.24 via protonolyis of the neutral Cp*2ScCH3
compound (first reported by Bercaw et al. in 198725) using
ammonium salts of tetraarylborate counteranions, [BAr4]− (Ar
= C6H5 or C6F5). The products are contact ion pairs, in which
the cationic scandium center is coordinated by meta and para
C−H or C−F groups of one of the borate aryl rings. For Ar =
C6H5, solvent-separated ion pairs are formed in the presence of
fluoroarene solvents, which serve as donors to the highly Lewis
acidic scandium center. These cations are thus highly
electrophilic, but the counteranion partners in these instances
are rather unreactive.
We were thus interested in the preparation of decamethyl-
scandocinium ion pairs in which the counteranion consisted of
potentially more reactive fluorarylhydridoborates. Although
protonolyis routes akin to those used by Hessen et al. are
feasible, the most convenient method for accomplishing this is
shown in Scheme 1. Treatment of decamethylscandocene
Figure 1. Bonding of carbon monoxide to a transition metal. (a)
Classical bonding with σ donation from C to M and π backbonding
from M to CO. (b) Nonclassical C-bound carbonyl in which the π
backbonding component is absent or negligible. (c) O-bonded, or
isocarbonyl, bonding in which the carbonyl oxygen atom is bonded to
the metal.
Migratory insertion and deinsertion of CO into metal alkyl and
metal hydrogen bonds is a primary reaction in organometallic
chemistry and is the key step in widely deployed commercial
processes such as olefin hydroformylation, the Monsanto acetic
acid process, and Pauson−Khand chemistry, to name a few.1 As
such, its mechanism has been studied in detail;10−13 it always
involves C-bound carbon monoxide complexes of the type
shown in Figure 1a,b, no exceptions. Thus, although the O-
bound isomer appears to be energetically and kinetically
accessible from the C-bound isomer in highly electropositive
systems,9 reactivity emanating from the O-bound isomer, which
would be expected to be quite different from that of the C-
bound form, has not been observed.7
Scheme 1
New reactivity patterns require new ways to activate small
molecules. In the mid 1990s, it was discovered that the strong
Lewis acid tris-(pentafluorophenyl)borane, B(C6F5)3, is capable
of catalyzing the hydrosilylation of organic carbonyl functions14
and imines15 via activation of the Si−H bond of the silane,16
rather than the Lewis basic organic function as might have
initially been predicted. Subsequently, it was discovered that
dihydrogen17,18 can similarly be activated by B(C6F5)3 in
combination with an appropriately bulky Lewis base in systems
described as “frustrated” Lewis pairs (FLPs),19 since they are
not able to form classical Lewis acid/Lewis base adducts. Such
systems turn this frustration into useful bond activations of a
fundamentally different nature than what was previously
known. In the past few years, many groups have explored the
scope of these reactions in terms of the molecules that can be
activated by the FLPs;20 however, CO appears to be reluctant
to undergo binding or activation by the typical B(C6F5)3/bulky
Lewis base combinations explored so far.21
26
chloride, Cp*2ScCl, with one equivalent of B(C6F5)3 or B(p-
27
We have become interested in extending the FLP concept to
include other strong Lewis acids, in particular cationic early
transition metal complexes22,23 partnered with a weakly
coordinating but potentially functional anion. Here, the metal
cation serves as the Lewis acid, and the counteranion is the
bulky Lewis base partner that, upon activation of a substrate
small molecule through coordination to the Lewis acidic
C6F4H)3 and a slight excess of triethylsilane in benzene
rapidly afforded the desired ion pairs as bright-yellow,
microcrystalline precipitates in excellent isolated yields
(>90%). Notably, no reaction is observed between Cp*2ScCl
and silane in the absence of borane. The triethylchlorosilane
byproduct was easily removed by washing the solid product
with hexanes and drying it under vacuum. Alternatively, a
10844
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
protonolysis route starting from the methyl derivative
Cp*2ScCH3 and employing the bulky iminium salt (2,6-di-iso-
propylphenyl)-1,2,2-trimethylpropenylideneammonium deuter-
io-tris-(pentafluorophenyl)borate cleanly produces the ion pair
d1-1-F. This was the most convenient route to the deuterated
ion pair, since the iminium reagent is readily available via FLP
activation of D2 using the free imine and B(C6F5)3.28 The imine
liberated in the protonolyis reaction does not coordinate the
scandium center and is removed by washing the product d1-1-F
with hexanes.
Cp*2ScH, and the free boranes B(p-C6F4R)3. Hydride
Cp*2ScH is a known species that is highly reactive toward
deuterium gas, D2, and α-olefins such as propene.25,29 Thus, we
treated bromobenzene solutions of 1-F with 4 atm of D2
1
2
(Scheme 2) and monitored the H and H NMR spectra for
Scheme 2
Ion pairs 1-R (R = H or F, as defined in Scheme 1) were
characterized by NMR spectroscopy and elemental analysis;
furthermore, the structure of 1-F was determined by X-ray
1
crystallography (Figure 2). In the 11B-decoupled H NMR
any indication of incorporation of deuterium into the
hydridoborate anion. None was observed over the course of
48 h at room temperature; a similar experiment utilizing d1-1-F
and H2 also gave no indication of isotope exchange between the
deuterioborate and free dihydrogen. Furthermore, when the ion
pairs were stirred in the presence of a large excess of propene,
no change in the NMR spectra was seen over the course of 12
h. These experiments show that the hydridoborate counter-
anion does not reversibly transfer hydride to the scandium
center and that the highly reactive Cp*2ScH species is not
present in chemically meaningful quantities in solutions of the
ion pairs.
Figure 2. X-ray structure of 1-F (thermal ellipsoids drawn to 50%
probability level). All hydrogens except H1 omitted for clarity.
Selected bond and nonbonded distances (Å): Sc−F1, 2.3261(14); Sc−
F2, 2.3396(16); B1−H1, 1.14(3); Sc−H1, 4.853.
spectra, in addition to the intense singlet for the Cp* ligand
protons, broad resonances at 4.13 (1-F) and 4.28 (1-H) were
detected for the B−H protons; 11B NMR spectra showed
resonances typical of the [HB(p-C6F4R)3]− anions at −25.3 and
−24.6 ppm. The room temperature 19F NMR spectra showed
broadened resonances for the ortho- and meta-F nuclei, (ν1/2
≈
145 Hz, 1-F; ν1/2 ≈ 490 Hz, 1-H), and a sharp triplet resonance
(3JF−F = 21.1 Hz) for the para-F nuclei in the case of 1-F. These
observations are consistent with κ2-F binding of the [HB(p-
C6F4R)3]− anions via the ortho and meta fluorine atoms, with
the scandocinium cation moving from one ring to another on a
time scale more rapid than that of the NMR spectroscopic
experiment at room temperature. Due to low solubility in
nonpolar solvents, this process could not be further defined via
low-temperature NMR spectroscopy. In agreement with the
solution NMR spectroscopic data, the solid-state structure of 1-
F shows that the contact between cation and anion occurs by
coordination of an ortho and meta fluorine of one of the
fluoraryl groups and not via the hydride, a situation likely
dictated by steric factors. The Sc−F1 and Sc−F2 bond
distances observed (2.3261(14) and 2.3396(16)Å, respectively)
are completely consistent with those observed by Hessen et
al.24 in the related complexes [Cp*2Sc(κ2-F-1,2-F2C6H4)]+[B-
(C6H5)4]− and [Cp*2Sc]+[(κ2-F-2,3-F2C6F4)B(C6F5)3]−
wherein the difluoroarene engages in a similar chelation of
the cationic scandium center. The nonbonded Sc−H1 distance
of 4.85(4)Å clearly shows that the hydride is not interacting
with the scandium center.
This was an important point to establish in light of the
observed reactivity between ion pairs 1-R with carbon
monoxide. Separate experiments in which Cp*2Sc-H is
generated at low temperature via hydrogenolysis of Cp*2Sc-
25
CH3 and reacted with CO gave a complex product mixture,
indicating that the putative formyl species is highly reactive.
The propensity of Cp*2Sc-H to form “tuck-in” complexes25 and
the “carbene-like” behavior of early metal formyl ligands30,31 are
likely factors in the intractable reaction chemistry of Cp*2Sc-H
with CO. It is thus significant that, even though this hydride is
not accessible from ion pairs 1-R, they react over the course of
12 h when exposed to 1 atm of CO gas to yield two distinct
products, 2-R and 3-R, whose ratio is dependent on the nature
of R as depicted in Scheme 3. By 1H NMR spectroscopy (vide
infra and Figure S1, Supporting Information [SI]), these
reactions are essentially quantitative, with only ∼5% of a
decomposition product of 1-R in evidence along with 2-R and
3-R in the final product mixture.
In solution, the compounds are sufficiently stable to be fully
1
characterized by H, 11B, 13C, and 19F NMR spectroscopy and
IR spectroscopy; the molecular structure of compound 3-F was
identified concretely by single-crystal X-ray diffraction (Figure
3). The assignments were aided through the preparation of 13C-
labeled materials prepared by performing the reaction with
isotopically enriched carbon monoxide, 13CO (99%), and
Despite the remote position of the hydride relative to the
scandium, it is conceivable that the ion pairs 1-R are in
equilibrium with the neutral decamethylscandocene hydride,
10845
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
bond coupling constant of 11.4 Hz is observed between the
formyl hydrogen and the 11B isotope of the borate boron.
Correspondingly, the 1H-coupled 11B NMR spectrum showed a
doublet coupling in the resonance at −13.3 ppm, in accord with
boron in a four-coordinate, anionic borate environment, and
distinctly different from the characteristic resonance for the
hydridoborate anion at −25.3 ppm. Also consistent with the
proposed structure is a band of moderate intensity at 1603
cm−1 in the IR spectrum of the product mixture, assignable to
the CO stretching frequency for the coordinated formyl
group on the basis of a shift to 1567 cm−1 in the 13C-labeled
isotopomer.
Scheme 3
The nature of product 3-F is less straightforwardly assigned
on the basis of its NMR spectroscopic data, although these data
are fully consistent with the structure as determined by X-ray
analysis on a small sample of single crystals obtained by
selective crystallization of 3-F from the product mixture. This
unusual species may be described as a contact ion pair between
[Cp*2Sc]+ and a borataepoxide anion derived from CO and the
hydridoborate anion of 1-F. As can be seen in the thermal
ellipsoid diagram of Figure 3, the core of the counteranion is a
three-membered ring formed from B1, C21, and O1; 13C-
2
labeling shows that C21 is derived from CO, while H-labeling
(starting from d1-1-F) shows that the proton on C21 comes
from the hydridoborate hydrogen. This methine unit gives rise
1
to resonances at 4.31 ppm in the H NMR spectrum and 60.5
1
ppm in the 13C NMR spectrum, with a large JCH coupling
constant of 161 Hz, characteristic of C−H groups in a three-
membered ring. The 11B resonance at −8.6 ppm is very broad,
but consistent with retention of a borate boron center, shifted
downfield due to the oxygen donor. As might be expected, the
19F NMR spectrum is complex, exhibiting significant broad-
ening in the room-temperature spectra. However, upon cooling
to 203 K a spectrum consistent with the molecule’s structure is
obtained, with two separate sets of resonances for the
diastereotopic C6F5 groups directly bonded to boron, and
three signals, integrating to 2:1:2 fluorines, for the C6F5 group
attached to C21.
Figure 3. X-ray structure of 3-F (thermal ellipsoids drawn to 50%
probability level). All hydrogens except H bonded to C21 are omitted
for clarity. Selected bond distances (Å) and angles (deg): Sc−F1,
2.320(4); Sc−O1, 2.187(4); B1−O1, 1.522(8); B1−C21, 1.559(9);
O1−C21, 1.478(7). B1−C21−O1, 60.1(4); C21−O1−B1, 62.6(4);
O1−B1−C21, 57.3(4).
While these products are unusual, particularly the borataep-
oxides 3-R, they have been mentioned as potential
intermediates to explain products obtained in the carbonylation
of alkali metal hydridoborate reagents such as [Li]+[HBEt3]−.
In an early study, Hubbard speculated32 that a lithium borate
with a borataepoxide structure must be present during the
carbonylation of [Li]+[HBEt3]− to account for the observation
of trimethylsilyl ethyl(1-ethylpropyl)borinate upon quenching
of the reaction with Me3SiCl. Here, the observation of a B−O
bond in the product, rather than the expected C−O bond can
be explained by the intermediacy of a borate with a B−O−C
ring, (i.e., a borataepoxide), but the species was only
through the use of d1-1-F. The spectroscopic data is similar for
both R = F and H, and only the data for compounds 2-F and 3-
F will be discussed in detail. In the case of compound 2-F, the
assignment of its structure as the contact ion pair, in which the
counteranion is the tris-(pentafluorophenyl)formylborate
shown, is based primarily on the NMR and IR spectroscopic
data obtained for the compound; the precise structural details
of the contact between the cation and anion are not known
experimentally. Although secondary Sc−F contacts are
plausible, no evidence for this exists in the 19F NMR spectrum
of the compound, which shows a trio of sharp multiplet signals
for the ortho, para, and meta fluorines of three equivalent C6F5
rings even at low temperature (203 K). Furthermore, the
computed structure (vide infra) does not indicate such
1
characterized by H-decoupled 11B NMR spectroscopy via a
signal at −2.7 ppmstrikingly similar to what we observe for
compounds 3-R. In other metal hydride-induced carbonylations
of trialkylboranes,33 formyl borates have been proposed to be
intermediates but have not been observed. Kabalka was able to
generate acyl borates through reactions of acyl lithium reagents
with trialkylboranes at −115 °C,34 but these rearrange rapidly
upon warming. Thus, compounds 2-R and 3-R are unique in
that they are stable in solution and provide confirmation of the
earlier ideas of Hubbard. However, in light of these earlier
precedents, it is significant that neither the [TMPH]+ (TMP =
2,2,6,6-tetramethylpiperidine)35 nor potassium salts of the
perfluoroarylhydridoborate anion [HB(C6F5)3]− react with
1
secondary stabilization. H and 13C NMR spectra, however,
are diagnostic for a formylborate counteranion. A downfield
signal at 11.80 ppm in the 1H NMR spectrum that correlates to
a resonance at 266.3 ppm in the 13C NMR spectrum (1JCH
=
151 Hz) is consistent with the presence of a formyl group that
is derived from the hydride of the borate counteranion in 1-F
and the added CO. This is corroborated by the absence of the
resonance at 11.80 ppm when d1-1-F is reacted with CO.
Furthermore, one-bond coupling between the formyl carbon
atom and boron-11 (1JCB = 50.5 Hz) is observed, showing that
this carbon is directly bonded to boron. Additionally, a two-
10846
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
CO in bromobenzene, indicating that the decamethylscando-
cinium cation is essential for the observed reactivity.
The mechanism by which these unusual products are formed
is of considerable interest. While the formyl borate ion pairs 2-
R could conceivably arise via “normal” CO insertion into the
scandium hydride bond of Cp*2Sc-H (via a C-bound
nonclassical CO complex) followed by abstraction of the
formyl group by borane, this is not possible on the basis of the
control experiments described above, since Cp*2Sc-H is not
present. The products therefore arise via direct reaction of CO
with ion pairs 1-R, implying that CO interacts in some way
with the electrophilic scandium center in the cation and is
activated to accept the hydride from the borohydride anion.
Since no intermediates were detected in the spectra of the
ongoing reactions, this mechanistic hypothesis was explored
using density functional theory (DFT) computations on the
fully fluorinated compound 1-F and CO.36 The picture that
emerges is that of these products arising from the O-bound
isocarbonyl adduct of the decamethylscandocinium ion as
depicted in Scheme 4; although the O-bound isomers 1-R·OC
Figure 4. Gibbs energy profile for the transformation of 1-F·CO in 2-F
and 3-F. The isomerization of 1-F·CO into 1-F·OC occurs most likely
by way of CO decoordination−coordination, and no transition state
could be located.
The calculated structure of 1-F is close to that obtained by X-
ray crystallography; in particular the Sc−F1 and Sc−F2
distances of 2.29 and 2.41 Å are close to the experimental
values of 2.3261(14) and 2.3396(16), respectively. Upon
coordination of CO to form 1-F·CO, the coordination mode
of the [HB(C6F5)3]− anion changes from κ2- as in 1-F to a κ1-
linkage in which only the ortho F1 is in the coordination sphere
of Sc with a Sc−F1 distance of 2.259 Å and an angle Sc−F1−
C(aryl) of 170°; the meta F2 is far from Sc (Figure S2 [SI]). In
1-F·CO, the Sc−C distance is 2.336 Å, and the binding energy
of C-bound CO is 6.7 kcal mol−1. The Gibbs energy of the O-
coordinated complex, 1-F·OC, is 8.5 kcal mol−1 higher than
that of the C-bound; the transition state connecting these two
isomers37 could not be identified, and thus it is presumed that
they interconvert via CO dissociation/association.
Scheme 4
A two-step reaction transforms 1-F·OC into 2-F. The O-
coordinated CO ligand abstracts a hydride from [HB(C6F5)3]−
to form an η2-formyl intermediate that is 10.6 kcal mol−1 above
the energy reference. The transition state, TSA, of the hydride
abstraction has a Gibbs energy of 19.5 kcal mol−1 and connects
1-F·OC with the formyl intermediate. The next elementary step
is the addition of the borane to the carbon of the formyl ligand.
The corresponding transition state, TSB, is 33.6 kcal mol−1
above the energy reference and connects the formyl
intermediate to observed product 2-F.
The geometries of TSA and INT are given in Figure 5. TSA
has a bent O-coordinated CO with an Sc−O bond length of
2.265 Å, an O−C bond length of 1.193 Å, and a C−O−Sc angle
of 102.6°. The H of the borohydride is 1.431 Å from C, and the
B−H−C angle is 167.9°. The hydride transfer occurs in the
plane bisecting the Cp*centroid−Sc−Cp*centroid equatorial plane.
Interestingly, the hydride that does not transfer from the
borohydride anion to Sc in 1-F does transfer to the O−C ligand
bound to Sc. In TSA, the position of the hydride relative to the
OC ligand shows that the hydride maximizes its interaction
with the π*CO orbital that is in the equatorial plane (H···C−O
angle of 109.3°) and minimizes its overlap with the carbon lone
pair that points away from the OC ligand. The transfer of the
hydride to CO occurs with a relatively low energy barrier
because the electrophilicity of this ligand is enhanced by its
coordination to the cationic, electron-deficient Sc center. In the
η2-formyl complex, INT, the Sc−O, Sc−C, and C−O bond
distances are 2.214, 2.160, and 1.247 Å, respectively. The C−
are less stable than the C-bond isomers 1-R·CO, no path that
originated from the C-bound isomers was identified. The O-
bound species, on the other hand, are close enough in energy to
be accessible and directly lead to transition states that furnish
the product 2-F. Moreover, a low-energy path from 2-F to 3-F
is also identifiable, and the computations imply that the two
compounds are in equilibrium. The Gibbs energy profile for the
reaction, with the energy of 1-F·CO as energy reference, is
depicted in Figure 4.
10847
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
metallocenium cations.40 This perhaps accounts for the lack of
secondary F→Sc interactions in this structure.
Calculations show that 3-F, which is 1.3 kcal mol−1 above 2-
F, is formed by isomerization of 2-F with a transition state,
TSC, whose Gibbs energy is −0.4 kcal mol−1 below the energy
reference and 22 kcal mol−1 above that of 2-F. The geometries
for the transition state between 2-F into 3-F, TSC, and 3-F are
shown in Figure 7. At TSC, one of the C6F5 group bridges the
Figure 5. Geometries of TSA (left) and INT (right). Sc green, C gray,
B pink, H white. See Figure 4 and Scheme 3 for labeling. Selected
geometrical parameters (distances in Å, angles in degrees): TSA: B−H
1.366, H···C 1.431, C−O 1.193, O−Sc 2.265, C−O−Sc 102.6, H···C−
O 109.3°. INT: B···H 3.241 Å, C−O 1.247, Sc−O, 2.214, Sc−C =
2.160, H−C−Sc 168.7, H−C−O 115.0.
O−Sc and O−C−Sc angles have the acute values of 71 and
75.8°, respectively. In contrast, the angles involving the
hydrogen of the formyl group shows some geometrical features
revealing of its reactivity behavior since the hydrogen is almost
“trans” to Sc (Sc−C−H angle of 168.7°) but the H−C−O
angle is 115°, indicative of an sp2-hybridized carbon. These
angles at the formyl carbon indicate that the carbon lone pair
does not point toward Sc but away from the three-membered
metallacycle ring. This is consistent with a nucleophilic
property of the formyl carbon. Similar structural features
were obtained in a computational study of the reaction of
Cp2CeH with CO. The intermediate, resulting form the
insertion of CO in the CeH bond, Cp2Ce(η2-HCO), cleaves
H2 in a heterolytic manner.38 Related geometrical features have
been obtained for the Cp′2Ce(η2-H2COMe) complex where the
two hydrogens, the C, and Ce are almost coplanar. This species,
which is an intermediate of reaction of the cerium hydride with
MeOMe, reacts at the carbon of the η2-H2COMe ligand with
the nonfluorinated Lewis acid BPh3.39
Figure 7. Geometries of TSC (left) and 3-F (right). Sc green, C gray,
B pink, H white. See Figure 4 and Scheme 3 for labeling. Selected
geometrical parameters (distances in Å, angles in degrees): TSC: Sc−
O 2.021, O−C1 1.329, B−C1 1.539, B−C2, 1.902, C1−C2 1.738,
B···O 2.584, B−C1−O 128°. 3-F: Sc−F 2.429, Sc−O 2.175, B−O
1.535, B−C 1.580, O−C 1.450.
C1−B bond, and the B−C1−O angle is 128° which is
associated with a B···O distance of 2.584 Å. Thus, in TSC, there
is not yet any interaction between B and O, and the B−O bond
is formed during the descent from the transition state. The
computed structure of 3-F agrees well with that determined by
X-ray crystallography, with the only significant discrepancy
being observed for the computed Sc−F1 distance that is 2.429
Å vs the experimental value of 2.320(4) Å; bonds that are weak
are often calculated to be too long.
The geometries of TSB and 2-F are given in Figure 6. In
TSB, the Sc−C bond is mostly broken, but the B−C bond is
While the computations suggest that 2-F and 3-F are in
equilibrium, with a barrier of 17−18 kcal mol−1 between them,
we have no conclusive experimental evidence for exchange
between the two compounds. Exchange NMR spectroscopy
experiments showed that the rate of exchange was outside of
the time scale window for this technique. The ratio of the two
compounds did not change significantly with varying temper-
ature, indicating that ΔS° for this equilibrium is close to zero.
When crystals of 3-F isolated from the product mixture were
redissolved in C6D5Br, small amounts of 2-F were present and
grew in slowly over time; however, the final ratio observed in
the forward ratio was not established, so this result too is
inconclusive.
Collectively, the computations show that, while the O-
bonded CO is not the preferred coordination between 1-F and
CO, the isocarbonyl complex 1-F·OC (which is a shallow
secondary minimum) still determines the reactivity of 1-R with
CO. Carbon monoxide has a lone pair of σ-symmetry on C and
O and two orthogonal π orbitals. In the usual C-bound
coordination of CO to a metal complex, the higher lying C-
localized σ lone pair is used for binding to the metal, and the
π*CO, which are more on C than O (see Figure 1), are used for
back-donation from the metal when the metal d orbitals are
occupied. The highest occupied and lowest unoccupied orbitals
of CO are thus used for the metal−CO interaction. However,
in the O-bonded CO complex, the C-localized lone pair is
directed away from the metal, and the coordination of the
Figure 6. Geometries of TSB (left) and 2-F (right). Sc: green, C gray,
B pink, H white. See Figure 4 and Scheme 3 for labeling. Selected
geometrical parameters (distances in Å, angles in degrees): TSB: Sc−
O 2.11, Sc−C 2.70, B···C 3.1, Sc−O−C 109.0, H−CO 109.4. 2-F: Sc−
O 2.09, O−C 1.246 Å, B−C 1.641; Sc−O−C 162.5, H−C−O 114.7,
H−C−B 118.0.
still forming, as suggested by the distances of 2.70 Å and 3.1 Å,
respectively. The Sc−O−C and H−C−O angles of ∼109.2° are
also indicative of the lack of Sc−C interaction and of a C lone
pair that points more toward B than Sc. The geometry of 2-F
shows a formylborate with no unexpected structural features, in
full agreement with that proposed on the basis of the
spectroscopic data (vide supra). The Sc−O−C angle of
162.5° and the Sc−O distance of 2.09 Å suggests substantial
π-donation from the carbonyl function to the cationic metal
center, as has been observed in other ketone adducts of
10848
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
mmol) and Et3SiH (140 mg, 1.20 mmol) was added to a solution of
Cp*2ScCl (344 mg, 0.98 mmol) in benzene (2.5 g; 20 mL screw cap
vial) by pipet at room temperature (rt). The color of the reaction
mixture changed instantaneously from dark to bright yellow. The
resulting solution is carefully layered with hexanes (∼10 mL), and the
vial sealed and kept at rt overnight. Bright yellow prismatic crystals
separated and were recovered by careful removal of the pale-yellow
mother liquor by pipet, washed with hexanes (2 mL, 3 times), and
dried under dynamic vacuum. Yield: 760 mg, 0.88 mmol, 89% 1-
F*0.5C6H6. 1H{11B} NMR (400 MHz, C6D5Br, 298 K): δ 7.20 (s, 3H,
0.5 equiv. C6H6), 4.13 (br s, ν1/2 ≈ 20 Hz, 1H, B-H), 1.58 (s, 30H,
C5(CH3)5); 19F NMR (376 MHz, C6D5Br, 298K): δ −133.9 (br s, ν1/2
≈ 145 Hz, 6F, o-F), −160.3 (t, J = 21.1 Hz, 3F, p-F), −166.8 (br s, ν1/2
∼145 Hz, 6F, m-F); 11B NMR (128 MHz, C6D5Br, 298K): δ −25.3
(br d, line width ≈ J); 13C{1H} NMR (101 MHz, C6D5Br, 298 K): δ
128.3 (C6H6), 126.5 (C5(CH3)5), 10.2 (C5(CH3)5) [13C resonances of
C6F5 rings were not detected]; IR (KBr): 2387 cm−1 (νB−H); analysis
(calcd, found for C41H34BF15Sc): C (56.77, 56.45), H (3.95, 4.02).
Yield for 1-H (0.14 mmol Sc): 110 mg (0.14 mmol), quant., thin
yellow needles. 1H{11B} NMR (400 MHz, C6D5Br, 298 K): δ 6.66 (tt,
J = 10.1, 7.5 Hz, 3H, p-C6F4H), 4.28 (br s, ν1/2 ∼20 Hz, 1H, B-H),
1.59 (s, 30H, C5(CH3)5); 19F NMR (376 MHz, C6D5Br, 298K): δ
oxygen to the metal increases the electron-withdrawing
character of the oxygen. This results in a lowering of the
energy of all orbitals, and in particular of the empty π*CO
,
which are localized on the carbon. The isocarbonyl complex
thus has a doubly occupied σ-type lone pair localized on carbon
and two empty π*CO-type orbitals with a strong carbon
contribution. This makes the O-bound CO ligand a strong
enough electrophile to engage in an interaction with a
nucleophile. The geometry of TSA highlights the Lewis acid
character of the O-bound CO since the hydride is positioned to
maximize the interaction with the π*CO orbital. While the
hydride does not transfer from the borohydride to the Sc center
probably because the distance Sc···H in 1-F is too long, the
presence of the CO ligand decreases the distance between the
Lewis base and the Lewis acid, making the hydride transfer a
possible reaction. This transfer is not very demanding in
energy, but all intermediates, 1-F·OC and INT, have a Gibbs
energy above that of 1-F·CO and also that of separated 1-F and
CO. Thus, none of the intermediates can accumulate
sufficiently to be detectable by spectroscopic means.
−134.2 (ν1/2 ∼490 Hz, 6F, o-F), −144.4 (ν1/2 ∼490 Hz, 6F, m-F); 11
B
The formation of 2-F is significantly exoergic so that the
overall reaction is favorable. The overall activation barrier of
33.6 kcal mol−1 above 1-F·CO is compatible with the
experimental conditions (12 h at rt). Calculations show that
3-F is obtained from isomerization of 2-F. The difference in
energy between the two isomers is small, and the accuracy of
the calculated difference in energy between 2-F and 3-F is not
sufficient to comment on the experimental ratio of 2-F and 3-F
and also on that of 2-H and 3-H. In all cases these energy
values are compatible with ratio of 2-R and 3-R that is small.
NMR (128 MHz, C6D5Br, 298K): δ −24.6 (br s, ν1/2 ∼120 Hz);
1H,13C-HSQC, HMBC spectroscopy (400 MHz, C6D5Br, 298K): δ
148.5 and 146.2 (Caryl-F), 125.9 (C5(CH3)5), 100.9 (Caryl−H), 10.1
(C5(CH3)5) [13C NMR resonance of B-bound ipso-C was not
detected]; IR (KBr): 2410 cm−1 (νB−H); analysis (calcd., found for
C38H34BF12Sc): C (58.93, 58.84), H (4.43, 4.45).
4.1.2. Protonolysis Route (d1-1-F). A 100 mL round-bottom flask
equipped with a magnetic stir bar was charged with Cp*2ScCH3 (111
mg, 0.34 mmol) and (2,6-di-iso-propylphenyl)-1,2,2-trimethylprope-
nylideneammonium deuterio-tris-(pentafluorophenyl)borate (260 mg,
0.34 mmol), and was attached to a swivel-frit assembly. The apparatus
was removed from the glovebox, evacuated on the vacuum line, and
the solids cooled to −78 °C (acetone/dry ice) for 20 min, and toluene
(20 mL) was finally condensed onto the solids. The cold bath was
removed 10 min after solvent transfer and the reaction mixture
allowed to slowly warm to rt, yielding an dark yellow solution. Toluene
was removed after 30 min at rt under dynamic vacuum and pentane
was condensed onto the semisolid residue at −78 °C. After warming
to rt in a water bath, the mixture was vigorously triturated until a
yellow powder separated (∼60 min). The latter solid was isolated by
filtration, washed with pentane (3 times) cycled within the swivel-frit
to completely remove the 1,1,1-trimethyl-2-(2,6-diisopropylphenyl)-
iminopropane byproduct, and finally dried under dynamic vacuum for
3 h. In the glovebox, the solid product was recrystallized from
benzene/hexanes at −35 °C, yielding bright-yellow prismatic crystals.
The mother liquor was removed by pipet, and the crystals were
washed with hexanes (2 mL, 3 times) and dried under dynamic
vacuum (30 min). Yield: 200 mg (0.23 mmol), 68% d1-1-F*0.5C6H6.
1H NMR data were found to match those of 1-F, except for the B−H
resonance. Analysis (calcd, found for C41H33DBF15Sc): C (56.70,
55.40), H (4.06, 4.11).
4.2. Reactions of 1-R with CO. In a glovebox, a 100 mL pressure-
resistant glass vessel equipped with a micro stir bar was charged with
1-F*0.5C6H6 (23 mg, 26 μmol) and C6D5Br (0.85 g, 0.54 mL),
yielding a bright-yellow solution. The vessel was sealed, removed from
the glovebox, and connected to the high vacuum line. The solution
was frozen in a −78 °C (acetone/dry ice) cold bath and the headspace
evacuated. The vacuum manifold of the line was charged with a ∼1:1
mixture of 13C-labeled and unlabeled CO, and the gases were allowed
to mix for 3 h. The obtained gas mixture was admitted to the
evacuated vessel cooled to −196 °C. The vessel was sealed and slowly
warmed to rt and the reaction mixture stirred overnight. After careful
removal of CO gas from the headspace, the vessel was taken into the
glovebox and the solution transferred into a sealable NMR tube. The
1H NMR spectrum (Figure S1 [SI]) revealed the formation of a 2:1
mixture of 2-F/3-F with a 13C/12C ratio of 1:1.5. A mixture of 2-F/3-F
can be isolated as a pale-yellow solid (>70% by mass balance) by
3. CONCLUSIONS
In conclusion, we have observed the activation and (albeit
stoichiometric) functionalization of carbon monoxide using a
strategy that employs an ion pair as an “ionic frustrated Lewis
pair” in which the Lewis acidic site remains accessible for the
incoming CO ligand and in which the weakly coordinating
anion is capable of contributing chemically to the activation
process. A key aspect of this chemistry is that the highly
electropositiveand oxophilicnature of the decamethylscan-
docinium cation employed allows for access of the ephemeral
O-bound isomer of a carbonyl complex and that it is from this
complex that the observed chemistry emanates. This is, to our
knowledge, a new mode of reactivity for transition metal
coordinated carbon monoxide and the first to arise from an O-
bound isocarbonyl complex. The ionic nature of compounds 1-
R opens this unusual reaction channel by keeping the reacting
B−H and B−C bonds in close proximity to the carbon of the
O-bound carbonyl ligand, which due to its energetic nature
must have a relatively short lifetime. In this sense, the reaction
is a homogeneous analog of other systems in which chemistry
occurs in a physically confined environment, for example the
pores of a zeolite, or the active site of an enzyme. Here, the
electrostatic pocket of the ion pair defines the confined but
reactive volume and orients the CO ligand through O-
bounding that allows for this unprecedented reaction to be
observed. The potential for this concept in catalytic processes is
currently being explored for the activation and conversion of
carbon monoxide and other small molecules.
4. EXPERIMENTAL SECTION
4.1. Preparation of Ion Pairs 1-R. 4.1.1. Silane Route. In a
glovebox, a benzene (2.5 g) solution of B(C6F5)3 (512 mg, 1.00
10849
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
removal of bromobenzene under dynamic vacuum and trituration of
the residual oil with pentane (1 mL) at −78 °C (acetone/dry ice),
filtration, and drying under dynamic vacuum. Single crystals of 3-F
suitable for X-ray diffraction were grown from layering a benzene
solution of the solid with hexanes, but quantities of these crystals
sufficient for elemental analysis were not obtainable. 13C-2-F: 1H{19F}
ACKNOWLEDGMENTS
■
Funding for the experimental work described was provided by
the Natural Sciences and Engineering Research Council of
Canada in the form of a Discovery Grant to WEP A.B. thanks
the Deutsche Forschungsgemeinschaft for financial support in
the form of a Postdoctoral Fellowship. L.C. thanks the
NMR (400 MHz, C6D5Br, 298 K): δ 11.80 (dq, JH−C = 151 Hz, JH−B
=
12 Hz, 1H, [O=C(H)B(C6F5)3]−), 1.43 (s, 30H, C5(CH3)5); 19F
NMR (282 MHz, C7D8, 203 K): δ −130.4 (m, 6F, o-F), −156.9 (m,
3F, p-F), −162.9 (m, 6F, m-F); 11B NMR (128 MHz, C6D5Br, 298 K):
δ −13.5 (br dd, JB−C = 51 Hz, JB−H = 12 Hz); 13C NMR (101 MHz,
C6D5Br, 298 K): δ 266.3 (dq, JC−H = 151 Hz, JC−B = 51 Hz), 124.0 (s,
C5(CH3)5), 10.3 (C5(CH3)5) [13C NMR resonances of C6F5 rings
were not detected]; IR (KBr): 1603 cm−1 (νO12C(H)B), 1567 cm−1
Minister
a Ph.D. fellowship. L.M. and O.E. thank the CNRS and the
Ministere de l’Enseignement Superieur et de la Recherche for
̀ ́
e de l’Enseignement Superieur et de la Recherche for
̀
́
funding. L.M. is a junior member of the Institut Universitaire de
France. The CALMIP (Toulouse) and CINES (French
National centre) are gratefully acknowledged for a generous
donation of computational time.
2
(νO13C(H)B). d1-2-F: H NMR (61 MHz, C6H5Br, 298 K): δ 11.77
1
(br, ν1/2 ≈ 17 Hz). 13C-3-F: H{19F} NMR (400 MHz, C6D5Br, 298
K): δ 4.31 (d, J = 160.8 Hz, 1H, cyclo-O-CH(C6F5)−B(C6F5)2), 1.67
and 1.55 (s each, 15H each, C5(CH3)5); 19F NMR (282 MHz, C7D8,
203 K): C-bound C6F5: δ −126.4 (m, 2F, o-F), −152.0 (m, 1F, p-F),
−161.9 (m, 2F, m-F), B-bound C6F5: (a) δ −132.5 (m, 1F, o-F),
−136.5 (m, 1F, o′-F), −156.2 (m, 1F, p-F), −161.2 (m, 1F, m-F),
−163.3 (m, 1F, m′-F), (b) δ −136.8 (m, 1F, o-F), −155.2 (m, 1F, p-F),
−159.4 (m, 1F, m-F), −163.2 (m, 1F, m′-F), −169.6 (m, 1F, o′-F); 11B
NMR (128 MHz, C6D5Br, 298 K): δ −8.6 (br, ν1/2 ≈ 280 Hz); 13C,
13C{1H} NMR (101 MHz, C6D5Br, 298 K): δ 144.5 and 137.7 (o-, m-
C cyclo-O-CH(C6F5)−B(C6F5)2), 124.2 and 123.8 (s each, C5(CH3)5),
118.6 (ipso-C cyclo-O-CH(C6F5)−B(C6F5)2), 60.5 (d, J = 161 Hz,
cyclo-O-CH(C6F5)−B(C6F5)2), 11.2 and 11.1 (s each, C5(CH3)5) [13C
NMR resonances of p-C (cyclo-O-CH(C6F5)−B(C6F5)2) and of B-
REFERENCES
■
(1) Hartwig, J. F. Organotransition Metal Chemistry: From Bonding to
Catalysis; University Science Books: Mill Valley, CA, 2010.
(2) Lupinetti, A. J.; Strauss, S. H.; Frenking, G. Progress in Inorganic
Chemistry; John Wiley & Sons, Inc.: New York, 2001; p 1.
(3) Antonelli, D. M.; Tjaden, E. B.; Stryker, J. M. Organometallics
1994, 13, 763.
(4) Guo, Z.; Swenson, D. C.; Guram, A. S.; Jordan, R. F.
Organometallics 1994, 13, 766.
(5) Hurlburt, P. K.; Anderson, O. P.; Strauss, S. H. J. Am. Chem. Soc.
1991, 113, 6277.
(6) Willner, H.; Schaebs, J.; Hwang, G.; Mistry, F.; Jones, R.; Trotter,
J.; Aubke, F. J. Am. Chem. Soc. 1992, 114, 8972.
(7) Frey, A. S.; Cloke, F. G. N.; Hitchcock, P. B.; Day, I. J.; Green, J.
C.; Aitken, G. J. Am. Chem. Soc. 2008, 130, 13816.
(8) Selg, P.; Brintzinger, H. H.; Schultz, M.; Andersen, R. A.
Organometallics 2002, 21, 3100.
2
bound C6F5 rings were not detected]. d1-3-F: H NMR (61 MHz,
C6H5Br, 298 K): δ 4.26 (br, ν1/2 ≈ 40 Hz).
5. COMPUTATIONAL DETAILS
The Stuttgart−Dresden−Bonn Relativistic Large Effective Core
Potential (RECP) was used to represent the inner shells of Sc with
its associated basis set.41 Fluorine atoms were treated with an effective
core potential in conjunction with its associated basis set augmented
by a set of d polarization functions.42 The atoms B, C, O, and H were
represented by an all-electron 6-31G(d, p) basis set.43 Calculations
were carried out at the DFT(B3PW91) level with Gaussian09.44 The
nature of the extrema (minimum or transition state) was established
with analytical frequencies calculations, and the intrinsic reaction
coordinate (IRC) was followed to confirm that the transition states
connect to reactants and products. The zero-point energy (ZPE) and
entropic contribution have been estimated within the harmonic
potential approximation. The Gibbs free energies, G, were calculated at
T = 298.15 K and p = 1 atm. The NBO analysis was carried out with
the NBO program attached to Gaussian09.45 Additional test
calculations on the CO insertion in the BC bond were carried out
with the DFT-D46 and the MO6 functionals47,48 to evaluate the
consequence of a better representation of the weak interactions on the
activation barriers. Similar results were obtained with all functionals.
(9) Maron, L.; Perrin, L.; Eisenstein, O.; Andersen, R. A. J. Am. Chem.
Soc. 2002, 124, 5614.
(10) Butler, I. S.; Basolo, F.; Pearson, R. G. Inorg. Chem. 1967, 6,
2074.
(11) Noack, K.; Calderazzo, F. J. Organomet. Chem. 1967, 10, 101.
(12) Bock, P. L.; Boschetto, D. J.; Rasmussen, J. R.; Demers, J. P.;
Whitesides, G. M. J. Am. Chem. Soc. 1974, 96, 2814.
(13) Wax, M. J.; Bergman, R. G. J. Am. Chem. Soc. 1981, 103, 7028.
(14) Parks, D. J.; Piers, W. E. J. Am. Chem. Soc. 1996, 118, 9440.
(15) Blackwell, J. M.; Sonmor, E. R.; Scoccitti, T.; Piers, W. E. Org.
Lett. 2000, 2, 3921.
(16) Parks, D. J.; Blackwell, J. M.; Piers, W. E. J. Org. Chem. 2000, 65,
3090.
(17) Welch, G. C.; Juan, R. R. S.; Masuda, J. D.; Stephan, D. W.
Science 2006, 314, 1124.
(18) Welch, G. C.; Stephan, D. W. J. Am. Chem. Soc. 2007, 129, 1880.
(19) Stephan, D. W. Org. Biomol. Chem. 2008, 6, 1535.
(20) Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 46.
(21) Dureen, M. A.; Stephan, D. W. J. Am. Chem. Soc. 2010, 132,
13559.
(22) Chapman, A. M.; Haddow, M. F.; Wass, D. F. J. Am. Chem. Soc.
2011, 133, 8826.
(23) Chapman, A. M.; Haddow, M. F.; Wass, D. F. J. Am. Chem. Soc.
2011, 133, 18463.
(24) Bouwkamp, M. W.; Budzelaar, P. H. M.; Gercama, J.; Del Hierro
Morales, I.; de Wolf, J.; Meetsma, A.; Troyanov, S. I.; Teuben, J. H.;
Hessen, B. J. Am. Chem. Soc. 2005, 127, 14310.
(25) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.;
Nolan, M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am.
Chem. Soc. 1987, 109, 203.
(26) Piers, W. E. Adv. Organomet. Chem. 2005, 52, 1.
(27) Welch, G. C.; Prieto, R.; Dureen, M. A.; Lough, A. J.; Labeodan,
O. A.; Holtrichter-Rossmann, T.; Stephan, D. W. Dalton Trans. 2009,
1559.
ASSOCIATED CONTENT
* Supporting Information
■
S
Crystallographic data files for 1-F and 3-F, additional
experimental and spectroscopic details and tables of coor-
dinates, energies E and Gibbs energies G in a.u. for all
calculated structures and the full list of authors for ref 44. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
wpiers@ucalgary.ca; Odile.Eisenstein@univ-montp2.fr
■
Notes
(28) Chase, P. A.; Jurca, T.; Stephan, D. W. Chem. Commun. 2008,
The authors declare no competing financial interest.
1701.
10850
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851

Journal of the American Chemical Society
Article
(29) Piers, W. E.; Shapiro, P. J.; Bunel, E. E.; Bercaw, J. E. Synlett
1990, 74.
(30) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E. J.
Am. Chem. Soc. 1978, 100, 2716.
(31) Fagan, P. J.; Moloy, K. G.; Marks, T. J. J. Am. Chem. Soc. 1981,
103, 6959.
(32) Hubbard, J. L. Heteroat. Chem. 1992, 3, 223.
(33) Hubbard, J. L.; Smith, K. J. Organomet. Chem. 1984, 276, c41.
(34) Kabalka, G. W.; Gotsick, J. T.; Pace, R. D.; Li, N.-S.
Organometallics 1994, 13, 5163.
(35) Sumerin, V.; Schulz, F.; Nieger, M.; Leskela, M.; Repo, T.;
̈
Rieger, B. Angew. Chem., Int. Ed. 2008, 47, 6001.
(36) The results are discussed using Gibbs energies in kcal mol−1,
calculated at 298 K and one atmosphere of pressure (see Supporting
Information for further details).
(37) Wesolowski, S. S.; Galbraith, J. M.; Schaefer, H. F. J. Chem. Phys.
1998, 108, 9398.
(38) Werkema, E. L.; Maron, L.; Eisenstein, O.; Andersen, R. A. J.
Am. Chem. Soc. 2007, 129, 2529.
(39) Werkema, E. L.; Andersen, R. A.; Yahia, A.; Maron, L.;
Eisenstein, O. Organometallics 2009, 28, 3173.
(40) Sun, Y. M.; Piers, W. E.; Yap, G. P. A. Organometallics 1997, 16,
2509.
(41) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys. 1987,
86, 866.
(42) Maron, L.; Teichteil, C. Chem. Phys. 1998, 237, 105.
(43) Hariharan, P. C.; Pople, J. A. Theor. Chem. Acc. 1973, 28, 213.
(44) Frisch, M. J.; et al. Gaussian 09, Revision A.2; Gaussian Inc.:
Wallingford CT, 2009. See SI for complete citation.
(45) NBO, version 3.1; Glendening E. D.; Reed A. E.; Carpenter, J.
E.; Weinhold, F.
(46) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
2010, 132, 154104.
(47) Zhao, Y.; Truhlar, D. G. J. Chem. Phys. 2006, 125, 194101.
(48) Zhao, Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157.
10851
dx.doi.org/10.1021/ja300591v | J. Am. Chem. Soc. 2012, 134, 10843−10851