Lewis base mediated β-elimination and lewis acid mediated insertion reactions of disilazido zirconium compounds

Article abstract of DOI:10.1021/ja407950e

The reactivity of a series of disilazido zirconocene complexes is dominated by the migration of anionic groups (hydrogen, alkyl, halide, OTf) between the zirconium and silicon centers. The direction of these migrations is controlled by the addition of two-electron donors (Lewis bases) or two-electron acceptors (Lewis acids). The cationic nonclassical [Cp₂ZrN(SiHMe ₂)₂]⁺ ([2]⁺) is prepared from Cp₂Zr{N(SiHMe₂)₂}H (1) and B(C ₆F₅)₃ or [Ph₃C][B(C ₆F₅)₄], while reactions of B(C ₆F₅)₃ and Cp₂Zr{N(SiHMe ₂)₂}R (R = Me (3), Et (5), n-C₃H₇ (7), CH? -CHSiMe₃ (9)) provide a mixture of [2]⁺ and [Cp₂ZrN(SiHMe₂)(SiRMe₂)]⁺. The latter products are formed through B(C₆F₅)₃ abstraction of a β-H and R group migration from Zr to the β-Si center. Related β-hydrogen abstraction and X group migration reactions are observed for Cp₂Zr{N(SiHMe₂)₂}X (X = OTf (11), Cl (13), OMe (15), O-i-C₃H₇ (16)). Alternatively, addition of DMAP (DMAP = 4-(dimethylamino)pyridine) to [2]⁺ results in coordination to a Si center and hydrogen migration to zirconium, giving the cationic complex [Cp₂Zr{N(SiHMe₂)(SiMe₂DMAP)}H] ⁺ ([19]⁺). Related hydrogen migration occurs from [Cp ₂ZrN(SiHMe₂)(SiMe₂OCHMe₂)] ⁺ ([18]⁺) to give [Cp₂Zr{N(SiMe ₂DMAP)(SiMe₂OCHMe₂)}H]⁺ ([22] ⁺), whereas X group migration is observed upon addition of DMAP to [Cp₂ZrN(SiHMe₂)(SiMe₂X)]⁺ (X = OTf ([12]⁺), Cl ([14]⁺)) to give [Cp ₂Zr{N(SiHMe₂)(SiMe₂DMAP)}X]⁺ (X = OTf ([26]⁺), Cl ([20]⁺)). The species involved in these transformations are described by resonance structures that suggest β-elimination. Notably, such pathways are previously unknown in early metal amide chemistry. Finally, these migrations facilitate direct Si-H addition to carbonyls, which is proposed to occur through a pathway that previously had been reserved for later transition metal compounds.

Full text of DOI:10.1021/ja407950e

Article
pubs.acs.org/JACS
Lewis Base Mediated β‑Elimination and Lewis Acid Mediated
Insertion Reactions of Disilazido Zirconium Compounds
KaKing Yan, Juan J. Duchimaza Heredia, Arkady Ellern, Mark S. Gordon, and Aaron D. Sadow*
Department of Chemistry, Iowa State University, 1605 Gilman Hall, Ames, Iowa 50011, United States
S
* Supporting Information
ABSTRACT: The reactivity of a series of disilazido
zirconocene complexes is dominated by the migration of
anionic groups (hydrogen, alkyl, halide, OTf) between the
zirconium and silicon centers. The direction of these
migrations is controlled by the addition of two-electron
donors (Lewis bases) or two-electron acceptors (Lewis acids).
The cationic nonclassical [Cp2ZrN(SiHMe2)2]+ ([2]+) is
prepared from Cp2Zr{N(SiHMe2)2}H (1) and B(C6F5)3 or
[Ph3C][B(C6F5)4], while reactions of B(C6F5)3 and Cp2Zr-
{N(SiHMe2)2}R (R = Me (3), Et (5), n-C3H7 (7), CHCHSiMe3 (9)) provide a mixture of [2]+ and [Cp2ZrN(SiHMe2)-
(SiRMe2)]+. The latter products are formed through B(C6F5)3 abstraction of a β-H and R group migration from Zr to the β-Si
center. Related β-hydrogen abstraction and X group migration reactions are observed for Cp2Zr{N(SiHMe2)2}X (X = OTf (11),
Cl (13), OMe (15), O-i-C3H7 (16)). Alternatively, addition of DMAP (DMAP = 4-(dimethylamino)pyridine) to [2]+ results in
coordination to a Si center and hydrogen migration to zirconium, giving the cationic complex [Cp2Zr{N(SiHMe2)-
(SiMe2DMAP)}H]+ ([19]+). Related hydrogen migration occurs from [Cp2ZrN(SiHMe2)(SiMe2OCHMe2)]+ ([18]+) to give
[Cp2Zr{N(SiMe2DMAP)(SiMe2OCHMe2)}H]+ ([22]+), whereas X group migration is observed upon addition of DMAP to
[Cp2ZrN(SiHMe2)(SiMe2X)]+ (X = OTf ([12]+), Cl ([14]+)) to give [Cp2Zr{N(SiHMe2)(SiMe2DMAP)}X]+ (X = OTf
([26]+), Cl ([20]+)). The species involved in these transformations are described by resonance structures that suggest β-
elimination. Notably, such pathways are previously unknown in early metal amide chemistry. Finally, these migrations facilitate
direct Si−H addition to carbonyls, which is proposed to occur through a pathway that previously had been reserved for later
transition metal compounds.
the continuum, such three-center−two-electron (3c-2e) inter-
actions of aromatic C−H bonds and electron-rich metal centers
may be viewed as arrested C−H bond oxidative additions.3
Similarly, β-agostic organometallics containing a low-valent
metal center may be viewed as intermediate between the metal
alkyl and a metallacyclopropane hydride resulting from
oxidative addition.4 On the other hand, high-valent metal
centers containing β-agostic C−H bonds are characterized as
arrested intermediates on the path to an isovalent metal hydride
and olefin.1c,4a,5 β-agostic main-group alkyls are at the other end
of the continuum, and electron-density analysis suggests the
metal−CH interaction is mainly electrostatic;4a β-hydrogen
elimination is the least facile in these main-group systems.
These electrostatic agostic structures are not established as
intermediates on pathways for insertion or elimination.
INTRODUCTION
■
β-elimination and its microscopic reverse, 1,2-migratory
insertion, are central to bond-forming and -breaking processes.
These reactions are well studied for many metal−ligand pairs;
however, the formation of new C−E bonds through insertion
reactions into M−E bonds (E = halide, OR, NR2) remains a
major challenge in chemistry. New elementary steps are
needed, as these could provide enabling strategies, including
catalytic methods, for the efficient synthesis of functionalized
organic compounds (e.g., enantioselective hydration, halogen-
ation) or the selective defunctionalization of organic com-
pounds (e.g., for the conversion of biorenewables).
When the migrating group is hydrogen, β-agostic species are
the proposed intermediates on this pathway.1 There have been
detailed structural and spectroscopic studies of these
compounds, which are suggested to provide a description of
the species on the reaction coordinate between the metal alkyl
and the metal hydride/olefin.
Thus, β-agostic species have a special relationship with the
pathways involving insertion of unsaturated organics into M−H
bonds and β-hydrogen elimination. However, strongly Lewis
acidic metal centers and polarizable E−E′ bonds (e.g., E−E′ =
Si−H, Si−C, B−H) are well-known to form side-on
interactions.2,6−8 These structures are not associated with
The bonding nature of agostic interactions, the chemical
interpretations offered to describe the interactions, and the
anticipated reactions associated with the structures, however,
vary with the relative position of a C−H bond with respect to
the metal (α, β, etc.), the metal center and its valence, and the
other elements present in the agostic ligand.2 On one end of
Received: August 1, 2013
Published: September 25, 2013
© 2013 American Chemical Society
15225
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
insertion/elimination reactions. For example, rare-earth dis-
ilazide compounds, such as [Me2Si(C5Me4)2]LaN(SiHMe2)2,
contain an unusual structure characterized by an obtuse Si−N−
the study of migration chemistry of β-OR, β-OTf, and β-Cl
transfer between zirconium and silicon centers.
1
RESULTS
Si angle, upfield SiH resonances (2.6−4 ppm) and low JSiH
■
values (ca. 130−150 Hz), and low-energy νSiH bands in the IR
spectra (1790−1845 cm−1).9 The bonding of [Me2Si-
(C5Me4)2]LaN(SiHMe2)2 may be similar to that of the
bis(catacolborane) compound Cp2Ti(η2-HBcat)2, which con-
tains two side-on HBcat ligands,10 and Cp2Zr{(HB-
(C6F5)2)2CH2}, which contains two Zr−H−B bridges.11
A key question associated with these bridging structures
involves their relationship to the insertion−elimination
reactivity. It has been suggested that the rarity of β-eliminations
for transition-metal amido compounds12 is related to the nature
of agostic β-CH structures of amide ligands that is distinct in
geometry and spectroscopy from the agostic alkyls.13 For
example, β-agostic amides generally feature long N−C bonds,
large (ca. 120°) ∠M−N−C angles, and short β-C−H
distances,13 whereas β-agostic alkyls contain short C−C
bonds, acute ∠M−C−C angles, and elongated β-C−H bonds.4b
As the microscopic reverse of β-elimination, the insertion of
olefins into more polar M−X bonds (e.g., M−F, M−Cl, M−
OR, M−NR2) varies from unknown to rare.14 Interestingly, a
Si−N bond formation was recently described in the reaction of
{PhC(N-2,6-(Me2HC)2C6H3)2}Sc{N(SiHMe2)2}2 and Ph3C+,
which gives Ph3CH and [Sc{N(SiHMe2)SiMe2N-
(SiHMe2)2}]+.15
Here, we present a study of the cationic disilazidozirconium
compound [Cp2ZrN(SiHMe2)2]+ ([2]+), which possesses
extreme spectroscopic and structural features attributed to the
side-on interaction of two SiH groups with a zirconium center.
The analogy of the side-on β-Si−H⇀Zr interaction with
agostic β-CH organometallic compounds is supported by
pathways to form [2]+ and its reactivity. This cationic
disilazidozirconium reacts with DMAP to give a zirconium
hydride through an apparent β-hydrogen elimination process.
Addition of the Lewis acid B(C6F5)3 to Cp2Zr{N(SiHMe2)2}R
results in Si−C bond formation through an apparent migratory
insertion reaction. In fact, the reactivity of the β-groups on the
disilazido ligand, in response to two-electron donors and two-
electron acceptors, provides connections to β-elimination and
insertion chemistry reminiscent of late transition-metal β-
agostic alkyl systems (Scheme 1).
Synthesis and Characterization of the Neutral
Precursor of [Cp2ZrN(SiHMe2)2]+. We recently communi-
cated the formation of Cp2Zr{N(SiHMe2)2}H (1) as a side
product in the reaction of Cp2Zr{N(SiHMe2)2}OTf and
LiN(SiHMe2)t-Bu.16 Identification of 1 in that reaction
required its independent synthesis, which was achieved by the
reaction of Cp2ZrHCl and LiN(SiHMe2)2. We have also briefly
communicated its solution-phase structure in the context of an
unusual γ-abstraction reaction.17 Surprisingly, [ZrCl{N-
(SiHMe2)2}2(μ-Cl)]2 is the only other reported zirconium
complex containing the N(SiHMe2)2 moiety,18 even though
this hydrosilazide ligand is widely used in group 3 and
lanthanide chemistry,19 and the related hexamethyldisilazido
ligand [N(SiMe3)2]− is important in transition-metal, lantha-
nide, and main-group chemistry.20 We describe this compound
here because the spectroscopy associated with the nonclassical
SiH zirconium interaction represents a starting point for
comparison to the unique compounds described here.
The 1H NMR spectrum of 1 contained the expected
resonances, including those for ZrH (5.60 ppm) and SiH
1
(3.78 ppm, JSiH = 161.0 Hz). The slightly upfield silicon
1
hydride resonance and moderately low JSiH value (cf.
1
HN(SiHMe2)2, JSiH = 170 Hz) suggest a possible side-on
Si−H interaction with Zr,21 however, the SiMe2 groups are
1
equivalent in the H, 13C{1H}, and 29Si NMR spectra acquired
at room temperature.
Therefore, 1H NMR spectra of 1 in toluene-d8 were recorded
from 298 to 191 K. The SiH and SiMe resonances broaden to
1
the coalescence point at 210 K. At 191 K, the H NMR
spectrum contains two SiH resonances and two SiMe2 signals.
The two SiH resonances were assigned as nonclassical SiH
(2.13 ppm, 1JSiH = 129.7 Hz) and terminal SiH (4.99 ppm, 1JSiH
= 179.6 Hz) on the basis of the 1JSiH coupling constants. At 191
K, the 29Si NMR spectrum was resolved to show resonances at
−23.1 and −62.9 ppm, which were assigned to silicon atoms
with terminal and bridging hydrogens, respectively, by a
1H−29Si HMBC experiment. A similar upfield 29Si NMR
resonance was reported for the nonclassical Cp2Zr{N(SiHMe2)
t-Bu}H (−73.4 ppm),22 while the downfield signal is
comparable to the 29Si NMR resonance of the SiH in
In addition, these compounds react with carbonyls, resulting
in hydrosilylation. The mechanism of this hydrosilylation is
shown to be related to the hydrogen shuttling between Zr and
Si centers. The pathway for this reaction is explored through
1
Y{N(SiHMe2)2}3(NHC)2 (δ −22.5, JSiH = 172 Hz; NHC =
1,3-dimethylimidazolin-2-ylidene), which contains short Y−Si
distances (Y−Si ≈ 3.13 Å; Y−N−Si = 105.1(3)°) and a low-
energy νSiH (2041 cm−1).23 Bands at 2047 and 1907 cm−1 in the
infrared spectrum of 1 further indicate a nonclassical
interaction, and IR also provides support for the ZrH group
(1559 cm−1).
Scheme 1. Two-Electron Donor/Acceptor Mediated Anionic
Group Transfer
Interestingly, the reaction of deuterium-labeled Cp2ZrDCl
and LiN(SiHMe2)2 provides an isotopically scrambled mixture
of Cp2Zr{N(SiDMe2)(SiHMe2)}H and Cp2Zr{N(SiHMe2)2}D
(eq 1).
As expected, the deuterium distribution favors its localization
in the most tightly bonded position on the basis of IR
analysis.24 Experiments to discern if exchange occurs prior to
formation of 1-d1 first involved treatment of 1 with D2 (1 atm)
or PhSiD3, with the expectation that ZrH/D2 and ZrH/SiD
exchange would provide Cp2Zr{N(SiHMe2)2}D. However,
these reaction mixtures, heated to 80 °C, give only
deuterium-free 1. Attempts by electrospray MS at determining
15226
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
the deuterium distribution (e.g., 1-d0, 1-d1, 1-d2, 1-d3) to clarify
possible crossover were also unsuccessful. We suspect that H/D
exchange between Cp2ZrDCl and LiN(SiHMe2)2 occurs prior
to Zr−N bond formation because the ZrH and D2 or PhSiD3
do not exchange under the conditions attempted.
Synthesis of Nonclassical [Cp2ZrN(SiHMe2)2]+. Cationic
[Cp2ZrN(SiHMe2)2]+ ([2]+) is synthesized as the [HB-
(C6F5)3]− or [B(C6F5)4]− salt from reactions of 1 and
B(C6F5)3 or [Ph3C][B(C6F5)4] (eq 2). The reactions are
Figure 1. ORTEP diagram of [Cp2ZrN(SiHMe2)2][HBC6F5)3]
([2][HB(C6F5)3]). Ellipsoids are drawn at the 35% probability level.
The dashed line between Si2 and F12 indicates the shortest contact
between the cation and anion. Significant interatomic distances (Å):
Zr1−N1, 2.193(2); Zr1−Si1, 2.8740(8); Zr1−H1s, 2.09(3); Si1−H1s,
1.69(3); Zr1−Si2, 2.8706(7); Zr1−H2s, 2.06(3); Si2−H2s, 1.54(4);
N1−Si1, 1.662(2); N1−Si2, 1.661(2). Selected interatomic angles
(deg): Zr1−N1−Si1, 95.4(1); Zr1−N1−Si2, 95.3(1); Si1−N1−Si2,
168.9(1); H1s−Zr1−H2s, 137(1); Zr1−N1−Si1−H1s, 0(1); Zr1−
N1−Si2−H2s, −3(1), H1si−Si1−Si2−H2s, 3(2).
than the Si−N distance associated with the terminal SiHMe2 in
1 (1.710(2) Å) and ca. 0.02 Å shorter than the Si−N distance
of 1.683(2) Å associated with the nonclassical β-SiHMe2 group.
The ∠Zr−N−Si angles of [2]+ are small (95.4(1) and
95.3(1)°), and the ∠Si−N−Si angle approaches linearity at
168.9(2)°.
Interestingly, the Zr−N distance of 2.193(2) Å is 0.05 Å
longer than that in 1. The hydrogen atoms on the SiHMe2 (as
well as the H atom in HB(C6F5)3 group) were located in the
Fourier difference map and refined; the Zr−N bond and the
two Si−H bonds are essentially coplanar, as indicated by the
torsion angles Zr1−N1−Si1−H1s (0(1)°), Zr1−N1−Si2−H2s
(−3(1)°), and H1s−Si1−Si2−H2s (3(2)°).
quantitative in bromobenzene-d5 on a micromolar scale, but
benzene is preferable as the solvent in preparatory-scale
reactions for easy separation of [2]+ as an insoluble oil.
The unusual NMR spectroscopic features of the N(SiHMe2)2
ligand in [2]+ are similar in both the [HB(C6F5)3]− and
[B(C6F5)4]− compounds, and the spectroscopic data here are
given for the [B(C6F5)4]− salt.25 The SiH group in [2]+ was
characterized by a far upfield 1H NMR signal at −0.48 ppm and
an unusually low 1JSiH value (89.3 Hz). For comparison, the 1H
NMR spectrum of the isoelectronic and C2-symmetric rac-
Me2Si(2-Me-Benz-Ind)2]YN(SiHMe2)2 contained an upfield
SiH signal (2.65 ppm) and a low JSiH value (133 Hz).9 The
The Zr1−H1s and Zr1−H2s distances in [2]+ (2.09(3) and
2.06(3) Å) are long in comparison to the Zr−H distance for the
zirconium hydride in 1 (1.90(3) Å), although they are
significantly shorter than the Zr−HSi distance of the non-
classical Zr↼H−Si in that compound (2.46(4) Å). The
terminal Si−H and the nonclassical Si−H distances in 1 are
1.53(4) and 1.47(4) Å, respectively. Furthermore, the HB-
(C6F5)3 counterion has minimal impact on the distances and
angles of the N(SiHMe2)2 ligand. The Zr−H (2.06(3) Å) and
Si−H (1.54(4) Å) distances of the SiHMe2 with the fluorine−
silicon close contact are within 3× of the esd of the distances in
the other SiHMe2 (2.09(3) and 1.69(3) Å, respectively). The
observed spectroscopic and structural features provide support
for significant nonclassical Zr−N(SiHMe2)2 interactions in the
electron-poor metal complex.9,19
1
spectroscopy of [2]+ is consistent with a C2v-symmetric
compound. One 29Si NMR signal was observed at −43 ppm;
the 29Si NMR resonance of [2]+ is upfield of terminal SiH
groups in 1 but downfield of the nonclassical SiH group in 1. In
the IR spectrum of [2][B(C6F5)4], the two bands that are
observed at 1738 and 1659 cm−1 were assigned to νSiH. These
energies are significantly lower than 2c-2e SiHs in the classical
[Cp2Zr{N(SiHMe2)2}OPEt3]+ (2122 cm−1; see below) or the
neutral mixed classical/nonclassical 1 (2047 and 1907 cm−1).
X-ray-quality crystals of [2][HB(C6F5)3] were obtained by
slow diffusion of pentane into a concentrated bromobenzene
solution at −30 °C (Figure 1). In the solid-state structure, the
cationic portion [Cp2Zr(N(SiHMe2)2]+ is separated from the
[HB(C6F5)3]− anion, and the shortest distance in the ion pair is
3.27 Å between an aryl fluoride and one of the silicon centers
Computational Model of [Cp2ZrN(SiHMe2)2]+ ([2]+).
More insight into the relationship between the structural
features of [Cp2ZrN(SiHMe2)2]+ ([2]+) and its spectroscopic
properties is provided by a computational study. The geometry
optimization and Hessian analysis were carried out using
Møller−Plesset second-order perturbation theory (MP2)28
with a model core potential triple-ζ basis (MCP-TZP)29 in
GAMESS.30,31 The optimized geometry of the cationic portion
of [Cp2ZrN(SiHMe2)2]+ shown in Figure S1 (Supporting
Information) is in good agreement with the coordinates
VDW
(∑SiF
= 3.57 Å). The solid-state structure contains a
significant, almost perfectly symmetrical distortion of the
N(SiHMe2)2 ligand. The short Zr−Si distances (Zr1−Si1 =
2.8740(8) Å and Zr1−Si2 = 2.8706(7) Å) and the N−Si
distances (1.661(2) and 1.662(2) Å) are equivalent within error
bounds. For comparison, the Zr−Si distance in the authentic
zirconium silyl species Cp2ZrSiMe3(S2CNEt2) is 2.815(1) Å,26
while the Zr−Si distance in Cp2Zr{η2-N(t-Bu)SiMe2}PMe3 is
2.654(1) Å.27 The Si−N distances in [2]+ are ca. 0.05 Å shorter
15227
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
obtained from an X-ray crystallographic structure determi-
nation. For example, the Zr−Si distances, calculated to be 2.861
Å, closely match the experimental Zr−Si distances of 2.8740(8)
and 2.8706(7) Å. The calculated Zr−N distance of 2.22 Å is
slightly longer than the experimental distance of 2.193(2) Å.
The bridging hydrogens are of particular interest, and the
calculated Zr−H and Si−H distances are 2.06 and 1.57 Å,
respectively.
The vibrational calculation verified that this structure is a
minimum on the potential energy surface. Two normal modes
are associated with the bridging Zr↼H−Si structure; these are
symmetric and asymmetric SiH stretching motions with
unscaled frequencies of 1800 and 1743 cm−1. These frequencies
compare well to the bands in the IR spectrum of [2]+;
furthermore, the motion is parallel to the Si−H bond vector
rather than along the Zr−H vector.
An orbital localization using the method developed by
Edmiston and Ruedenberg32 reveals two pairs of orbitals of
interest (see Figures S2 and S3 in the Supporting Information).
One pair of orbitals displays a clear bonding interaction
between the three atoms in the Zr−H−Si bridge, while the
second pair of orbitals shows a four-atom interaction among Zr,
N, and Si. These orbitals suggest the existence of significant
interactions between the metal and the silyl groups.
HMQC experiment contained a cross peak between the
resonance at −14.2 ppm and a broad H NMR resonance at
1
1.11 ppm (3 H) assigned to a [MeB(C6F5)3] group. Integration
of the resonances for [2]+ and [MeB(C6F5)3] is the basis for
assignment of the major product as [2][MeB(C6F5)3]. The
minor product [4][HB(C6F5)3], meanwhile, contained a
3
doublet (0.31 ppm, JHH = 2.4 Hz, 6 H) and a singlet (0.27
ppm, 9 H) assigned to SiMe2 and SiMe3 groups and a multiplet
at 0.64 ppm (1JSiH = 94.4 Hz) assigned to a nonclassical β-Si−
H⇀Zr interaction.
The proposed pathway to the minor product [4]+ involves β-
hydrogen abstraction from the disilazido ligand to give a
silylium center followed by migration of the methyl from
zirconium. On the basis of this idea, zirconium disilazido
compounds containing sterically hindered alkyl groups should
react with B(C6F5)3 more readily by β-hydrogen abstraction
than by alkyl group abstraction. Competition experiments were
designed to test this, and the compounds Cp2Zr{N-
(SiHMe2)2}R (R = Et (5), n-C3H7 (7), CHCHSiMe3 (9))
were allowed to react with B(C6F5)3 to give mixtures of [2]+
and [Cp2ZrN(SiHMe2)(SiRMe2)]+ (R = Et (6), n-C3H7 (8),
CHCHSiMe3 (10)) (Scheme 3). Importantly, the ratio of
[Cp2ZrN(SiHMe2)(SiRMe2)]+ to [2]+ increases as the alkyl
group is varied, following the trend Me < Et < n-C3H7 < CH
CHSiMe3.
Reactions of Cp2Zr{N(SiHMe2)2}R with B(C6F5)3. Two
pathways that could provide [2]+ from the interaction of 1 and
B(C6F5)3 are shown in Scheme 2: (A) abstraction of the ZrH
Scheme 3. Competition between [Zr]R Abstraction and β-
Hydrogen Abstraction in Reactions of Mixed Alkyl
Disilazido Zirconium Compounds and B(C6F5)3
Scheme 2. Possible Pathways for Formation of [2]+: (A) ZrH
Abstraction or (B) SiH Abstraction
Although [4]+, [6]+, [8]+, and [10]+ could not be separated
from the side product [2]+, the assignments of Si−C bond
and (B) β-hydrogen abstraction followed by ZrH migration.
Although labeled Cp2Zr{N(SiHMe2)2}D could potentially
resolve this issue, attempted synthesis of 1-d1 from Cp2ZrDCl
and LiN(SiHMe2)2 provides a mixture with Cp2Zr{N-
(SiHMe2)2-d1}H as noted above. Instead, abstraction reactions
of alkyl disilazido zirconium compounds were studied to
distinguish the abstraction pathways. Furthermore, variation of
the alkyl group is a means to control the nucleophilic site in the
zirconium compounds, as probed by reactions with B(C6F5)3.
Treatment of the zirconium methyl species Cp2Zr{N-
(SiHMe2)2}Me (3) with B(C6F5)3 gives [2][MeB(C6F5)3] as
the major product (70%); however, [Cp2ZrN(SiHMe2)-
(SiMe3)][HB(C6F5)3] ([4][HB(C6F5)3]) is also formed in
30% yield (eq 3). In the reaction mixture, two 11B NMR
1
formation are unambiguously supported by H−29Si HMBC
and COSY experiments. For example, in compound [10]+, a
cross peak is detected between the 29Si NMR signal at −12.9
1
ppm and the H NMR vinylic signals at 6.43 and 6.75 ppm.
Interestingly, [HB(C6F5)3]− is the only counterion in the
reaction mixture resulting from interaction of B(C6F5)3 and the
ethyl, n-propyl, and trimethylsilylvinyl zirconium compounds.
Meanwhile, ethylene, propylene, and trimethylsilylacetylene are
formed as byproducts, and these data are consistent with β-H
abstraction from the alkyl group. Thus, β-hydrogen abstraction
(as part of either SiH or CH groups) is favored with respect to
alkyl group abstraction.33 Previously, we observed a concen-
tration dependence on β-hydrogen abstraction vs alkyl group
abstraction in reactions of ZnR2 and bis(4,4-dimethyl-2-
oxazolinyl)phenylborane.34 However, in the current system
1
carbon−boron bond formation is below H and 11B NMR
detection limits, and the product ratios are similar in reactions
performed at concentrations from 2.4 to 9.6 mM.
Reactions of Cp2Zr{N(SiHMe2)2}X with B(C6F5)3. Zirco-
nium alkyls, and likely a zirconium hydride, migrate to the β-
silicon center of the silazido ligand upon addition of Lewis
resonances were observed at −14.2 ppm (singlet; major) and
1
−24.8 ppm (doublet, JBH = 81.5 Hz; minor). A 1H−11B
15228
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
acids. Therefore, we were interested in studying the migration
of other anionic groups, such as OR, Cl, and OTf.
NMR resonances at 21.4 ppm (SiMe2OTf) and −25.9 ppm
1
(SiHMe2) were assigned by H−29Si HMBC experiments. For
comparison, the Me3SiOTf 29Si NMR chemical shift is +43.5
ppm and the Me(Et2N)2SiOTf 29Si NMR chemical shift is −19
ppm.35
X-ray-quality crystals of [12][HB(C6F5)3] were obtained
from a concentrated bromobenzene solution layered with
pentane cooled to −30 °C. A single-crystal X-ray diffraction
study shows the OTf− is bridging between Zr and a β-Si center
(Figure 2). In addition, there is a short Zr1−Si2 distance of
The compounds Cp2Zr{N(SiHMe2)2}X (X = OTf (11), Cl
(13), OMe (15), OCHMe2 (16)) react with B(C6F5)3 or
[Ph3C][B(C6F5)4] in benzene to give [Cp2ZrN(SiHMe2)-
(SiMe2-μ-X)]+ (X = OTf ([12]+), Cl ([14]+), OMe ([17]+),
OCHMe2 ([18]+)) (Scheme 4).
Scheme 4. β-Hydrogen Abstraction Reactions of Zirconium
Disilazido Triflate, Chloride, and Alkoxide Compounds
Figure 2. ORTEP diagram of [Cp2ZrN(SiHMe2)(SiMe2-μ-κ2-OTf)]-
[HB(C6F5)3] ([12][HB(C6F5)3]) with ellipsoids plotted at the 35%
probability level. The cationic portion of the structure is illustrated,
and only the bridging hydrogen atom is plotted. All other hydrogen
atoms, the HB(C6F5)3 counterion, and a disordered C6H5Br molecule
are not included for clarity.
In all cases, hydrogen abstraction is supported by the
formation of [HB(C6F5)3]− (e.g., [12][HB(C6F5)3]: 11B NMR
−24.8 ppm, 1JBH = 87.1 Hz) or Ph3CH. The series of products
[12]+, [14]+, [17]+, and [18]+ contain similar spectroscopic
features for the N(SiHMe2)(SiMe2X) ligand. The cationic
[Cp2ZrN(SiHMe2)(SiMe2X)]+ is Cs symmetric, as two SiMe2
and one C5H5 resonance were observed in the 1H NMR
2.890(1) Å. The Zr1−H1g distance of 2.20(3) Å for the
nonclassical SiH is between the related distances in neutral 1
and cationic [2]+. However, [12]+ does not display the other
unusual structural features of [2]+; namely, the Si1−N1−Si2
angle is normal (127.7(2)°) and the Zr1−Si1 distance is long
(3.59 Å).
1
spectrum. All the compounds feature upfield-shifted H NMR
1
resonances for the β-SiH, extremely low JSiH values, and 29Si
NMR resonances for the β-SiH group ranging from −26 to
−32 ppm (Table 1). These assignments are supported by
1H−29Si HMBC experiments.
The upfield SiH chemical shifts and 1JSiH values for both salts
of [12]+ provide support for the nonclassical structure. The 29Si
The Zr1−O10 distance of 2.313(2) Å is slightly longer than
the Zr−OTf distance in [Cp2Zr(κ1-OTf)(μ-H)]2 (2.205(2)
1
Table 1. H and 29Si NMR Data of Cationic β-SiH Containing Dihydrodisilazido Zirconium Compounds
a
compd
SiH (ppm)
1JSiH (Hz)
SiHMe2 (ppm)
SiMe2X (ppm)
[Cp2ZrN(SiHMe2)2][HB(C6F5)3] ([2][HB(CF5)3])
[Cp2ZrN(SiHMe2)2][B(C6F5)4] ([2][B(C6F5)4])
[Cp2ZrN(SiHMe2)(SiMe3)]+ ([4]+)
−0.44
−0.48
0.64
0.5
107
89
94
94
87
89
107
99
89
96
94
−42.7
−43.7
8.2
n.a.
n.a.
−0.2
5.6
[Cp2ZrN(SiHMe2)(SiMe2Et)]+ ([6]+)
7.4
[Cp2ZrN(SiHMe2)(SiMe2-n-C3H7)]+ ([8]+)
[Cp2ZrN(SiHMe2)(SiMe2CHCHSiMe3)]+ ([10]+)
[Cp2ZrN(SiHMe2)(SiMe2-μ−κ2-OTf)][HB(C6F5)3] ([12][HB(C6F5)3])
[12][B(C6F5)4]
[Cp2ZrN(SiHMe2)(SiMe2-μ-Cl)]+ ([14]+)
[Cp2ZrN(SiHMe2)(SiMe2-μ-OMe)]+ ([17]+)
[Cp2ZrN(SiHMe2)(SiMe2-μ-OCHMe2)]+ [18]+
[Cp2Zr{N(SiHMe2)(SiMe2DMAP)}H]+ ([19]+)
[Cp2Zr{N(SiHMe2)(SiMe2DMAP)}Cl]+ ([20]+)
[Cp2Zr{N(SiHMe2)(SiMe2py)}OTf]+ ([25]+)
[Cp2Zr{N(SiHMe2)(SiMe2DMAP)}OTf]+ ([26]+)
[Cp2Zr{N(SiHMe2)(SiMe2OPEt3)}OTf]+ ([27]+)
0.46
0.56
0.22
0.18
0.65
0.34
0.41
0.95
3.04
1.27
1.37
1.35
5.2
6.0
−11.5
−25.9
−25.2
−32.5
−29.7
−33.0
−63.6
−34.2
−24.0
−26.6
−29.1
−5.2
21.4
21.4
29.6
20.4
13.2
−0.6
5.7
b
118
155
115
115
115
14.4
7.7
2.4
a
b
All compounds are [HB(C6F5)3] salts unless otherwise noted. Acquired at −88 °C.
15229
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
Å).36 The Si1−O1 distance of 1.788(2) Å is within the sum of
covalent radii.37 These distances, the 29Si NMR chemical shift
of the SiOTf center, and the short Zr−H and Zr−Si distances
argue for greater positive charge localization on Zr rather than
on the Si center.
Similarly, the reaction of Cp2Zr{N(SiHMe2)2}Cl and
B(C6F5)3 gives [Cp2ZrN(SiHMe2)(SiMe2-μ-Cl)]+ ([14][HB-
(C6F5)3]) via β-hydrogen abstraction. The 29Si NMR chemical
shift for the SiMe2Cl moiety is 29.6 ppm; for comparison, the
value for Me3SiCl is 30.2 ppm38 and the value for
Me2NSiMe2Cl is 13.9 ppm.39
(0.51 and 0.21 ppm), ZrH (4.43 ppm), SiH (1.27 ppm), and
aromatic (pyridine) CH resonances (7.93 and 6.78 ppm) were
readily assigned (all assignments are supported by COSY
experiments, chemical shift, and integration). The resonances
assigned to C5H5 (5.81 ppm) and NMe2 (3.21 ppm) were
sharp. However, we were unable to detect 29Si NMR
resonances at room temperature using 29Si INEPT experiments.
The low-temperature NMR spectra of [19]+ were sharper
and better resolved. At 185 K, two 29Si resonances were
observed at −0.5 and −63.6 ppm. Only the signal at −63.6 ppm
(1JSiH = 118 Hz) was observed in a 29Si INEPT experiment
optimized for JSiH = 120 Hz, and the signal at −0.5 ppm was
only detected in a 29Si INEPT experiment optimized for long-
Finally, reactions of 1 and paraformaldehyde or acetone
provide the alkoxide (disilazido)zirconium compounds Cp2Zr-
{N(SiHMe2)2}OR (R = Me (15), CHMe2 (16)). Notably,
reactions of 1 and excess ketone or aldehyde do not result in
insertion into the Si−H bonds of the N(SiHMe2)2 group under
these conditions (see reactions of [2]+ and carbonyls below for
such transformations). The bonding of the N(SiHMe2)2 group
in neutral Cp2Zr{N(SiHMe2)2}OR is classical, as evidenced by
downfield SiH chemical shifts (δ 4.71 and 4.77), high SiH
coupling constants (1JSiH = 180 and 182 Hz), and high-energy
νSiH bands (15, 2078 cm−1; 16, 2113 and 2051 cm−1). The
structure is further supported by a single-crystal X-ray
diffraction study of 16 (see the Supporting Information). The
∠Zr1−N1−Si1 and ∠Zr1−N1−Si2 angles are 121.4(1) and
122.2(1)°, respectively; the Zr1−Si1 and Zr1−Si2 distances are
3.3714(9) and 3.3824(8) Å. In addition, the Zr1−O1 distance
is 1.937(2) Å while the O1−Si1 distance is 3.353(2) Å.
The reaction of methoxyzirconium species 15 and B(C6F5)3
or [Ph3C][B(C6F5)4] provides [Cp2ZrN(SiHMe2)(SiMe2-μ-
OMe)]+ ([17]+). Likewise, the reaction of isopropoxyzirco-
nium species 16 and B(C6F5)3 or [Ph3C][B(C6F5)4] yields
([18]+) (Scheme 4). Compounds [17]+ and [18]+, remarkably,
also feature spectroscopic features associated with nonclassical
Zr↼H−Si structures. This, perhaps, is most notable because
the neutral precursors 15 and 16 contain only 2c-2e SiH
moieties. Abstraction of a β-hydrogen and formation of a Zr−
O−Si bridging interaction might be expected to geometrically
limit possible Zr↼H−Si interactions. Instead, the structure of
[Cp2ZrN(SiMe2OCHMe2)2]+ suggests that the N-
(SiMe2OCHMe2) moiety is best described as a β-silyl ether,
and this permits a long Zr−O distance (see below).
1
range proton coupling (JSiH = 7 Hz). In a H−29Si HMBC
experiment optimized for long-range 1H−29Si coupling (JSiH = 7
Hz) shown in Figure 3, the −63.6 ppm 29Si resonance was
Figure 3. 1H−29Si HMBC experiment on [Cp2Zr{N(SiHMe2)-
(SiMe2DMAP)}H][HB(C6F5)3] ([19][HB(C6F5)3]) in methylene
chloride-d2 optimized for JSiH = 7 Hz and acquired at 185 K. The
experiment is optimized for long-range bonding; thus, there is no cross
peak for the SiH and the cross peak between ZrH and SiMe2DMAP
shows small scalar coupling.
1
correlated with a H NMR signal at 0.21 ppm assigned to
SiMe2. In that experiment, cross peaks from the 29Si signal at
1
Overall, the addition of Lewis acids to Cp2Zr{N(SiHMe2)2}-
X-type compounds results in β-hydrogen abstraction and Si−X
bond formation, accompanied by the formation of nonclassical
Zr↼H−Si structures.
Reactions with Lewis Bases and Hydride Migration.
Remarkably, the cationic disilazide compounds [2][HB-
(C6F5)3] and [2][B(C6F5)4] react with 4-(dimethylamino)-
pyridine (DMAP) to give [Cp2 Zr{N(SiHMe2 )-
(SiMe2DMAP)}H]+ ([19]+; X = HB(C6F5)3, B(C6F5)4) (eq
4); the product contains only one β−Si−H⇀Zr group and a
new zirconium hydride.
−0.5 ppm were correlated in the H dimension to the second
SiMe2 group (0.51 ppm), the signal at 3.95 ppm assigned to the
ZrH, and the aromatic resonances assigned to DMAP (7.93
ppm). The correlations between the silicon atom and the
aromatic signals provide convincing evidence that DMAP is
coordinated to silicon rather than zirconium.
Further evidence for Si−N bond formation is provided by a
1H−15N HMBC experiment, which showed a correlation
between the pyridine nitrogen and the SiMe2 at 0.51 ppm.
1
We assign the H NMR signal at 3.95 ppm to a zirconium
hydride on the basis of a correlation in a COSY experiment
between that signal and the resonance at 0.95 ppm assigned to
β-Si−H⇀Zr. The extent of interaction between the ZrH and
the silicon center in the SiMe2DMAP group is considerably less
than that in [2]+ (1JSiH = 89 Hz). The chemical shift of 3.95
ppm is upfield relative to the signal for the zirconium hydride in
1 (5.60 ppm). However, zirconium hydrides have been assigned
to signals as upfield as 3.12 ppm in Cp2Zr(H)NH2BH3.40
Although spectra were obtained for [19]+ in methylene
chloride-d2, after solutions are heated to 120 °C for 45 min
in that solvent, a reaction occurs to provide [Cp2Zr{N-
A 1H NMR spectrum of [19]+ acquired at room temperature
in methylene chloride-d2 was slightly broad, but the SiMe2
15230
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
(SiMe2DMAP)(SiHMe2)}Cl]+ ([20]+). The identity of [20]+ is
supported by its independent synthesis (see below).
SiMe2 signal in COSY experiments. Two 29Si NMR resonances
were detected at 16.6 and −11.0 ppm with INEPT experiments.
1H−29Si HMBC experiments (optimized for JSiH = 7 Hz)
contained cross peaks between the 29Si signal at 16.6 ppm and
the resonances assigned to one SiMe2 group, OCHMe2, and
ZrH; the upfield 29Si resonance (−11.0 ppm) showed long-
range correlations to the other SiMe2 group, ZrH, and the α-
CH protons of DMAP. The last cross peak supports the
structural assignment involving a DMAP−silicon interaction.
Moreover, the long-range correlations between the Zr−H
proton and both Si atoms (SiMe2DMAP and SiMe2OCHMe2)
in the 1H−29Si HMBC experiment provide support for a Zr−H
group that is formed via DMAP-induced hydrogen elimination.
Interestingly, this migration is chemically reversible. Addition
of B(C6F5)3 to a methylene chloride solution of [22][HB-
(C6F5)3] provides [18][HB(C6F5)3] and (DMAP)B(C6F5)3.
Furthermore, compound [22]+ reacts similarly to related
zirconium hydride 1, as well as Cp2Zr{N(SiMe3)2}H, by
undergoing γ-hydrogen abstraction.17,41 Thus, thermolysis of
[22]+ in a sealed glass tube at 140 °C for 4 h affords H2 and
[Cp2ZrN(SiMe2OCHMe2)SiMe(DMAP)CH2]+ ([23]+; eq 6).
The broad, room-temperature 1H NMR spectrum for
compound [19]+ suggests a slow exchange process involving
the silyl groups, the zirconium hydride, and DMAP. In an EXSY
experiment performed at room temperature, a cross peak
between the SiMe2 and SiHMe2 groups and a cross peak
between ZrH and SiH showed exchange involving hydrogen
transfer between Zr and both silicon centers. The coalescence
temperature for this exchange process is above room
temperature and has not been observed, and the process also
slows down at low temperature. The EXSY experiment
indicates that migration of hydrogen from silicon to zirconium
is reversible. Despite this, the reactivity of [19]+, such as its
conversion to [20]+ and a cyclometalation described later,
occur without apparent loss of DMAP.
In contrast to the chemistry with DMAP, reactions of [2]+
and Et3PO afford [Cp2Zr{N(SiHMe2)2}OPEt3]+ ([21]+; eq 5).
The structure of [21]+ is readily distinguished from that of
1
the DMAP adduct [19]+ by NMR spectroscopy. In the H
NMR spectrum of [21]+, a single resonance assigned to
equivalent SiMe groups was observed at 0.11 ppm (3JHH = 4.2
Hz; 12 H), and a signal at 4.09 ppm (1JSiH = 181.8 Hz; 2 H)
was assigned as a SiH on the basis of its correlation to the
SiMe2 signal in a COSY experiment. The typical chemical shift
1
The H NMR spectrum of [23]+ contained signals assigned
3
to a diastereotopic isopropyl group (1.32 ppm, JHH = 6.1 Hz;
3
1.29 ppm, JHH = 6.1 Hz) and a diastereotopic ZrCH2 group
(1.14 ppm, 2JHH = 13.4 Hz; 0.95 ppm, 2JHH = 13.3 Hz), as well
1
and JSiH values suggest that [21]+ contains terminal silicon
as three SiMe signals (0.44, 0.17, and 0.14 ppm). A H−29Si
1
hydride groups. A single 29Si NMR signal was detected at −19.2
ppm. In addition, compound [19]+ and OPEt3 react to give
[21]+ and free DMAP, while starting materials are observed in
the reaction of [21]+ and DMAP.
HMBC experiment revealed correlations from the stereogenic
Si center to the aromatic α-CH proton of DMAP and the
diastereotopic CH2, which unambiguously confirms the
connectivity between ZrCH2Si and DMAP.
On the basis of the interesting results in reactions of [2]+
with two-electron donors, we also examined reactions of [18]+
with coordinating ligands. The reaction of [18]+ with DMAP in
CH2Cl2 forms [Cp2Zr{N(SiMe2OCHMe2)(SiMe2DMAP)}H]+
([22]+; Scheme 5).
As in the reaction of [2]+ and OPEt3, phosphine oxide
coordinates to the zirconium center in [18]+ to give
[Cp2Zr{N(SiHMe2)(SiMe2OCHMe2)}OPEt3]+ ([24]+; eq 7).
Scheme 5. Reversible Hydrogen Migration between Zr and
Si Controlled by the Addition or Removal of a Two-Electron
Donor
The NMR spectroscopy of [24]+, particularly the SiH group
(δ 4.04, 1JSiH = 187 Hz), distinguishes its connectivity from that
of the silicon-coordinated DMAP species [22]+. In a COSY
experiment, the 1H NMR resonance at 0.12 ppm (6 H)
assigned to the SiMe2 correlated with the SiH resonance at 4.04
ppm. Two 29Si resonances were detected in 29Si INEPT and
1H−29Si HMBC experiments at −5.1 and −21.9 ppm. In the
latter experiments, cross peaks between the 29Si resonance at
−5.1 ppm and the methine proton (3.74 ppm) of the OCHMe2
group and OSiMe2 protons (0.06 ppm, 6 H) provided evidence
for an Si−O−C linkage in the SiOCHMe2 moiety.
1
In contrast to the NMR spectroscopy of [19]+, the H, 13C,
and 29Si NMR spectra of [22]+ were sharp at room
temperature. The H NMR resonances of ZrH, SiMe2DMAP,
and SiMe2OCHMe2 of [22][HB(C6F5)3] were observed at 4.44
(1 H), 0.39 (6 H), and −0.06 ppm (6 H), respectively. The
ZrH signal at 4.44 ppm did not show any correlation to either
1
Halide and Pseudohalide Migration. Reactions of the
nonclassical SiH and bridging triflate compound [12]+ with
pyridine, DMAP, or OPEt3 induce migration of OTf− to
zirconium to provide the series of compounds [Cp2Zr{N-
15231
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
(SiHMe2)(SiMe2L}OTf]+ (L = py ([25]+), DMAP ([26]+),
OPEt3 ([27]+); eq 8). The β-hydrogen does not migrate, and
the nonclassical Zr↼H−Si structure is evident in the products.
The preparation of [28]+ requires elevated temperature to
dissolve and depolymerize paraformaldehyde. In contrast, the
reaction of [2]+ and 2 equiv of acetone occurs within 10 min at
room temperature in bromobenzene or methylene chloride to
give [Cp2ZrN(SiMe2OCHMe2)2]+ ([29]+) (Scheme 6).
Scheme 6. Hydrosilylation of Acetone by [2]+ Followed by a
−
Slower Hydroboration with HB(C6F5)3
1
The H NMR spectra of [25]+−[27]+ contain resonances in
the region 1.27−1.37 ppm assigned to SiH groups; the upfield
1
chemical shift and the JSiH values of 115 Hz establish the
nonclassical Zr↼H−Si structures. 1H−29Si HMBC experiments
revealed two silicon signals for [25]+−[27]+; correlations
between the further downfield signals and aromatic resonances
of pyridine and DMAP established Si−L bond formation.
These correlations are not available for OPEt3, and in that case
we relied on the similarity in other spectral features of [27]+
with those of [25]+ and [26]+. The upfield 29Si NMR
1
The H NMR spectrum of [29]+ contained a multiplet at
4.14 ppm (2 H) and a doublet at 1.51 ppm (3JHH = 6.5 Hz, 12
H) that are characteristic of the isopropoxy group. One 29Si
resonance was detected by 29Si INEPT experiments at 5.2 ppm,
1
resonances correlated in H−29Si HMBC experiments to the
1
upfield H NMR peaks assigned to the SiH in [25]+−[27]+.
1
The 19F NMR chemical shifts of the signals assigned to OTf−
for compounds [25]+−[27] varied only 0.3 ppm from −78.5 to
−78.8 ppm, whereas the signal for bridging OTf− in [12]+
appears at −75.0 ppm. The monodentate OTf− in 11 is −78.2
ppm. Thus, of three possible outcomes that include
coordination to zirconium, coordination to silicon with
hydrogen migration, and coordination to silicon with OTf−
migration, the observed products are consistent with the last
pathway.
and this signal correlated to H NMR signals of SiMe2 (12 H)
and OCHMe2 in 1H−29Si HMBC experiments. Thus, the
spectroscopy unambiguously identified the SiOCHMe2 moiety.
X-ray-quality crystals of [29][HB(C6F5)3] are obtained from
a concentrated methylene chloride solution cooled to −30 °C
(Figure 4). Hydrosilylation of the acetone is confirmed, and the
resulting isopropoxy groups bridge between the silicon and
zirconium centers. The Zr1−O1 and Zr1−O2 distances of
2.385(4) and 2.333(3) Å are similar to other distances in three-
coordinated oxygen centers bonded to silicon and zirconium
Likewise, reaction of chloride-bridged [14]+ and DMAP
gives Cp2Zr{N(SiHMe2)(SiMe2DMAP)}Cl]+ ([20]+). As
mentioned above, [20]+ is also prepared quantitatively through
the thermolysis of [19]+ in methylene chloride. Similarly, the
reaction of [19]+ and MeOTf gives [26]+ and methane.
Hydrosilylation of Carbonyls for the Synthesis of
Cationic [Cp2ZrN(SiMe2OR)2]+. The implications of the
nonclassical structure of [2]+ on reactivity was further explored
with carbonyl reagents. Compound [2]+ and 2 equiv of
paraformaldehyde react at 80 °C in bromobenzene to yield
[Cp2ZrN(SiMe2OMe)2]+ ([28]+; eq 9).
In the 1H NMR spectrum of [28]+, a resonance at 2.99 ppm
Figure 4. ORTEP diagram of the cationic portion of [29]+ with
ellipsoids plotted at the 50% probability level. HB(C6F5)3 and a
CH2Cl2 solvent molecule are not illustrated, nor are the hydrogens on
the C5H5 and methyl groups. Significant interatomic distances (Å):
Zr1−N1, 2.121(4); Zr1−O1, 2.385(4); Zr1−O2, 2.333(3); N1−Si1,
1.642(4); N1−Si2, 1.712(4); Si1−O1, 1.711(4); Si2−O2, 1.699(4).
Selected interatomic angles (deg): Zr1−N1−Si1, 108.5(2); Zr1−N1−
Si2, 104.88(19); Si1−N1−Si2, 146.6(2); O1−Zr1−O2, 130.7(1);
Zr1−N1−Si1−O1, −1.4(2); Zr1−N1−Si2−O2, −2.3(2), O1−Si1−
Si2−O2, −4.0(2).
1
(6 H) in the H NMR spectrum is characteristic of a methoxy
group. The observed single 29Si NMR signal at 9.6 ppm was
>50 ppm farther downfield than the corresponding resonance
at −43 ppm in the starting material [2]+. The formation of a
1
SiOMe group was unambiguously supported by a H−29Si
HMBC experiment that contained a cross peak between the
methoxy group and the silicon center. The methoxy group,
however, is likely bridging silicon and zirconium, as shown
below in the structure of [Cp2ZrN(SiMe2OCHMe2)2]+.
15232
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
where the O center is unambiguously described as an L-type
ligand,42 such as {(C6H11NSiMe2)2O}Zr(CH2Ph)2, which
features a Zr−O distance of 2.381(2) Å.43 The Si1−O1 and
Si2−O2 distances of 1.711(4) and 1.699(4) Å are slightly
longer than the other distances in that compound. For
comparison, the Zr−O distance in neutral Cp2Zr{N-
(SiHMe2)2}OCHMe2 is 1.937(2) Å, while the Zr−O and Si−
O distances in Cp2Zr(OSiMe2CH2Cl)Cl are 1.943(3) and
1.609(3) Å, respectively.44 On the basis of these structural
comparisons, [29]+ is probably best described as a N-
(SiMe2OCHMe2)2 tridentate L2X-type ligand coordinated to
the Zr center through an amide and two silyl ether groups.
Addition of excess acetone (>3 equiv) to [2][HB(C6F5)3]
yields [Cp2ZrN(SiMe2OCHMe2)2][(Me2HCOB(C6F5)3]
([29][Me2HCOB(C6F5)3]) over 2 h (Scheme 6). The
formation of [29][HB(C6F5)3] as an intermediate occurs
within 5 min, followed by slow conversion of [HB(C6F5)3] into
[Me2HCOB(C6F5)3]−. The 1H NMR spectrum of [29]-
[Me2HCOB(C6F5)3] contained two sets of isopropoxy
An alternative resonance structure of the cationic transient
shows localization of charge on the electropositive Zr center, in
which case the intermediate is described as a cationic
zirconium-coordinated silanimine complex (Scheme 7, struc-
ture I.B.). Support for this description is provided by reactions
of β-SiH containing alkyl moieties MC(SiHMe2)3 and
B(C6F5)3, which produce M−H−B(C6F5)3 and disilacyclobu-
tane (Scheme 7, part II).49
The latter species is postulated to form via 2π + 2π
cyclodimerization of the silene intermediate that forms upon β-
hydrogen abstraction. Further support for this description is
given by abstraction of a β-hydrogen from an alkyl ligand in
Cp2Zr{N(SiHMe2)2}R that provides the olefin, as noted above.
Finally, the intermediacy of a coordinated Me2SiNSiHMe2 in
the present system is supported by the selective dimerization
observed in the absence of a reactive M−X group.
Both 1,1-insertions and 1,2-insertions are better described as
migrations of an X-type ligand to an electrophilic, metal-
coordinated carbon center (Scheme 7, part III), as evidenced by
stereochemical studies of insertion reactions.50,51 Thus,
formation of Si−X bonds from the polarized silanimine/
silylium functionality in [Cp2Zr{N(SiMe2)(SiHMe2)}X]+ is
best understood with both resonance structures (I.A. and I.B.)
that highlight polarization of unsaturated moieties as an
important component of insertion reactions.
The influence of the alkyl group R on reactions of
Cp2Zr{N(SiHMe2)2}R and Lewis acids provides data to
compare the interpretation of abstraction. In these reactions,
four pathways may be postulated on the basis of structures of
the products: (a) alkyl group abstraction by the Lewis acid, (b)
β-hydrogen abstraction from the alkyl group, (c) β-hydrogen
abstraction from the disilazido group followed by β-hydrogen
abstraction by the transient silylium electrophile, and (d) β-
hydrogen abstraction from the disilazido group followed by
alkyl group migration (Scheme 8).
resonances in a 2:1 ratio assigned to Me2HCOSi (1.51 ppm,
3
3JHH = 6.5 Hz; 4.14 ppm) and Me2HCOB (0.88 ppm, JHH
=
5.9 Hz; 3.60 ppm). Furthermore, the resonance in the 11B
NMR spectrum at −3.4 ppm is characteristic of formation of an
alkoxyborate moiety.45 Interestingly, the rate of insertion
involving the nonclassical SiH group (10 min) is much faster
than that involving the BH (2 h). Furthermore, conversion of
[HB(C6F5)3] in [29]+ into the [Me2HCOB(C6F5)3] anion is
not detected in the absence of excess acetone.
DISCUSSION
■
Migrations from Zr to Si. The migration of an anionic
group (H, alkyl, halide, triflate, alkoxide) from zirconium to a β-
silicon center can be described as an X group abstraction by a
transient cationic β-silylium electrophile generated through
hydrogen abstraction by a Lewis acid. This description is based
on a resonance structure of the transient where charge is
localized on the silicon center (Scheme 7, structure I.A.).
Additionally, silylium cations are known as strong Lewis acids
in hydride and halide abstractions.46−48
Scheme 8. Possible Pathways for the Interaction of
Cp2Zr{N(SiHMe2)2}R and B(C6F5)3
Scheme 7. Comparisons of (I) Borane-Induced X Group
Migration through a Silanimine, (II) β-Hydrogen
Abstraction To Generate Transient ESiR2, and (III)
Insertion of an Olefin into a M−C Bond
With larger R groups (e.g., n-C3H7) the pathway that
provides Si−C bond formation (alkyl migration) is the most
favored, and alkyl group abstraction products (pathway a) are
not detected. A minor amount of β-CH abstraction (pathways
b and c) is evident by the formation of [2]+ and the
corresponding olefin. It is important that alkyl group migration
is favored over β-CH abstraction. The dominance of alkyl
migration suggests that the silanimine resonance structure is a
more important contributor than the silylium structure in
Scheme 8; the latter would be expected to react by β-hydrogen
abstraction because the well-defined Lewis acid B(C6F5)3 reacts
by β-hydrogen abstraction in this system. Furthermore, the
much larger concentration of charge at Zr than at Si that is
inferred by experiment is supported by the MP2 calculations of
15233
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
[Cp2ZrN(SiHMe2)2]+. The latter predict an order of magnitude
larger positive charge on Zr (1.77) than on Si (0.19), on the
basis of Mulliken populations.
elimination gives saturated products (as opposed to unsaturated
products). While [Cp2ZrMe2AlMe2]+ is proposed to dissociate
AlMe3 to generate active [Cp2ZrMe]+ catalytic species, the only
donors present in such polymerizations are olefins.60 Addition-
ally, a yttrium hydridoaluminate is converted to a [Y(μ-H)]2
species in the presence of excess DIBAH.61 That conversion
occurs without an additional two-electron donor. However, it
should also be noted that the nucleophilicity of a hydrosilane is
proposed to be enhanced by coordination of a Lewis base such
as F−, for example, to facilitate the hydrosilylation of a carbonyl
compound.62 This analogy may extend to the Lewis base-
facilitated hydrogen transfer to an electrophilic Zr center.
Thus, the present example of DMAP coordination to the β-
silicon of the disilazido ligand is distinguished from typical
Lewis base chemistry of bridging aluminates or borates. Despite
the strong nonclassical interactions, it should be noted that β-
hydrogen elimination in transition-metal amido compounds is
rare,12 and β-eliminations of amides from d0 transition-metal
and rare-earth-metal centers are unknown. Still, there are
similarities in the present system to the β-elimination chemistry
of transition-metal alkyls.4a
A related X group migration was also observed for
Cp2Zr{N(SiHMe2)2}X and B(C6F5)3, where X is chloride,
triflate, methoxide, and isopropoxide. In the final products,
bridging Zr−X−Si structures are obtained (compounds [12]+,
[14]+, [17]+, and [18]+, respectively). In these bridging
compounds, the X group on X−SiMe2 behaves as a L-type,
two-electron donor to the electron-deficient Zr center. This
assignment is supported by 29Si NMR spectroscopy and X-ray
diffraction studies of [29]+.
In addition, a recent report documents a related N(SiHMe2)2
migration from a scandium center to a β-Si upon addition of a
Lewis acid, and a similar sequential hydrogen abstraction/
silazide migration sequence was proposed.15 These similarities
indicate that the reactivity pattern described in the current
contribution is not simply limited to the Cp2Zr system.
Migrations from Si to Zr. The entry points to the Si to Zr
migration chemistry in all cases are bridging Si−X−Zr
structures. DMAP induces migration of a monovalent group
from silicon to the zirconium center. The migrating group may
be hydrogen, chloride, or triflate. In the systems in which both
β-H and β-X (X = Cl, OTf) are present, migration of X rather
than H is observed, whereas competition between β-H and β-
OR results in hydrogen migration. Other nucleophilic ligands
coordinate, such as OPEt3 and pyridine, and the binding site
varies between Si and Zr depending on the identity of the
migrating X group. There are similarities in these donor-
assisted migrations to β-elimination as well as to Lewis base
induced cleavage reactions. The comparison between β-
elimination and Lewis acid/base chemistry is evaluated through
analogies to main-group and transition-metal chemistry and
through analysis of the microscopic reverse reaction.
The unsaturated β-elimination products from reactions of
low-valent transition-metal alkyls are stabilized by two-electron
π back-donation to give metallacyclopropyl resonance
structures. In the cationic zirconium disilazido compounds,
two electrons from DMAP serve as a surrogate for metal-based
π back-donation to stabilize the silanimine (Scheme 9).
Scheme 9. Comparison of H Migration Products in (A)
DMAP-Induced Reactions and (B) Low-Valent β-
Eliminations
A related Lewis base induced hydride transfer is reported for
the reaction of the symmetrical hydrogen-bridged, cationic
disilyl species [(μ-H)(Me2Si(CH2)3SiMe2]+ and acetonitrile
that gives [(Me2HSi(CH2)3SiMe2(NCMe)]+.48
In contrast to the [Si−H−Si]+ system, the atoms in the
bridging Zr↼H−Si moieties are inequivalent. In these
structures, a coordinating ligand may interact with the
zirconium center or the silicon center and disrupt the M−
H−Si interaction. These bridging structures, and their
interactions with ligands, may also be compared to M−H−
BRn adducts, which are known for rare-earth-metal, main-
group-metal, and transition-metal complexes. In rare-earth
examples, Cp*2LaHBEt3 and (C5H4CMe3)2SmHBEt3 are
coordinated by THF without displacing the Ln−H−B
interaction, whereas (C5H3(CMe3)2)2CeHBPh3 spontaneously
dissociates BPh3.54 In Zr−H−BRn compounds, it has been
pointed out that Zr−H or B−H cleavage may result, either
spontaneously or through the assistance of a donor such as
Et2O.55 The main-group compound KHB(C6F5)3 reacts with
TMEDA to give (tmeda)KHB(C6F5)3,56 whereas
(C5(SiMe3)3H2)CaHBEt3 is inert to PMe3.57 Thus, the cleavage
site of the M−H−B bridging interaction depends on M and
substituents on the boron center.
β-hydrogen elimination and insertion into a M−H bond are
related by the principle of microscopic reversibility. Likewise,
this Si to Zr migration and its microscopic reverse of Zr to Si
migrations are controlled by the addition of Lewis acids or
Lewis bases. Thus, the zirconium hydride DMAP adduct [22]+
and B(C6F5)3 react to re-form β-SiH-containing [18]+ and
(DMAP)B(C6F5)3.
The migration of halide and triflate from Si to Zr should also
be considered in the context of abstraction vs elimination.
While the lone-pair electrons on X argue for the Lewis acid
model, several other comparisons are worth noting. First, β-Cl
eliminations from metal chloroalkyls limit vinyl chloride
polymerizations.63 Moreover, similar β-X elimination reactions
were observed from Cp*2ScCH2CH2X (X = PPh2, OEt, F).64
In these systems, as well as in the current cationic zirconium
compounds, β-X elimination products are formed instead of β-
H, at least as the isolated products.
52
53
Alternatively, rare-earth-metal tetraalkylaluminate adducts
react with pyridine to form metal alkyls.58,59 Although
aluminum−pyridine adducts are formed in these reactions,
both a [M]−Me bond and an [Al]−Me bond are broken
during the transformation; in addition, both bridging methyl
groups are engaged in electron-deficient bonding, and therefore
Finally, it is remarkable that the nonclassical Zr↼H−Si
moiety in [2]+ behaves as if it is an intermediate on the reaction
coordinate for β-hydrogen elimination. This comparison further
reinforces the description of the reaction as a Lewis base
15234
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
mediated β-hydrogen elimination and highlights the similarities
between these bridging SiH groups and β-agostic alkyls.
(SiHMe2)2}OPEt3]+ ([21]+) with paraformaldehyde or acetone
provided the hydrosilylation products, albeit qualitatively more
slowly than in the absence of OPEt3. Still, OPEt3 is displaced
from zirconium in the final product.
Carbonyl Hydrosilylation. Given the observations of
hydrogen migration between silicon and zirconium centers,
several pathways for the formation of silyl ethers from the
carbonyl compounds may be considered. First (pathway A), a
carbonyl oxygen may coordinate to the zirconium center,
disrupting the nonclassical Zr↼H−Si structure, as was
observed in the interaction of [2]+ and OPEt3. Transfer of
hydrogen to the carbonyl carbon, followed by alkoxide
migration to silicon, would then provide the hydrosilylated
product. Alternatively (pathway B), the carbonyl may
coordinate to the silicon center to induce migration of
hydrogen from silicon to form a ZrH group, which
subsequently is transferred to the carbon of the coordinated
carbonyl. A third possible mechanism (pathway C) again
invokes carbonyl-assisted ZrH formation which is then followed
by insertion of a second carbonyl into the Zr−H bond to give a
zirconium alkoxide. Dissociation of the coordinated carbonyl
from the silicon center, followed by migration of the alkoxide to
silicon, would then provide the product. In general, Lewis acid
assisted hydrosilylation reactions involve hydrogen abstraction
from silane, coordination of the carbonyl oxygen to a silylium
center, and transfer of hydride from silicon to the resulting
carbocation.65 Note that the cationic components of complexes
containing [HB(C6F5)3]− or [B(C6F5)4]− react similarly, and
the hydroboration of the carbonyls by [HB(C6F5)3] is slow
relative to the hydrosilylation. These observations rule out a
B(C6F5)3-catalyzed process and hydride transfer from [HB-
(C6F5)3]−, as proposed by Piers (not shown in Scheme 10).47
Similarly, reactions of DMAP adduct [19]+ and paraformal-
dehyde provide [28]+. Again, these reactions are slower than
the corresponding reaction of [2]+ and the carbonyls,
suggesting that the silicon site is involved in the hydrosilylation
pathway. Interestingly, the additional product [Cp2ZrN-
(SiMe2OCMeCH2)2]+ and H2 are obtained (from reactions
of DMAP adduct [19]+ and acetone); these products
correspond to deprotonation of acetone and formation of an
enolate (the ratio of enolate to insertion is 1.8:1). Enolate
formation is not observed in reactions of [2]+ and acetone or in
reactions of OPEt3 adduct [Cp2Zr{N(SiHMe2)2}OPEt3]+ and
acetone. This observation suggests that the ZrH can react with
acetone through two routes, insertion or deprotonation/enolate
formation. Because enolate formation is not observed in
reactions of [2]+, it seems unlikely that pathway C is dominant.
These experiments suggest that both sites are necessary for
the hydrosilylation. In accordance with this idea, compounds
[25]+−[27]+, which contain OTf bonded to the zirconium
center, a Lewis base coordinated to a silicon center, and a
nonclassical β-SiH are unchanged in the presence of acetone,
even at elevated temperature (85 °C). In those compounds,
excess DMAP does not induce migration of hydrogen to
zirconium.
Pathway A should be available for compounds such as [25]+
that feature a nonclassical Zr↼H−Si structure that could be
disrupted by carbonyl coordination to zirconium. However, no
hydrosilylation is observed with [25]+, and thus we rule out
pathway A. Furthermore, the OPEt3 adduct [21]+, as a model
for the intermediate on pathway A, features classical 2c-2e SiH
moieties that should not be reactive in hydrosilylation.47
Although coordination of a carbonyl to a Lewis acid Mo center
was recently proposed as a pathway for hydrosilylation on the
basis of an isotopic labeling experiment,66 this mechanism
seems unlikely in the current case.
Scheme 10. Possible Pathways for the Hydrosilylation of
Carbonyl Compounds by [2]+
Instead, a pathway in which hydride migrates to zirconium is
favored, as suggested by pathways B and C. The formation of
enolates from the cationic ZrH DMAP adduct [19]+ suggests
that pathway C should also feature enolate formation in
general. Because [2]+ and acetone react to give enolate-free
products, we rule out pathway C for the hydrosilylation of
carbonyls by [2]+.
The remaining mechanism, pathway B, features formation of
a Si−O bond and Zr−H bond, followed by hydrogen transfer
to the electrophilic carbonyl site. The proposed pathway is
similar to that proposed by Abu-Omar and co-workers for the
Re(V)-catalyzed hydrosilylation of carbonyls, on the basis of
elimination of a mechanism involving carbonyl insertion into a
ReV−H bond.67 A related proposed mechanism is described by
Brookhart that features an IrIII−(η1-H-SiR3) group that
transfers R3Si+ to the carbonyl oxygen.68 Notably, the
zirconium center in [2]+ is d0 and thus π back-donation is
not available to stabilize the M−(η2-HSiR3) interaction or to
populate the σ* orbital to assist in the Si−H bond cleavage.
Attempts to distinguish these pathways through kinetic
studies were not successful, because the reaction of [2]+ and
acetone is finished before it can be measured even at 200 K.
Furthermore, kinetic studies were limited by the heterogeneous
nature of the reaction between [2]+ and paraformaldehyde.
Other carbonyl compounds, including bulky aldehydes and
ketones, also were not useful for kinetic studies.
Thus, alternative tests are needed to probe the pathway(s) by
which hydrosilylation occurs. While the Zr−H group in 1 reacts
with acetone and paraformaldehyde to form 15 and 16, the β-
SiH groups in these molecules do not react with either
substrate (at least under the conditions tested). We then
attempted to block the zirconium center by OPEt3 coordina-
tion. In fact, reactions of the OPEt3 adduct [Cp2Zr{N-
CONCLUSION
■
The nonclassical Zr↼H−Si group in compounds [2]+ and its
derivatives provide a connection between the insertion/β-
agostic CH/elimination organometallic reaction pathways and
inorganic Lewis acid/Lewis base chemistry of silylium, boranes,
15235
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
(5) Dawoodi, Z.; Green, M. L. H.; Mtetwa, V. S. B.; Prout, K. J.
Chem. Soc., Chem. Commun. 1982, 802−803.
(6) (a) Nikonov, G. I. In Advances in Organometallic Chemistry; West,
R., Hill, A. F., Stone, F. G. A., Eds.; Academic Press: New York, 2005;
Vol. 53, pp 217−309. (b) Corey, J. Y. Chem. Rev. 2011, 111, 863−
1071.
and trialkylaluminum/tetraalkylaluminate adducts. This com-
parison is important, particularly with respect to the transfer of
anionic hydride and alkyl ligands between Lewis acidic centers.
Furthermore, the addition of a two-electron donor to facilitate
hydrogen migration shows the connection between the two
seemingly unrelated organometallic systems in terms of
reactivity. The transformation from a nonclassical Zr↼H−Si
group to a trapped β-eliminated product is noteworthy.
In addition, the nonclassical structures in [2]+ are central to
the observed carbonyl hydrosilylation reaction. Here, it is seen
that the nonclassical structure facilitates the overall addition
reaction without generating an enolate side product, as is
observed with the cationic zirconium hydride [19]+. The
similarities between the chemistry of [2]+ and the proposed
catalytic mechanisms based on ReV and IrIII catalysts further
support the generality of the nonclassical pathway across the
transition-metal series.
Finally, it is worth considering that the tetramethyldisilazido
ligand used here is commonly employed in f-element chemistry,
as well as in group 2 chemistry. The amine HN(SiHMe2)2 is
relatively acidic,9 and relatively inert compounds are often
obtained. Although the precursors and LnMN(SiHMe2)2
compounds are more easily handled than corresponding alkyls,
they often do not provide direct access to catalytically active
species. The Lewis acid/Lewis base treatment strategy
developed here provides an effective route for the synthesis
of catalytically reactive hydrides from stable precursors.
(7) Tilley, T. D.; Andersen, R. A.; Zalkin, A. Inorg. Chem. 1984, 23,
2271−2276.
(8) (a) Lin, Z. In Contemporary metal boron chemistry. I, Borylenes,
boryls, borane σ-complexes, and borohydrides; Marder, T. B., Lin, Z.,
́
Eds.; Springer: Berlin, 2008; pp 151−202. (b) Besora, M.; Lledos, A.
In Contemporary metal boron chemistry. I, Borylenes, boryls, borane σ-
complexes, and borohydrides; Marder, T. B., Lin, Z., Eds.; Springer:
Berlin, 2008; Vol. 130, pp 203−214.
(9) Eppinger, J.; Spiegler, M.; Hieringer, W.; Herrmann, W. A.;
Anwander, R. J. Am. Chem. Soc. 2000, 122, 3080−3096.
(10) (a) Hartwig, J. F.; Muhoro, C. N.; He, X.; Eisenstein, O.;
Bosque, R.; Maseras, F. J. Am. Chem. Soc. 1996, 118, 10936−10937.
(b) Muhoro, C. N.; He, X.; Hartwig, J. F. J. Am. Chem. Soc. 1999, 121,
5033−5046. (c) Lam, W. H.; Lin, Z. Organometallics 2000, 19, 2625−
2628.
(11) (a) Spence, R. E. v. H.; Parks, D. J.; Piers, W. E.; MacDonald,
M.-A.; Zaworotko, M. J.; Rettig, S. J. Angew. Chem., Int. Ed. Engl. 1995,
34, 1230−1233. (b) Radius, U.; Silverio, S. J.; Hoffmann, R.; Gleiter,
R. Organometallics 1996, 15, 3737−3745. (c) Spence, R. E. v. H.; Piers,
W. E.; Sun, Y.; Parvez, M.; MacGillivray, L. R.; Zaworotko, M. J.
Organometallics 1998, 17, 2459−2469.
(12) Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7010−7011.
(13) Scherer, W.; Wolstenholme, D. J.; Herz, V.; Eickerling, G.;
Bruck, A.; Benndorf, P.; Roesky, P. W. Angew. Chem., Int. Ed. 2010, 49,
̈
2242−2246.
ASSOCIATED CONTENT
* Supporting Information
■
(14) (a) Hanley, P. S.; Hartwig, J. F. J. Am. Chem. Soc. 2011, 133,
15661−15673. (b) Neukom, J. D.; Perch, N. S.; Wolfe, J. P.
Organometallics 2011, 30, 1269−1277. (c) Hanley, P. S.; Markovic,
D.; Hartwig, J. F. J. Am. Chem. Soc. 2010, 132, 6302−6303.
(d) Neukom, J. D.; Perch, N. S.; Wolfe, J. P. J. Am. Chem. Soc.
2010, 132, 6276−6277.
S
Text, tables, and CIF files giving experimental procedures,
crystallographic data, and computational details. This material is
available free of charge via the Internet at http://pubs.acs.org.
(15) Chen, F.; Fan, S.; Wang, Y.; Chen, J.; Luo, Y. Organometallics
2012, 31, 3730−3735.
AUTHOR INFORMATION
Corresponding Author
sadow@iastate.edu
■
(16) Yan, K.; Ellern, A.; Sadow, A. D. J. Am. Chem. Soc. 2012, 134,
9154−9156.
Notes
(17) Yan, K.; Sadow, A. D. Chem. Commun. 2013, 49, 3212−3214.
(18) Herrmann, W. A.; Huber, N. W.; Behm, J. Chem. Ber. 1992, 125,
1405−1407.
The authors declare no competing financial interest.
(19) (a) Herrmann, W. A.; Eppinger, J.; Runte, O.; Spiegler, M.;
Anwander, R. Organometallics 1997, 16, 1813−1815. (b) Klimpel, M.
G.; Gorlitzer, H. W.; Tafipolsky, M.; Spiegler, M.; Scherer, W.;
Anwander, R. J. Organomet. Chem. 2002, 647, 236−244.
(20) (a) Harris, D. H.; Lappert, M. F. In Organometallic Chemistry
Reviews: Organosilicon Reviews; Elsevier: Amsterdam, 1976; p 13.
(b) Bradley, D. C.; Chisholm, M. H. Acc. Chem. Res. 1976, 9, 273−280.
(c) Ghotra, J. S.; Hursthouse, M. B.; Welch, A. J. J. Chem. Soc., Chem.
Commun. 1973, 669−700. (d) Andersen, R. A.; Templeton, D. H.;
Zalkin, A. Inorg. Chem. 1978, 17, 2317−2319. (e) Bradley, D. C.;
Ghotra, J. S.; Hart, F. A.; Hursthouse, M. B.; Raithby, P. R. J. Chem.
Soc., Chem. Commun. 1972, 1225−1226.
ACKNOWLEDGMENTS
■
A.D.S. and K.Y. gratefully thank the National Science
Foundation (CHE-0955635 and MRI-1040098) for financial
support. J.J.D.H. and M.S.G. acknowledge support from
National Science Foundation grant CHE-1147446 and an
AGEP supplement to that grant.
REFERENCES
■
(1) (a) Schmidt, G. F.; Brookhart, M. J. Am. Chem. Soc. 1985, 107,
1443−1444. (b) Doherty, N. M.; Bercaw, J. E. J. Am. Chem. Soc. 1985,
107, 2670−2682. (c) Burger, B. J.; Thompson, M. E.; Cotter, W. D.;
Bercaw, J. E. J. Am. Chem. Soc. 1990, 112, 1566−1577.
(2) Clot, E.; Eisenstein, O. In Principles and Applications of Density
Functional Theory in Inorganic Chemistry II; Kaltsoyannis, N., McGrady,
J. E., Eds.; Springer: Berlin; 2004; Vol. 113, pp 1−36.
(21) Anwander, R.; Runte, O.; Eppinger, J.; Gerstberger, G.;
Herdtweck, E.; Spiegler, M. J. Chem. Soc., Dalton Trans. 1998, 847−
858.
(22) Procopio, L. J.; Carroll, P. J.; Berry, D. H. J. Am. Chem. Soc.
1994, 116, 177−185.
(23) Herrmann, W. A.; Munck, F. C.; Artus, G. R. J.; Runte, O.;
Anwander, R. Organometallics 1997, 16, 682−688.
(24) Wolfsberg, M. Acc. Chem. Res. 1972, 5, 225−233.
(25) See the Supporting Information for full characterization of
[HB(C6F5)3]− and [B(C6F5)4]− salts.
(26) Tilley, T. D. Organometallics 1985, 4, 1452−1457.
(27) Procopio, L. J.; Carroll, P. J.; Berry, D. H. J. Am. Chem. Soc.
1991, 113, 1870−1872.
(3) (a) Green, M. L. H.; Ohare, D. Pure Appl. Chem. 1985, 57, 1897−
1910. (b) Ryabov, A. D. Chem. Rev. 1990, 90, 403−424. (c) Shilov, A.
E.; Shul’pin, G. B. Chem. Rev. 1997, 97, 2879−2932.
(4) (a) Scherer, W.; Herz, V.; Bruck, A.; Hauf, C.; Reiner, F.;
̈
Altmannshofer, S.; Leusser, D.; Stalke, D. Angew. Chem., Int. Ed. 2011,
50, 2845−2849. (b) Brookhart, M.; Green, M. L. H.; Parkin, G. Proc.
Natl. Acad. Sci. U.S.A. 2007, 104, 6908−6914. (c) Crabtree, R. H. In
The Organometallic Chemistry of the Transition Metals, 2nd ed.; Wiley:
New York, 1994; pp 53−54, 188−190.
15236
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237

Journal of the American Chemical Society
Article
(28) (a) Pople, J. A.; Binkley, J. S.; Seeger, R. Int. J. Quantum Chem.
1976, S10, 1−19. (b) Møller, C.; Plesset, M. S. Phys. Rev. 1934, 46,
618−622.
(29) (a) Sakai, Y.; Miyoshi, E.; Klobukowski, M.; Huzinaga, S. J.
Chem. Phys. 1997, 106, 8084−8092. (b) Osanai, Y.; Soejima, E.; Noro,
T.; Mori, H.; Ma San, M.; Klobukowski, M. Chem. Phys. Lett. 2008,
463, 230−234.
(30) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.;
Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.;
Su, S. J.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput.
Chem. 1993, 14, 1347−1363.
(31) Gordon, M. S.; Schmidt, M. W. In Theory and Applications of
Computational Chemistry; Dykstra, C. E., Frenking, G., Kim, K. S.,
Scuseria, G. E., Eds.; Elsevier: Amsterdam, 2005; pp 1167−1189.
(32) Edmiston, C.; Ruedenberg, K. Rev. Mod. Phys. 1963, 35, 457−
465.
(33) (a) Walker, D. A.; Woodman, T. J.; Hughes, D. L.; Bochmann,
M. Organometallics 2001, 20, 3772−3776. (b) Milione, S.; Grisi, F.;
Centore, R.; Tuzi, A. Organometallics 2006, 25, 266−274. (c) Garner,
L. E.; Zhu, H.; Hlavinka, M. L.; Hagadorn, J. R.; Chen, E. Y. X. J. Am.
Chem. Soc. 2006, 128, 14822−14823.
(34) Zhu, J.; Mukherjee, D.; Sadow, A. D. Chem. Commun. 2012, 48,
464−466.
(35) Uhlig, W.; Tretner, C. J. Organomet. Chem. 1994, 467, 31−35.
(36) Luinstra, G. A.; Rief, U.; Prosenc, M. H. Organometallics 1995,
14, 1551−1552.
(54) Werkema, E. L.; Andersen, R. A.; Yahia, A.; Maron, L.;
Eisenstein, O. Organometallics 2009, 28, 3173−3185.
(55) Liu, F.-C.; Liu, J.; Meyers, E. A.; Shore, S. G. Inorg. Chem. 1999,
38, 2169−2173.
(56) Yan, K.; Schoendorff, G.; Upton, B. M.; Ellern, A.; Windus, T.
L.; Sadow, A. D. Organometallics 2013, 32, 1300−1316.
(57) Harvey, M. J.; Hanusa, T. P.; Pink, M. Chem. Commun. 2000,
489−490.
(58) Holton, J.; Lappert, M. F.; Ballard, D. G. H.; Pearce, R.; Atwood,
J. L.; Hunter, W. E. J. Chem. Soc., Dalton Trans. 1979, 54−61.
(59) (a) Watson, P, L.; Herskovitz, T. In Initiation of Polymerization;
Bailey, F. E., Vandenberg, E. J., Blumstein, A., Bowden, M. J., Arthur, J.
C., Lal, J., Ottenbrite, R. M., Eds.; American Chemical Society:
Washington, DC, 1983; ACS Symposium Series 212, pp 459−479.
(b) Watson, P. L.; Parshall, G. W. Acc. Chem. Res. 1985, 18, 51−56.
(60) Bochmann, M.; Lancaster, S. J. Angew. Chem., Int. Ed. 1994, 33,
1634−1637.
(61) Klimpel, M. G.; Sirsch, P.; Scherer, W.; Anwander, R. Angew.
Chem., Int. Ed. 2003, 42, 574−577.
(62) Corriu, R. J. P.; Perz, R.; Rey
1009.
́
e, C. Tetrahedron 1983, 39, 999−
(63) Stockland, R. A., Jr.; Jordan, R. F. J. Am. Chem. Soc. 2000, 122,
6315−6316.
(64) Burger, B. J. Ph.D. Thesis, California Institute of Technology,
1987.
(65) Parks, D. J.; Piers, W. E. J. Am. Chem. Soc. 1996, 118, 9440−
9441.
(37) Cordero, B.; Gom
Echeverría, J.; Cremades, E.; Barragan
2008, 2832−2838.
́
ez, V.; Platero-Prats, A. E.; Reves
́
, M.;
(66) Shirobokov, O. G.; Kuzmina, L. G.; Nikonov, G. I. J. Am. Chem.
Soc. 2011, 133, 6487−6489.
́
, F.; Alvarez, S. Dalton Trans.
(67) (a) Du, G.; Abu-Omar, M. M. Organometallics 2006, 25, 4920−
4923. (b) Du, G.; Fanwick, P. E.; Abu-Omar, M. M. J. Am. Chem. Soc.
2007, 129, 5180−5187.
(38) Chaudhry, S. C.; Kummer, D. J. Organomet. Chem. 1988, 339,
241−252.
(39) Passarelli, V.; Zanella, P. Eur. J. Inorg. Chem. 2004, 4439−4446.
(40) Forster, T. D.; Tuononen, H. M.; Parvez, M.; Roesler, R. J. Am.
Chem. Soc. 2009, 131, 6689−6691.
(41) Simpson, S. J.; Andersen, R. A. Inorg. Chem. 1981, 20, 3627−
3629.
(68) (a) Yang, J.; White, P. S.; Schauer, C. K.; Brookhart, M. Angew.
Chem., Int. Ed. 2008, 47, 4141−4143. (b) Park, S.; Brookhart, M.
Organometallics 2010, 29, 6057−6064.
(42) Green, M. L. H. J. Organomet. Chem. 1995, 500, 127−148.
(43) Male, N. A. H.; Thornton-Pett, M.; Bochmann, M. J. Chem. Soc.,
Dalton Trans. 1997, 2487−2494.
(44) Enders, M.; Fink, J.; Maillant, V.; Pritzkow, H. Z. Anorg. Allg.
Chem. 2001, 627, 2281−2288.
(45) Kidd, R. G. In Boron-11 in NMR of Newly Accessible Nuclei;
Laszlo, P., Ed.; Academic Press: New York, 1983; Vol. 2, Chapter 3.
(46) (a) Lambert, J. B.; Kania, L.; Zhang, S. Chem. Rev. 1995, 95,
1191−1201. (b) Reed, C. A. Acc. Chem. Res. 1998, 31, 325−332.
(c) Scott, V. J.; Celenligil-Cetin, R.; Ozerov, O. V. J. Am. Chem. Soc.
2005, 127, 2852−2853. (d) Reed, C. A. Acc. Chem. Res. 2009, 43,
121−128.
(47) (a) Parks, D. J.; Blackwell, J. M.; Piers, W. E. J. Org. Chem. 2000,
65, 3090−3098. (b) Gevorgyan, V.; Rubin, M.; Benson, S.; Liu, J.-X.;
Yamamoto, Y. J. Org. Chem. 2000, 65, 6179−6186. (c) Rubin, M.;
Schwier, T.; Gevorgyan, V. J. Org. Chem. 2002, 67, 1936−1940.
(48) (a) Muller, T. Angew. Chem., Int. Ed. 2001, 40, 3033−3036.
̈
(b) Panisch, R.; Bolte, M.; Muller, T. J. Am. Chem. Soc. 2006, 128,
̈
9676−9682. (c) Hoffmann, S. P.; Kato, T.; Tham, F. S.; Reed, C. A.
Chem. Commun. 2006, 767−769.
(49) Yan, K.; Upton, B. M.; Ellern, A.; Sadow, A. D. J. Am. Chem. Soc.
2009, 131, 15110−15111.
(50) (a) Collman, J. P.; Hegedus, L. S. In Principles and Applications
of Organotransition Metal Chemistry; Kelly, A., Ed.; University Science
Books: Sausalito, CA, 1980. (b) Calderazzo, F. Angew. Chem., Int. Ed.
Engl. 1977, 16, 299−311. (c) Berke, H.; Hoffman, R. J. Am. Chem. Soc.
1978, 100, 7224−7236.
(51) Brintzinger, H. H.; Fischer, D.; Mulhaupt, R.; Rieger, B.;
̈
Waymouth, R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143−1170.
(52) Evans, W. J.; Perotti, J. M.; Ziller, J. W. Inorg. Chem. 2005, 44,
5820−5825.
(53) Baudry, D.; Dormond, A.; Lachot, B.; Visseaux, M.; Zucchi, G. J.
Organomet. Chem. 1997, 547, 157−165.
15237
dx.doi.org/10.1021/ja407950e | J. Am. Chem. Soc. 2013, 135, 15225−15237