Interaction and activation of carbon-heteroatom π bonds with a Zr/Co heterobimetallic complex

Article abstract of DOI:10.1021/om500217a

Single-electron transfer from the Zr^IVCo^-I heterobimetallic complex (THF)Zr(MesNPⁱPr₂) ₃Co-N₂ (1) to benzophenone was previously shown to result in the isobenzopinacol product [(Ph₂CO)Zr(MesNPⁱPr ₂)₃Co-N₂]₂ (4) via coupling of two ketyl radicals. Thermolysis of 4 led to cleavage of the C=O bond to generate a Zr/Co μ-oxo species featuring an unusual terminal Co=CPh₂ carbene linkage (3). In this work monomeric ketyl radical complexes have been synthesized, and the reactivity of these compounds has been explored. The electronic preference for the formation of a ketyl radical complex or a coordination complex has been investigated computationally. Furthermore, thione substrates were allowed to react with 1, generating new complexes that bind the thione to the Co rather than undergoing single-electron transfer (12, 14). The preference of thiones to coordinate to Co can be overridden if the Co is ligated by CO, in which case a thioketyl radical complex forms (13) analogous to 4. The reaction between 1 and imines resulted in N-H bond activation, affording a μ-methyleneamido Co-H complex (16) that can undergo cyclometalation and loss of H₂ (15).

Full text of DOI:10.1021/om500217a

Article
pubs.acs.org/Organometallics
Interaction and Activation of Carbon−Heteroatom π Bonds with a Zr/
Co Heterobimetallic Complex
Seth L. Marquard, Mark W. Bezpalko, Bruce M. Foxman, and Christine M. Thomas*
Department of Chemistry, Brandeis University, 415 South Street, Waltham, Massachusetts 02454, United States
S
* Supporting Information
ABSTRACT: Single-electron transfer from the ZrIVCo−I
heterobimetallic complex (THF)Zr(MesNPiPr2)3Co-N2 (1)
to benzophenone was previously shown to result in the
isobenzopinacol product [(Ph2CO)Zr(MesNPiPr2)3Co-N2]2
(4) via coupling of two ketyl radicals. Thermolysis of 4 led
to cleavage of the CO bond to generate a Zr/Co μ-oxo
species featuring an unusual terminal CoCPh2 carbene
linkage (3). In this work monomeric ketyl radical complexes
have been synthesized, and the reactivity of these compounds
has been explored. The electronic preference for the formation
of a ketyl radical complex or a coordination complex has been
investigated computationally. Furthermore, thione substrates were allowed to react with 1, generating new complexes that bind
the thione to the Co rather than undergoing single-electron transfer (12, 14). The preference of thiones to coordinate to Co can
be overridden if the Co is ligated by CO, in which case a thioketyl radical complex forms (13) analogous to 4. The reaction
between 1 and imines resulted in N−H bond activation, affording a μ-methyleneamido Co−H complex (16) that can undergo
cyclometalation and loss of H2 (15).
highly reduced heterobimetallic complex (THF)Zr-
(MesNPiPr2)3Co-N2 (1).13−15 We demonstrated that the
CO bond of carbon dioxide is cleaved, affording the μ-
oxo/Co-carbonyl complex 2,13 and recently we have shown
that ketones undergo a similar transformation, yielding the μ-
oxo/Co-carbene complex 3 (Scheme 1).16 The rapid reaction
of 1 with CO2 to cleave a CO bond is of obvious importance
in the context of sustainable strategies for utilizing CO2 as a
useful C1 feedstock. However, the mechanism by which this
INTRODUCTION
■
Understanding and predicting the activation of strong σ and π
bonds by transition-metal complexes remains a formidable
synthetic challenge. Activation of C−H, N−H, O−H, and other
σ bonds has been explored as fundamental steps in cross-
coupling reactions as well as alkane and alkene functionaliza-
tions.1−4 The activation of CC, CN, CO, and CS
bonds could expand the understanding of bond cleavage
reactions and expand the scope of synthetic methods to utilize
these functional groups. Deoxygenation of CO to a surface-
bound methylene is a proposed step in the Fischer−Tropsch
synthesis of hydrocarbons, but the mechanism of this
heterogeneous reaction is not well understood.5 Homogeneous
deoxygenation of ketones has been demonstrated by Ti(III)
complexes,6 tungsten complexes,7,8 and a Zr−Fe heterobime-
tallic species.9 Desulfurization of CS bonds is less well-
known, but a few examples with early and late transition metals
are known. For example, early transition metals such as
titanium(III) desulfurize thiobenzophenone to afford tetraphe-
nylethylene,6 and the oxidative addition of the CS bond
across a Mo−Mo or W−W homobimetallic complex has been
reported.10 The late transition metals ruthenium and iron have
also been shown to react with CS bonds. Tobita reported a
silylene ruthenium complex that reacts with isothiocyanates
resulting in CS bond cleavage,11 and desulfurization of N,N-
dimethylthioformamide to afford an iron carbene was reported
by Nakazawa.12
Scheme 1
Over the past several years our group has focused on the
functionalization and activation of small molecules by the
Received: March 2, 2014
Published: April 9, 2014
© 2014 American Chemical Society
2071
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
Scheme 2
reaction occurs is not well understood and could involve either
a concerted two-electron pathway or a stepwise single-electron-
transfer mechanism. The activation of CO bonds in ketones
may serve as a useful model to understand the activation of
CO2 and may also lead to new manifolds for synthetic
chemistry. In an effort to further explore the activation of
carbon−heteroatom double bonds by our Zr/Co complex, we
have turned our attention to the reaction of 1 with substrates
containing CE (E = O, N, S) bonds to interrogate the effect
of the heteroatom and its substituents on the bond activation
products.
complex 5 is 2059 cm−1, which is higher than that in 4 (2046
cm−1), indicative of an oxidized Co center.
X-ray crystallographic analysis of single crystals of 5 revealed
the monomeric pseudo-C3-symmetric structure, shown in
Figure 1. The Zr−O (2.0169(11) Å) and C−O (1.3265(19)
RESULTS AND DISCUSSION
■
Reactivity of 1 with Ketones. We previously reported that
the stoichiometric addition of benzophenone to 1 afforded the
isobenzopinacol-bridged tetrametallic complex 4 arising
through radical coupling (Scheme 1).17 Similar reactions had
been previously reported and well-studied with well-defined
TiIII complexes.18,19 Under thermal conditions, 4 is proposed to
dissociate into a monomeric ketyl radical complex before the
formation of a μ-oxo/Co-carbene complex (3). We sought to
prepare and isolate a monomeric ketyl radical complex to
validate this hypothesis and to investigate the chemistry of such
Zr-bound ketyl radical intermediates. We were pleased to find
that a monomeric ketyl radical compound could be isolated by
the reaction of complex 1 with fluorenone, affording complex 5
Figure 1. (left) Displacement ellipsoid (50%) representation of
complex 5. Hydrogen atoms have been omitted for clarity. Selected
interatomic distances (Å): Zr1−Co1, 2.6324(3); Zr1−O1,
2.0169(11); O1−C46, 1.3265(19). (right) Computed unpaired spin
density surface of complex 5.
Å) bond distances of 5 are analogous to those in the Zr
fluorenone ketyl radical complex Cp*2ZrCl(fluorenone)
reported by Hou and co-workers (Zr−O = 2.00 Å and C−O
= 1.32 Å).20 The Zr−Co interatomic distance in 5 is 2.6324(3)
Å in comparison to 2.36 Å in 1, suggesting a much weaker
metal−metal interaction upon one-electron oxidation of the
Zr/Co unit. A computational analysis of the electronic structure
of complex 5 at the BP86/LANL2TZ(f)/6-311+G(d)/D95 V
level of theory reveals that one unpaired electron is distributed
on the Co center (1.07e) and one unpaired electron is
delocalized throughout the aromatic π system of the
1
as a dark red crystalline solid (Scheme 2). The H NMR
spectrum of complex 5 was broad and exhibited eight
paramagnetically shifted peaks, suggesting a reasonably
symmetric complex that had not dimerized in a fashion similar
to 4. Consistent with a triplet ground state, the solution
magnetic moment of complex 5 measured using the Evans’
method is 2.5 μB. The stretching frequency of the N2 ligand in
2072
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
fluorenone, with a small amount of spin on Zr (0.09e). A plot
of the spin density is shown in Figure 1.
it becomes weaker, as evidenced by an elongation of the Zr−Co
distance in complex 6 (2.6958(5) Å) in comparison to complex
5 (2.6324(3) Å).
Although the isobenzopinacol-bridged tetrametallic complex
undergoes thermal CO cleavage to afford a Co carbene,16
complex 5 does not undergo the same transformation. This key
difference could be a result of the fused-ring system present in
fluorenone but absent in benzophenone. We have noted other
fundamental differences in the reactivity of other benzophe-
none and fluorenone derivatives (vide infra). We had previously
suggested that a ketyl radical intermediate such as 5 was on the
reaction pathway for catalytic hydrosilylation of ketones and
that C−H bond formation proceeded via abstraction of an H
atom from the PhSiH3 reductant.17 Consistent with this
On the basis of the fundamental differences in the reactivity
between fluorenone and benzophenone, we began to explore
the reaction between 1 and several other benzophenone
derivatives with para substituents in place to prevent radical
coupling. Allowing 1 to react with 4,4′-dimethylbenzophenone
immediately afforded a dark red solution of the paramagnetic (S
= 1) monomeric ketyl radical complex 7. Complex 7 is only
1
transiently stable, and H NMR spectra of samples of 7 were
consistently contaminated with the alkoxide 8 (S = 1/2).
Complex 8 is formed from 7 by hydrogen atom abstraction
from the stoichiometric quantity of THF present from complex
1 (Scheme 2),22 and due to this facile process, we were unable
to isolate complex 7 even at low temperature (−35 °C).
In contrast to the ketone substrates explored above, the
reaction between 1 and bis(4,4′-dimethylamino)benzophenone
afforded the diamagnetic ketone adduct 9 (Scheme 2).
hypothesis, a considerable kinetic isotope effect (kH/kD
=
2.4) was observed. To further support that H atom abstraction
was possible for a ketyl radical, complex 5 was allowed to react
with 1,4-cyclohexadiene at room temperature, resulting in
hydrogen atom abstraction to afford the yellow alkoxide
complex 6 (Scheme 2). In contrast, the reaction of
isobenzopinacol complex 4 with 1,4-cyclohexadiene does not
occur at room temperature and at elevated temperatures
proceeds to carbene complex 3 rather than the alkoxide.
The 1H NMR spectrum of 6 shows nine broad paramagneti-
cally shifted resonances, consistent with a pseudo-3-fold
symmetric structure. The newly formed C−H bond in the
1
Complex 9 was characterized by H, 31P{1H}, and 13C{1H}
NMR experiments, which revealed the replacement of the THF
ligand present in 1 with the diaryl ketone. The IR spectrum of
complex 9 has a diagnostic N2 stretch at 2044 cm−1, confirming
that N2 remains bound to cobalt, as well as a stretch for the
bound ketone at 1601 cm−1, implying that the carbon−oxygen
double bond remains intact. While we were unable to obtain X-
ray-quality crystals of 9, the geometry derived computationally
(vide infra) features a Zr−Co distance (2.53 Å) much shorter
than those in the S = 1 complexes discussed previously.
Thermolysis of 9 did not afford a carbene complex analogous to
3, suggesting that formation of a ketyl radical is a prerequisite
for the CO bond cleavage reaction.
1
alkoxide is clearly visible as a singlet at 9.94 ppm in the H
NMR spectrum. Evans’ method data collected on a solution of
6 reveal a decreased magnetic moment, consistent with a single
unpaired electron remaining on the Co0 center. The stretching
frequency of the N2 ligand present in 6 is 2044 cm−1, which is
consistent with those for other alkoxide complexes.21 X-ray
diffraction was carried out on single crystals of complex 6,
revealing the solid-state structure shown in Figure 2.
From the above results, it is evident that electronic
parameters of the substrate play an important role in dictating
the formation of an S = 1 ketyl radical complex or a simple S =
0 ketone adduct. Evidence for the formation of a Zr−ketone
adduct in the absence of electron transfer (complex 9) suggests
that initial coordination of the ketone to Zr is a key step. The
carbonyl analogue of 1, (THF)Zr(MesNPiPr2)3Co-CO (10), in
which the CO ligand has been found to be far less labile than
N2,23 was used to probe this hypothesis. Allowing benzophe-
none to react with the reduced carbonyl complex 10 afforded a
new paramagnetic yellow complex (11) with a 1H NMR
spectrum similar to that of the isobenzopincaol coupling
product 4 (Scheme 3). The stretching frequency of the CO
ligand is 1889 cm−1 for complex 11, which is in the range of
those for previously reported (alkoxide)Zr(MesNPiPr2)3Co-
CO complexes.23 This result demonstrates that initial
coordination of the substrate occurs at the Zr side of the
molecule. Subsequent electron transfer would then proceed
through the metal−metal interaction to the Zr-bound ketone.
In cases where the diaryl ketone is electron-rich, such as
bis(4,4′-dimethylamino)benzophenone, electron transfer is
disfavored, leading to a diamagnetic ketone adduct (9).
Figure 2. Displacement ellipsoid representation (50%) of complex 6.
Hydrogen atoms and a cocrystallized 2,4,6-trimethylaniline molecule
have been omitted for clarity. A disordered p-Me group of one of the
mesityl rings was adequately modeled, and for clarity only one of two
possible positions is shown (see the Supporting Information for more
crystallographic details). Selected interatomic distances (Å): Zr1−Co1,
2.6958(5); Zr1−O1, 1.9446(18); O1−C46, 1.420(3).
To further investigate the substituent effects on single-
electron transfer to ketones, a series of (Ar2CO)Zr-
(MesNPiPr2)3CoN2 complexes were computationally inves-
tigated to determine the differences in energy between the
singlet ketone adduct (Ar2CO)ZrIV(MesNPiPr2)3Co−IN2
and the triplet ketyl radical (Ar2C•-O)ZrIV(MesNPiPr2)3Co0N2.
Consistent with our experimental observations, the triplet states
of the benzophenone and dimethylbenzophenone derivatives
were calculated to be 3.9 and 2.7 kcal/mol lower in energy than
Comparison of the Zr−O and C−O bond distances in the
solid-state structures of 5 and 6 reveals a contraction of the Zr−
O bond (2.0169(11) Å in 5; 1.9446(18) Å in 6) and an
increase in the C−O bond (1.3265(19) Å in 5; 1.420(3) Å in
6), consistent with the formation of an alkoxide with a C−O
single bond and a more covalent Zr−O interaction. As the Zr−
O interaction becomes stronger, the Zr−Co interaction trans to
2073
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
Scheme 3
the singlet, respectively. The singlet state of the dimethylamino-
substituted derivative, however, was predicted to be favored
over the triplet by 2.8 kcal/mol, in good agreement with our
experimental observations. Several other derivatives were
calculated in silico to explore the trend, revealing that a triplet
state is preferred by 7.0 kcal/mol with p-CONH2 substituents
and the singlet ketone adduct is very slightly favored by 0.6
kcal/mol for the p-NH2 derivative. A plot of the difference in
energy between the S = 0 ketone adduct and the S = 1 ketyl
radical complex as a function of the Hammett parameter of the
ketone para substituent (σp)24 reveals a linear correlation
(Figure 3).
derivative is now only favored by 2.1 kcal/mol, and the singlet
ketone adduct is favored for the p-NMe2 derivative by 4.8 kcal/
mol.
The singlet/triplet preference can easily be explained by
trends in the one-electron-reduction potential of benzophenone
derivatives as a function of para substituent. Single-electron
transfer occurs to generate the triplet ketyl radical complex
when the reduction potential of the ketone is more positive,
while electron transfer does not occur when the ketone is more
electron rich and more difficult to reduce (E1/2 of
benzophenone, −1.84 V; E1/2 of p-NMe2-benzophenone,
−2.16 V). A plot of the difference in energy between the S =
0 and S = 1 states vs the diaryl ketone redox potential also
reveals a linear correlation (Figure 4).25
Reactivity of 1 with Thiones. From the differences noted
between simple ketone derivatives, we decided to investigate
the effect the heteroatom has on the CY activation reaction.
Treatment of 1 with thiobenzophenone afforded the purple
1
diamagnetic complex 12 (Scheme 3). The H NMR spectrum
of 12 reveals an asymmetric product with inequivalent
Figure 3. Plot of the calculated difference in energy between the
singlet ketone adduct ((p-XAr)2CO)ZrIV(MesNPiPr2)3Co−I-L and
the triplet ketyl radical complex (((p-XAr)2 C• -O)-
ZrIV(MesNPiPr2)3Co0-L vs the Hammett parameter (σp)24 of the
para substituent (X), where L = CO (R2 = 0.99), N2 (R2 = 0.99). See
the Experimental Section and Supporting Information for computa-
tional details.
A similar linear correlation between the energetic preferences
for the singlet vs triplet state was observed for the hypothetical
Co carbonyl complexes (Ar2CO)Zr(MesNPiPr2)3Co-CO (Fig-
ure 3). In contrast to the N2-bound complexes, a stronger
preference for the singlet state is observed when the stronger π
acceptor CO is bound to cobalt, resulting in a trend line with a
similar slope but a decreased y intercept. For example, once CO
is bound, the ketyl radical state of the parent benzophenone
Figure 4. Plot of the calculated difference in energy between the
singlet ketone adduct ((p-XAr)2CO)ZrIV(MesNPiPr2)3Co−I-L and
the triplet ketyl radical complex (((p-XAr)2 C• -O)-
ZrIV(MesNPiPr2)3Co0-L vs the E1/2 value of the diaryl ketone,25
where L = CO (R2 = 1.00), N2 (R2 = 0.99).
2074
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
phosphinoamide ligands. The 31P{1H} NMR spectrum of 12
reveals two signals at 49.54 and 15.90 ppm, corresponding to
two ligands bridging the Zr and Co centers, and one ligand
dissociated from Co and η2-bound to the Zr center.
To elucidate the structure of 12, single crystals suitable for X-
ray diffraction were grown. The solid-state structure shown in
Figure 5 was obtained, confirming an asymmetric structure with
sulfur atom, which may contribute to the preference for the
formation of 12, rather than a thioketyl radical.33
Although cleavage of a CS bond is known to occur with
Mo and W complexes, affording μ-sulfide carbene complexes,10
thermolysis of complex 12 did not result in CS bond
cleavage to form a carbene analogous to 3. The thermal stability
and connectivity of 12 contrast sharply with those of the
benzophenone analogue and illustrate the importance of
sequential one-electron-transfer reactions during the CO
bond cleavage reactions facilitated by 1.
The inherent preference for the thiobenzophenone to react
at the Co center can be overridden if the Co center has a
strongly binding CO ligand in place of the more weakly bound
N2 ligand. Allowing thiobenzophenone to react with the
reduced carbonyl complex 10 affords a new paramagnetic
1
yellow complex (13) with a H NMR spectrum similar to that
of the isobenzopinacol coupling product 11 (Scheme 3). This
result demonstrates that one-electron transfer to form a
thioketyl radical complex is possible; however, without the
carbonyl ligand present the preferred site of reactivity is at the
cobalt rather than the zirconium side of the molecule.
Figure 5. Displacement ellipsoid (50%) representation of complexes
12 and 14. Hydrogen atoms, except for H481 and an ether solvate
molecule in the asymmetric unit of 12, have been omitted for clarity.
Selected interatomic distances (Å) for 12: Zr1−Co1, 2.3857(2); Zr1−
S1, 2.6866(4); Co1−S1, 2.2062(4); Co1−C46, 2.0203(14); C46−S1,
1.8073(15); Co1−H481, 2.32. Selected interatomic distances (Å) for
14: Zr1−Co1, 2.7072(5); Zr1−S1, 2.5677(9); Co1−S1, 2.2252(9);
Co1−C46, 2.099(3); C46−S1, 1.787(3).
To explore whether the η2-thiobenzophenone adduct 12
could be driven to isomerize to thioketyl-derived product 13
upon exposure to CO, a solution of complex 12 was exposed to
an atmosphere of carbon monoxide. An immediate color
change from purple to brown occurred, and the new
diamagnetic complex 14 was formed rather than complex 13.
1
an η2-phosphinoamide ligand. Rather than binding to the Zr
center, the thione binds η2 to Co through the CS π bond,
with the S atom in a position bridging the Zr and Co centers.
The C−S bond is elongated (1.8073(15) Å) in comparison to
that in free thiobenzophenone (1.636(9) Å)26 and is within the
range of C−S distances found in Ni, Pt, and V thiobenzophe-
none π complexes reported in the literature (1.76−1.81
Å).27−30 The geometry about the thioketone carbon C46 is
pyramidalized, with an angle of 137° between the C−S bond
vector and the C−C46−C plane. The Zr−S distance
(2.6866(4) Å) is much longer than the Co−S distance
(2.2062(4) Å), suggesting stronger binding to Co, as would
be expected on the basis of hard/soft acid base predictions. In
addition to binding to Co through the C and S atoms, the solid-
state structure reveals a weak agostic interaction between Co
and one of the ortho hydrogen atoms of the thione (Co−H481
= 2.32 Å). The Zr−Co distance in complex 12 remains quite
short (2.3867(2) Å), implying that a substantial Co→Zr
interaction remains despite coordination and partial activation
of the CS bond.
The H NMR spectrum of 14 revealed different phosphinoa-
mide ligand environments indicative of a structure in which one
phosphinoamide ligand remains η2-bound to the Zr center,
similar to the case for complex 12. Further, the IR spectrum of
complex 14 reveals a Co−CO stretch at 1904 cm−1, which is
substantially higher than that observed for 13 (1888 cm−1).
The solid-state structure of 14 confirms its connectivity as a
CO adduct structurally similar to 12 (Figure 5). Notably, the
Zr−Co bond length increases to 2.7072(5) Å in 14 upon
binding of the π-acidic CO ligand. A similar elongation of the
Co−Zr distance has been observed when comparing the
coordinatively unsaturated (THF)Zr(MesNPiPr2)3Co complex
with the CO- and N2-bound derivatives 10 and 1.23,34 The
remainder of the geometrical parameters of 14 are largely
similar to those of 12, with the noticeable absence of an agostic
interaction and a slight contraction of the C−S bond (1.787(3)
Å) as the electron density at the Co center is diminished.
Notably, thermolysis of complex 13 does not lead to the
formation of 14, and likewise, thermolysis of 14 does not lead
to the formation of 13.
Reactivity of 1 with Imines. Given the marked differences
in reactivity between ketones and thioketones, we chose to
expand our investigation to include the reactions between 1 and
imines. In addition to a CN bond polarized toward nitrogen,
imines contain an N−H bond which may undergo activation
upon exposure to the highly reduced Zr/Co complex. Allowing
1 to react with benzophenone imine afforded the dark green
diamagnetic complex 15 (Scheme 4). Much like complex 12,
The isolation of complex 12 rather than a thioketyl radical
complex was unexpected, because the one-electron reduction
potential of thiobenzophenone is ∼0.7 V more positive than
that of benzophenone (E1/2 of thiobenzophenone, −1.17 V;
E
1/2 of benzophenone, −1.84 V).25 However, the preference of
thiobenzophenone to bind to Co may dictate the different
reactivity observed. The decreased nucleophilicity of sulfur in
thiobenzophenone is also a well-known phenomenon in the
organometallic chemistry of thiobenzophenone. Addition of
nucleophiles such as alkyllithiums and Grignard reagents
proceeds via nucleophilic attack at sulfur rather than the
carbon in thioketone derivatives.31,32 Such polarization of the
CS bond would disfavor binding of sulfur to the Lewis acidic
Zr center, favoring addition to the more electron rich Co center
of 1 instead. Furthermore, the radical anions of thioketones
have been shown to have significant spin delocalization on the
1
the H NMR and 31P{1H} NMR spectra of 15 exhibited a 2:1
ratio of inequivalent ligand resonances, indicating an
asymmetric structure with a phosphinoamide ligand η2-bound
to the Zr center. The aryl rings of the benzophenone imine
ligand were spectroscopically inequivalent and no N−H or
Co−hydride signals were observed, suggesting that the
benzophenone ligand may have undergone cyclometalation at
Co.
2075
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
literature (1.89−2.07 Å).37 This elongation is likely the result of
constraints imparted by the bridging nature of the imine anion.
The reaction pathway to form 15 likely involves activation of
the imine N−H bond by Co to form a putative Co−H bond,
followed by cyclometalation at the ortho position of one of the
aryl rings and extrusion of H2. The intermediacy of a Co−H
during the formation of 15 is supported by the reaction
between 1 and fluorenoneimine, which affords a red-purple
diamagnetic complex 16 containing a Co−hydride observable
Scheme 4
1
by H NMR spectroscopy at −14.03 ppm (Scheme 4). The
spectroscopic features of 16 are largely similar to those of 15,
with the exception of the hydride resonance, suggesting a
bridged structure resulting from activation of the imine N−H
bond without cyclometalation. Cyclometalation also does not
occur upon thermolysis of complex 16, likely due to the
decreased flexibility of the fused-ring system.
The structure of complex 16 was determined using single-
crystal X-ray diffraction, confirming the proposed connectivity
(Figure 6). The C−N bond in 16 is identical with that of 15,
with steric constraints imparting a less symmetrically bridged
geometry with slightly elongated Zr−N distance (2.150(4) Å)38
and contracted Co−N distance (1.901(3) Å) in comparison to
that in complex 15. In addition to a Co-bound hydride ligand,
the coordination sphere of Co is completed by an agostic
interaction from one of the ortho hydrogen atoms of the
fluorenone ring with a Co−H distance of ∼2.01 Å. This
structural snapshot reveals a likely intermediate on the pathway
for the ortho-metalation process that occurs with benzophe-
none imine.
Support for the proposed connectivity of 15 was provided by
a solid-state structure determined by X-ray crystallography,
which revealed that cyclometalation of the aryl ring had indeed
occurred (Figure 6). The C−N bond distance in 15 (1.315(6)
From these results, it appears that activation of the N−H σ
bond of the imine substrates is the preferred mode of reaction
rather than the formation of ketiminyl radicals on the Zr side of
the heterobimetallic complex. Allowing 1 to react with an imine
substrate without an N−H bond, N-phenylfluorenone imine, in
an attempt to induce C−N bond cleavage did not result in the
formation of a new complex, and the unreacted starting
materials were observed by 1H NMR spectroscopy. The
increased steric demand of N-phenylfluorenone imine may
make displacement of THF and coordination unfavorable, and
likewise the formation of a η2-bound imine structure similar to
12 may be sterically disfavored, resulting in no reaction with
this substrate.
Figure 6. Displacement ellipsoid (50%) representation of complexes
15 and 16. Hydrogen atoms have been omitted for clarity. Selected
interatomic distances for 15 (Å): Zr1−Co1, 2.6166(9); Zr1−N4,
2.094(4); Co1−N4, 1.945(4); Co1−C48, 2.079(5); N4−C46,
1.315(6). Only one of the two crystallographically independent but
structurally similar molecules in the asymmetric unit of complex 16 is
shown. Selected interatomic distances for 16 (Å, with distances in
second independent molecule in brackets): Zr1−Co1, 2.5087(5)
[2.5217(5)]; Zr1−N4, 2.149(2) [2.123(2)]; Co1−N4, 1.901(2)
[1.903(2)]; C46−N4, 1.306(4) [1.315(4)]; Co1−H571, 2.010
[2.059]; Co1−H1, 1.43 [1.38(4)].
CONCLUSIONS
■
In summary, we have discovered that single-electron reduction
of ketones by the reduced heterobimetallic complex (THF)-
Zr(MesNPiPr2)3Co-N2 (1) cleanly affords ketyl radical
complexes either as transient intermediates en route to radical
coupling products such as 4 or as observable species such as 5
and 7. The electronic parameters of the diaryl ketones dictate
the formation of a simple ketone adduct or a ketyl radical
complex, with more electron donating aryl substituents
disfavoring electron transfer. In contrast to the simple one-
electron reduction observed for some CO bonds, the
reaction between 1 and CS and CN bonds does not
afford thioketyl or ketiminyl radical products. Rather,
thioketones coordinate through the π bond to the cobalt
center (12). If, however, the Co side of the molecule is blocked
by CO, the formation of a thioketyl radical dimer product can
be observed (13). Addition of CO to complex 12 does not
result in conversion to 13, but instead a simple CO adduct 14 is
observed. Finally, we discovered that imines undergo
preferential N−H bond activation over one-electron reduction.
Å) is between that of a single and double bond, in comparison
to a zirconocene complex containing both a [Ph2CN]−
ligand (C−N distance 1.257(5) Å) and a [Ph2HC−NH]−
ligand (C−N distance 1.372(6) Å),35 and is similar to the
C−N distances in a Ti/Co heterobimetallic complex featuring
two bridging [NCHPh]− ligands (1.31(2) and 1.30(2) Å).36
The Zr−N and Co−N distances associated with the
deprotonated imine functionality are similar (2.094(4) and
1.945(4) Å, respectively), suggesting a relatively symmetric
bridging mode. The distance between the Co center and the
bound aryl carbon atom (C48) is 2.079(5) Å, which is long
relative to the typical range of Co−Caryl bonds reported in the
2076
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
mmol) was dissolved in 2 mL of THF. To this solution was added
a solution of (4-Me-C6H4)2CHOH (22 mg, 0.10 mmol) in 2 mL of
THF. After 5 min of swirling, the solvent was removed from the
brown-yellow solution in vacuo. The residue was extracted into
pentane and the extract filtered through a plug of Celite. The brown-
yellow filtrate was concentrated to a final volume of 1 mL and then
was allowed to stand at −35 °C for 12 h. The brownish supernatant
was decanted from the crystalline solid and discarded. The yellow solid
was then dried in vacuo to afford 47 mg (40%) of complex 8. 1H NMR
(400 MHz, C6D6): δ 9.28 (s, 1H), 8.00 (d, J = 8.0 Hz, 4H), 7.53 (s,
6H), 7.15 (d, J = 8.0 Hz, 4H), 5.76 (s, 18H), 3.45 (br s, 6H), 2.72 (s,
18H), 2.33 (s, 6H), 2.17 (s, 9H), −1.71 (s, 18H). Evans’ method (298
K, C6D6): 1.67 μB. UV−vis (C6H6; λ, nm (ε, M−1 cm−1)): 289 (1.6 ×
104), 358 (5.0 × 103). IR (C6H6): 2043 cm−1 (Co-N2). Satisfactory
elemental analysis was collected on complex 8 without the labile N2
ligand, which presumably dissociates prior to analysis. Anal. Calcd for
C60H90N3OP3CoZr: C, 64.78; H, 8.15; N, 3.78. Found: C, 64.67; H,
8.18; N, 3.78.
Activation of the N−H bond occurs to form a Co−H complex
in the case of 16, and this is a proposed intermediate on the
pathway for the formation of cyclometalated benzophenonei-
mine complex 15.
EXPERIMENTAL SECTION
■
General Considerations. All manipulations were carried out
under an inert atmosphere using a nitrogen-filled glovebox or standard
Schlenk techniques unless otherwise noted. All glassware was oven- or
flame-dried immediately prior to use. Diethyl ether was obtained as
HPLC grade without inhibitors; pentane and benzene were obtained
as ACS reagent grade. All protio solvents were degassed by sparging
with ultrahigh-purity argon and dried via passage through columns of
drying agents using a Seca solvent purification system from Pure
Process Technologies. Benzene-d6 was degassed and dried over 4 Å
molecular sieves before use. All 1H NMR spectra were obtained at 400
MHz and recorded relative to residual protio solvent. Complexes 1−4
and 10 were synthesized according to literature procedures.16,17,23,34
All other reagents and solvents were obtained from commercial
sources and used without further purification. Infrared spectra were
recorded on a Varian 640-IR spectrometer controlled by Resolutions
Pro software. UV−vis spectra were recorded on a Cary 50 UV−vis
spectrophotometer using Cary WinUV software. Elemental micro-
analyses were performed by Complete Analysis Laboratories, Inc.,
Parsippany, NJ.
(NMe2-C6H4)2CO-Zr(MesNPiPr2)3Co-N2 (9). In a 20 mL
scintillation vial, (THF)Zr(MesNPiPr2)3CoN2 (67 mg, 0.067 mmol)
was dissolved in 2 mL of tetrahydrofuran. To this solution was added a
solution of (4-NMe2-C6H4)2CO (18 mg, 0.067 mmol) in 2 mL of
tetrahydrofuran. After 5 min of swirling, the solvent was removed in
vacuo from the brown-yellow solution. The residue was triturated with
1
pentane to afford 66 mg (82%) of dark brown complex 9. H NMR
(400 MHz, C6D6): δ 6.74 (s, 6H), 6.29 (br s, 4H), 3.15 (br s, 4H),
2.54 (s, 18H), 2.49 (s, 18H), 2.18 (s, 9H), 1.87 (br s, 18H), 1.62 (brm,
18H). 31P{1H} NMR (162 MHz, C6D6): δ 50.3 (br s). 13C{1H} NMR
(100 MHz, THF-d8): δ 153.37, 149.23, 134.78, 132.26, 131.08, 130.20,
129.18, 127.57, 111.09, 45.44, 39.99, 23.22, 23.15 (two signals
overlapping), 20.38. UV−vis (C6H6; λ, nm (ε, M−1 cm−1)): 298 (3.2 ×
104), 347 (4.8 × 104). IR (C6H6): 2044 cm−1 (Co−N2), 1601 cm−1
(CO). Repeated attempts to obtain satisfactory elemental analysis
results were unsuccessful with this complex, likely the result of its
instability.
(C13H8)-O-Zr(MesNPiPr2)3Co-N2 (5). In a 20 mL scintillation vial,
(THF)Zr(MesNPiPr2)3CoN2 (88 mg, 0.088 mmol) was dissolved in 2
mL of diethyl ether. To this solution was added a solution of
fluorenone (19 mg, 0.11 mmol) in 2 mL of diethyl ether. After 5 min
of swirling, the green solution was allowed to stand at −35 °C for 12 h.
The greenish brown supernatant was decanted from the crystalline
solid and discarded. The red-brown solid was washed once with cold
pentane and then dried in vacuo to afford 76 mg (77%) of crystalline
1
complex 5. H NMR (400 MHz, C6D6): δ 57.72 (s, 2H), 54.70 (s,
[(Ph2CO)Zr(MesNPiPr2)3Co(CO)]2 (11). In a 20 mL scintillation
vial, complex 11 (61 mg, 0.061 mmol) was dissolved in 2 mL of
diethyl ether. To this solution was added a solution of benzophenone
(13 mg, 0.071 mmol) in 2 mL of diethyl ether. After 5 min of swirling
at room temperature, the yellow-orange solution was filtered through a
plug of Celite and the filtrate was allowed to stand at −35 °C for 12 h.
The brownish yellow supernatant was decanted from the yellow
crystalline solid and discarded. The yellow solid was dried in vacuo,
affording 28 mg (40%) of complex 11. 1H NMR (400 MHz, C6D6): δ
9.74 (s), 9.40 (s), 9.03 (s), 8.84 (s), 8.72(s), 8.67 (s), 8.44 (s), 8.18
(s), 8.07 (s), 7.96 (s), 7.87 (s), 7.78 (s), 7.68−7.58 (m), 7.42 (s), 7.29
(s), 6.93 (d, J = 8.0 Hz), 6.53 (br s), 5.93 (d, J = 8.0 Hz), 5.39 (s), 5.24
(br s), 4.85 (s), 4.58 (s), 4.19 (s), 3.56 (d), 3.49 (s), 3.37 (d, J = 8.0
Hz), 3.07 (s), 2.40 (s), 2.22 (s), 2.14 (s), 1.91 (s), 1.80 (s), 0.61 (br s),
0.30 (br s), −0.09 (br s), −0.48 (br s), −1.10 (br s), −2.72 (br s),
−3.15 (br s), −4.00 (br s), −4.80 (br s). UV−vis (C6H6; λ, nm (ε,
2H), 6.71 (s, 6H), 6.34 (s, 18H), 3.86 (s, 5H), 3.00 (s, 2H), 1.97 (s,
9H), −0.84 (s, 18H). UV−vis (C6H6; λ, nm (ε, M−1 cm−1)): 350 (8.0
× 103), 365 (8.3 × 103), 399 (3.9 × 103), 505 (1.4 × 103). Evans’
method (298 K, C6D6): 2.46 μB. IR (C6H6): 2059 cm−1 (Co−N2).
Anal. Calcd for C58H83N5OP3CoZr: C, 62.79; H, 7.54; N, 6.31. Found:
C, 62.70; H, 7.49; N, 6.25.
(C13H9)-O-Zr(MesNPiPr2)3Co-N2 (6). In a 20 mL scintillation vial,
complex 5 (32 mg, 0.029 mmol) was dissolved in 2 mL of
cyclohexadiene. After the mixture was stirred at 25 °C for 16 h, the
initial red color changed to yellow. The volatiles were removed in
vacuo. The yellow residue was dissolved in diethyl ether and the
solution filtered through a plug of Celite. The filtrate was concentrated
to dryness in vacuo, affording 31 mg (99%) of 6 as a yellow powder.
1H NMR (400 MHz, C6D6): δ 9.94 (s, 1H), 7.63−7.76 (m, 6H), 7.56
(s, 6H), 7.45 (m, 2H), 5.77 (s, 18H), 3.06 (br s, 6H), 2.64 (s, 18H),
2.15 (s, 9H), −1.59 (br s, 18H). UV−vis (C6H6; λ, nm (ε, M−1
cm−1)): 288 (6.4 × 104), 311 (1.2 × 104), 347 (6.0 × 103). Evans’
method (298 K, C6D6): 2.02 μB. IR (C6H6): 2044 cm−1 (Co−N2).
Anal. Calcd for C58H84N5OP3CoZr: C, 62.74; H, 7.63; N, 6.31;
Found: C, 62.63; H, 7.74; N, 6.19.
M
−1 cm−1)): 335 (3.4 × 104). Evans method (298 K, C6D6): 2.28 μB.
IR (PhMe): 1890 cm− 1 (Co-CO). Anal. Calcd for
C118H170N6O4P6Co2Zr2: C, 63.76; H, 7.71; N, 3.78; Found: C,
63.75; H, 7.64; N, 3.81.
(η2-MesNPiPr2)Zr(MesNPiPr2)2Co(SCPh2) (12). In a 20 mL
scintillation vial, (THF)Zr(MesNPiPr2)3CoN2 (103 mg, 0.103 mmol)
was dissolved in 2 mL of diethyl ether. To this solution was added a
solution of thiobenzophenone (21 mg, 0.10 mmol) in 2 mL of diethyl
ether. After 10 min of swirling, the initial burgundy color of the
solution changed to dark purple-red. The ether was removed in vacuo,
and the residue was extracted into pentane. The residue was then
extracted into benzene, and the solution was filtered through a plug of
Celite. The filtrate was then frozen, and the benzene was lyophilized in
vacuo, affording 100 mg (88%) of complex 12. X-ray-quality crystals
(4-Me-C6H4)2CO-Zr(MesNPiPr2)3Co-N2 (7). In a 20 mL scintilla-
tion vial, (THF)Zr(MesNPiPr2)3CoN2 (22 mg, 0.022 mmol) was
dissolved in 0.5 mL of C6D6. To this solution was added a solution of
(4-CH3-C6H4)2CO (4.6 mg, 0.022 mmol) in 1 mL of C6D6. After 30 s
of swirling, the red solution was analyzed immediately due to the
inherent instability of complex 7. After prolonged standing, red
complex 7 was transformed into yellow complex 8. 1H NMR (selected
signals, spectrum is contaminated with the alkoxide generated by H
atom abstraction from the THF present in the starting compound, 400
MHz, C6D6): δ 8.19 (s, 1H), 7.40 (s, 2H), 7.06 (s, 6H), 4.16 (br s,
12H), 2.74 (s, 18H), 2.34 (s, 9H), 2.18 (s, 6H), 0.21 (br s, 24H). IR
(C6H6): 2045 cm−1 (Co-N2). Due the instability of complex 7,
satisfactory elemental analysis data was not obtained.
1
were grown from a concentrated solution of 12 in pentane. H NMR
(400 MHz, C6D6): δ 7.43 (d, J = 8.0 Hz, 2H), 7.32 (d, J = 8.0 Hz, 2H),
7.09 (m, 2H), 7.03 (d, J = 8.0 Hz, 1H), 6.99 (br s, 2H), 6.92 (m, 2H),
6.87 (br s, 1H), 6.76 (br s, 1H), 6.71 (br s, 2H), 3.14 (d, J = 8.0 Hz,
1H), 2.98 (s, 3H), 2.96 (s, 3H), 2.75 (s, 3H), 2.61 (s, 3H), 2.36 (s,
(4-Me-C6H4)2CHO-Zr(MesNPiPr2)3Co-N2 (8). In a 20 mL
scintillation vial, (THF)Zr(MesNPiPr2)3CoN2 (103 mg, 0.102
2077
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
1H), 2.18 (s, 6H), 2.14 (s, 3H), 2.11 (s, 3H), 1.78 (s, 3H), 1.66 (dd, J
= 20.0, 8.0 Hz, 3H), 1.52 (m, 3H), 1.43 (dd, J = 20.0, 8.0 Hz, 3H),
1.34 (dd, J = 20.0, 8.0 Hz, 3H), 1.23 (m, 3H), 1.10 (m, 3H), 1.04−
0.98 (m, 6H), 0.96−0.79 (m, 14H), 0.47 (dd, J = 20.0, 8.0 Hz, 3H).
31P{1H} NMR (162 MHz, C6D6): δ 49.54 (br s), 15.90 (s). 13C{1H}
NMR (100 MHz, THF-d8): δ 165.37 (d, J = 2.0 Hz), 153.29, 150.19,
148.10 (d, J = 13.0 Hz), 147.70, (d, J = 14.0 Hz), 146.62 (d, J = 14.0
Hz), 133.71 (d, J = 3.0 Hz), 133.61 (d, J = 3.0 Hz), 133.19 (d, J = 3.0
Hz), 132.76, 132.12 (d, J = 3.0 Hz), 130.96, 130.81 (d, J = 2.0 Hz),
130.37 (d, J = 2.0 Hz), 130.13 (d, J = 2.0 Hz), 128.15, 127.99 (d, J =
13.0 Hz), 127.78 (d, J = 10.0 Hz), 127.04, 126.52, 125.66, 125.53,
121.53, 39.48, 35.55(d, J = 4.0 Hz), 34.91 (d, J = 10.0 Hz), 33.14 (d, J
= 10.0 Hz), 33.13, 32.67 (d, J = 16.0 Hz), 32.36 (d, J = 7.0 Hz), 22.56
(dd, J = 30.0, 4.0 Hz), 22.33, 21.95, 21.77 (d, J = 11.0 Hz), 21.28,
21.18 (d, J = 11.0 Hz), 20.92, 20.74 (d, J = 7.0 Hz), 20.41, 20.40 (d, J
= 18.0 Hz), 19.95 (dd, J = 30.0, 8.0 Hz), 19.48 (d, J = 12.0 Hz), 18.81,
18.64 (d, J = 7.0 Hz), 18.44 (d, J = 9.0 Hz), 17.90 (d, J = 9.0 Hz),
12.47. UV−vis (C6H6; λ, nm (ε, M−1 cm−1)): 350 (8.7 × 103), 500
(1.7 × 103). Anal. Calcd for C58H85N3P3SCoZr: C, 63.36; H, 7.79; N,
3.82. Found: C, 63.19; H, 7.84; N, 3.56.
MHz, C6D6): δ 78.04 (br s), 54.85 (br s), 22.75 (s). 13C{1H} NMR
(100 MHz, THF-d8): δ 214.47, 155.80, 149.20 (d, J = 13.0 Hz),
134.79, 134.03, 133.12, 132.82, 130.29, 130.11, 129.90, 129.33, 129.06,
128.10, 124.67, 73.12 (t, J = 9.0 Hz), 35.13 (d, J = 8.0 Hz), 34.90,
34.31, 28.53 (d, J = 18.0 Hz), 23.32 (dd, J = 32.0, 7.0 Hz), 23.03, 22.46
(d, J = 7.0 Hz), 21.61, 20.47, 20.33, 18.92 (d, J = 18.0 Hz), 17.57 (d, J
= 8.0 Hz), 15.51, 14.21. IR (C6H6): 1904 cm−1 (Co-CO). UV−vis
(C6H6; λ, nm (ε, M−1 cm−1)): 329 (1.1 × 104), 418 (3.9 × 103). Anal.
Calcd for C59H85N3OP3SCoZr: C, 62.85; H, 7.60; N, 3.73. Found: C,
62.71; H, 7.74; N, 3.80.
(η2-MesNPiPr2)Zr(μ-NC(C12H9))(MesNPiPr2)2Co (15). In a 20
mL scintillation vial, (THF)Zr(MesNPiPr2)3CoN2 (79 mg, 0.079
mmol) was dissolved in 2 mL of diethyl ether. To this solution was
added a solution of benzophenone imine (16 mg, 0.088 mmol) in 2
mL of diethyl ether. After 5 min of swirling, the blue-green solution
was concentrated to dryness. The residue was then extracted into
benzene, and the solution was filtered through a plug of Celite. The
filtrate was then frozen, and the benzene was lyophilized in vacuo,
affording 62 mg (72%) of complex 15. X-ray-quality crystals were
1
grown from a concentrated solution of 15 in pentane. H NMR (400
[(Ph2CS)Zr(MesNPiPr2)2Co(CO)] (13). Complex 13 was synthe-
sized from a solution of (THF)Zr(MesNPiPr2)3CoCO (10) generated
in situ using the following procedure. In a 100 mL resealable Schlenk
tube containing a stir bar, (THF)Zr(MesNPiPr2)3CoN2 (156 mg,
0.156 mmol) was dissolved in 40 mL of tetrahydrofuran.
Approximately half of the volatiles were removed in vacuo, resulting
in a color change from the red N2-ligated species to the blue-green N2-
free species. The reaction vessel was resealed, frozen, and back-filled
with excess CO(g). As the solution thawed and was stirred, it turned
to a red-orange color characteristic of complex 10. The solution was
stirred for 10 min before being refrozen and evacuated. After the
reaction mixture was rethawed, volatiles were removed in vacuo. The
residue was dissolved in pentane and filtered through a plug of Celite.
To the red filtrate was added a blue solution of thiobenzophenone (32
mg, 0.16 mmol) in 2 mL of pentane. After 5 min of swirling at room
temperature, the brown-yellow solution was allowed to stand at −35
°C for 24 h to complete crystallization. The dark brown supernatant
was decanted from the yellow-orange crystalline solid and discarded.
The solid was dried in vacuo, affording 72 mg (41%) of complex 13.
1H NMR (400 MHz, C6D6): δ 9.36, 9.13, 8.34, 8.18, 8.07, 7.91, 6.91,
6.78, 6.56, 6.30, 5.40, 5.18, 4.42, 4.34, 3.81, 3.50, 3.17, 3.08, 3.00, 2.73,
2.49, 2.28, 2.13, 1.83, 1.70, 1.56, 0.28, −0.62, −1.91, −2.19, −2.74,
−3.43, −4.94. Evans’ method (298 K, THF-d8): 1.70 μB. UV−vis
(C6H6; λ, nm (ε, M−1 cm−1)): 315 (3.6 × 104), 418 (4.6 × 103). IR
( C 6 H 6 ) : 1 8 8 8 c m − 1 ( C o - C O ) . A n a l . C a l c d f o r
C118H170N6O2P6S2Co2Zr2: C, 62.85; H, 7.60; N, 3.73. Found: C,
62.71; H, 7.65; N, 3.64.
MHz, C6D6): δ 8.39 (br s, 1H), 7.92 (br s, 1H), 7.47 (br s, 1H), 7.28
(d, J = 8.0 Hz, 2H), 7.24 (m, 1H), 7.07 (m, 1H), 6.93 (m, 2H), 6.87
(s, 2H), 6.75 (s, 2H), 6.47 (s 2H), 2.75 (s, 6H), 2.29 (s, 6H), 2.17 (s,
3H), 2.16 (s, 6H), 2.09 (s, 6H), 1.63 (m, 6H), 1.49 (m, 2H), 1.26 (m,
1H), 0.97 (m, 6H), 0.82−0.68 (m, 21H), 0.38 (m, 6H). 31P{1H} NMR
(162 MHz, C6D6): δ 69.95 (br s), 22.35 (s). 13C{1H} NMR (100
MHz, THF-d8): δ 171.62 (dt, J = 3.0, 8.0 Hz), 151.43, 150.07 (t, J =
7.0 Hz), 149.64 (d, J = 11.0 Hz), 143.44, 142.62 (d, J = 10.0 Hz),
135.44, 134.35 (d, J = 24.0 Hz), 133.22, 132.52 (d J = 23.0 Hz),
131.29, 130.37, 130.17 (d, J = 7.0 Hz), 129.94 (d, J = 15.0 Hz), 129.59,
129.18, 128.71, 128.65 (d, J = 3.0 Hz), 128.34 (d, J = 24.0 Hz) 127.90,
127.69, 126.71, 126.05 (d, J = 9.0 Hz), 34.46 (d J = 7.0 Hz), 32.43 (t, J
= 11.0 Hz), 31.86, 28.67 (d, J = 17.0 Hz), 24.36, 23.32 (d, J = 32.0
Hz), 22.53, 21.82 (d, J = 16.0 Hz), 20.61 (d, J = 14.0 Hz), 20.44,
20.13, 19.08 (d, J = 20.0 Hz), 18.91, 18.40, 17.74 (d, J = 10.0 Hz).
UV−vis (C6H6; λ, nm (ε, M−1 cm−1)): 288 (1.5 × 104), 344 (5.2 ×
103). Anal. Calcd for C58H84N4P3CoZr: C, 64.48; H, 7.84; N, 5.19.
Found: C, 64.50; H, 7.95; N, 5.11.
(η2-MesNPiPr2)Zr(μ-NC(C12H8))(MesNPiPr2)2Co(H) (16). In a
20 mL scintillation vial, (THF)Zr(MesNPiPr2)3CoN2 (87 mg, 0.087
mmol) was dissolved in 2 mL of diethyl ether. To this solution was
added a solution of fluorenone imine (16 mg, 0.089 mmol) in 2 mL of
diethyl ether. After 10 min of swirling, the initially blue solution
changed to dark purple. The ether was removed in vacuo, and the
residue was extracted into pentane. The residue was then extracted
into benzene, and the solution was filtered through a plug of Celite.
The filtrate was then frozen, and the benzene was lyophilized in vacuo,
affording 62 mg (72%) of complex 16. X-ray-quality crystals were
(η2-MesNPiPr2)Zr(MesNPiPr2)2Co(SCPh2)(CO) (14). In a 100
mL resealable Schlenk tube containing a stir bar, (THF)Zr-
(MesNPiPr2)3CoN2 (72 mg, 0.072 mmol) was dissolved in 40 mL
of diethyl ether. To this solution was added a solution of
thiobenzophenone (16 mg, 0.078 mmol) in 2 mL of diethyl ether.
After 10 min of swirling, the initial burgundy color of the solution
changed to dark purple-red. Approximately half of the volatiles were
removed in vacuo, and the flask was sealed and removed from the
glovebox. The reaction vessel was then frozen in N2(l) and back-filled
with excess CO(g). As the solution thawed and was stirred, it turned
brown. The solution was stirred for 20 min at room temperature
before being refrozen and evacuated. After thawing, the reaction
mixture was reintroduced to the glovebox, and the volatiles were
removed in vacuo. The brown residue was dissolved in pentane, and
the solution was filtered through a plug of Celite. The filtrate was
concentrated to a final volume of about 2 mL of pentane and then
allowed to stand at −35 °C for 16 h to complete crystallization. The
light brown supernatant was decanted from the dark brown crystalline
material, affording 39 mg (48%) of complex 14. 1H NMR (400 MHz,
THF-d8): δ 7.47 (br s, 3H), 6.83 (s, 2H), 6.80 (br s, 2H), 6.75 (s, 2H),
6.64 (s, 3H) 2.65 (s, 6H), 2.55 (m, 3H), 2.29 (s, 2H), 2.20 (s, 4H),
2.13 (s, 7H), 1.48 (m, 7H), 1.32−1.21 (m, 14H), 1.12−1.05 (m, 3H),
1.00 (m, 8H), 0.87 (m, 13H), 0.67 (br s, 6H). 31P{1H} NMR (162
1
grown from a concentrated solution of 16 in pentane. H NMR (400
MHz, C6D6): δ 7.54 (d, J = 8.0 Hz, 1H), 7.49 (m, 2H), 7.36 (m, 1H),
7.22 (brm, 2H), 6.87 (s, 2H), 6.86 (s, 3H), 6.76 (s, 2H), 3.35 (d, J =
8.0 Hz, 1H), 2.66 (s, 6H), 2.43 (s, 6H), 2.32 (s, 6H), 2.29 (s, 3H),
2.15 (s, 6H), 1.64−1.58 (m, 8H), 0.96−0.91 (m, 7H), 0.90−0.84 (m,
13H), 0.74−0.68 (m, 14H), −14.03 (t, J = 64 Hz, 1H). 31P{1H} NMR
(162 MHz, C6D6): δ 76.36 (br s), 24.34 (s). 13C{1H} NMR (100
MHz, THF-d8): δ 166.85, 150.01 (d, J = 10.0 Hz), 149.56 (t, J = 6.0
Hz), 141.13, 140.62, 139.67, 134.38 (t, J = 5.0 Hz), 134.51, 134.16,
133.70 (d, J = 2.0 Hz), 132.38, 131.94, 130.50, 129.80, 128.86, 128.10,
126.32, 125.69, 125.10, 120.44 (d, J = 21.0 Hz), 119.30, 34.88, 34.46,
30.66 (d, J = 13.0 Hz), 30.46, 23.70 (d, J = 4.0 Hz), 23.16, 23.04,
22.61, 21.98, 21.42 (d, J = 15.0 Hz), 20.92 (t, J = 6.0 Hz), 20.69, 20.63,
20.49, 18.07 (d, J = 14.0 Hz), 14.23. UV−vis (C6H6; λ, nm (ε, M−1
cm−1)): 336 (4.5 × 103), 467 (2.0 × 103), 494 (1.9 × 103). Note: a
Co-H stretching vibration was not observed by IR spectroscopy in
solution (C6H6). Anal. Calcd for C58H84N4P3CoZr: C, 64.48; H, 7.84;
N, 5.19. Found: C, 64.38; H, 7.76; N, 5.13.
X-ray Crystallography. All operations were performed on a
Bruker-Nonius Kappa Apex2 diffractometer, using graphite-mono-
chromated Mo Kα radiation. All diffractometer manipulations,
including data collection, integration, scaling, and absorption
2078
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079

Organometallics
Article
corrections, were carried out using the Bruker Apex2 software.39
Preliminary cell constants were obtained from 3 sets of 12 frames.
Fully labeled diagrams and data collection and refinement details are
included in Tables S1 and S2 and on pages S14−S27 of the Supporting
Information.
Computational Details. All calculations were performed using
Gaussian09, Revision A.02, for the Linux operating system.40 Density
functional theory calculations were carried out using a combination of
Becke’s 1988 gradient-corrected exchange functional41 and Perdew’s
1986 electron correlation functional42 (BP86). A mixed-basis set was
employed, using the LANL2TZ(f) triple-ζ basis set with effective core
potentials for cobalt and zirconium,43 Gaussian09’s internal 6-
311+G(d) set for heteroatoms (nitrogen, oxygen, and phosphorus),
and Gaussian09’s internal LANL2DZ basis set (equivalent to D95V44)
for carbon and hydrogen. Using crystallographically determined
geometries as a starting point, the geometries were optimized to a
minimum, followed by analytical frequency calculations to confirm that
no imaginary frequencies were present.
(13) Krogman, J. P.; Foxman, B. M.; Thomas, C. M. J. Am. Chem. Soc.
2011, 133, 14582−14585.
(14) Napoline, J. W.; Bezpalko, M. W.; Foxman, B. M.; Thomas, C.
M. Chem. Commun. 2013, 49, 4388−4390.
(15) Thomas, C. M.; Napoline, J. W.; Rowe, G. T.; Foxman, B. M.
Chem. Commun. 2010, 46, 5790−5792.
(16) Marquard, S. L.; Bezpalko, M. W.; Foxman, B. M.; Thomas, C.
M. J. Am. Chem. Soc. 2013, 135, 6018−6021.
(17) Zhou, W.; Marquard, S. L.; Bezpalko, M. W.; Foxman, B. M.;
Thomas, C. M. Organometallics 2013, 32, 1766−1772.
(18) Covert, K. J.; Wolczanski, P. T.; Hill, S. A.; Krusic, P. J. Inorg.
Chem. 1992, 31, 66−78.
(19) Covert, K. J.; Wolczanski, P. T. Inorg. Chem. 1989, 28, 4565−
4567.
(20) Hou, Z.; Fujita, A.; Koizumi, T.-a.; Yamazaki, H.; Wakatsuki, Y.
Organometallics 1999, 18, 1979−1985.
(21) Napoline, J. W.; Krogman, J. P.; Shi, R.; Kuppuswamy, S.;
Bezpalko, M. W.; Foxman, B. M.; Thomas, C. M. Eur. J. Inorg. Chem.
2013, 3874−3882.
(22) To test this hypothesis, a control reaction was conducted by the
addition of 1 drop of d8-THF to a benzene-d6 solution of 7, which
resulted in immediate conversion to 8 with a substantially diminished
1H NMR signal for the alkoxide at 9.28 ppm.
ASSOCIATED CONTENT
* Supporting Information
■
S
Text, figures, and tables giving experimental procedures, 1H and
13C{1H} NMR spectra of 5−9 and 11−16, computational
details, and X-ray crystallographic data collection and refine-
ment details, CIF files giving crystallographic data, and a text
file of all computed molecule Cartesian coordinates in a format
for convenient visualization. This material is available free of
charge via the Internet at http://pubs.acs.org.
(23) Krogman, J. P.; Bezpalko, M. W.; Foxman, B. M.; Thomas, C.
M. Inorg. Chem. 2013, 52, 3022−3031.
(24) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165−195.
(25) Loutfy, R. O.; Loutfy, R. O.; Taft, R. W. Chem. Rev. 1991, 91,
165−195.
(26) Rindorf, G.; Carlsen, L. Acta Crystallogr., Sect. B 1979, 35,
1179−1182.
(27) Weisheit, T.; Petzold, H.; Gorls, H.; Mloston, G.; Weigand, W.
̈
AUTHOR INFORMATION
Corresponding Author
*E-mail for C.M.T.: thomasc@brandeis.edu.
■
Eur. J. Inorg. Chem. 2009, 2009, 3545−3551.
(28) Beck, R.; Sun, H.; Li, X.; Camadanli, S.; Klein, H.-F. Eur. J. Inorg.
Chem. 2008, 2008, 3253−3257.
Notes
(29) Weigand, W.; Wunsch, R.; Robl, C.; Mloston, G.; Noth, H.;
̈
̈
Schmidt, M. Z. Naturforsch., B 2000, 55, 453−458.
The authors declare no competing financial interest.
(30) Pasquali, M.; Leoni, P.; Floriani, C.; Chiesi-Villa, A.; Guastini, C.
Inorg. Chem. 1983, 22, 841−844.
ACKNOWLEDGMENTS
■
(31) Beak, P.; Worley, J. W. J. Am. Chem. Soc. 1970, 92, 4142−4143.
(32) Beak, P.; Worley, J. W. J. Am. Chem. Soc. 1972, 94, 597−604.
(33) Alberti, A.; Benaglia, M.; Macciantelli, D.; Marcaccio, M.;
Olmeda, A.; Pedulli, G. F.; Roffia, S. J. Org. Chem. 1997, 62, 6309−
6315.
(34) Greenwood, B. P.; Rowe, G. T.; Chen, C.-H.; Foxman, B. M.;
Thomas, C. M. J. Am. Chem. Soc. 2010, 132, 44−45.
(35) Lefeber, C.; Arndt, P.; Tillack, A.; Baumann, W.; Kempe, R.;
Burlakov, V. V.; Rosenthal, U. Organometallics 1995, 14, 3090−3093.
(36) Zippel, T.; Arndt, P.; Ohff, A.; Spannenberg, A.; Kempe, R.;
Rosenthal, U. Organometallics 1998, 17, 4429−4437.
(37) Based on a February 2014 search of the Cambridge structural
database.
(38) For clarity, the distances in only one of the two crystallo-
graphically independent molecules in the asymmetric unit of 15 are
discussed.
(39) Apex 2, Version 2 User Manual, M86-E01078; Bruker Analytical
X-ray Systems, Madison, WI, 2006.
(40) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Braone, V.; Mennucci,
B.; Petersson, G. A., et al. Gaussian09, Revision A.02; Gaussian, Inc.,
Wallingford, CT, 2009. See the Supporting Information for the full
reference.
(41) Becke, A. D. Phys. Rev. A 1988, 38, 3098−3100.
(42) Perdew, J. P. Phys. Rev. B 1986, 33, 8822−8824.
(43) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299−310.
(44) Dunning, T. H., Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., Ed.; Plenum: New York, 1976; pp 1−28.
We acknowledge financial support from the U.S. Department of
Energy under Award DE-SC0004019. C.M.T. is grateful for a
2011 Sloan Research Fellowship. Funding for S.L.M. was
provided by the Camille and Henry Dreyfus Postdoctoral
Program in Environmental Chemistry.
REFERENCES
■
(1) Milstein, D.; Calabrese, J. C.; Williams, I. D. J. Am. Chem. Soc.
1986, 108, 6387−6389.
(2) Blum, O.; Milstein, D. J. Am. Chem. Soc. 2002, 124, 11456−
11467.
(3) Li, Y.; Xie, Y.; Zhang, R.; Jin, K.; Wang, X.; Duan, C. J. Org. Chem.
2011, 76, 5444−5449.
(4) Sevov, C. S.; Zhou, J.; Hartwig, J. F. J. Am. Chem. Soc. 2014, 136,
3200−3207.
(5) Sage, V.; Burke, N. Catal. Today 2011, 178, 137−141.
(6) McMurry, J. E.; Fleming, M. P.; Kees, K. L.; Krepski, L. R. J. Org.
Chem. 1978, 43, 3255−3266.
(7) Chisholm, M. H.; Folting, K.; Klang, J. A. Organometallics 1990,
9, 602−606.
(8) Bryan, J. C.; Mayer, J. M. J. Am. Chem. Soc. 1987, 109, 7213−
7214.
(9) Lutz, M.; Haukka, M.; Pakkanen, T. A.; Gade, L. H.
Organometallics 2001, 20, 2631−2634.
(10) Budzichowski, T. A.; Chisholm, M. H.; Folting, K. Chem. Eur. J.
1996, 2, 110−117.
(11) Ochiai, M.; Hashimoto, H.; Tobita, H. Organometallics 2012, 31,
527−530.
(12) Fukumoto, K.; Sakai, A.; Oya, T.; Nakazawa, H. Chem. Commun.
2012, 48, 3809−3811.
2079
dx.doi.org/10.1021/om500217a | Organometallics 2014, 33, 2071−2079