Evidence for the existence of terminal scandium imidos: Mechanistic studies involving imido-scandium bond formation and c-h activation reactions

Article abstract of DOI:10.1021/ja307279r

The anilide-methyl complex (PNP)Sc(NH[DIPP])(CH₃) (1) [PNP ^- = bis(2-diisopropylphosphino-4-tolyl)amide, DIPP = 2,6-diisopropylphenyl] eliminates methane (k_avg = 5.13 × 10^-4 M^-1s^-1 at 50 °C) in the presence of pyridine to generate the transient scandium imido (PNP)Sc=[DIPP](NC ₅H₅) (A-py), which rapidly activates the C-H bond of pyridine in 1,2-addition fashion to form the stable pyridyl complex (PNP)Sc(NH[DIPP])(η²-NC₅H₄) (2). Mechanistic studies suggest the C-H activation process to be second order overall: first order in scandium and first order in substrate (pyridine). Pyridine binding precedes elimination of methane, and α-hydrogen abstraction is overall-rate-determining [the kinetic isotope effect (KIE) for 1-d₁ conversion to 2 was 5.37(6) at 35 C and 4.9(14) at 50 C] with activation parameters ΔH^? = 17.9(9) kcal/mol and ΔS^? = -18(3) cal/(mol K), consistent with an associative-type mechanism. No KIE or exchange with the anilide proton was observed when 1-d₃ was treated with pyridine or thermolyzed at 35 or 50 °C. The post-rate-determining step, C-H bond activation of pyridine, revealed a primary KIE of 1.1(2) at 35 °C for the intermolecular C-H activation reaction in pyridine versus pyridine-d₅. Complex 2 equilibrated back to the imide A-py slowly, as the isotopomer (PNP)Sc(ND[DIPP])(η²-NC₅H₄) (2-d ₁) converted to (PNP)Sc(NH[DIPP])(η²-NC ₅H₃D) over 9 days at 60 °C. Molecular orbital analysis of A-py suggested that this species possesses a fairly linear scandium imido motif (169.7 ) with a very short Sc-N distance of 1.84 A?. Substituted pyridines can also be activated, with the rates of C-H activation depending on both the steric and electronic properties of the substrate.

Full text of DOI:10.1021/ja307279r

Article
pubs.acs.org/JACS
Evidence for the Existence of Terminal Scandium Imidos: Mechanistic
Studies Involving Imido−Scandium Bond Formation and C−H
Activation Reactions
Benjamin F. Wicker, Hongjun Fan,† Anne K. Hickey, Marco G. Crestani, Jennifer Scott,‡ Maren Pink,
and Daniel J. Mindiola*
Department of Chemistry, Indiana University, Bloomington, Indiana 47405, United States
S
* Supporting Information
ABSTRACT: The anilide−methyl complex (PNP)Sc(NH-
[DIPP])(CH3) (1) [PNP− = bis(2-diisopropylphosphino-4-
tolyl)amide, DIPP = 2,6-diisopropylphenyl] eliminates meth-
ane (kavg = 5.13 × 10−4 M−1s−1 at 50 °C) in the presence of
pyridine to generate the transient scandium imido (PNP)Sc
N[DIPP](NC5H5) (A-py), which rapidly activates the C−H
bond of pyridine in 1,2-addition fashion to form the stable
pyridyl complex (PNP)Sc(NH[DIPP])(η2-NC5H4) (2).
Mechanistic studies suggest the C−H activation process to
be second order overall: first order in scandium and first order
in substrate (pyridine). Pyridine binding precedes elimination of methane, and α-hydrogen abstraction is overall-rate-determining
[the kinetic isotope effect (KIE) for 1-d1 conversion to 2 was 5.37(6) at 35 °C and 4.9(14) at 50 °C] with activation parameters
ΔH⧧ = 17.9(9) kcal/mol and ΔS⧧ = −18(3) cal/(mol K), consistent with an associative-type mechanism. No KIE or exchange
with the anilide proton was observed when 1-d3 was treated with pyridine or thermolyzed at 35 or 50 °C. The post-rate-
determining step, C−H bond activation of pyridine, revealed a primary KIE of 1.1(2) at 35 °C for the intermolecular C−H
activation reaction in pyridine versus pyridine-d5. Complex 2 equilibrated back to the imide A-py slowly, as the isotopomer
(PNP)Sc(ND[DIPP])(η2-NC5H4) (2-d1) converted to (PNP)Sc(NH[DIPP])(η2-NC5H3D) over 9 days at 60 °C. Molecular
orbital analysis of A-py suggested that this species possesses a fairly linear scandium imido motif (169.7°) with a very short Sc−N
distance of 1.84 Å. Substituted pyridines can also be activated, with the rates of C−H activation depending on both the steric and
electronic properties of the substrate.
dinuclear scandium imido, which was formed by the insertion
of phenylnitrile into the Sc−C bond of the complex shown in
Scheme 1, was reported by Hessen and co-workers.9a They
proposed that after the insertion, the alkylimino ring system
underwent a cyclization to produce a transient mononuclear
scandium imide, which readily dimerized.
While Hessen’s system did not engage in C−H activation
reactions, presumably because of its dimeric nature, numerous
early transition metal imido complexes can intermolecularly
activate C−H bonds of alkyl and aryl groups3f,10,11 via a 1,2-C−
H bond addition mechanism.12 In general, examples of rare-
earth and lanthanide metal ions supported by multiply bonded
ligands have seldom been reported,2a,3f,7−9,11,13−15 and to our
knowledge, an in-depth study into the mechanism of formation
and reactivity of these rare species has not been reported, most
likely because of their scarcity.16 One advantage of studying
rare-earth or lanthanide complexes containing metal−ligand
multiple bonds is that there are many other well-studied
transition metal imido systems that are also known to perform
1. INTRODUCTION
Imide ligands have become ubiquitous in organometallic
chemistry, given that every transition metal triad (with the
exception of the groups 9 and 10 triads) has examples of
monomeric imido complexes.1,2 This ligand class has proven to
be very useful: its roles are necessary or have been implicated in
many catalytic reactions, including multicomponent coupling
processes,1,3,4 or simply as a robust ancillary ligand in important
catalysts.1d,i,5 Consequently, some of these imido complexes
require in-depth studies in order to discern their formation and
mode of reactivity.
Our group has explored 3d early transition metal complexes
that are both unsaturated and have terminal imido ligands, since
we are interested in catalytic transformations involving C−H
bond activation and functionalization.3e,f,6 One would anticipate
that using more electropositive metal centers such as the group
3 triad would result in even more reactive imido ligands, in view
of the polarized nature of the ScNR moiety toward a more
basic canonical form such as Sc+−N−R.7,8 While complexes of
some early transition metals (e.g., Ti) having imido ligands
have been well-studied,1d,j scandium has been resistant to form
complexes with such a popular ligand. The first example of a
Received: July 24, 2012
Published: October 26, 2012
© 2012 American Chemical Society
20081
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
Scheme 1. Formation of Hessen’s Bridging Scandium Imido
Complex
Scheme 2. Alternate Pathways to the Proposed Scandium
Imido Intermediate
a
a
C−H activation and/or catalytic reactions from which we can
glean mechanistic insights.10 While there are many early
transition metal imido systems on which one could focus, the
most analogous d0 metal centers that possess ionic radii and
Pauling electronegativities comparable to those of Sc(III) (74.5
pm, 1.36) are Zr(IV) (72 pm, 1.33) and Hf(IV) (71 pm,
1.30).17 For this reason, we focused our attention on Sc(III)
imido complexes, as they are rare but comparable to well-
known group 4 transition metal imido complexes.
For clarity, the pyridine adduct of the transient imido (PNP)Sc
N[DIPP] (A) is not shown.
in C6D6.11 Our proposed intermediate A, a terminal scandium
imido (or an adduct form that can be prepared by three
independent routes) would be soon supported by the isolation
of a Sc(III) imido by Chen and co-workers.2a,15a−c Using a
tridentate β-diketiminate ligand, they were able to generate the
imido by α-hydrogen abstraction (Scheme 3a), which was most
All imido systems that intermolecularly activate C−H bonds
are proposed to accomplish this by 1,2 addition of the substrate
across the metal imido multiple bond.12 Bergman and
Wolczanski have extensively studied unsaturated imido
complexes via kinetic and isotope labeling experiments and
established that the C−H activation reaction is zeroth order in
substrate, with the formation of the imido as the rate-
determining step (RDS).10a,b,f−j Kinetic isotope effect (KIE)
experiments showed primary and secondary KIEs for the loss of
alkane (α-hydrogen abstraction) to yield the imido interme-
diate.10g,h
Scheme 3. Generation of Chen’s Sc(III) Imidos via α-
Hydrogen Abstraction with the Aid of (a) DMAP or (b) a
Pendant Amine Group
Since our group reported transient titanium alkylidyne (or
vanadium alkylidene) complexes using the same PNP ligand18
that can activate a wide array of hydroaromatic and aliphatic
C−H bonds, including those in methane,19 we wished to
extend this work to scandium. We reported a Sc(III) system,
(PNP)Sc(NH[DIPP])(CH3) (1) [PNP− = bis(2-diisopropyl-
phosphino-4-tolyl)amide, DIPP = 2,6-diisopropylphenyl], that
undergoes methane loss in the presence of pyridine to yield the
pyridyl complex (PNP)Sc(NH[DIPP])(η2-NC5H4) (2) by
virtue of a C−H bond activation process across the transient
ScN[DIPP] ligand (Scheme 2).11,3f This type of reaction was
unprecedented in group 3 and lanthanide metal chemistry.
As shown in Scheme 2, our group first demonstrated that
from 1, the transient imido (PNP)ScN[DIPP] (A) (or its
pyridine adduct A-py, which is not shown for the purpose of
clarity) can activate the C−H bond of pyridine at the 2-position
to yield 2.3f,11 Although the imido intermediate was not
isolated, isotopic labeling of the pyridine substrate as well as the
isolation of the Al(CH3)3-trapped imido zwitterion in the form
(PNP)Sc(μ2-N[DIPP])(μ2-CH3)[Al(CH3)2] supported our
assertion that an imido is involved in the C−H activation
step.11 Likewise, we showed that an intermediate such as A (or
an adduct thereof) is most likely responsible for the C−H bond
likely promoted by using 4-N,N-dimethylaminopyridine
(DMAP) as a Lewis base. Chen and co-workers also prepared
a DMAP-free scandium imido by using a pendant dimethyla-
mino group (Scheme 3b).15b To the best of our knowledge, the
latter system is surprisingly inert toward intermolecular C−H
bonds with high pKa’s,15c presumably because of the congested
environment about the Sc(III) ion. Very recently, Cui and co-
workers observed intramolecular C−H bond activation of a
pendant arene by a scandium imido complex.15d In the latter
case, a Lewis base such as DMAP was also necessary to
promote formation of the imido ligand.
In this paper, we present convincing synthetic and
mechanistic evidence for the formation of a transient
mononuclear Sc(III) imido. Along with isotopic labeling and
KIE studies, we also extracted thermodynamic parameters for
the formation of the pyridyl complex via a transient imido
adduct, which was also corroborated by theoretical studies. We
investigated the role of pyridine in the C−H activation reaction
pathway by analyzing competing mechanistic scenarios
t
activation of benzene, as (PNP)Sc(NH[DIPP])(CH2 Bu) and
(PNP)Sc(NH[DIPP])(CH2SiMe3) form H3CtBu or SiMe4,
respectively, and (PNP)Sc(ND[DIPP])(C6D5) when dissolved
20082
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
drying over CaH2 and distilled under reduced pressure at 120 °C with
a short-path distillation apparatus. Room-temperature water was used
in the distillation column, and the receiving flask was immersed in a
−78 °C bath. The oil was degassed, taken into the glovebox, and used
for subsequent reactivity.
(whether methane loss occurs before or after pyridine binding)
as well as the effect of substituted pyridines on the rate of the
reaction, including some showcase reactivity. Our study
represents the first detailed investigation of metal imido
formation and C−H bond reactivity for any rare-earth or
lanthanide system.
(PNP)Sc(CD3)(I). A solution of CD3MgI (0.69 mL, 1.0 M) in diethyl
ether was added dropwise to a stirring suspension of (PNP)ScCl2
(376.0 mg, 0.69 mmol) in toluene (15 mL) at −35 °C. The reaction
mixture was stirred for 5 h, after which 1 mL of anhydrous 1,4-dioxane
was added to promote the precipitation of magnesium salts. The
resulting cloudy, yellow suspension was filtered through Celite, and
the filtrate was concentrated to <1 mL and placed in the freezer
overnight. Yellow crystals of (PNP)Sc(CD3)(I) were isolated upon
standing at −35 °C overnight (256 mg, 0.36 mmol, 61% yield). The
NMR spectroscopic data were similar to those for the previously
reported complex (PNP)Sc(CH3)(Br),11 except for the absence of the
2. EXPERIMENTAL SECTION
General Considerations. Unless otherwise stated, all operations
were performed in an M. Braun Lab Master double drybox under an
atmosphere of purified nitrogen or using standard high-vacuum
Schlenk techniques under an argon atmosphere.20 Anhydrous n-
pentane, n-hexane, and toluene were purchased from Aldrich in Sure-
Sealed reservoirs (18 L) and dried by passage through a column of
activated alumina and then a Q-5 column.21 1,4-Dioxane was dried by
stirring in small pieces of Na metal for 24 h, followed by filtration
through activated alumina. C6D6 (purchased from Cambridge Isotope
Laboratories) was degassed and dried over 4 Å molecular sieves for at
least 24 h before use. Celite, alumina, and 4 Å molecular sieves were
activated under vacuum overnight at 200 °C. (PNP)ScCl2,11
(PNP)Sc(CH3)(Br),11 and complexes 1,11 2,11 8,3f and 93f were
synthesized according to literature protocols. DMAP was purchased
from Aldrich and cycled into the box under vacuum. 2-Picoline, 4-
picoline, and 2-fluoropyridine were vacuum-distilled from CaH2 and
stored in the drybox over 4 Å molecular sieves. LiND[DIPP] and
LiNH[DIPP] were prepared by addition of equimolar amounts of 2.5
M nBuLi in n-hexane to D2N[DIPP] and H2N[DIPP], respectively, in
n-hexane. The suspensions were stirred overnight and then filtered,
and the solids were washed with hexanes and dried under vacuum.
D2N[DIPP] and H2N[DIPP] were dried over CaH2 and distilled
before being brought into the glovebox; both were passed through
alumina and stored over 4 Å molecular sieves. The complex
(PNP)Sc(ND[DIPP])(Cl) was prepared analogously to its isotopo-
logue (PNP)Sc(NH[DIPP])(Cl).11 1H, 13C, and 31P NMR spectra
1
methyl resonance. Mp: 230−240 °C (decomp). H NMR (25 °C,
399.93 MHz, C6D6): δ 7.24 (s, 1H, C6H3), 7.02 (br, 1H, C6H3), 6.90−
6.72 (m, 4H, C6H3), 2.14 (s, 3H, C6H3−CH3), 2.10 (s, 3H, C6H3−
3
CH3), 2.09−1.94 (m, 3H, CHMe2), 1.89 (sept, JH−H = 6.8 Hz, 1H,
CHMe2), 1.40−1.00 (m, 18H, PCH(CH3)2), 0.85 (m, 6H, PCH-
(CH3)2). 13C NMR (25 °C, 100.57 MHz, C6D6): δ 160.6 (d, C6H3),
159.2 (d, C6H3), 133.8 (s, C6H3), 133.7 (s, C6H3), 133.6 (s, C6H3),
133.5 (s, C6H3), 127.73 (s, C6H3), 129.69 (s, C6H3), 121.5 (s, C6H3),
121.4 (s, C6H3), 118.4 (s, C6H3), 118.3 (s, C6H3), 24.5 (t,
PCH(CH3)2), 21.1 (d, PCH(CH3)2), 20.1−18.6 (m, PCH(CH3)2/
PCH(CH3)2), 16.9 (d, PCH(CH3)2). 31P NMR (25 °C, 161.98 MHz,
C6D6): δ 3.16 and 2.37 (br, overlapping singlets, Δν1/2 = 253 Hz).
(PNP)Sc(ND[DIPP])(CH3) (1-d1). A solution of LiND[DIPP] (153.9
mg, 0.83 mmol) in n-hexane (3 mL), with a few drops of ether added
to solubilize the salt, was added dropwise to a stirring suspension of
(PNP)Sc(CH3)(Br) (475 mg, 0.83 mmol) in n-hexane (15 mL). The
reaction mixture was stirred for 5 h. The resulting cloudy, yellow
suspension was filtered through Celite, and the filtrate was
concentrated to 5 mL and placed in the freezer overnight. Yellow
crystals of 1-d1 were isolated upon standing at −35 °C overnight
(258.1 mg, 0.40 mmol, 48% yield). The NMR spectra obtained for 1-
d1 were identical to those reported for 1, except for the notable
1
were recorded on Varian 400 and 300 MHz NMR spectrometers. H
and 13C NMR spectral data are reported with reference to solvent
resonances (residual C6D5H in C6D6, 7.16 and 128.0 ppm,
respectively), and 31P NMR chemical shifts are reported with respect
to external H3PO4 (aqueous solution, 0.0 ppm). Multinuclear NMR
spectral data for all complexes are included in the Supporting
Information (SI).22 Repeated attempts to obtain satisfactory micro-
analysis of the compounds failed because of their extreme sensitivity to
air and moisture and quite possibly their incomplete combustion. We
have therefore provided high-resolution 1D and 2D NMR spectra as
proof of their purity and in lieu of combustion elemental analysis data.
Melting points for selected compounds were measured in flame-sealed
capillaries on an Electrothermal Mel Temp apparatus. IR spectra were
measured using a Thermo Scientific Nicolet 6700 spectrometer.
Samples were prepared by crushing and mixing the compounds with
KBr (FT-IR grade, ≥99% trace metals basis, Sigma-Aldrich) and
pressing the mixture into a pellet. All of the IR spectral data are
included in the SI.
1
absence of the anilide proton resonance at 6.01 ppm in the H NMR
spectrum.
(PNP)Sc(NH[DIPP])(CD3) (1-d3). A solution of LiNH[DIPP] (30.9
mg, 0.17 mmol) in n-hexane (3 mL), with a few drops of ether added
to solubilize the salt, was added dropwise to a stirring suspension of
(PNP)Sc(CD3)(I) (104.5 mg, 0.17 mmol) in n-hexane (15 mL). The
reaction mixture was stirred for 5 h. The resulting cloudy, yellow
suspension was filtered through Celite, and the filtrate was
concentrated to <1 mL and placed in the freezer overnight. Yellow
crystals of 1-d3 were isolated upon standing at −35 °C overnight (45
mg, 0.07 mmol, 40% yield). Mp: 138−142 °C. The NMR spectra
obtained for 1-d3 were identical to those reported for 1, except for the
notable absence of the scandium methyl resonance at 0.29 ppm in the
1H NMR spectrum. IR (KBr): νNH 3312 cm−1.
(PNP)Sc(ND[DIPP])(CD3) (1-d4). A solution of CD3MgI (1.0 M,
0.68 mL) dissolved in dioxane (1 mL) was added to a solution of
(PNP)Sc(ND[DIPP])(Cl) (234.3 mg, 0.34 mmol) dissolved in
toluene (10 mL). The reaction mixture was stirred for 2 h and then
filtered through a pad of Celite, after which the filtrate was
concentrated to 1 mL. Crystals of 1-d4 were isolated upon standing
at −35 °C overnight (90.5 mg, 0.14 mmol, 40% yield). The NMR
spectra obtained for 1-d4 were identical to those reported for 1, except
for the notable absences of the anilide proton resonance at 6.01 ppm
Syntheses. (CH3)2CN[DIPP]. A round-bottom flask was charged
with 6.0 mL (5.64 g, 31.8 mmol) of H2N[DIPP], 20.0 mL (17.0 g,
163.2 mmol) of 2,2-dimethoxypropane, and 155 mg (1.6 mmol) of
sulfuric acid in 200 mL of ethanol. The solution was refluxed for 12 h
under nitrogen and then cooled. The volatiles were removed via rotary
evaporator, yielding the product as a brown oil (6.2 g, 28.6 mmol, 90%
yield). The 1H and 13C NMR spectra were consistent with those from
a previously reported synthesis of (CH3)2CN[DIPP].23
1
and the scandium methyl resonance at 0.29 ppm in the H NMR
D2N[DIPP]. A round-bottom flask was charged with (CH3)2C
N[DIPP] (6.2 g, 28.6 mmol), 75.0 mL (83.03 g, 4145.0 mmol) of
D2O, and 1 mL of 3.0 M CF3CO2D in D2O [prepared by adding 5.29
mL of (CF3CO)2O to 25 mL of D2O]. The oil remained immiscible
with the D2O while the reaction mixture was refluxed for 6 h under
nitrogen. After the mixture was cooled, the product was extracted into
n-hexane, dried with Mg2SO4, and filtered through Celite. The volatiles
in the filtrate were removed by a rotary evaporator, yielding a yellow
oil (4.7 g, 26.3 mmol, 92% yield). The oil was further purified by
spectrum.
(PNP)Sc(ND[DIPP])(η2-NC5H4) (2-d1). In a vial, a solution of 2-
bromopyridine (39.1 mg, 0.25 mmol) in tetrahydrofuran (THF) (3
mL) was cooled to −100 °C (dry ice/ethanol bath), and to this cold
solution was added 0.1 mL of 2.5 M n-BuLi in n-hexane. The solution
was stirred for 10 min and then added dropwise cold to a stirring
suspension of (PNP)Sc(ND[DIPP])(Cl) (171.0 mg, 0.25 mmol) in
THF (15 mL) at −100 °C. The reaction mixture was stirred for 1 h,
20083
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
after which the THF was evaporated under vacuum. The residue was
extracted into n-pentane, and the resulting cloudy, yellow suspension
was filtered through Celite. The filtrate was concentrated to <1 mL
and allowed to stand at −35 °C overnight, after which yellow crystals
of 2-d1 were isolated (88.8 mg, 0.12 mmol, 49.3% yield). The ratio of
From the filtrate, yellow crystals of 4 were isolated upon standing at
−35 °C overnight (35.5 mg, 0.048 mmol, 68.5% yield). Mp: 146−170
°C (decomp). 1H NMR (25 °C, 300.1 MHz, C6D6): δ 8.28 (d, 3JH−H
=
3
5.3 Hz, 1H, NC5H3), 7.81 (s, 1H, NC5H3), 7.28 (dd, JH−H = 8.3, 4.4
Hz, 1H, C6H3), 7.25−7.17 (m, 3H, C6H3), 7.02−6.82 (m, 5H, C6H3),
6.59 (d, 3JH−H = 5.2 Hz, 1H, NC5H3), 6.16 (s, 1H, NH), 3.39 (br, 2H,
CH(CH3)2), 2.23, (s, 3H, C6H3−CH3), 2.18 (s, 3H, C6H3−CH3), 1.94
(s, 3H, NC5H3−CH3), 2.09−1.84 (m, 3H, PCH(CH3)2), 1.72 (sept,
1
2-d1 to protonated anilide was 3:1. H NMR (25 °C, 399.93 MHz,
C6D6): δ 8.33 (d, 3JH−H = 5.2 Hz, 1H, 6-H in NC5H4), 7.88 (d, 3JH−H
3
= 7.3 Hz, 1H, 3-H in NC5H4), 7.28 (dd, JH−H = 8.3, 4.5 Hz, 1H,
3
3
3JH−H = 7.0 Hz, 1H, PCH(CH3)2), 1.36 (pseudo t, JH−H = 6.4 Hz,
C6H3), 7.23−7.16 (m, 3H, C6H3), 7.09 (td, JH−H = 7.4, 1.0 Hz, 1H,
3
NC5H4), 6.91 (d, JH−H = 7.6 Hz, 2H, C6H3), 6.89−6.83 (m, 3H,
12H, CH(CH3)3), 1.27−1.15 (m, 6H, PCH(CH3)2), 1.09−0.95 (m,
3
6H, PCH(CH3)2), 0.94−0.84 (m, 6H, PCH(CH3)2), 0.76 (dd, 3JH−H
=
C6H3), 6.67 (ddd, JH−H = 6.0, 5.3, 1.0 Hz, 1H, NC5H4), 6.18 (s, 1H,
3
NH), 3.37 (br s, 2H, Ar−CHMe2), 2.22 (s, 3H, C6H3−CH3), 2.16 (s,
3H, C6H3−CH3), 2.06−1.82 (m, 3H, PCHMe2), 1.65 (m, 1H,
PCHMe2), 1.35 (d, 3JH−H = 6.4 Hz, 12H, Ar−CH(CH3)2), 1.26−1.12
(m, 6H, PCH(CH3)2), 1.06−0.80 (m, 12H, PCH(CH3)2), 0.74−0.66
(m, 3H, PCH(CH3)2), 0.46−0.36 (m, 3H, PCH(CH3)2). 13C NMR
(23 °C, 100.6 MHz, C6D6): δ 217.2 (s, Sc−C, η2-NC5H4), 161.1 (d,
C6H3), 160.4 (d, C6H3), 151.5 (s, C6H3), 145.7 (s, C6H3), 134.0 (s,
NC5H4), 132.9 (s, C6H3), 132.8 (s, C6H3), 132.6 (s, C6H3), 132.4 (s,
C6H3), 130.2 (s, NC5H4), 128.1 (s, C6H3), 127.8 (s, C6H3), 126.1 (d,
C6H3), 122.7 (s, NC5H4), 121.9 (s, NC5H4), 120.5 (d, C6H3), 119.7 (s,
C6H3), 119.5 (s, C6H3), 119.4 (s, C6H3), 119.2 (s, C6H3), 118.3 (d,
C6H3), 116.0 (s, C6H3), 28.8 (s, Ar−CHMe2), 25.2 (d, PCHMe2), 24.3
(s, Ar−CH(CH3)2), 23.9 (s, Ar−CH(CH3)2), 23.5 (d, PCHMe2), 20.8
(s, C6H3−CH3), 20.7 (s, C6H3−CH3), 20.1 (d, PCHMe2), 19.9 (d,
PCHMe2), 19.5 (d, Ar−CHMe2), 19.5 (s, PCH(CH3)2), 19.4 (d,
PCH(CH3)2), 19.1 (s, PCH(CH3)2), 19.0 (s, PCH(CH3)2), 18.8 (d,
PCH(CH3)2), 18.75 (d, PCH(CH3)2), 17.8 (s, PCH(CH3)2), 16.5 (s,
15.3, 6.8 Hz, 3H, PCH(CH3)2), 0.50 (dd, JH−H = 14.6, 7.0 Hz, 3H,
PCH(CH3)2). 13C{1H} NMR (25 °C, 75.5 MHz, C6D6): δ 216.0 (br,
Sc−C, η2-NC5H3), 161.1 (d, C6H3), 160.6 (d, C6H3), 151.6 (s, C6H3),
145.1 (s, C6H3), 143.7 (s, NC5H3), 134.1 (br, C6H3), 132.8 (s,
NC5H3), 132.7 (s, C6H3), 132.4 (s, C6H3), 131.0 (s, C6H3), 126.1 (d,
C6H3), 123.4 (s, NC5H3), 122.7 (s, C6H3), 120.5 (d, C6H3), 119.8 (d,
C6H3), 119.5 (d, C6H3), 118.5 (d, C6H3), 115.9 (s, NC5H3), 28.9 (br,
anilide CH(CH3)2), 25.3 (d, CH(CH3)2), 24.4 (s, Ar−CH3), 24.0 (s,
Ar−CH3), 23.6 (d, CH(CH3)2), 20.9 (d, CH(CH3)2), 20.6 (NC5H3−
CH3), 20.4−18.7 (m, CH(CH3)2/CH(CH3)2/anilide CH(CH3)2),
17.9 (s, CH(CH3)2) 16.6 (d, CH(CH3)2). 31P{1H} NMR (25 °C,
162.0 MHz, C6D6): δ 4.7 (Δν1/2 = 86.6 Hz), 3.9 (Δν1/2 = 84.5 Hz). IR
(KBr): νNH 3311 cm−1.
(PNP)Sc(NH[DIPP])(η2-NC5H3-2-CH3) (5) and (PNP)Sc(NH[DIPP])-
(κ2-C,N-NC5H4-2-CH2) (6). 1 (40.3 mg, 0.061 mmol) was placed in a J-
Young tube and dissolved in C6D6. To this was added 2 drops of 2-
picoline. The mixture was heated at 65 °C for 12 h, and the solvent
and excess 2-picoline removed under reduced pressure. The residue
was extracted into n-hexane (3 mL) and filtered through a small pad of
Celite. From the filtrate, yellow crystals of a mixture of 5 and 6 were
isolated upon standing at −35 °C overnight (24.1 mg, 0.032 mmol,
PCH(CH3)2). 31P NMR (23 °C, 121.5 MHz, C6D6): δ 5.1 (br, Δν1/2
131 Hz), 4.5 (br, Δν1/2 = 131 Hz).
=
(PNP)Sc(NH[DIPP])(η2-NC5H3-4-N(CH3)2) (3). A J-Young tube was
charged with 1 (54.3 mg, 0.082 mmol). To this was added a solution
of DMAP (10.0 mg, 0.082 mmol) in C6D6 (1.5 mL). The reaction
quickly changed color from yellow to orange. The mixture was heated
at 65 °C for 30 min, and the solvent was removed under reduced
pressure. The residue was extracted with 1 mL of toluene and filtered
through a pad of Celite. The Celite was then washed with 1 mL of n-
hexane. From the filtrate, yellow crystals of 3 were isolated upon
standing at −35 °C overnight (34.1 mg, 0.044 mmol, 55.4% yield).
1
53.9% yield). Mp: 118−202 °C (decomp). For the 5/6 mixture: H
NMR (25 °C, 399.93 MHz, C6D6): δ 7.75 (d, 7.1 Hz), 7.38 (dd, JH−H
= 8.3, 4.1 Hz), 7.29−7.13 (m), 7.13−7.03 (m), 7.02−6.83 (m), 6.74
(td, JH−H = 7.6, 1.5 Hz), 6.62−6.51 (m), 6.17 (s, N−H), 6.08 (s, N−
H), 5.85 (t, JH−H = 6.1 Hz), 3.47 (br), 3.20 (br), 2.65 (sept, 6.6 Hz),
2.45 (s), 2.24 (s), 2.23 (s), 2.20 (s), 2.19 (s), 2.17−1.78 (m), 1.99 (s),
1.66 (sept, JH−H = 7.1 Hz), 1.45 (d, 6.6 Hz), 1.40 (d, 6.6 Hz), 1.37−0.5
(m), 0.22 (dd, JH−H = 14.7, 7.0 Hz). 13C NMR (25 °C, 100.57 MHz,
C6D6): δ 216.9 (br, Sc−C, η2-NC5H3), 146.8 (s, NC6H3), 133.7 (s,
NC6H3), 133.2 (s, C6H3), 132.4 (s, C6H3), 132.2 (s, C6H3), 132.15 (s,
C6H3), 132.0 (s, C6H3), 131.8 (s, C6H3), 130.1 (s, NC6H3), 127.9 (s,
NC6H3), 126.0 (s, C6H3), 122.6 (s, C6H3), 122.4 (s, C6H3), 121.8 (s,
NC6H3), 119.9 (d, C6H3), 119.3 (s, NC6H3), 117.9 (d, C6H3), 117.7
(d, C6H3), 116.8 (s, NC6H3), 116.3 (s, NC6H3), 115.6 (s, NC6H3),
110.4 (s, NC6H3), 53.2 (s, Sc−CH2−NC5H4), 28.9 (s, NC5H3−CH3),
25.5 (d, PCH(CH3)2)), 25.0 (s, PCH(CH3)2), 23.9 (s, PCH(CH3)2),
23.8 (s, PCH(CH3)2), 23.2 (s, PCH(CH3)2), 22.3 (s, PCH(CH3)2),
20.5 (Ar−CH3), 20.1 (s, PCH(CH3)2), 17.2 (s, PCH(CH3)2). 31P{1H}
NMR (25 °C, 161.98 MHz, C6D6): δ 5.58 (Δν1/2 = 118 Hz), 4.83
(Δν1/2 = 132 Hz). IR (KBr): νNH 3313 cm−1.
1
Mp: 126−164 °C (decomp). H NMR (25 °C, 300.1 MHz, C6D6): δ
3
3
8.20 (d, JH−H = 6.1 Hz, 1H, NC5H3), 7.31 (dd, JH−H = 8.3, 4.5 Hz,
1H, C6H3), 7.25 (dd, 3JH−H = 8.7, 4.4 Hz, 1H, C6H3), 7.21 (d, 3JH−H
=
3
7.6 Hz, 2H, C6H3), 7.18 (d, JH−H = 2.3 Hz, NC5H3), 7.01−6.90 (m,
4H, C6H3), 6.78 (t, 3JH−H = 7.5 Hz, 1H, C6H3), 6.15 (dd, 3JH−H = 6.2,
2.7 Hz, 1H, NC5H3), 6.08 (s, 1H, NH), 3.50 (br, 2H, CH(CH3)2),
2.37 (s, 6H, N(CH3)2), 2.24, (s, 3H, C6H3−CH3), 2.18 (s, 3H, C6H3−
3
CH3), 2.14−1.89 (m, 3H, PCH(CH3)2), 1.83 (sept, JH−H = 6.8 Hz,
1H, PCH(CH3)2), 1.42 (d, 3JH−H = 6.8 Hz, 6H, CH(CH3)3), 1.39 (d,
3JH−H = 6.7 Hz, 6H, CH(CH3)2), 1.31 (dd, JH−H = 14.7, 7.0 Hz, 3H,
3
3
PCH(CH3)2), 1.26 (dd, JH−H = 14.2, 7.0 Hz, PCH(CH3)2), 1.11−
1.02 (m, 6H, PCH(CH3)2), 0.98−0.89 (m, 6H, PCH(CH3)2), 0.83
3
3
(dd, JH−H = 15.3, 6.8 Hz, 3H, PCH(CH3)2), 0.67 (dd, JH−H = 14.6,
7.0 Hz, 3H, PCH(CH3)2). 13C{1H} NMR (25 °C, 100.6 MHz, C6D6):
δ 213.8 (br, Sc−C, η2-NC5H3), 161.4 (d, C6H3), 161.0 (d, C6H3),
152.1 (s, C6H3), 151.9 (s, C6H3), 144.6 (s, NC5H3), 134.1 (br, C6H3),
132.9 (s, NC5H3), 132.6 (s, C6H3), 132.5 (s, C6H3), 132.4 (C6H3),
127.3 (d, C6H3), 125.7 (d, C6H3), 122.7 (s, C6H3), 120.5 (d, C6H3),
119.9 (d, C6H3), 119.7 (d, C6H3), 118.4 (d, C6H3), 115.6 (s, NC5H3),
111.2 (s, NC5H3), 106.8 (s, NC5H3), 38.6 (N(CH3)2), 28.7 (br, anilide
CH(CH3)2), 25.3 (d, CH(CH3)2), 24.6 (s, CH(CH3)2), 24.1 (s, Ar−
CH3), 23.7 (d, CH(CH3)2), 20.9 (d, CH(CH3)2), 20.4 (d, CH-
(CH3)2), 20.1 (d, CH(CH3)2), 19.9−19.2 (m, CH(CH3)2/CH-
(CH3)2), 18.0 (s, CH(CH3)2) 16.7 (d, CH(CH3)2). 31P{1H} NMR
(25 °C, 162.0 MHz, C6D6): δ 4.7 (Δν1/2 = 84.0 Hz), 3.8 (Δν1/2 = 66.0
Hz). IR (KBr): νNH 3309 cm−1.
(PNP)Sc(NH[DIPP])(κ2-C,N-NC5H3-2-CH2-6-CH3) (7). 1 (36.3 mg,
0.055 mmol) was placed in a J-Young tube and dissolved in C6D6. To
this was added 2 drops of 2,6-lutidine. The mixture was heated at 65
°C for 12 h, and the solvent and excess 2,6-lutidine were removed
under reduced pressure. The residue was extracted into hexanes (3
mL) and filtered. The filtrate was concentrated to ∼1 mL and allowed
to stand at −35 °C for 12 h to yield X-ray-quality crystals (27.0 mg,
1
0.036 mmol, 64.9%). Mp: 168−196 °C (decomp). H NMR (25 °C,
300.1 MHz, C6D6): δ 7.29−7.11 (m, 4H, NC5H3/C6H3), 6.98−6.82
3
3
(m, 5H, C6H3), 6.77 (t, JH−H = 7.5 Hz, 1H, C6H3), 6.49 (d, JH−H
=
3
8.1 Hz, 1H, NC5H3), 6.30 (s, 1H, NH), 5.87 (d, JH−H = 7.0 Hz, 1H,
NC5H3), 3.24 (br, 2H, CH(CH3)2), 2.66 (br, 2H, Sc−CH2−NC5H3),
2.19 (s, 7H, C6H3−CH3/PCH(CH3)2), 2.04−1.84 (m, 3H,
PCH(CH3)2), 1.70 (s, 3H, NC5H3−CH3), 1.44−0.64 (m, 36H,
PCH(CH3)2/CH(CH3)2). 13C{H} NMR (25 °C, 75.5 MHz, C6D6): δ
169.4 (s, NC5H3), 160.4 (d, C6H3), 159.4 (d, C6H3), 157.9 (s, NC5H3),
151.1 (s, C6H3), 136.8 (s, NC5H3), 135.5 (br, C6H3), 132.8 (s,
NC5H3), 132.7 (s, C6H3), 132.6 (s, C6H3), 126.8 (d, C6H3), 123.2 (s,
(PNP)Sc(NH[DIPP])(η2-NC5H3-4-CH3) (4). 1 (48.5 mg, 0.073 mmol)
was placed in a J-Young tube and dissolved in C6D6. To this was added
2 drops of 4-picoline. The mixture was heated at 65 °C for 12 h, and
the solvent and excess 4-picoline were removed under reduced
pressure. The residue was extracted into hexanes (3 mL) and filtered.
20084
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
C6H3), 120.8 (d, C6H3), 120.0 (d, C6H3), 119.2 (d, C6H3), 119.0 (d,
C6H3), 117.2 (s, C6H3), 117.0 (s, C6H3), 110.9 (s, NC5H3), 57.4 (br,
Sc−CH2−NC5H3), 32.1 (s, CH3−NC5H3), 28.0 (br, anilide CH-
(CH3)2), 26.2 (d, CH(CH3)2), 25.7 (s, CH(CH3)2), 25.1 (s,
CH(CH3)2), 23.1 (d, CH(CH3)2), 12.3−19.9 (m, CH(CH3)2/
CH(CH3)2/Ar−CH3), 19.5 (s, CH(CH3)2), 19.1 (s, CH(CH3)2),
18.6 (d, CH(CH3)2), 16.9 (d, CH(CH3)2). One PNP resonance and
one pyridyl resonance were obscured by residual solvent. 31P{H}
NMR (25 °C, 162.0 MHz, C6D6): δ 5.6 (Δν1/2 = 96.7 Hz), 3.6 (Δν1/2
= 102.0 Hz). IR (KBr): νNH 3315 cm−1.
reported the first study to provide evidence for the existence of
such species11 via a series of trapping and deuteration studies
using the robust and sterically encumbering pincer scaffold
bis(2-diisopropylphosphino-4-tolyl)amide (PNP−), which was
developed by Ozerov.26 Soon after, Piers proposed a transient
monomeric scandium imide intermediate formed via degrada-
tion of the imine residue of the β-diketiminate ligand.9b
Concurrent with our study, a terminal scandium imido formed
using a cyclometalated species and a Lewis base was reported
by Piers and co-workers.9c
Kinetic Studies. The reaction of 1 to give 2 was monitored by 1H
NMR spectroscopy. The rate constant (kobs) for the decay of 1 was
measured by the integration of two separate resonances over time, the
N−H resonance of the anilide group (s, 6.01 ppm) and the Sc−CH3
resonance (t, 0.29 ppm), using as an external reference a capillary
containing 1,4-dioxane in C6D6. The rate constant was measured every
5 °C over the range 35−75 °C, with triplicate runs at each
temperature. This yielded six separate measurements for each
temperature (triplicate runs for both N−H and Sc−CH3), and their
average is shown in Table 1. To obtain kobs at each temperature, a
As reported in an earlier communication, the starting
material (PNP)ScCl2 can be readily prepared in 95% yield
from Li(PNP) and ScCl3(THF)3 in toluene.13b A close
analogue of (PNP)ScCl2 was previously described by Fryzuk
using a more flexible and sterically encumbering PNP variant,
[tBu2PCH2Si(Me)2]2N−.27 (PNP)ScCl2 can be converted to
(PNP)Sc(NH[DIPP])(CH3) (1) via a series of transmetalation
steps with LiNH[DIPP] and then LiMe.11 Compound 1 is
remarkably stable up to 50 °C over 24 h without significant
evidence of decomposition. However, upon thermolysis of 1 at
70 °C in C6(H/D)6, 1 is completely consumed within 12 h, and
multiple products are formed, one of which we identified as
(PNP)Sc(NH[DIPP])(C6H5) on the basis of 31P and 1H NMR
spectroscopy via comparison to a sample independently
prepared from (PNP)Sc(NH[DIPP])(Cl) and LiC6H5.11,22
Since α-hydrogen abstraction is likely the rate-determining step
on the basis of kinetic studies performed for the neighboring
complex (PNP)TiCHtBu(R),19 we anticipated a much
quicker reaction with more steric crowding in the departing
alkyl group. Lowering the barrier to α-hydrogen abstraction
could also disfavor other deleterious pathways. Accordingly, the
Table 1. Rate Constants for the Reaction 1 → 2 over the
Temperature Range 35−75 °C
T (°C)
actual T (°C)
T (K)
average kobs (s−1
)
75
70
65
60
55
50
45
40
35
76.91
72.19
67.07
61.96
56.64
51.72
46.61
41.99
36.77
350.06
345.34
340.22
335.11
329.79
324.87
319.76
315.14
309.92
6.29 × 10−3
3.75 × 10−3
2.25 × 10−3
1.34 × 10−3
9.88 × 10−4
5.17 × 10−4
3.64 × 10−4
3.26 × 10−4
1.83 × 10−4
t
metastable complex (PNP)Sc(NH[DIPP])(CH2 Bu) and the
stable complex (PNP)Sc(NH[DIPP])(CH2SiMe3) activate the
C−H bond of benzene relatively cleanly under similar
conditions (∼70 °C). Using C6D6, we clearly showed that
second-order analysis was used. The decay of pyridine was followed by
integrating the pyridine 2,6-protons (d, 8.56 ppm) over time. The
concentration of 1 in these measurements was 0.15 M. At each
temperature, the actual temperature of the probe was recorded using
an ethylene glycol standard [T = (4.637 − Δ)/0.009967, where Δ =
t
the NH[DIPP] ligand in (PNP)Sc(NH[DIPP])(CH2 Bu) is
converted to its isotopologue ND[DIPP], affording (PNP)Sc-
(ND[DIPP])(C6D5) with concurrent H3CtBu formation
(Scheme 2).11 Likewise, reactions to trap the elusive imido
could be accomplished via the Tebbe and Schrock route: using
AlMe3 with 1 cleanly formed the imido zwitterion (PNP)Sc(μ2-
N[DIPP])(μ2-CH3)[Al(CH3)2] and CH4 (Scheme 2).11 This
zwitterionic complex resembles the alkylidyne zwitterion
(PNP)Ti(μ2-CtBu)(μ2-CH3)[Al(CH3)2] prepared analogously
δOH − δ for 100% ethylene glycol).
CH2
Single-Crystal X-ray Diffraction. The data collection was carried
out using Mo Kα radiation (graphite monochromator) with a frame
time of 5 s and a detector distance of 5.0 cm. A collection strategy was
calculated and complete data to a resolution of 0.73 Å with a
redundancy of 4 were collected using APEX2 software. Six sections of
frames were collected with 0.50° ϕ and ω scans.24 Data to a resolution
of 0.77 Å were considered in the reduction. Final cell constants were
calculated from the xyz centroids of 1859 strong reflections from the
actual data collection after integration. The intensity data were
corrected for absorption using SADABS.25 Additional crystallographic
details are available in the SI in the form of CIF files and a description
of the data collection, solution, and refinement procedures for each
structure.22
via treatment of (PNP)TiCHtBu(CH2 Bu) or (PNP)Ti
t
CHtBu(C6H5) with AlMe3.28 To our surprise, however, we
found that the AlMe3 moiety of (PNP)Sc(μ2-N[DIPP])(μ2-
CH3)[Al(CH3)2] can be readily removed with excess pyridine
to form the pyridyl−anilide complex (PNP)Sc(NH[DIPP])(η2-
NC5H4) (2), which itself is a synthon of a terminal scandium
imido complex, as shown via deuteration studies with pyridine-
d5 to form (PNP)Sc(ND[DIPP])(η2-NC5D4) (2-d5) (Scheme
2). Likewise, complex 1 can also react rapidly with pyridine-d5
to produce 2-d5.11 In both cases, deuteration at the α-N of the
anilide suggested that α-hydrogen abstraction precedes the C−
H activation step, therefore discarding the more common σ-
bond metathesis pathway generally associated with rare-earth
metal or lanthanide C−H bond activation.29,30 This combina-
tion of results implied that a transient scandium imido likely
was generated in all of these reactions [except for the reaction
of 1 with Al(CH3)3]11 and that the roles of the Lewis base and
Lewis acid are critical in the α-hydrogen abstraction pathway.
3. RESULTS AND DISCUSSION
Synthesis of Precursors to Scandium Imidos. Unlike
early transition metals belonging to groups 4−6, the electro-
positive nature coupled with the large ionic radius of the Sc3+
ion has made finding (or generating) terminal scandium imidos
a challenging task.7−9,16 In most cases, oligomerization of the
highly nucleophilic imide results in the formation of kinetically
stable complexes.9 Teuben and Hessen’s proposed transient
monomeric scandium complex suggested that it was only a
matter of ligand choice before a terminal imido of a rare-earth
ion could be generated and investigated in detail.9a In 2008, we
20085
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
Scheme 4. Proposed Paths for the Transformation of 1 into 2
C−H Activation of Pyridine by a Transient Scandium
Imido and Kinetic Isotope Effects. Previously, our group
had established by isotope labeling studies that C−H activation
of pyridine involves a transient scandium imido,3f,11 presumably
via 1,2-C−H bond addition analogous to that in other reported
imido systems.10 Since the ligands, PNP, and the peripherals of
the anilide group are not deuterated, our isotopic labeling
studies suggest that formation of the imido likely proceeds by
one or a combination of the three paths shown in Scheme 4.
Path I is analogous to that proposed for Chen and Cui’s
imidos2a,15d in that the σ donor must first bind to Sc before the
α-hydrogen abstraction proceeds. Path II is not Lewis base-
induced but instead involves direct α-hydrogen abstraction,
akin to the results of mechanistic studies provided by Bergman
and Wolczanski on transient group 4 imides.10a,b,e−h Path III
involves a concerted step wherein C−H abstraction and C−H
bond activation of pyridine occur in an asynchronous fashion
without the need to form a Sc−N multiple bond. This last path
is not to be confused with σ-bond metathesis involving only the
methyl and o-pyridine protons, which is clearly not operative on
the basis of isotopic labeling studies. Nevertheless, path III
should still depend on a pre-equilibrium step involving pyridine
binding. Distinguishing path III from path I or II might not be
trivial because the role of pyridine and α-hydrogen abstraction
may not be decoupled (interchange mechanism). However,
paths I and II can be discriminated since the main difference is
the role of pyridine in the reaction: binding occurs before or
after methane release and imido formation, respectively.
Our group has also shown that pyridine-d5 exchanges with 2
over 8 h at 90 °C to yield the isotopologue 2-d5, hinting that
the formation of the imide (PNP)ScN[DIPP](NC5H5) (A-
py) is reversible, therefore discrediting path III (Scheme 4).11
When 1 was treated with pyridine-d5 at 35 °C and the rate
compared with that for pyridine, a KIE of 1.1(2) was
observed,22 indicating that C−H activation of pyridine is not
rate-determining in the formation of the pyridyl. To establish
whether α-hydrogen abstraction is the slow step, the
isotopologue (PNP)Sc(ND[DIPP])(CH3) (1-d1) was pre-
pared. At 35 °C, a KIE of 5.37(6) was observed for three
independent runs (kavg = 3.41 × 10−5 M−1 s−1). When the latter
complex was heated to 50 °C, a KIE (or temperature-
dependent or equilibrium KIE) of 4.9(14) compared with the
rate for 1 → 2 was observed.22 The conversion of 1 to 2 (with
concomitant release of CDH3) without any deuterium
incorporation into the anilide nitrogen also suggests α-
hydrogen abstraction to be virtually irreversible under these
conditions and argues against hypothetical σ-complex inter-
mediates such as (PNP)Sc(N[DIPP])(σ-CDH3) or (PNP)Sc-
(N[DIPP])(σ-CDH3)(NC5H5) being stable or even present in
significant quantities along the reaction coordinate to allow for
exchange with the imide.19a,10m This was verified by the fact
that no isotope scrambling was observed when 1-d1 was
thermolyzed in deuterobenzene (3 days at 50 °C).22 The
isotopologue (PNP)Sc(NH[DIPP])(CD3) (1-d3), which was
synthesized from salt metathesis with LiNH[DIPP] and
(PNP)Sc(CD3)(I), also showed no evidence of isotopic
scrambling upon thermolysis (3 days at 50 °C).22 In summary,
our primary KIE data at 35 and 50 °C are consistent with α-
hydrogen abstraction being overall-rate-determining and in
accord with the formation of a scandium imido moiety. The
data collected at 35 °C argue that equilibrium isotope effects do
not make a major contribution to our values (see below). Our
reported KIE is also consistent with those for other α-H-
abstraction-based mechanisms reported with the pincer ligand
PNP.19
Dependence on Pyridine. The role of pyridine in the
formation of the scandium imide A-py is critical, since we,11
Chen,2a,15a−c Cui,15d and recently Piers9c have relied on a Lewis
base of this type presumably to promote α-hydrogen
abstraction. For this reason, we explored the role of pyridine
in the conversion of 1 to 2. From our proposed set of paths
described in Scheme 4, we can differentiate I from II via kinetic
analysis of the reaction. For instance, path I invokes a pre-
equilibrium between 1 and pyridine (Keq1 = k1/k−1) to generate
the intermediate (PNP)Sc(NH[DIPP])(CH3)(py) (1-py),
which then decomposes with rate constant k2 to generate A-
py and methane (eq 1). Path II relies on the loss of methane
from 1 with rate constant k3 to generate unsaturated imido
intermediate A, which should bind pyridine reversibly to form
A-py (Keq4 = k4/k−4; eq 2). On the basis of our KIE for
pyridine/pyridine-d5 (see above), A-py should be in equilibrium
with 2 in the post-rate-determining step (Keq5 = k5/k−5; eq 3).
However, at room temperature, this equilibrium should heavily
favor 2, as it is the only species observed spectroscopically. As a
result, eq 3 should not be part of our rate expressions derived
using eqs 1 and 2.
k1
k2
path I: 1 + py XooY 1−py ⎯⎯⎯⎯⎯→ A−py
−CH4
k−1
(1)
(2)
(3)
k4 ,+py
k3
path II: 1 ⎯⎯⎯⎯⎯→ A XoooooooY A−py
−CH4
k−4 ,−py
k5
tautomerization: A−py XooY 2
k−5
20086
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
By applying the pre-equilibrium approximation to path I, we
obtained the rate expression d[2]/dt = kobs[1][py], with kobs
=
Keq1k2.22 According to this analysis, if C−H activation of
pyridine occurs via path I, we should observe a first-order
dependence on pyridine and a first-order dependence on 1.
Likewise, for path I, as the concentration of pyridine
approaches infinity, the rate equation should then simplify to
d[2]/dt = k′2[1], where k2′ = k2[py]0 (i.e., pseudo-first-order
kinetics). Since path II does not involve a pre-equilibrium step,
we had to invoke the steady-state approximation, from which
the rate equation d[2]/dt = k3[1] was obtained.22 For this path,
the rate of the reaction should be independent of the pyridine
concentration. However, this analysis holds true only if there
are no other intermediates in solution. In addition to the
absence of side products, our plot in Figure 1 shows that the
decay of 1 and the formation of 2 were clean, without any side
Figure 2. Zeroth-order plot with respect to pyridine for the 1 → 2
transformation. The pyridine concentration was determined by
monitoring the integral of the NMR peak for the pyridine 2,6-
protons. The reaction was conducted at 40 °C.
1
reactions or detectable intermediates, as judged by H or 31P
NMR spectroscopy.
Figure 3. First-order plot with respect to pyridine for the 1 → 2
transformation. The pyridine concentration was determined by
monitoring the integral of the NMR peak for the pyridine 2,6-
protons. The reaction was conducted at 40 °C.
Figure 1. Plot showing the decay of 1 and formation of 2 at 40 °C
(obtained by monitoring the N−H resonances of 1 and 2). Kinetic
data for only ∼2 half-lives (t1/2 = 5.8 h) are shown. The reaction was
complete after ∼36 h.
If the reaction is zeroth order in substrate [as in the proposed
path II or our (PNP)Ti(IV) alkylidyne system, (PNP)Ti
CtBu],19 then a plot of [py] versus time would exhibit a linear
dependence. Alternatively, if the reaction is first order in
substrate, a plot of ln[py] versus time would be linear. When
we plotted both [py] versus time and ln[py] versus time, we
observed nonlinear fits (Figures 2 and 3). However, when a
second-order plot [ln([py]/[1]) vs time] was used, a linear
relationship (R2 = 0.998) was observed (Figure 4); to ensure
consistency in our measurements, both the N−H and Sc−CH3
protons were monitored to determine the concentration of 1.
Thus, our data are more in accord with the reaction being first
order in pyridine and first order in 1 (i.e., path I).
The strongest evidence that path I is the most likely route
was observed when we examined the effect of the pyridine
concentration on the rate of the reaction. To determine more
precisely the role of pyridine in the reaction, we conducted
concentration-dependent studies of the overall rate of the 1 →
2 conversion with [py] ranging from 0.07 to 1.0 M using C6D6
and NC5D5.22 As shown in Figure 5 and Table 2, the rate of the
reaction decayed as the concentration of pyridine was
decreased. No saturation effect was observed, which is
inconsistent with path II.
Figure 4. Second-order plot with respect to pyridine for the 1 → 2
transformation. The pyridine concentration was determined by
monitoring the integral of the NMR peak for the pyridine 2,6-
protons, and the concentration of 1 was determined by monitoring the
integrals of both the Sc−NH and Sc−CH3 resonances. The reaction
was conducted at 40 °C.
intermediates 1-py, A-py, and A. To that end, a solution of 1
and pyridine in toluene-d8 was cooled to −60 °C in order to
decrease the rate of binding of pyridine, which would allow us
to observe the adduct 1-py or the product arising from α-
hydrogen abstraction. Unfortunately, no evidence for 1-py was
Finally, to lend further support to the mechanism proposed
in path I, we examined evidence for the existence of the
20087
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
but after the complex was heated at 60 °C for 9 days, the
integrals of the anilide resonance and the pyridyl proton
resonance at 8.33 ppm had equilibrated to 0.75H and 0.67H,
respectively, with the integral of the pyridyl resonance at 7.77
ppm remaining at 1.0H. This exchange provides hard evidence
that A-py is accessible but formed slowly at higher temperatures
and that the imido is not accessible at room temperature,
consistent with the observed reactivity of complex 1.
To probe the possibility of the formation of A from 1,
complex 1 was thermolyzed at 70 °C in C6D6 in the absence of
other substrates such as pyridine. After 12 h, 1 was completely
consumed, and the reaction afforded several products, including
(PNP)Sc(NH[DIPP])(C6H5) [or, when C6D6 was used,
(PNP)Sc(ND[DIPP])(C6D5)] as the major product.11 As a
result, we cannot completely refute path II as a competing
reaction coordinate taking place at higher temperatures.
However, in the presence of pyridine, complex 2 is formed
1
quantitatively (as determined by H and 31P NMR spectros-
copy) with a t1/2 of 17.8 min at 70 °C, suggesting that path II is
unfeasible at room temperature or in the time frame for
completion of the reaction at higher temperatures. Although
path II has analogous precedent in our group when the
alkylidyne precursor (PNP)TiCHtBu(R) (R− = CH3,
19
t
CH2 Bu, CH2SiMe3, C6H5) was used, path I is more likely
for a reaction involving a donor substrate because of the vacant
coordination site available in 1 coupled with the much larger
ionic radius of Sc3+ (74.5 pm) compared with Ti4+(60.5 pm).17a
This path would also be more in accord with the necessity of a
Lewis base in the formation of Chen and Cui’s scandium
imides.2a,15
Activation Parameters for the Formation of a
Transient Scandium Imide. Since our system follows a
second-order rate expression, we can monitor the decays of
both 1 and pyridine over time. The integrated rate law
expressed in a linear form is shown in eq 4.
Figure 5. Plots showing the pyridine concentration dependence of the
rate of the 1 → 2 conversion. The top plot was obtained by
1
monitoring the decay of the anilide N−H H NMR resonance of 1,
and the bottom plot was obtained by monitoring the decay of the Sc−
CH3 1H NMR resonance of 1. [1]0 = 0.15 M in both plots.
Table 2. Values of kobs Showing the Pyridine Concentration
Dependence of the Rate of the 1 → 2 Conversion at 35 °C
⎛
⎜
⎝
⎞
⎟
⎠
[1]0 [py]
kobs
[py]0 − [1]0
[py] (M)
kobs (s−1
)
ln
=
t
0.07
0.22
1.0
1.42 × 10−6
3.74 × 10−5
1.62 × 10−4
8.63 × 10−4
2.66 × 10−3
[py] [1]
(4)
0
By monitoring the decays of both 1 and pyridine, we
determined the rate constant of the reaction at different
temperatures, thereby allowing us to obtain values of the
activation parameters. Accordingly, from the Eyring plot in
Figure 6 we were able to extract a ΔH⧧ value of 17.9(9) kcal/
mol, which is lower than those for Wolczanski’s imido system,
(tBu3SiNH)2Zr=NSitBu3 [25.9(4) kcal/mol],10h and our
group’s alkylidyne system, (PNP)TiCtBu [24(7) kcal/
mol].19d,f This lower enthalpy of activation could translate
into the transition state of 1-py (1-py-TS) being relatively early,
with both the N−H and Sc−CH3 bonds being essentially
preserved. The same type of transition state was observed in
the calculated pathway for the 2 → A-py tautomerization (see
below). In the case of ΔG⧧, however, the 1 → 2 conversion
yielded an estimated value of 23.5 kcal/mol at 308.15 K, which
is comparable to those for (tBu3SiNH)3Zr(CH3) →
(tBu3SiND)3Zr(C6D5) (∼28.5 kcal/mol at 308.15 K)10h and
5.0
10.0
detected by either 1H or 31P NMR spectroscopy, and the direct
observation of A-py or A was futile, as both intermediates occur
after the α-hydrogen abstraction step, which is overall-rate-
determining. Spectroscopic observation of such an intermediate
might be compromised by the fact that such species would have
spectroscopic signatures very similar to that of 1. However,
indirect evidence for formation of A-py was observed by the
thermolysis of (PNP)Sc(ND[DIPP])(η2-NC5H4) (2-d1), the
isotopologue of complex 2 containing ∼75% deuterium
incorporation at the anilide position (Scheme 5). Complex 2-
d1 was prepared from (PNP)Sc(ND[DIPP])(Cl) and Li-
(NC5H4) at −100 °C.31 The amount of proton/deuterium
incorporation in 2-d1 was determined by monitoring the
integrals of the N−H resonance and the proton resonances of
the pyridyl moiety. After 2 days at room temperature in C6D6,
there was no change in the integrals of the anilide resonance
(6.18 ppm, ∼0.36H) and the resonances of the 3- and 6-
protons on the pyridyl ring (7.77 and 8.33 ppm, respectively,
each resonance being a doublet integrating to 1.0H; Scheme 5),
(PNP)TiCHtBu(CH2 Bu) → (PNP)TiCDtBu(C6D5)
t
(∼24.6 kcal/mol at 308.15 K).19d,f The reason why the Gibbs
free energy of our (PNP)ScN[DIPP] system is comparable
to those of Wolczanski’s zirconium imido and our titanium
alkylidyne, despite its much lower enthalpy of activation, is that
the scandium system has a negative entropy of activation [ΔS⧧ =
−18(3) cal/(mol K)]. In fact, the entropic component accounts
20088
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
Scheme 5. Isotopic Labeling Studies Showing the Formation of the Imide Adduct A-py upon Heating at 60 °C for 9 days (Red
Spectrum, t = 0; Blue Spectrum, t = 9 days at 60 °C)
Scheme 6. Proposed Transition State Involved in the
Formation of the Imido A-py from 1-py
might play a dual role, as it could also behave as a Brønsted
base that promotes H+ transfer to the methyl. This mechanism,
though, is disfavored because the rate of activation of 2-
substituted pyridines is lower than that of unsubstituted
pyridine (DMAP > 4-picoline > pyridine > 2-picoline > 2-
iminopyridine or 2,6-lutidine; see below).
To understand further the steps involving pyridine binding,
α-hydrogen abstraction, and pyridine C−H bond activation, we
can break the pathway into two distinct stages. The first stage
Figure 6. Eyring plot for the 1 → 2 transformation in C6D6 over the
temperature range 35−74 °C.
for almost 24% of the total Gibbs free energy, which is
consistent with a highly ordered transition state along the
reaction coordinate. Both the zirconium and titanium systems
mentioned are zeroth order in substrate, so their major entropic
penalty is likely due to alignment in the molecule to allow for
α-proton abstraction [−7(1) cal/(mol K) for Zr;10h −2(3) cal/
(mol K) for Ti19d,f].
involves the conversion of 1 to the imido A-py via pyridine
binding and subsequent α-hydrogen abstraction, while the
second is the tautomerization of A-py to give 2. This second
step is what can allow for the formation of 2-d5 from 2, and it
also permits the study of the C−H bond activation event
without too much entropic dominance. Consequently, the
profile of the tautomerization step was calculated using density
functional theory. Calculations at the B3LYP/LACVP** level
were performed to elucidate the reaction profile but also to
discard other plausible mechanistic scenarios not involving the
formation of a scandium imido (A-py or A).22
Figure 7 depicts the most likely reaction profile for the
formation of the imide A-py via a tautomerization step
involving the pyridyl complex 2. As anticipated, the resting
state in complex 2 has the hydrogen atom of the anilide
nitrogen pointed away from the Sc−C bond of the pyridyl
ligand (2-anti). This orientation is consistent with the X-ray
Our highly negative ΔS⧧ is consistent with an associative
process involving the binding of pyridine to 1 to promote the
release of methane concomitant with formation of the imido,
which most likely is A-py. The highly negative ΔS⧧ also
suggests a highly constrained geometry preceding methane
elimination, which likely involves the four-centered transition
state 1-py-TS shown in Scheme 6. We assume that the role of
pyridine is to coordinate to the scandium metal center in order
to induce more steric crowding, which would place the methyl
moiety closer to the departing H+ of the anilide, but pyridine
20089
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
[2.385(2) Å].11 The transferred proton in A-py-TS is about
halfway between the pyridyl carbon and the imido nitrogen
(H−Nimide, 1.342 Å; H−Cpyridyl, 1.454 Å), indicating the TS to
be at the midway point between A-py and 2-syn.
A-py has a similar energy to 2, being only 1.7 kcal/mol
higher in energy than 2-anti, which implies that this species is
relatively stable. Figure 7 does not depict the entire energy
diagram involving equilibration of A-py back to 2, since this
involves the microscopic reverse (and hence, A-py is not
isolable). Although A-py is expected to be thermodynamically
stable relative to 2-syn, the energy required to form this species
results in the overall re-formation of 2-syn and ultimately 2-anti
by virtue of microscopic reversibility. This type of tautomeriza-
tion (at temperatures high enough to overcome the
thermodynamic barrier) has been exploited by our group to
generate iminopyridine ligands, in some cases catalytically.3f
The most notable feature of the reaction profile shown in
Figure 7 is the optimized computed structure of the imido A-
py. This putative intermediate has a seemingly short ScN
bond length of 1.84 Å, and the Sc−N−Cipso angle is 169.7°.
The short bond length and large bond angle imply that the
ScN[DIPP] bond should actually be considered as a Sc
N[DIPP] bond (Figure 7), in agreement with the bond
distance of 1.90 Å computed for a Sc−N triple bond using the
method put forth by Pauling (eq 5),17b
Figure 7. Computed reaction coordinate for the 2 → A-py
tautomerization. The rest of the reaction coordinate is not shown
since it represents the microscopic reverse, A-py → 2.
structure of 2 previously reported by our group as well as the X-
ray structure of 1.11 Prior to tautomerization, rotation of the
Sc−N bond occurs (with a small penalty of 10.3 kcal/mol) in
order to align the N−H bond with the basic pyridyl carbon (2-
syn). The barrier for this step mainly arises from the large steric
repulsion between the DIPP ring and the isopropyl groups of
the PNP ligand in the transition state 2-TS, wherein the anilide
hydrogen is rotated to give a H−N−Sc−C torsion angle of
105.6° instead of the more planar values of 173.0° for 2-anti
and 4.1° for 2-syn. The computed Sc−Nanilide bond distances in
2-anti (2.066 Å) and 2-syn (2.057) are similar to those found in
other Sc−Nanilide complexes (∼2.07 Å). In the computed
structure of 2-TS, the Sc−P distances of 2.751 and 2.779 Å are
slightly longer than those in complex 2 [2.7251(10) and
2.7397(10) Å], which supports our assertion that steric clashing
among these groups likely disfavors one isomer relative to the
other.
In addition to a small barrier for isomerization, the rotamer
2-syn is slightly endergonic by 4.7 kcal/mol, which implies that
fast interconversion of these species takes place at room
temperature. However, the slow step in our reaction coordinate
is the tautomerization involving 2-syn and A-py. Intramolecular
α-hydrogen migration costs an additional 20.0 kcal/mol (24.7
kcal/mol overall from 2-anti). As a result, the latter step is rate-
determining, akin to the formation of 2 from 1-py. The
computed transition state involved in the tautomerization step,
A-py-TS, involves a four-centered species in which the
migrating hydrogen is locked between the imide nitrogen and
the pyridyl carbon. The computed bond distances are
consistent with an early transition state. The Sc−Nimide bond
distance for A-py-TS is computed to be 1.907 Å, which is close
to the Sc−Nanilide distance reported for the AlMe3-trapped
imido complex (PNP)Sc(μ2-N[DIPP])(μ2-CH3)[Al(CH3)2]
[1.9366(14) Å] and the Sc−Nanilide bond distances of anilide
complexes (∼2.07 Å). However, this distance is slightly longer
than the ScNAr distances observed in the structurally
characterized imidos reported by Chen [1.881(5) and
1.8591(18) Å].2a,15b The Sc−C bond distance in A-py-TS
(2.520 Å) is rather long compared with both the average Sc−
Cpyridyl bond distance (∼2.22 Å)3f,11,14,31,32 and the Sc−CH3
bond length in (PNP)Sc(μ2-N[DIPP])(μ2-CH3)[Al(CH3)2]
DSc−N = rSc + rN − c|χSc − χN |
(5)
where DSc−N is the bond distance, c is the Schomaker−
Stevenson coefficient, and ri and χi are the covalent radius and
electronegativity, respectively, of atom i. Using eq 5, we
obtained a bond distance of 1.98 Å for a Sc−N double bond
(slightly longer than the ScN[DIPP] bond distance of 1.94 Å
in (PNP)Sc(μ2-N[DIPP])(μ2-CH3)[Al(CH3)2]) and a bond
distance of 2.06 Å for a Sc−N single bond (similar to the
observed average bond distance of 2.06 Å for our Sc−Nanilide
bonds). Additionally, these computed metrical parameters are
comparable to those for Chen’s structurally characterized
scandium imide [Sc−N = 1.881(5) and 1.8591(18) Å and Sc−
N−Cipso = 169.6(5)° and 167.90(17)°, respectively].2a,15b The
evidence that Sc−N multiple bonding in A-py takes place is
further augmented by the two orthogonal Sc−N π molecular
orbitals HOMO and HOMO−2 shown in Figure 8. These
orbitals suggest the imido ligand in complex A-py to be the
most basic site. The calculated Mayer bond order (MBO) of
the imido intermediate (1.56) is consistent with a Sc−N
interaction between those of a double bond and a triple bond21
and is similar to the MBO of 1.41 for the related phosphinidene
complex (PNP)Sc(μ2-P[DMP])(μ2-Br)[Li(THF)2] (DMP =
Figure 8. Most important frontier orbitals for the computed structure
of A-py. Isodensity is shown at 0.05 au.
20090
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
2,6-dimesitylphenyl), which also exhibits a pseudo-triple-bond
interaction.13e
Scheme 7. Reactions of 1 with Substituted Pyridines To
Form Complexes 3−9
Complexes bearing a scandium−ligand multiple bond have
been reported for ligands containing phosphorus, the heavier
congener of nitrogen.13e Our phosphinidene-ate complex
(PNP)Sc(μ2-P[DMP])(μ2-Br)[Li(THF)2], which contains the
same (PNP)Sc(III) framework, also shows pseudo-triple-bond
character between scandium and phosphorus, albeit more
polarized toward the phosphorus.13e In the present case,
however, the imido moiety in A-py undergoes 1,2-C−H bond
addition across the scandium−nitrogen multiple bond, and this
is presumably due to the better overlap of the smaller, harder
nitrogen atom with the rare-earth ion’s less diffuse 3d orbitals.
In contrast, (PNP)Sc(μ2-P[DMP])(μ2-Br)[Li(THF)2] engages
in group transfer reactions, although the reaction of the
phosphinidene complex with pyridine leads to decomposition,
resulting the formation of several products that we have been
unable to characterize.13e Such a contrast might be due to the
fact that the phosphinidene complex has an inert bromide ion
coordinating where the pyridine would have to bind to allow
for a 1,2-addition reaction. Additionally, the HOMO of
(PNP)Sc(μ2-P[DMP])(μ2-Br)[Li(THF)2] is orthogonal to
the plane of the substrate (enforced by the bulky DMP
group) (Figure 9), whereas the HOMO of A-py is directed
toward the C−H bond of pyridine, which facilitates the bond
activation step (Figure 8).
environment around the metal. This was indicated by a broad
methyne proton (3.43 ppm) and two isopropyl methyl
2
1
doublets (1.36 and 1.33 ppm, JHH = 6.8 Hz) in the H
NMR spectrum. The 13C NMR spectrum was consistent with a
pyridyl carbon at 213.8 ppm (broad), and the 31P NMR
spectrum exhibited two resonances at 3.8 and 4.4 ppm,
analogous to those of other scandium complexes having a
pyridyl moiety. The salient NMR resonances of 3 and other
complexes are summarized in Table 3.
Table 3. Salient NMR Spectroscopic Signatures (in ppm) of
1H, 31P, and 13C Nuclei for Complexes 1−9
complex
N−H
PNP
α-C
a
1
6.01
6.18
6.01
6.16
6.08
6.17
6.30
6.34
6.28
4.2; 2.5
4.5; 5.1
3.8; 4.4
3.9; 4.7
4.8; 5.6
4.8; 5.6
3.6; 5.6
4.8; 6.1
4.0; 5.9
30.4 (CH3)
217.2 (pyridyl)
213.8 (pyridyl)
216.0 (pyridyl)
216.9 (pyridyl)
53.3 (CH2)
Figure 9. Most important frontier orbitals for the computed structure
of (PNP)Sc(μ2-P[DMP])(μ2-Br)[Li(THF)2].13e The Li ion and THF
ligands have been omitted for clarity. Isodensity is shown at 0.05 au.
a
2
3
4
5
6
7
Effect of Other Pyridine Substrates on the Overall
Rate of the Reaction. We examined the effect of substituted
pyridine rings in the C−H bond activation step. According to
our proposed path I, the binding of pyridine to 1 promotes the
loss of methane to form A-py, which then promotes C−H
activation at the 2-position of pyridine to generate 2. Since
pyridine does have an effect on the C−H abstraction step and is
clearly involved in the TS leading to 2, we thermolyzed 1 with
various substituted pyridines to observe their effect on the rate.
Accordingly, when DMAP was added to 1 in C6D6 and the
mixture was monitored over time, the reaction was complete
within 30 min (k = 1.25 × 10−2 M−1 s−1 at 50 °C) and yielded
the 4-N,N-dimethylaminopyridyl complex (PNP)Sc(NH-
[DIPP])(η2-NC5H3-4-NMe2) (3) quantitatively (Scheme 7).
The connectivity in complex 3 was established on the basis of a
57.4 (CH2)
b
8
b
216.0 (pyridyl)
215.4 (pyridyl)
9
a
b
Data from ref 11. Data from ref 3f.
The reaction of 4-methylpyridine with 1 to generate
(PNP)Sc(NH[DIPP])(η2-NC5H3-4-Me) (4) (Scheme 7; k =
1.05 × 10−3 M−1 s−1 at 50 °C) was an order of magnitude
slower than the DMAP reaction but faster than C−H bond
activation by pyridine to yield 2 (k = 5.13 × 10−4 M−1 s−1 at 50
°C). The 1H NMR spectrum of 4 had features similar to those
of other C−H-activated pyridine products including the anilide
proton (6.16 ppm), the [DIPP] group exhibiting hindered
1
combination of H, 13C, and 31P NMR spectral data. For
rotation (iPr-methine: 3.40 ppm, broad; Pr-methyl: 1.36 ppm,
virtual triplet from two overlapping doublets, JH−H = 6.4 Hz),
and asymmetric PNP tolyl groups (2.18 and 2.23 ppm). The
31P NMR spectrum of complex 4 was also consistent with this
assignment, having two resonances at 3.9 and 4.7 ppm. Lastly,
i
1
example, the H NMR spectrum revealed the anilide N−H
3
resonance at 6.01 ppm, while the dimethylamino group showed
one singlet at 2.31 ppm consistent with a fluxional −NMe2
group. The [DIPP] aryl group exhibited slow rotation relative
to the NMR time scale due to the sterically hindered
20091
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
the pyridyl carbon appeared in the 13C NMR spectrum at 216.0
ppm. Akin to complex 2, single-crystal X-ray diffraction analysis
of complex 4 revealed a pseudo-octahedral Sc(III) ion having
an η2-bound pyridyl ligand [Sc1−N2, 2.192(2) Å; Sc1−C27,
2.216(2) Å] and an anilide ligand [Sc1−N3, 2.0497(18) Å]. No
α-H agostic interaction of the isotropically refined anilide
hydrogen (Sc1···H4, 2.363 Å) was observed, and the anilide
hydrogen points away from the pyridyl carbon (Figure 10).
polarization transfer (DEPT-135) NMR experiment was run on
the mixture, an oppositely phased signal was observed at 53.3
ppm,22 indicating the presence of a CH2 moiety. This would be
possible only if the methyl group of 2-picoline had been
activated. This feature, along with the presence of the o-H
resonance observed as a doublet at 7.75 ppm (3JHH = 7.1 Hz),
led us to conclude that one of the products was derived from
C−H bond activation at the 2-methyl carbon while the other
product was a pyridyl product analogous to 2 (Scheme 7). On
the basis of a combination of multinuclear NMR spectroscopic
data, we determined that the two products produced from C−
H bond activation were the complexes (PNP)Sc(NH[DIPP])-
(η2-NC5H3-2-CH3) (5) and (PNP)Sc(NH[DIPP])(κ2-C,N-
NC5H4-2-CH2) (6) (Scheme 7).22 Although the mixture of
complexes was inseparable because of their similar solubilities,
we resorted to single-crystal X-ray diffraction studies to
establish conclusively the site of C−H bond activation and
the binding mode of the pyridine motif in compound 6. To our
surprise, complex 5 cocrystallized with 6 in a ∼30:70 ratio in
the solid state (Figure 11).
The PNP and anilide ligands in 6 and 5 are virtually
crystallographically equivalent. The only disorder occurs in the
picoline ligand, and it is due to the two binding modes (η2 and
η3) shown in Figure 11, which are averaged over the entire
lattice.22 This allows for the direct measurement of both the sp2
and sp3 C−H activated products of the reaction of 2-picoline.
Since complexes 5 and 6 cocrystallized, their metrical
parameters correspond to the average values for the two
complexes [Sc1−N1, 2.190(2) Å; Sc1−N2, 2.059(2) Å; Sc1−
P1, 2.7337(7) Å; Sc1−P2, 2.7086(11) Å; P1−Sc1−P2,
144.60(3)°; Sc1−N2−Cipso, 158.22(19)°]. Given that the
disorder is centered around the activated 2-picoline, the
products corresponding to the different activation sites are
labeled separately in each structural diagram. The major
component, 6, has the picoline atoms labeled with just a
number, while those in the minor component, 5, are designated
with a D after the atom number (i.e., N3 for the picoline
nitrogen in 6 and N3D for the same nitrogen in 5). Even
though the solid-state X-ray structure of 5 resembles the
geometrical features observed in 2 and 4 [Sc1−N3D, 2.186(14)
Å; Sc1−C39D, 2.176(10) Å],11 the structure of complex 6
differs significantly. First, the coordinated pyridine group [Sc1−
N3, 2.338(7) Å] is intact, with activation of the C−H bond in
this case occurring at the 2-CH3 position. The binding mode of
the picolyl ligand in 6 appears to be η3 (Figure 11 right),
although the Sc1−C43 bond distance of 2.720(5) Å is longer
than those of other crystallographically characterized allyl
species (avg Sc−CAC distance = 2.44 Å, where CAC denotes the
central allylic carbon).33,34 This bond distance, however is
within the range observed for scandium−arene interactions
(2.50−2.84 Å).35 Presumably, the steric clash between the
picoline ring and the isopropyl groups of the PNP ligand
prevents the full allyl-type interaction. The dual reactivity of 2-
picoline is unique, because in general, C−H bond activation of
pyridine occurs in the sp2-CH and not the sp3-CH position in
picolines.33,36 Only in rare cases has activation of only the
methyl C−H bonds been observed.33,36b
Figure 10. Molecular structure of 4 with thermal ellipsoids at the 50%
probability level. Pr methyl groups on the PNP ligand, solvent
i
molecules, and H atoms (except for the α-hydrogen on N3) have been
omitted for clarity. Selected distances (Å) and angles (deg): Sc1−N1,
2.338(7); Sc1−C43, 2.720(5); Sc1−C44, 2.391(4); N3−C39,
1.355(10); N3−C43, 1.377(8); C43−C44, 1.408(7); Sc1−N2−C27,
73.14(14); N2−C27−Sc1, 71.21(13); Sc1−N3−Cipso, 160.64(17);
P1−Sc1−P2, 142.40(3).
When pyridine was substituted at the 4-position, the C−H
activation occurred at the 2-position, analogous to the case for
pyridine, albeit more rapidly for better σ donors. However,
when the pyridine was substituted at the 2-position, divergent
reactivity was observed. As anticipated for a more hindered
Lewis base, the reaction of 1 with 2-picoline (k = 5.7 × 10−5
M−1 s−1 at 50 °C) was much slower than that of unsubstituted
pyridine even though the pKa is higher than that of pyridine.
Only minimal reactivity was observed after 4 h at 50 °C. This
lack of reactivity is presumably due to the more sterically
crowded Lewis base. Regardless, when the mixture was heated
at 70 °C for 36 h, complex 1 was fully consumed as assayed by
31P NMR spectroscopy. To our surprise, however, there were
two separate anilide N−H resonances (6.08 and 6.17 ppm),
indicating that two products were formed. On the basis of the
integration of these resonances, the ratio of products was 46:54.
Further heating of the mixture did not alter the ratio of the two
products, suggesting that they were not equilibrating. Although
1
two products were observed in the H NMR spectrum, there
was no evidence of multiple products in the 31P NMR
spectrum. The 31P NMR spectra of the two compounds were
close to coincidental, showing two broad resonances at 4.8 ppm
(ν1/2 = 118 Hz) and 5.6 ppm (ν1/2 = 132 Hz) similar to those
of 1 (4.5 and 5.1 ppm). The peak width at half height (Δν1/2
)
To probe further the impact of the sterics at the 2- and 6-
positions on the rate of substrate activation, 1 was reacted with
2,6-lutidine, which was expected to interact with the metal
center significantly less than the other substrates while
maintaining a basicity similar to those of the other pyridine
derivatives. Accordingly, when 1 was allowed to react with 2,6-
for these resonances was broader than normal (80−90 Hz),
which is the only indication in the 31P NMR spectrum that
there were two overlapping resonances. A Sc−Cpy 13C
resonance was observed at 216.9 ppm, indicating that one of
the complexes was the expected C−H-activated product.
Interestingly, when a 135° distortionless enhancement by
20092
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
1
also observed in the H NMR spectrum (singlet at 1.70 ppm),
albeit significantly downfield-shifted from that of free lutidine
(2.35 ppm), and the anilide proton was observed as well (6.30
ppm). The fact that this reaction was so slow (t1/2 >10 days)
supports our assertion that the pyridine binds the metal center
before methane loss and imido formation. This also disfavors a
proton shuttling mechanism, as the steric bulk of the lutidine
would have a much less pronounced effect on its ability to
deprotonate. This reactivity is consistent with our proposal that
the pyridine substrate must bind to the metal center before the
complex releases methane and forms the imido, since among
the substituted pyridines presented here, DMAP is the most
basic [pKa values: pyridine, 5.23; 4-picoline, 5.99; 2-picoline,
6.00; 2,6-lutidine, 6.65; 4-aminopyridine (instead of DMAP),
9.11] and therefore the most nucleophilic.37 Notably, the
stronger the Lewis base, the swifter the α-hydrogen abstraction,
unless the base is too sterically hindered. This was also
observed with our previously reported activation of iminopyr-
idines.3f,7b At room temperature, less than 5% of 1 was
converted to the iminopyridine adduct. However, at 70 °C, the
reaction was complete in 14 h. Table 4 summarizes the rates
Table 4. Rates Measured at 50 °C for the Conversion of 1 to
Its Respective C−H-Activated Products Using Various
Substituted Pyridines
Figure 11. Molecular structures of (top) 5 and (bottom) 6 with
i
thermal ellipsoids at the 50% probability level. Pr methyl groups on
the PNP ligand, solvent molecules, and H atoms (except for the α-
hydrogen on N2) have been omitted for clarity. Selected distances (Å)
and angles (deg) for 5: Sc1−N3D, 2.186(14); Sc1−C39D, 2.176(10);
N3D−C39D, 1.370(17); N3D−C43D, 1.369(15); C43D−C44D,
1.423(14); Sc1−N3D−C39D, 71.3(7); N3D−C39D−Sc1, 72.1(7).
For 6: Sc1−N3, 2.338(7); Sc1−C43, 2.720(5); Sc1−C44, 2.391(4);
N3−C39, 1.355(10); N3−C43, 1.377(8); C43−C44, 1.408(7); N3−
C43−C44−Sc1, 31.3.
lutidine at room temperature, the reaction was >40% complete
after 10 days. Under thermolytic conditions (65 °C), the
reaction reached completion, generating the complex (PNP)-
Sc(NH[DIPP])(κ2-C,N-NC5H3-2-CH2-6-CH3) (7), which
shares some similar NMR spectral features with complex 6
(Scheme 7).22 The 31P NMR spectrum showed two resonances
at 3.6 and 5.6 ppm similar to the ones observed for the 5/6
mixture (4.8 and 5.6 ppm). Also, the DEPT-135 NMR
spectrum exhibited a resonance at 57.4 ppm,22 consistent
with activation of one of the lutidine methyl groups.30,33,36 This
α-carbon resonance corresponded by heteronuclear multiple-
bond coherence (HMQC) to the broad proton resonance at
3.24 ppm.22 The unactivated methyl resonance of lutidine was
measured for the conversion of 1 to the respective C−H
activation products obtained using various substituted
pyridines. It is quite possible that the formation of complexes
5−7 proceeds via an unsaturated imide A (or a mixture of A
and A-substrate, where the substrate is 2-picoline or 2,6-
lutidine), given the higher temperatures needed. This might
help explain why two C−H bond activation products were
observed in the case of 2-picoline.
20093
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
The reaction of 1 with 2-fluoropyridine was unfortunately
very slow and yielded several products that were detected by
1H and 31P NMR spectroscopy.22 We found that complex 1
also reacts with other σ donors, such as triethylamine (TEA),
trimethylphosphine, quinuclidine, and 2,2′-bipyridine, but
lamentably, these reactions yielded multiple (PNP)Sc(III)
products that precluded definitive characterization. Resonances
in the anilide N−H region of the 1H NMR spectrum that could
not be attributed to the starting material 1 were observed for
these reactions, indicating that C−H activation likely occurred
in parallel with other competing reactions.
ASSOCIATED CONTENT
■
S
* Supporting Information
X-ray crystallographic information (CIF), isotopic and kinetic
studies, kinetic analysis, and spectral data. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
mindiola@indiana.edu
Present Addresses
†Dalian Institute of Chemical Physics, Chinese Academy of
Sciences, Dalian 116023, China.
4. CONCLUSIONS
‡Department of Chemistry and Chemical Engineering, Royal
Military College of Canada, Kingston, Ontario, Canada K7K
7B4.
In this work, we have shown that the Sc(III) scaffold (PNP)Sc
can support a terminal imido group. However, the reactive
nature of the imide results in C−H bond activation of a
substrate such as pyridine. In the absence of a Lewis base,
activation of the solvent (C6H6) occurs, albeit sluggishly and
not cleanly. We have also demonstrated that a Lewis base such
as pyridine (or a close derivative thereof) promotes α-hydrogen
abstraction in complex 1. The Lewis base with the lowest pKa
results in the fastest α-hydrogen abstraction. We propose the
role of the pyridine to be twofold: (i) binding of the Lewis base
forces the methyl and primary anilide to be closer in space,
therefore lowering the barrier for proton transfer, and (ii) the
pyridine ligand allows for convenient trapping of the polarized
basic imide group, resulting in the formation of a pyridyl ligand
by 1,2-C−H bond addition. Evidence for direct participation of
pyridine (as a Brønsted base) in the proton transfer was
disproved by the fact that bulkier substrates such as 2-picoline
and 2,6-lutidine retarded the α-hydrogen abstraction. As a
result, our proposed pathway proceeds via an associative-type
mechanism in which scandium imide formation most likely
involves a pre-equilibrium step between 1 and the pyridine
adduct 1-py, followed by the α-hydrogen abstraction step,
which is overall-rate-determining. Although the imido moiety
can be accessed at higher temperatures without the aid of a
donor such as pyridine, this leads to multiple products and
occurs over a longer time frame than the C−H activation of
pyridine at the same temperature.
Via a deuteration study, we have shown indirectly that the
microscopic reverse reaction, the formation of A-py from 2,
occurs slowly at higher temperatures. As shown previously with
Sc(III), the phosphine arms of the PNP ligand can be
hemilabile,14 which could allow for binding of more sterically
encumbered pyridine derivatives such as 2,6-lutidine and 2-
iminopyridines. At present, we cannot discard the possibility
that at least one phosphine pendant group dissociates during
the binding of the Lewis base or that it even promotes H+
transfer. Kinetic isotope effects strongly imply imido formation
and loss of methane to be the slow step, as opposed to C−H
bond activation of pyridine. The present work has also provided
the first evidence of aliphatic C−H activation by a scandium
imido, since complex 1 can indiscriminately activate the methyl
group of 2-picoline versus the ortho position. Likewise, the
methyl group of 2,6-lutidine can be activated, albeit sluggishly.
Although complex 2 can tautomerize to A-py, attempts to
observe chemistry reminiscent of an imido (metathesis or imide
group transfer) were compromised by the reactive nature of the
pyridyl Sc−C bond.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support of this research was provided by the National
Science Foundation (CHE-0848248 and CHE-1152123). J.S.
acknowledges postdoctoral financial support from NSERC, and
M.G.C. acknowledges CONACYT for a postdoctoral fellow-
ship.
REFERENCES
■
(1) For reviews, see: (a) Wigley, D. E. Prog. Inorg. Chem. 1994, 42,
239. (b) Nugent, W. A.; Haymore, B. L. Coord. Chem. Rev. 1980, 31,
123. (c) Cummins, C. C. Prog. Inorg. Chem. 1998, 47, 685. (d) Hazari,
N.; Mountford, P. Acc. Chem. Res. 2005, 38, 839. (e) Berry, J. F.
Comments Inorg. Chem. 2009, 30, 28. (f) Mehn, M. P.; Peters, J. C. J.
Inorg. Biochem. 2006, 100, 634. (g) Sharp, P. R. Comments Inorg. Chem.
1999, 21, 85. (h) Smith, J. M. Comments Inorg. Chem. 2008, 29, 189.
(i) Abu-Omar, M. M. Dalton Trans. 2011, 40, 3435. (j) Fout, A. R.;
Kilgore, U. J.; Mindiola, D. J. Chem.Eur. J. 2007, 13, 9428.
(k) Cundari, T. R. J. Am. Chem. Soc. 1992, 114, 7879. (l) Eikey, R. A.;
Abu-Omar, M. M. Coord. Chem. Rev. 2003, 243, 83. (m) Mindiola, D.
J. Acc. Chem. Res. 2006, 39, 813. (n) Mindiola, D. J.; Bailey, B. C.;
Basuli, F. Eur. J. Inorg. Chem. 2006, 3135. (o) Che, C. M. Pure Appl.
Chem. 1995, 67, 225. (p) Zarubin, D. N.; Ustynyuk, N. A. Russ. Chem.
Rev. (Engl. Transl.) 2006, 75, 671. (q) Nugent, W. A.; Mayer, J. M. In
Metal−Ligand Multiple Bonds; Wiley-Interscience: New York, 1988.
(r) Holland, P. L. Acc. Chem. Res. 2008, 41, 905.
(2) Examples of terminal imidos at the extreme ends of the transition
metal series (group 3 and groups 8 and 9) have been reported. See:
(a) Lu, E.; Li, Y.; Chen, Y. Chem. Commun. 2010, 46, 4469.
(b) Efremenko, I.; Poverenov, E.; Martin, J.; Milstein, D. J. Am. Chem.
Soc. 2010, 132, 14886. (c) Poverenov, E.; Efremenko, I.; Frenkel, A.;
Ben-David, Y.; Shimon, L. J. W.; Leitus, G.; Konstantinovski, L.;
Martin, J.; Milstein, D. Nature 2008, 455, 1093. (d) Mindiola, D. J.;
Hillhouse, G. L. J. Am. Chem. Soc. 2001, 123, 4623. (e) Laskowski, C.
A.; Miller, A. J. M.; Hillhouse, G. L.; Cundari, T. R. J. Am. Chem. Soc.
2011, 133, 771. (f) Waterman, R.; Hillhouse, G. L. J. Am. Chem. Soc.
2008, 130, 12628. (g) Waterman, R.; Hillhouse, G. L. J. Am. Chem. Soc.
2003, 125, 13350. (h) Laskowski, C. A.; Hillhouse, G. L. Organo-
metallics 2009, 28, 6114. (i) Harrold, N. D.; Waterman, R.; Hillhouse,
G. L.; Cundari, T. R. J. Am. Chem. Soc. 2009, 131, 12872. (j) Kogut, E.;
Wiencko, H. L.; Zhang, L.; Cordeau, D. E.; Warren, T. H. J. Am. Chem.
Soc. 2005, 127, 11248. (k) Wiese, S.; McAfee, J. L.; Pahls, D. R.;
McMullin, C. L.; Cundari, T. R.; Warren, T. H. J. Am. Chem. Soc. 2012,
134, 10114. (l) Bai, G.; Stephan, D. W. Angew. Chem., Int. Ed. 2007,
46, 1856. (m) Ackermann, L. Organometallics 2003, 22, 4367.
(n) Lorber, C.; Choukroun, R.; Vendier, L. Organometallics 2004,
23, 1845. A copper nitrene dimer and transient monomer have been
20094
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
reported. See: (o) Badiei, Y. M.; Krishnaswamy, A.; Melzer, M. M.;
Warren, T. H. J. Am. Chem. Soc. 2006, 128, 15056.
was recently reported. See: (b) Conroy, K. D.; Piers, W. E.; Parvez, M.
Organometallics 2009, 28, 6228. (c) Chu, T.; Piers, W. E.; Dutton, J.
L.; Parvez, M. Organometallics 2012, DOI: 10.1021/om300913d.
(10) (a) Walsh, P. J.; Hollander, F. J.; Bergman, R. G. J. Am. Chem.
Soc. 1988, 110, 8729. (b) Cummins, C. C.; Baxter, S. M.; Wolczanski,
(3) (a) Duncan, A. P.; Bergman, R. G. Chem. Rec. 2002, 2, 431.
(b) Odom, A. L. Dalton Trans. 2005, 225. (c) Thomson, R. K.;
Bexrud, J. A.; Schafer, L. L. Organometallics 2006, 25, 4069.
(d) Eisenberger, P.; Schafer, L. L. Pure Appl. Chem. 2010, 82, 1503.
(e) Mindiola, D. J. Comments Inorg. Chem. 2008, 29, 73. (f) Wicker, B.
F.; Scott, J.; Fout, A. R.; Maren, P.; Mindiola, D. J. Organometallics
2011, 30, 2453. (g) Cantrell, G. K.; Meyer, T. Y. J. Am. Chem. Soc.
1998, 120, 8035. (h) Pohlki, F.; Doye, S. Chem. Soc. Rev. 2003, 32,
104. (i) Bytschkov, I.; Doye, S. Eur. J. Org. Chem. 2003, 935.
(j) Pohlki, F.; Doye, S. Angew. Chem., Int. Ed. 2001, 40, 2305.
(k) Anderson, L. L.; Schmidt, J. A. R.; Arnold, J.; Bergman, R. G.
Organometallics 2006, 25, 3394. (l) Hill, J. E.; Profilet, R. D.; Fanwick,
P. E.; Rothwell, I. P. Angew. Chem., Int. Ed. Engl. 1990, 29, 713.
(4) For other selected examples of imides used in catalysis, see:
(a) Anderson, L. L.; Arnold, J.; Bergman, R. G. Org. Lett. 2004, 6,
2519. (b) Cao, C.; Shi, Y.; Odom, A. L. J. Am. Chem. Soc. 2003, 125,
2880. (c) Li, Y.; Shi, Y.; Odom, A. L. J. Am. Chem. Soc. 2004, 126,
1794. (d) Shi, Y.; Ciszewski, J. T.; Odom, A. L. Organometallics 2001,
́
P. T. J. Am. Chem. Soc. 1988, 110, 8731. (c) Royo, P.; Sanchez-Nieves,
J. J. Organomet. Chem. 2000, 597, 61. (d) De With, J.; Horton, A. D.
Angew. Chem., Int. Ed. Engl. 1993, 32, 903. (e) Hoyt, H. M.; Bergman,
R. G. Angew. Chem., Int. Ed. 2007, 46, 5580. (f) Cundari, T. R.;
Klinckman, T. R.; Wolczanski, P. T. J. Am. Chem. Soc. 2002, 124, 1481.
(g) Schaller, C. P.; Bonanno, J. B.; Wolczanski, P. T. J. Am. Chem. Soc.
1994, 116, 4133. (h) Schaller, C. P.; Cummins, C. C.; Wolczanski, P.
T. J. Am. Chem. Soc. 1996, 118, 591. (i) Hanna, T. E.; Keresztes, I.;
Lobkovsky, E.; Bernskoetter, W. H.; Chirik, P. J. Organometallics 2004,
23, 3448. (j) Walsh, P. J.; Hollander, F. J.; Bergman, R. G.
Organometallics 1993, 12, 3705. (k) De With, J.; Horton, A. D.;
Orpen, A. G. Organometallics 1993, 12, 1493. (l) De With, J.; Horton,
A. D.; Orpen, A. G. Organometallics 1990, 9, 2207. (m) Schafer, D. F.;
Wolczanski, P. T. J. Am. Chem. Soc. 1998, 120, 4881.
(11) Scott, J.; Basuli, F.; Fout, A. R.; Huffman, J. C.; Mindiola, D. J.
20, 3967. (e) Vujkovic, N.; Ward, B. D.; Maisee-Franco̧ is, A.;
Angew. Chem., Int. Ed. 2008, 47, 8502.
(12) Labinger, J. A.; Bercaw, J. E. Nature 2002, 417, 507.
Wadepohl, H.; Mountford, P.; Gade, L. H. Organometallics 2007, 26,
5522. (f) Bolton, P. D.; Mountford, P. Adv. Synth. Catal. 2005, 347,
355. (g) Ong, T.-G.; Yap, G. P. A.; Richeson, D. S. Organometallics
2002, 21, 2839. (h) Dubberly, S. R.; Friedrich, A.; Willman, D. A.;
Mountford, P.; Radius, U. Chem.Eur. J. 2003, 9, 3634. (i) Roizen, J.
L.; Harvey, M. E.; Du Bois, J. Acc. Chem. Res. 2012, 45, 911. (j) Du
Bois, J.; Tomooka, C. S.; Hong, J.; Carreira, E. M. Acc. Chem. Res.
(13) (a) Masuda, J. D.; Jantunen, K. C.; Ozerov, O. V.; Noonan, K. J.
T.; Gates, D. P.; Scott, B. L.; Kiplinger, J. L. J. Am. Chem. Soc. 2008,
130, 2408. (b) Scott, J.; Fan, H.; Wicker, B. F.; Fout, A. R.; Baik, M.-
H.; Mindiola, D. J. J. Am. Chem. Soc. 2008, 130, 14438. (c) Cui, P.;
Chen, Y.; Xu, X.; Sun, J. Chem. Commun. 2008, 5547. (d) Zimmer-
mann, M.; Takats, J.; Kiel, G.; Tornroos, K. W.; Anwander, R. Chem.
̈
1997, 30, 364. (k) Muniz, K. Chem. Soc. Rev. 2004, 33, 166. (l) Romao,
Commun. 2008, 612. (e) Wicker, B. F.; Scott, J.; Andino, J. G.; Gao, X.;
Park, H.; Pink, M.; Mindiola, D. J. J. Am. Chem. Soc. 2010, 132, 3691.
(f) Mills, D. P.; Soutar, L.; Lewis, W.; Blake, A. J.; Liddle, S. T. J. Am.
Chem. Soc. 2010, 132, 14379. (g) Fustier, M.; Le Goff, X. F.; Le Floch,
P.; Mezailles, N. J. Am. Chem. Soc. 2010, 132, 13108. (h) Zimmermann,
M.; Rauschmaier, D.; Eichele, K.; Tornroos, K. W.; Anwander, R.
Chem. Commun. 2010, 46, 5346. (i) Cui, P.; Chen, Y.; Borzov, M. V.
Dalton Trans. 2010, 39, 6886. (j) Korobkov, I.; Gambarotta, S.
Organometallics 2009, 28, 5560. (k) Zhang, W.-X.; Wang, Z. T.;
Nishiura, M.; Xi, Z. F.; Hou, Z. M. J. Am. Chem. Soc. 2011, 133, 5712.
(l) Liddle, S. T.; Mills, D. P.; Wooles, A. J. Chem. Soc. Rev. 2011, 40,
2164. In the gas phase, lanthanide compounds having metal−ligand
multiple bonds have been detected. See: (m) Gong, Y.; Wang, X.;
Andrews, L.; Chen, M.; Dixon, D. A. Organometallics 2011, 30, 4443.
(14) Wicker, B. F.; Pink, M.; Mindiola, D. J. Dalton Trans. 2011, 40,
9020.
̃
̃
C. C.; Kuhn, F. E.; Herrmann, W. A. Chem. Rev. 1997, 97, 3197.
̈
(m) Muller, T. E.; Hultzsch, K. C.; Yus, M.; Foubelo, F.; Tada, M.
̈
Chem. Rev. 2008, 108, 3795. (n) Majumder, S.; Gipson, K. R.; Staples,
R. J.; Odom, A. L. Adv. Synth. Catal. 2009, 351, 2013. (o) Vujkovic, N.;
Fillol, J. L.; Ward, B. D.; Wadepohl, H.; Mountford, P.; Gade, L. H.
Organometallics 2008, 27, 2518. (p) Leung, S. K.-Y.; Tsui, W.-M.;
Huang, J.-S.; Che, C.-M.; Liang, J.-L.; Zhu, N. J. Am. Chem. Soc. 2005,
127, 16629. (q) Walsh, P. J.; Baranger, A. M.; Bergman, R. G. J. Am.
Chem. Soc. 1992, 114, 1708. (r) McGrane, P. L.; Jensen, M.;
Livinghouse, T. J. Am. Chem. Soc. 1992, 114, 5459. (s) Johnson, J. S.;
Bergman, R. G. J. Am. Chem. Soc. 2001, 123, 2923. (t) Lian, B.;
Spaniol, T. P.; Horrillo-Martínez, P.; Hultzsch, K. C.; Okuda, J. Eur. J.
Inorg. Chem. 2009, 429. (u) Ruck, R. T.; Zuckerman, R. L.; Krska, S.
W.; Bergman, R. G. Angew. Chem., Int. Ed. 2004, 43, 5372. (v) Aneetha,
H.; Basuli, F.; Bollinger, J.; Huffman, J. C.; Mindiola, D. J.
Organometallics 2006, 25, 2402. (w) King, E. R.; Hennessy, E. T.;
Betley, T. A. J. Am. Chem. Soc. 2011, 133, 4917.
(15) (a) Lu, E.; Zhou, Q.; Li, Y.; Chu, J.; Chen, Y.; Leng, X.; Sun, J.
Chem. Commun. 2012, 48, 3403. (b) Lu, E.; Chu, J.; Chen, Y.; Borzov,
M. V.; Li, G. Chem. Commun. 2011, 47, 743. (c) Chu, J. X.; Lu, E. L.;
Liu, Z.; Chen, Y.; Leng, X.; Song, H. Angew. Chem., Int. Ed. 2011, 50,
7677. (d) Jian, Z.; Rong, W.; Mou, Z.; Pan, Y.; Xie, H.; Cui, D. Chem.
Commun. 2012, 48, 7516.
(16) (a) Knight, L. K.; Piers, W. E.; McDonald, R. Organometallics
2006, 25, 3289. (b) Knight, L. K.; Piers, W. E.; Fleurat-Lessard, P.;
Parvez, M.; McDonald, R. Organometallics 2004, 23, 2087.
(5) (a) Schrock, R. R. Chem. Rev. 2009, 109, 3211. (b) Wright, W. R.
H.; Batsanov, A. S.; Messinis, A. M.; Howard, J. A. K.; Tooze, R. P.;
Hanton, M. J.; Dyer, P. W. Dalton Trans. 2012, 41, 5502. (c) La Pierre,
H. S.; Arnold, J.; Toste, F. D. Angew. Chem., Int. Ed. 2011, 50, 3900.
(d) Bolton, P. D.; Mountford, P. Adv. Synth. Catal. 2005, 347, 355.
(e) Coles, M. P.; Dalby, C. I.; Gibson, V. C.; Clegg, W.; Elsegood, M.
R. J. J. Chem. Soc., Chem. Commun. 1995, 1709. (f) Coles, M. P.;
Gibson, V. C. Polym. Bull. 1994, 33, 529. (g) Nomura, K. In New
Developments in Catalysis Research; Bevy, L. P., Ed.; Nova Science
Publishers: New York, 2005; p 199. (h) Nomura, K.; Zhang, W. Chem.
Sci. 2010, 1, 161. (i) Nomura, K.; Sagara, A.; Imanishi, Y.
Macromolecules 2002, 35, 1583. (j) Wang, W.; Nomura, K.
Macromolecules 2005, 38, 5905. (k) Wang, W.; Nomura, K. Adv.
Synth. Catal. 2006, 348, 743. (l) Onishi, Y.; Katao, S.; Fujiki, M.;
Nomura, K. Organometallics 2008, 27, 2590. (m) Zhang, S.; Katao, S.;
Sun, W.-H.; Nomura, K. Organometallics 2009, 28, 5925.
(17) (a) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. (b) Pauling,
L. The Nature of the Chemical Bond, 3rd ed.; Cornell University Press:
Ithaca, NY, 1960.
(18) (a) Andino, J. G.; Kilgore, U. J.; Pink, M.; Ozarowski, A.;
Kyzstek, J.; Telser, J.; Baik, M.-H.; Mindiola, D. J. Chem. Sci. 2010, 1,
351. (b) Basuli, F.; Kilgore, U. J.; Hu, X.; Meyer, K.; Pink, M.;
Huffman, J. C.; Mindiola, D. J. Angew. Chem., Int. Ed. 2004, 43, 3156.
(19) (a) Flores, J. A.; Cavaliere, V. N.; Buck, D.; Pinter, B.; Chen, G.;
Crestani, M. G.; Baik, M.-H.; Mindiola, D. J. Chem. Sci. 2011, 2, 1457.
(b) Cavaliere, V. N.; Crestani, M. G.; Pinter, B.; Pink, M.; Chen, C.-
H.; Baik, M.-H.; Mindiola, D. J. J. Am. Chem. Soc. 2011, 133, 10700.
(c) Fout, A. R.; Scott, J.; Miller, D. L.; Bailey, B. C.; Pink, M.;
Mindiola, D. J. Organometallics 2009, 28, 331. (d) Bailey, B. C.; Fan,
H.; Huffman, J. C.; Baik, M.-H.; Mindiola, D. J. J. Am. Chem. Soc. 2007,
129, 8781. (e) Bailey, B. C.; Huffman, J. C.; Mindiola, D. J. J. Am.
Chem. Soc. 2007, 129, 5302. (f) Bailey, B. C.; Fan, H.; Baum, E. W.;
(6) Basuli, F.; Annetha, H.; Huffman, J. C.; Mindiola, D. J. J. Am.
Chem. Soc. 2005, 127, 17992.
(7) (a) Mindiola, D. J.; Scott, J. Nat. Chem. 2011, 3, 15. (b) Scott, J.;
Mindiola, D. J. Dalton Trans. 2009, 8463.
(8) Giesbrecht, G. R.; Gordon, J. C. Dalton Trans. 2004, 2387.
(9) (a) Beetstra, D. J.; Meetsma, A.; Hessen, B.; Teuben, J. H.
Organometallics 2003, 22, 4372. Another example of a bridging imido
20095
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096

Journal of the American Chemical Society
Article
Huffman, J. C.; Baik, M.-H.; Mindiola, D. J. J. Am. Chem. Soc. 2005,
127, 16016.
(20) For a general description of the equipment and techniques used
in carrying out this chemistry, see: Burger, B. J.; Bercaw, J. E. ACS
Symp. Ser. 1987, 357, 79.
(21) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.
(22) See the Supporting Information.
(23) Masuda, J. D.; Wei, P.; Stephan, D. W. Dalton Trans. 2003,
3500.
(24) APEX2; Bruker Analytical X-Ray Systems: Madison, WI, 2009.
(25) Blessing, R. Acta Crystallogr. 1995, A51, 33.
(26) (a) Fan, L.; Foxman, B. M.; Ozerov, O. V. Organometallics 2004,
23, 326. (b) Fan, L.; Yang, L.; Guo, C.; Foxman, B. M.; Ozerov, O. V.
Organometallics 2004, 23, 4778. (c) Ozerov, O. V.; Guo, C.; Papkov,
V. A.; Foxman, B. M. J. Am. Chem. Soc. 2004, 126, 4792.
(27) Fryzuk, M. D.; Giesbrecht, G.; Rettig, S. J. Organometallics 1996,
15, 3329.
(28) Bailey, B. C.; Fout, A. R.; Fan, H.; Tomaszewski, J.; Huffman, J.
C.; Mindiola, D. J. Angew. Chem., Int. Ed. 2007, 46, 8246.
(29) (a) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.;
Nolan, M. C.; Santasiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am.
Chem. Soc. 1987, 109, 203. (b) Shapiro, P. J.; Bunel, E.; Schaefer, W.
P.; Bercaw, J. E. Organometallics 1990, 9, 867. (c) Watson, P. L.;
Parshall, G. W. Acc. Chem. Res. 1985, 18, 51. (d) Watson, P. L. J. Chem.
Soc., Chem. Commun. 1983, 276. (e) Watson, P. L. J. Am. Chem. Soc.
1983, 105, 6491. (f) Sadow, A. D.; Tilley, T. D. J. Am. Chem. Soc.
2005, 127, 643. (g) Sadow, A. D.; Tilley, T. D. J. Am. Chem. Soc. 2003,
125, 7971. (h) Barros, N.; Eisenstein, O.; Maron, L.; Tilley, T. D.
Organometallics 2008, 27, 2252. (i) Sadow, A. D.; Tilley, T. D. Angew.
Chem., Int. Ed. 2003, 42, 803. (j) Fendrick, C. M.; Marks, T. J. J. Am.
Chem. Soc. 1984, 106, 2214. (k) Bruno, J. W.; Smith, G. M.; Marks, T.
J.; Fair, C. K.; Schultz, A. J.; Williams, J. M. J. Am. Chem. Soc. 1986,
108, 40.
(30) For some examples of pyridine C−H activation by an early
transition metal or rare-earth metal, see: (a) Bi, S; Lin, Z.; Jordan, R. F.
Organometallics 2004, 23, 4882. (b) Guram, A. S.; Jordan, R. F.
Organometallics 1991, 10, 3470. (c) Jordan, R. F.; Guram, A. S.
Organometallics 1990, 9, 2116. (d) Jordan, R. F.; Taylor, D. F.;
Baenziger, N. C. Organometallics 1990, 9, 1546. (e) Deelman., B.-J.;
Stevels, W. M.; Teuben, J. H.; Lakin, M. T.; Spek, A. L. Organometallics
1994, 13, 3881. (f) den Haan, K. H.; Wielstra, Y.; Teuben, J. H.
Organometallics 1987, 6, 2053.
(31) (a) Bailey, B. C.; Fan, H.; Huffman, J. C.; Baik, M.-H.; Mindiola,
D. J. J. Am. Chem. Soc. 2006, 128, 6798. (b) Parham, W. E.; Piccirilli, R.
M. J. Org. Chem. 1977, 42, 257.
(32) For comparison, titanium pyridyl complexes have also been
structurally characterized. See: Bolton, P. D.; Clot, E.; Adams, N.;
Dubberley, S. R.; Cowley, A. R.; Mountford, P. Organometallics 2006,
25, 2806.
(33) Activation of the 2-picolene CH3 group has been reported and a
similar η3 binding mode has been observed with a yttrium complex.
See: Duchateau, R.; Brussee, E. A. C.; Meetsma, A.; Teuben, J. H.
Organometallics 1997, 16, 5506.
(34) (a) Standfuss, S.; Abinet, E.; Spaniol, T. P.; Okuda, J. Chem.
Commun. 2011, 47, 11441. (b) Yu, N.; Nishiura, M.; Li, X.; Xi, Z.;
Hou, Z. Chem.Asian J. 2008, 3, 1406.
(35) (a) Huang, W.; Khan, S. I.; Diaconescu, P. L. J. Am. Chem. Soc.
2011, 133, 10410. (b) Hayes, P. G.; Piers, W. E.; Parvez, M. Chem.
Eur. J. 2007, 13, 2632.
(36) (a) Pool, J. A.; Scott, B. L.; Kiplinger, J. L. J. Am. Chem. Soc.
2005, 127, 1338. (b) Deelman, B.-J.; Stevels, W. M.; Teuben, J. H.;
Lakin, M. T.; Spek, A. L. Organometallics 1994, 13, 3881.
(37) CRC Handbook of Chemistry and Physics, 92nd ed.; CRC Press:
Boca Raton, FL, 2011.
20096
dx.doi.org/10.1021/ja307279r | J. Am. Chem. Soc. 2012, 134, 20081−20096