Yttrium-Catalyzed Amine-Silane Dehydrocoupling: Extended Reaction Scope with a Phosphorus-Based Ligand

Article abstract of DOI:10.1021/acs.organomet.5b00607

The scope of the catalytic dehydrocoupling of primary and secondary amines with phenylsilanes has been investigated using [Y{N(SiMe₃)₂}₃] and a four-coordinate analogue bearing a cyclometalated phosphonium methylide ligand

Full text of DOI:10.1021/acs.organomet.5b00607

Article
pubs.acs.org/Organometallics
Yttrium-Catalyzed Amine−Silane Dehydrocoupling: Extended
Reaction Scope with a Phosphorus-Based Ligand
Adi E. Nako, Wenyi Chen, Andrew J. P. White, and Mark R. Crimmin*
Department of Chemistry, Imperial College London, South Kensington SW7 2AZ, U.K.
S
* Supporting Information
ABSTRACT: The scope of the catalytic dehydrocoupling of primary and
secondary amines with phenylsilanes has been investigated using [Y{N-
(SiMe3)2}3] and a four-coordinate analogue bearing a cyclometalated
phosphonium methylide ligand. Inclusion of the phosphorus-based ligand
on yttrium results in increased substrate scope in comparison to the
tris(amide) analogue. While reversible C−H bond activation of the
cyclometalated ligand was observed in stoichiometric experiments, D-
labeling experiments and DFT calculations suggest that reversible ligand
activation is not involved in silazane formation under catalytic conditions.
We suggest that the extended reaction scope with the four-coordinate yttrium phosphonium methylide complex relative to the
three-coordinate yttrium (tris)amide complex is a result of differences in the ease of amine inhibition of catalysis.
occurs through either a concerted σ-bond metathesis step
between M−N and Si−H bonds or by nucleophilic attack of
the nitrogen atom of a metal amide onto the electrophilic Si
center to generate a hypervalent silicate intermediate that then
decomposes through hydrogen elimination.13,14
We recently introduced the cyclometalated complex 1. This
complex incorporates the three elements of ligand design
outlined in Figure 1namely appended X- and L-type ligands
INTRODUCTION
■
The dehydrocoupling of amines and silanes to form silazanes is
a reaction that has received considerable attention over the last
5 years.1 While early investigations into silazanes focused on the
formation of polysilazanes and their pyrolysis to form silicon
nitride (Si3N4),2 in recent years the Si−N moiety has found
extensive use as a protecting group in organic synthesis or as
part of kinetically stabilizing ligands in coordination chem-
istry.3,4 Catalysts based upon elements from across the periodic
table have been reported to effect the cross-dehydrocoupling of
amines and silanes. Although catalysts incorporating Pt,5 Rh,6
Ru,7 Ti,8 Cu,9 and U10 along with those based on Lewis acids11
and Lewis bases12 have been known for some time, recently
amide and alkyl complexes of the s-block and rare-earth
elements have emerged as highly efficient and inexpensive
mediators of this reaction.
Figure 1. Catalyst design for small molecule activation with rare-earth
metals.
For example, in 2007 Harder and co-workers reported that
the dehydrocoupling of amines and silanes could be efficiently
catalyzed by [(η2-Ph2CNPh)M(HMPA)3] (M = Ca, Yb).13a
Sadow and co-workers demonstrated improved reaction
selectivity and scope through use of the magnesium complex
[{ToM}MgMe] (ToM = tris(4,4-dimethyl-2-oxazolinyl)-
phenylborate).13b Bis(amide) and bis(alkyl) complexes of Ca,
Sr, Ba, and the related divalent lanthanide Yb, including
[M{N(SiMe3)2}2] (M = Ca, Sr, Ba) and [(IMes)Yb{N-
(SiMe3)2}2] (IMes = 1,3-bis(2,4,6-trimethylphenyl)imidazol-2-
ylidene) have been reported as catalysts for Si−N bond
formation by the groups of Hill, Cui, and Carpentier.13,14 The
turnover frequency and turnover number for one of the most
active barium catalysts, [Ba{CH(SiMe3)2}2(THF)3], reach
3600 h−1 and 396, respectively, for the dehydrocoupling of
pyrrolidine and triphenylsilane.13d Divergent mechanistic
arguments have materialized from these studies. Depending
on the catalyst, it has been proposed that Si−N bond formation
along with reactive σ-bonded substiuents.15 Here we show that
1 is a highly effective catalyst for the dehydrocoupling of amines
and silanes. The inclusion of the phosphonium methylide
ligand on yttrium results in an increase in the scope and
efficiency of catalysis relative to the parent tris(amide) complex
[Y{N(SiMe3)2}3] (2).
Substrate activation through a reversible reaction of a
chelating ligand has long been considered in the organometallic
chemistry of the rare-earth elements. For example, the potential
for cyclopentiadienyl systems to undergo reversible ligand
activation has been appreciated since the early studies of
Watson, Bercaw, and others on the addition of H−H and C−H
and bonds to [(Cp*)2M(Me)] (M = Lu, Sc).16−18 This
observation of ligand-based reactivity is not limited to tuck-in
Received: July 14, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.organomet.5b00607
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
complexes or related species bearing appended X-type ligands:
in 2010, Arnold and Turner reported the 1,2-addition of E−X
(E = Si, Sn, P, B; X = Cl, N3) bonds across the M−C bond of
N-heterocyclic carbene adducts of scandium, yttrium, cerium,
and uranium, along with elimination and subsequent carbon−
heteroatom and nitrogen−heteroatom bond formation from
isolated zwitterionic intermediates.19 Parallels may be drawn
with frustrated Lewis pairs based on zirconium and yttrium
reported by Wass and co-workers.20 Broadly both systems
could be classified as a reactive metal with an appended L-type
ligand (Figure 1).
Table 1. Reaction Scope of Amine−Silane Dehydrocoupling
Catalyzed by 1 or 2
As part of the current contribution, we have investigated the
potential of 1 to react with substrates by reversible ligand
activation under both stoichiometric and catalytic conditions.
Although substrate activation by participation of both the
appended L- and X-type ligands is possible under stoichio-
metric conditions, through the isolation of catalytic inter-
mediates, D-labeling experiments, inhibition experiments, and
DFT calculations, we conclude that reversible ligand activation
is not involved in silazane formation under catalytic conditions.
The extended reaction scope of catalyst 1 with respect to
catalyst 2 is rationalized through a mechanism in which amine
inhibition is more significant for the three-coordinate tris-
(amide) 2 than for the four-coordinate complex 1.
RESULTS
■
i
Reaction Scope. The selective dehydrocoupling of Pr2NH
and PhSiH3 may be catalyzed by 5 mol % [Y{N(SiMe3)2}3]
i
(2), producing Pr2NSiH2Ph in 93% conversion after 164 h at
80 °C in C6D6.7,8 Upon modification of the precatalyst to 1,
preparations proceed to high conversion within 24 h at 25 °C.
Control experiments with 2 or mixtures of 2 and Ph3PCH2
revealed only trace formation of iPr2NSiH2Ph after weeks at 25
°C (Scheme 1).
Scheme 1. Catalytic Dehydrocoupling of Diisopropylamine
with Phenylsilane
Reactions conducted with a 1:1 mixture of silane and amine. Yields
1
were recorded by H NMR spectroscopy against hexamethylbenzene
a
b
or ferrocene as an internal standard. 57% after 2 h. 83% after 0.25 h.
c
d
e
f
80% after 2.5 h. 60% after 19 h. 17% after 21 h. 22% after 19 h.
g
12% after 19 h.
reveals near-identical activities for both catalysts with no
observable induction period for either (Figures S1 and S2, in
the Supporting Information). A handful of catalytic systems are
known to be competent for the dehydrocoupling of sterically
demanding secondary amines with silanes.13,14 It is for these
substrates that stark differences are noted for catalysts 1 and 2.
While 1 rapidly dehydrocouples dicyclohexylamine, 2,6-
dimethylpiperidine, and isopropylcyclohexylamine with phenyl-
silane either at 25 °C or within a few hours at 80 °C, catalyst 2
is either inactive or requires more than 1 week at 80 °C to
reach moderate to acceptable yields of the corresponding
silazanes.
On the basis of this observation we investigated 1 and 2 as
catalysts for the cross-dehydrocoupling of a series of amines
and silanes. The catalysts are selective for the 1:1 reaction of a
number of primary and secondary amines with a series of
organosilanes. The reaction scope along with a comparison of
the two precatalysts is presented in Table 1. The products have
been characterized by 1H, 29Si, 15N, and 13C NMR spectroscopy
(see the Supporting Information).
Catalyst Resting States. The J coupling between 89Y and
31P nuclei observed in NMR offers a useful spectroscopic
handle to investigate the catalytic resting states. Monitoring
1
catalytic preparations by 31P{1H} and H NMR allowed the
Catalysts 1 and 2 show comparable activities for the addition
of primary amines and unhindered secondary amines to
phenylsilane and diphenylsilane. For example, comparison of
the initial rates of the reaction of piperidine with diphenylsilane
identification of two distinct regimes that are dependent on the
amine employed. For sterically unhindered primary amines
such as N,N-dimethylhydrazine, n-butylamine, and cyclohexyl-
amine (regime 1, Table 1, top), free CH2PPh3 and varying
B
DOI: 10.1021/acs.organomet.5b00607
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
amounts of HN(SiMe3)2 are observed by 31P and H NMR
spectroscopy within the first 15 min of the reaction, suggesting
facile protonolysis and displacement of the ylide from the
coordination sphere of the catalyst. Similarly, for piperidine and
tert-butylamine, 1 is consumed within the first point of analysis,
in this case generating the corresponding triphenylphospho-
nium methylide adducts 3a,b (Scheme 2).
that may lose 1 equiv of ylide and generate species similar to
those derived from 1 under catalytic conditions.
1
During preparations of 1 we identified the ylide adduct 2·
CH2PPh3 (δ +30.5 ppm, C6D6) as an intermediate. Consistent
with the hypothesis outlined above, VT NMR studies on
isolated samples of 2·CH2PPh3 in toluene-d8 demonstrated that
ylide coordination is reversible, with the equilibrium lying
toward 2 + CH2PPh3 at higher temperatures (Figure 3).
In contrast when hindered secondary amines are employed
as substrates (regime 2, Table 1, bottom), ring opening of the
P-based ligand of 1 or displacement of the ylide from yttrium is
not observed. For diisopropylamine, dicyclohexylamine, iso-
propylcyclohexylamine, and cis-2,6-dimethylpiperidine, 1 (δ
+28.3 ppm) is observed as the catalyst resting state throughout
the reaction. It is for these substrates that a significant
difference between the activities of 1 and 2 is recorded in
catalytic silazane formation (Table 1).
Scheme 2. Synthesis of Catalyst Resting States 3a,b
A stoichiometric reaction between 1 and piperidine in C6D6
resulted in clean formation of the in situ observed product. The
reaction still occurs rapidly when 2, Ph3PCH2, and excess amine
(1.2 equiv) are dissolved in C6D6. Scale-up of this and a related
reaction using tert-butylamine gave 3a,b as crystalline solids in
Reversible Ligand Activation under Stoichiometric
Conditions. In order to probe whether the reaction of 1 with
amines could be reversible, a series of deuterium labeling
t
experiments were conducted. The reaction of BuND2 with
either 1 or 3b resulted in significant D incorporation into the
C−H bonds of both the methylide position and the aryl ring, as
1
50% and 69% yields, respectively (Scheme 2). The H NMR
spectra of 3a,b both show a distinct sharp doublet of doublets
1
2
evidenced by H and H NMR spectroscopy (Scheme 3). A
similar labeling experiment in which 1 was reacted with excess
iPr2ND neither led to ring opening of the phosphonium
methylide ligand nor resulted in significant D incorporation
into the aryl ring. This experiment is consistent with the
observation of 1 as a catalytic resting state during the
2
31P−1H
for the methylide hydrogens (3a, δ 1.20 ppm, J
= 17.6
2
2
2
89
1
31
1
89Y−1H
Hz, J Y− H = 2.4 Hz; 3b, δ 1.13 ppm, J P− H = 17.6 Hz, J
= 2.4 Hz). For 3b the N−H resonance can be seen as a doublet
2
at δ 2.18 ppm (1H, J Y− H = 2.4 Hz). The coupling to 89Y was
confirmed by running a 1H{31P} experiment. A single
phosphorus environment was detected in each case (3a, δ
89
1
i
2
2
dehydrocoupling of Pr2NH with PhSiH3. In both labeling
89Y−1H
89Y−1H
+32.4 ppm, J
= 5.0 Hz; 3b, δ +32.5 ppm, J
= 4.9
experiments D incorporation was observed into the methylide
position, and control experiments in which triphenylphospho-
Hz). Single-crystal X-ray diffraction experiments confirmed the
structures (Figure 2). The Y−C bond lengths in 3a
(2.5441(18) Å), 3b (2.531(2) Å), and 2·CH2PPh3 (2.554(3)
Å) (vide infra) are within experimental error of one another.
Complex 3a proved kinetically competent for the dehy-
drocoupling of piperidine and diphenylsilane at an initial rate
similar to those of 1 and 2. Similarly, 3b was found to be
catalytically competent for the reaction of phenylsilane and tert-
butylamine, giving the corresponding silazane in 99% yield after
1 h at 25 °C. On the basis of this observation and the data
presented in Table 1, we propose that for the addition of 1°
amines and unhindered 2° amines to arylsilanes a common
catalyst may be generated. Complexes 1, 2, and 3a,b give rise to
similar species under catalytic conditions. Hence, protonolysis
of 2 with 1 equiv of amine yields observable intermediates 3a,b
t
nium methylide was reacted with BuND2 show facile H/D
exchange between the N−D and C−H groups. As such, little
mechanistic information can be gained from D incorporation
into the methylide position of complexes 1 and 3b.
The experiments were supported by DFT calculations using
a sterically unhindered model. The addition of the N−H bond
of dimethylamine across the Y−Cylide and aryl Y−CAryl bonds of
A-1 (a model of 1 in which the − N(SiMe3)2 ligands have been
replaced with NMe2) is represented in Figure 4. Both reactions
were found to occur by an energetically accessible σ-bond
metathesis transition state, but the former reaction not only is
less exergonic than the latter but is also expected to be
reversible, on the basis of the energy of the transition state for
Figure 2. Crystal structures of (a) 2·CH2PPh3, (b) 3a, and (c) 3b. Selected bond angles (deg) and bond lengths (Å): 2·CH2PPh3, Y−N(21)
2.246(2), Y−N(41) 2.274(2), Y−N(31) 2.277(2), Y−C 2.554(3), C−P 1.739(3;. 3a, Y−C 2.5441(18), P−C 1.7366(17), Y−N(21) 2.1781(15), Y−
N(31) 2.2810(14), Y−N(41) 2.676(14), Y−C−P 141.33(10);3b, Y−C 2.531(2), P−C 1.730(2), Y−N 2.158(2), Y−N 2.273(2), Y−N 2.2633(19),
Y−C−P 144.39(14).
C
DOI: 10.1021/acs.organomet.5b00607
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 3. 31P{1H} NMR stack plot for a 39 mM C7D8 solution of 2·CH2PPh3 between −80 and +80 °C.
addition of the amine across the aryl Y−CAryl position. Here 1 is
likely to be both kinetically and thermodynamically stable with
respect to the ring-opened products.
Scheme 3. Deuterium Isotope Tracer Experiments
Proposed Catalytic Mechanism. To summarize the
stoichiometric experiments and observations under catalytic
conditions, during reactions of primary or sterically unhindered
secondary amines with 1 the ylide is readily displaced from the
coordination sphere of yttrium. Complexes of the form
[Y(NR2)3(CH2PPh3)] are potential intermediates in this
reaction. In the cases where these adducts are isolable, evidence
for the reversible addition of the N−H bond of the amine
across both the Y−CAryl and Y−Cylide positions of the
phosphonium methylide ligand has been gathered. For
sterically hindered secondary amines, no data to support the
displacement of the ylide from 1 or reversible ligand activation
the microscopic reverse, amine elimination (Figure 4). Full
details of these calculations, including the structures of the
stationary points, can be found in the Supporting Information.
While we have not extended this model to more sterically
demanding substrates, for hindered secondary amines the D-
labeling experiments show no conclusive evidence for the
Figure 4. DFT calculations of the potential energy surfaces for addition of Me2NH to A-1. Electronic energies along with solvent-corrected (PCM,
benzene) energies from single-point calculations are provided. Values are given in kcal mol−1. The solvent-corrected values are given in parentheses.
D
DOI: 10.1021/acs.organomet.5b00607
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 5. DFT calculations of the potential energy surface for amine−silane dehydrocoupling. Pathway 1 is catalyzed by A-1, and pathway 2 is
catalyzed by [Y(NMe2)3]. Electronic energies along with solvent-corrected (PCM, benzene) energies from single-point calculations are provided.
Values are given in kcal mol−1. The solvent-corrected values are given in parentheses.
under either stoichiometric or catalytic conditions were
collected.
catalytic manifold. In contrast, pathway 1 is weakly inhibited
and amine coordination to A-1 is thermodynamically less
favorable than for [Y(NMe2)3].23
In order to gain more insight into catalysis and to explore
possible explanations for the difference between catalysts 1 and
2, a series of DFT calculations were conducted. The potential
energy surfaces for a series of reaction mechanisms for the
addition of PhSiH3 to Me2NH were calculated for [Y(NMe2)3]
and A-1 by DFT methods. To further exclude the role of ligand
activation in catalysis, the conventional σ-bond metathesis
pathways were compared against amine activation by addition
across both the Cylide and CAryl ligands of A-1. Despite repeated
attempts to optimize silicate structures, there was no evidence
to suggest that yttrium silicates play a role in low-energy
reaction pathways. The calculated intermediates and diamond-
like transition states conform to geometries that are well
established in rare-earth chemistry (see the Supporting
Information).21
The lowest energy reaction pathways occur for the addition
of the amine across the Y−N bond of [Y(NMe2)3] and the Y−
N bond of A-1 (Figure 5). While feasible reaction pathways
were located for silazane formation involving amine activation
using the X- or L-type ligand of A-1, in this model system these
pathways incorporate transition states that are considerably
higher in energy in comparison to those represented in Figure 5
(see the Supporting Information). These findings suggest that
reversible ligand activation is not a satisfactory explanation for
the improved catalytic performance of 1 over 2. Full details of
these calculations, including the structures of the stationary
points, can be found in the Supporting Information.
Although the suitability of the computational model is
limited due to to the reduced size of the alkyl groups on amine
and amide moieties relative to the real system, inhibition
experiments in which catalytic or stoichiometric quantities of
i
HN(SiMe3)2 were added to the reaction of Pr2NH with
PhSiH3 catalyzed by 1 slowed the reaction but failed to prevent
silazane formation. Under the same conditions, complex 2
shows no activity for amine silane dehydrocoupling (see Figure
S5 in the Supporting Information).
Although we cannot unambigously rule out a role of the
phosphorus-based ligand in supporting or solubilizing yttrium
hydride clusters, we suggest that the significant difference
between 1 and 2 in catalysis is a reflection of the strength of
amine binding to the precatalyst and the degreee of catalyst
inhibition. The difference between the two catalyst systems is
borne out for the most sterically hindered amines: those which
would be expected to show the largest sensitivity to the steric
environment at yttrium and those for which 1 is observed as a
resting state throughout the reaction.13
CONCLUSIONS
■
In summary, the cyclometalated complex 1 is an effective
catalyst for the dehydrogenative coupling of amines and silanes.
For the addition of a series of sterically demanding secondary
amines to phenylsilane, this catalyst is dramatically more
effective than [Y{N(SiMe3)2}3]. While reversible activation of
the cyclometalated ligand has been observed in stoichiometric
experiments with amines, D-labeling experiments and DFT
calculations suggest that reversible ligand activation is not
involved in silazane formation under catalytic conditions.
Constitent with these theoretical data, additional deuterium
t
labeling experiments in which 3b, BuND2, and PhSiH3 were
i
mixed in a 1:10:10 ratio or 1, Pr2ND, and PhSiH3 were mixed
in a 1:10:10 ratio did not lead to significant D incorporation
into the ortho position of the aryl group of the ligand. Hence,
under catalytic conditions reversible activation of the
phosphonium methylide ligand by substrate addition across
the Y−CAryl bond is not significant.
The DFT studies suggest that catalytic reactions should
suffer from amine inhibition, with the effect being more
significant for the three-coordinate precatalyst 2 than for the
four-coordinate precatalyst 1.22 Pathway 2 would be expected
to suffer from strong catalyst inhibition due to a significant
stabilization of the system upon amine coordination to
[Y(NMe2)3]. Amine coordination would be expected to raise
the overall activation energy to Si−N bond formation due to
the need for a dissociative step prior to entering into the
EXPERIMENTAL SECTION
■
General Procedure, Exemplified for Bn2SiH2Ph. In a glovebox,
dibenzylamine (22.4 μL, 0.12 mmol), phenylsilane (14.4 μL, 0.12
mmol), and the internal standard were dissolved in C6D6 (450 μL)
and transferred to a Youngs tap NMR tube. The tube was removed
1
from the glovebox, a baseline H NMR spectrum was recorded, and
the tube was returned to the glovebox before the addition of 3 (10 mol
1
%) in C6D6 (150 μL). The reaction was monitored in situ by H and
1
31P NMR spectroscopy and gave 99% yield after 3.5 h at 25 °C. H
NMR (400 MHz, C6D6): δ 3.85 (s, 4H), 5.35 (s, 2H), 7.05−7.19 (m,
13H), 7.62−7.64 (m, 2H). 13C NMR (100 MHz, C6D6): δ 51.2, 126.8,
128.1, 128.2, 128.3, 130.1, 134.3, 135.1, 139.8. 15N NMR (51 MHz,
C6D6): δ +24.1. 29Si NMR (99 MHz, C6D6): δ −20.0.
E
DOI: 10.1021/acs.organomet.5b00607
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Synthesis of 2·CH2PPh3. In a glovebox, [Y{N(SiMe3)2}2]2 (460 mg,
0.81 mmol) and Ph3PCH2 (0.22 g, 0.81 mmol) were weighed out
separately and transferred to a Schlenk. The Schlenk was sealed,
removed from the box, and attached to a vacuum line, where dry
diethyl ether (5 mL) was added under a purge of argon. The mixture
was agitated and left to stand for 1 h at 25 °C. The solution was
removed under vacuum, and the product was extracted into n-hexane
(5 mL). The solution was filtered, reduced in volume, and stored at
−20 °C to produce yellow crystals. The crystals were isolated through
filtration and dried under vacuum to give 2·CH2PPh3 (404 mg, 0.48
Computed structures (MOL)
Crystallographic data for 2·CH2PPh3 and 3a,b (CIF)
AUTHOR INFORMATION
Corresponding Author
*E-mail for M.R.C.: m.crimmin@imperial.ac.uk.
Funding
M.R.C. acknowledges the Royal Society for provision of a
University Research Fellowship. A.E.N. is grateful to Imperial
College London for sponsoring a Ph.D. studentship and the
EPSRC for provision of a prize fellowship. We are also grateful
to the Nuffield Foundation for a summer bursary for W.C.
■
1
mmol, 59%). H NMR (400 MHz, C6D6, 298 K): δ 0.45 (s, 54H),
2
31
1
1.23 (br d, 2H, J P− H = 13.6 Hz), 7.00−7.05 (m, 9H), 7.52−7.61 (m,
6H). 13C NMR (101 MHz, C6D6, 298 K): δ 6.5, 8.5 (br), 129.0 (d,
J
= 11.7 Hz), 132.4, 133.0 (d, J
= 9.7 Hz). 31P NMR (162
31P−13
C
31P−13
C
MHz, C6D6, 298 K): δ + 30.5. Anal. Calcd for C37H71N3PSi6Y: C,
52.51; H, 8.46; N, 4.96. Found: C, 52.31, H, 8.45; N, 4.98.
Notes
Synthesis of 3a. In a glovebox, [Y{N(SiMe3)2}2]2 (400 mg, 0.70
mmol) and Ph3PCH2 (194 mg, 0.70 mmol) were weighed into a 20
mL scintillation vial. Dry diethyl ether (5 mL) was added, followed by
piperidine (83 μL, 0.84 mmol, 1.2 equiv). The mixture was agitated
and left to stand for 2 h at 25 °C. The volatiles were removed under
vacuum, and the product was extracted into n-hexane (5 mL). The
solution was filtered, reduced in volume, and stored at −20 °C to
produce pale yellow crystals. The crystals were isolated through
filtration and dried under vacuum to give 3a (269 mg, 0.35 mmol,
The authors declare no competing financial interest.
REFERENCES
■
(1) Brook, M. A. Silicon in Organic, Organometallic and Polymer
Chemistry; Wiley: Hoboken, NJ, 2000.
(2) Lekishvili, N.; Kerherashvili, M.; Lekishvili, G. Asian J. Chem.
2010, 22, 4141−4149.
(3) Grellier, M.; Ayed, T.; Barthelat, J.-C.; Albinati, A.; Mason, S.;
Vendier, L.; Coppel, Y.; Sabo-Etienne, S. J. Am. Chem. Soc. 2009, 131,
7633−7640.
1
50%). H NMR (400 MHz, C6D6, 298 K): δ 0.47 (s, 36H), 1.20 (dd,
2 31
1
2
89
1
2H, J P− H = 17.6 Hz, J Y− H = 2.8 Hz), 1.52−1.58 (m, 4H), 1.67−
(4) Kocienski, P. J. Protecting Groups; Georg Thieme Verlag:
Stuttgart, Germany, 1994.
1.72 (m, 2H), 3.36 (m, 4H), 7.01−7.03 (m, 9H), 7.45−7.50 (m, 6H).
13C NMR (101 MHz, C6D6, 298 K): δ 5.8, 9.1 (dd, 1J
C = 1J
89Y−13
31P−13
C
(5) (a) Saam, J. C.; Speier, J. L. J. Org. Chem. 1959, 24, 119−120.
(b) Sommer, L. H.; Citron, J. D. J. Org. Chem. 1967, 32, 2470−2472.
(6) Wang, W.-D.; Eisenberg, R. Organometallics 1991, 10, 2222−
2227.
(7) (a) Blum, Y.; Laine, R. M. Organometallics 1986, 5, 2081−2086.
(b) Biran, C.; Blum, Y. D.; Glaser, R.; Tse, D. S.; Youngdahl, K. A.;
Laine, R. M. J. Mol. Catal. 1988, 48, 183−197. (c) Gutsulyak, D. V.;
Vyboishchikov, S. F.; Nikonov, G. I. J. Am. Chem. Soc. 2010, 132,
31P−13
= 31.4 Hz), 27.0, 29.3, 51.9, 127.7, 128.0, 129.4 (d, J
132.6 (d, J
MHz, C6D6, 298 K): δ +32.4 (d, J
C = 11.8 Hz),
C = 9.7 Hz). 31P NMR (162
= 5.2 Hz). Anal. Calcd for
Y
31P−13
31P−13
31P−89
C = 2.4 Hz), 132.9 (d, J
2
C36H62N3PSi4Y: C, 56.22; H, 8.13; N, 5.46. Found: C, 56.56; H, 8.24;
N, 5.40.
Synthesis of 3b. In a glovebox, [Y{N(SiMe3)2}2]2 (150 g, 0.26
mmol) and Ph3PCH2 (73 mg, 0.26 mmol) were weighed into a 20 mL
scintillation vial. Dry diethyl ether (5 mL) was added, followed by tert-
butylamine (33 μL, 0.32 mmol, 1.2 equiv). The mixture was agitated
and left to stand for 2 h at 25 °C. The volatiles were removed under
vacuum, and the product was extracted into n-hexane (5 mL). The
solution was filtered, reduced in volume, and stored at −20 °C to
produce pale yellow crystals. The crystals were isolated through
filtration and dried under vacuum to give 3b (138 mg, 0.18 mmol,
5950−5951. (d) Konigs, C. D. F.; Muller, M. F.; Aiguabella, N.; Klare,
̈
H. F. T.; Oestreich, M. Chem. Commun. 2013, 49, 1506−1508.
(8) (a) Liu, H. Q.; Harrod, J. F. Organometallics 1992, 11, 822−827.
(b) He, J.; Liu, H. Q.; Harrod, J. F.; Hynes, R. Organometallics 1994,
13, 336−343. (c) Liu, X.; Wu, Z.; Peng, Z.; Wu, Y.-D.; Xue, Z. J. Am.
Chem. Soc. 1999, 121, 5350−5351. (d) Gauvin, F.; Woo, H. G. Adv.
Organomet. Chem. 1998, 42, 363−405.
1
69%). H NMR (400 MHz, C6D6, 298 K): δ 0.46 (s, 36H), 1.13 (dd,
2
2
31P−1H
89Y−1H
= 2.4 Hz), 1.37 (s, 9H), 2.18 (d, 1H,
(9) Liu, H. Q.; Harrod, J. F. Can. J. Chem. 1992, 70, 107−110.
(10) (a) Dash, A. K.; Wang, J. X.; Berthet, J. C.; Ephritikhine, M.;
Eisen, M. S. J. Organomet. Chem. 2000, 604, 83−98. (b) Wang, J. X.;
Dash, A. K.; Berthet, J. C.; Ephritikhine, M.; Eisen, M. S. J. Organomet.
Chem. 2000, 610, 49−57.
(11) (a) Tsuchimoto, T.; Iketani, Y.; Sekine, M. Chem. - Eur. J. 2012,
18, 9500. (b) Greb, L.; Tamke, S. Chem. Commun. 2014, 50, 2318−
2320.
2H, J
J
= 17.6 Hz, J
Y− H = 2.4 Hz), 7.01−7.04 (m, 9H), 7.52−7.58 (m, 6H). 13C NMR
2
89
1
(101 MHz, C6D6, 298 K): δ 6.0, 9.0 (dd, 1J
C = 1J
31P−13
89Y−13
C = 30.5 Hz),
31P−13
31P−13
C = 2.4 Hz), 133.0 (d,
36.2, 129.3 (d, J
C = 11.8 Hz), 132.6 (d, J
J
= 9.7 Hz). 31P NMR (162 MHz, C6D6, 298 K): δ +32.5 (d,
31P−13
C
31P−89
2J
= 4.9 Hz). Due to the air-sensitive nature of this compound,
Y
repeated attempts to obtain satisfactory elemental analysis failed.
DFT Studies. Calculations were conducted in Gaussian09. All
minima were confirmed by frequency calculations, and where
applicable solid-state data were used as an input for the atom
coordinates. Geometry optimizations were performed using the hybrid
Becke three-parameter functional with Lee−Yang−Parr correlation
(B3LYP). A hybrid 6,31G+(d,p) (C, H, N, Si, P) and LanL2DZ (Y)
basis set was used. A (benzene) solvent correction was applied by
calculating single-point energies of the optimized structures employing
the polarizable continuum model in Gaussian 09.
(12) Corriu, R. J. P.; Leclercq, D.; Mutin, P. H.; Planeix, J. M.; Vioux,
A. J. Organomet. Chem. 1991, 406, C1−C4.
(13) (a) Buch, F.; Harder, S. Organometallics 2007, 26, 5132−5135.
(b) Dunne, J. F.; Neal, S. R.; Engelkemier, J.; Ellern, A.; Sadow, A. D. J.
Am. Chem. Soc. 2011, 133, 16782−16785. (c) Hill, M. S.; Liptrot, D. J.;
MacDougall, D. J.; Mahon, M. F.; Robinson, T. P. Chem. Sci. 2013, 4,
4212−4222. (d) Bellini, C.; Carpentier, J.-F.; Tobisch, S.; Sarazin, Y.
Angew. Chem., Int. Ed. 2015, 54, 7679−7683.
(14) (a) Takaki, K.; Kamata, T.; Miura, Y.; Shishido, T.; Takehira, K.
J. Org. Chem. 1999, 64, 3891−3895. (c) Xie, W.; Hu, H.; Cui, C.
Angew. Chem., Int. Ed. 2012, 51, 11141−11144.
(15) (a) Crimmin, M. R.; White, A. J. P. Chem. Commun. 2012, 48,
1745−1747. (b) Nako, A. E.; White, A. J. P.; Crimmin, M. R. Chem.
Sci. 2013, 4, 691−695.
(16) (a) Hong, S.; Marks, T. J. Acc. Chem. Res. 2004, 37, 673−686.
(b) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.; Nolan,
M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am. Chem.
Soc. 1987, 109, 203−219.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.organo-
met.5b00607.
Full experimental and computational details and
characterization data (PDF)
F
DOI: 10.1021/acs.organomet.5b00607
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(17) Arnold, P. L.; McMullon, M. W.; Rieb, J.; Kuhn, F. E. Angew.
̈
Chem., Int. Ed. 2015, 54, 82−100.
(18) (a) Watson, P. L.; Parshall, G. W. Acc. Chem. Res. 1985, 18, 51−
56. (b) Werkema, E. L.; Messines, E.; Perrin, L.; Maron, L.; Eisenstein,
O.; Andersen, R. A. J. Am. Chem. Soc. 2005, 127, 7781−7795.
(c) Kefalidis, C. E.; Perrin, L.; Burns, C. J.; Berg, D. J.; Maron, L.;
Andersen, R. A. Dalton Trans. 2015, 44, 2575−2587.
(19) (a) Turner, Z. R.; Bellabraba, R.; Tooze, R. P.; Arnold, P. L. J.
Am. Chem. Soc. 2010, 132, 4050−4051. (b) Arnold, P. L.; Turner, Z.
R.; Bellabarba, R.; Tooze, R. P. J. Am. Chem. Soc. 2011, 133, 11744−
11756. (c) Arnold, P. L.; Cadenbach, T.; Marr, I. H.; Fyfe, A. A.; Bell,
N. L.; Bellabarba, R.; Tooze, R. P.; Love, J. B. Dalton Trans. 2014, 43,
14346−14358.
(20) (a) Flynn, S. R.; Wass, D. F. ACS Catal. 2013, 3, 2574−2581.
(b) Chapman, A. M.; Haddow, M. F.; Wass, D. F. J. Am. Chem. Soc.
2011, 133, 18463−18478. (c) Chapman, A. M.; Haddow, M. F.; Wass,
D. F. J. Am. Chem. Soc. 2011, 133, 8826−8829.
(21) See for example: (a) Tobisch, S. Dalton Trans. 2012, 41, 9182−
9191. (b) Maron, L.; Perrin, L.; Eisenstein, O. J. Chem. Soc., Dalton
Trans. 2002, 534−539.
(22) The effect of excess amine on the dielectric constant of the
solvent or explicit amine solvation of the yttrium center by one or
more molecules of amine was not explored at every stage of the
catalytic cycle.
(23) Inhibition of pathway 2 could also occur by protonolysis and
ring opening of the cyclometalated ylide with the amine (Supporting
Information). On the basis of the observation of 1 as a resting state
throughout the reaction with these substrates, we propose that this
reaction is disfavored for sterically hindered 2° amines.
G
DOI: 10.1021/acs.organomet.5b00607
Organometallics XXXX, XXX, XXX−XXX