ACS Catalysis
Research Article
Scheme 2. Selective Synthesis of Either
Oxasilacyclopentanes and Oxasilacyclohexanes via Grubbs-
Type Ruthenium-Catalyzed Hydrosilylative Cyclization
Scheme 3. Insightful Observations of the Formation of
Vinylsilanes (7a and 7c) and Chlorosilane (8a)
19
metathesis, the metathesis of 9a and silane 1 may furnish
either 9g via 9e (metal-out: unproductivethe silicon never
goes to the ruthenium metal center) or 9d via 9f (metal-in:
productive). It could be an analogous situation where the
outcome of competitive cross-metathesis (CM) [i.e., productive
CM: an ethylene-producing process (cf., 9f to 9d) and
unproductive CM: a degenerate metathesis (cf., 9e to 9g)] is
substantially dependent upon the steric hindrance of olefins, as
that the regioselective synthesis of either oxasilacyclopentanes
(e.g., 2a or 2b) via a 5-exo-trig (for homoallylic silyl ethers) or a
5
-endo-trig (for allylic silyl ethers) hydrosilylative cyclization or
oxasilacyclohexanes (e.g., 2c) via a 6-exo-trig cyclization is
feasible.
Several speculative mechanisms of Grubbs-type ruthenium-
complex-catalyzed alkyne hydrosilylation have been proposed
by Cox and Cossy. Cox proposed either (i) a sequence of an
28
well as the ligand set of the ruthenium catalyst.
To sort out these two mechanistic hypotheses for the initial
stage of the catalysis, we first performed a control experiment
(Scheme 5A); we speculated that bulkier substituents such as a
t-Bu group at silicon (i.e., 1a-t-Bu) could be favored for
unproductive σ-bond metathesis to yield chlorosilyl ether 8a-t-
initial addition of R Si−H across the π-bond of a Ru-
3
2
0
benzylidene (Chauvin mechanism), silylruthenation, α-
21
elimination (to metal alkylidene/hydride), and reductive
elimination or (ii) the traditional organometallic route
oxidative addition, migratory insertion, and reductive elimi-
nation. Cossy conjectured either hydroruthenation or
silylruthenation. However, neither study provided full
̀
Bu (via 10a vis-a-vis 10b). However, the formation of 2a-t-Bu
3
via the sequential addition of R Si−H across RuCHAr and
3
5
reductive elimination is unlikely because an addition of di-tert-
butylsilane 1a-t-Bu to Ru-2 is greatly hindered, as seen in a
experimental details regarding any such elemental processes.
During our initial study, we made an observation that addressed
the initial stage of Grubbs-type ruthenium-complex-catalyzed
hydrosilylation within a ruthenium coordination sphere. In
detail, upon treatment with Ru-2, silane 1a produced
29
Grubbs’ classification of general reactivity patterns of olefins.
When 1a-t-Bu was subjected to the reaction conditions
employing 100 mol % of Ru-2, only 8a-t-Bu (1a-t-Bu/8a-t-Bu
= 19:81) was observed without any notable cyclization,
corroborating our mechanistic hypothesis for the σ-bond
metathesis. In an effort to support this hypothesis, a
deuterium-labeling experiment was carried out using deuter-
iosilane 11-D and Ru-4 (Scheme 5B). The benzylidene proton
within the resulting putative ruthenium complex 12 remained
intact; we did not detect deuterium incorporation at this
22
vinylsilane 7a (ca. 0.5%) and chlorosilane 8a (ca. 1%),
which were detected and characterized by GC/MS analysis
(
(
Scheme 3A). We were able to isolate cyclic vinylsilane 7c
20% isolated yield) from hydrosilylation of 1c (Scheme 3B).
The formation of the vinylsilane suggests that Grubbs-type
ruthenium-complex-catalyzed alkene hydrosilylation likely
proceeds through a modified Chalk−Harrod mechanism (i.e.,
2
position by H NMR spectroscopy. This result suggests that the
1
2b
silylruthenation)
rather than the Chalk−Harrod (i.e.,
R Si−H addition across the RuCHAr and HCl elimination
3
1
2a,23
hydroruthenation) pathway.
cascade is improbable.
The formation of the chlorosilane 8a offers indirect
information for the initial stage of the hydrosilylation. The
result suggests two plausible mechanisms (Scheme 4): (i) A
The experiment performed to directly detect a ruthenium
silane complex is shown in Figure 1. In this prototype, the use
of an essentially equimolar ratio of Ru-4 and silane 11-H, which
does not bear an alkene moiety, resulted in full conversion to a
putative ruthenium silane complex (e.g., 9d in Scheme 4 or 12
in Scheme 5). Over time, the Si−H bond disappeared, yet
two-step sequence, namely, an addition of R Si−H across the π-
3
3
bond of a Ru-benzylidene to give 9b, followed by HCl
2
4
elimination
(
to form a putative Ru−Si complex 9d
2
5
7
productive), could be responsible for alkene hydrosilylation
benzylidene proton (H ) and other protons within the catalyst
7
11
to furnish 2 or reductive elimination to afford the chlorosilane
byproduct 8 (e.g., 8a) and 9c (unproductive) (path A). (ii)
Direct σ-bond metathesis between Si−H and Ru−Cl via a four-
centered transition state could be involved in the activation of
Ru-4 (H to H ) remained intact; an isopropoxy group was
12
still anchored to the ruthenium center (H , 4.74 ppm).
Interestingly, all protons in substrate 11-H were shifted
downfield, particularly, those at the ortho-position of
26,27
1
the ruthenium complex by silanes (path B).
Based primarily
diphenylsilyl substituents (H , shifted downfield by 0.12
3
upon a bottom-face olefin coordination mechanism for olefin
ppm) and methylene protons (H , shifted downfield by 0.063
3
191
ACS Catal. 2015, 5, 3189−3195