Ruthenium-Catalyzed Alkyne trans-Hydrometalation: Mechanistic Insights and Preparative Implications

Article abstract of DOI:10.1021/jacs.6b12517

[Cp*RuCl]₄ (1) has previously been shown to be the precatalyst of choice for stereochemically unorthodox trans-hydrometalations of internal alkynes. Experimental and computational data now prove that the alkyne primarily acts as a four-electron

Full text of DOI:10.1021/jacs.6b12517

Article
pubs.acs.org/JACS
Ruthenium-Catalyzed Alkyne trans-Hydrometalation: Mechanistic
Insights and Preparative Implications
Dragos-
and Alois Furstner*
̧ Adrian Rosç
a, Karin Radkowski, Larry M. Wolf,‡ Minal Wagh,§ Richard Goddard, Walter Thiel,
̈
Max-Planck-Institut fur Kohlenforschung, D-45470 Mulheim/Ruhr, Germany
̈
̈
S
* Supporting Information
ABSTRACT: [Cp*RuCl]4 (1) has previously been shown to be the
precatalyst of choice for stereochemically unorthodox trans-hydro-
metalations of internal alkynes. Experimental and computational data
now prove that the alkyne primarily acts as a four-electron donor ligand
to the catalytically active metal fragment [Cp*RuCl] but switches to
adopt a two-electron donor character once the reagent R3MH (M = Si,
Ge, Sn) enters the ligand sphere. In the stereodetermining step the
resulting loaded complex evolves via an inner-sphere mechanism into a
ruthenacyclopropene which swiftly transforms into the product. In accord
with the low computed barriers, spectral and preparative data show that
the reaction is not only possible but sometimes even favored at low
temperatures. Importantly, such trans-hydrometalations are distinguished
by excellent levels of regioselectivity when unsymmetrical alkynes are
used that carry an −OH or −NHR group in vicinity of the triple bond. A nascent hydrogen bridge between the protic substituent
and the polarized [Ru−Cl] unit imposes directionality onto the ligand sphere of the relevant intermediates, which ultimately
accounts for the selective delivery of the R3M− group to the acetylene C-atom proximal to the steering substituent. The
interligand hydrogen bonding also allows site-selectivity to be harnessed in reactions of polyunsaturated compounds, since
propargylic substrates bind more tightly than ordinary alkynes; even the electronically coupled triple bonds of conjugated 1,3-
diynes can be faithfully discriminated as long as one of them is propargylic. Finally, properly positioned protic sites lead to a
substantially increased substrate scope in that they render even 1,3-enynes, arylalkynes, and electron-rich alkynylated heterocycles
amenable to trans-hydrometalation which are otherwise catalyst poisons.
order to ensure faithful delivery of the R3M unit (M = Si, Ge,
Sn) to the alkyne C-atom proximal to the steering substituent
(Scheme 1).11,17 This directing effect is particularly pronounced
in trans-hydrostannation and has already served natural product
total synthesis well.18,19
The synergy between a protic substituent on the substrate A
and the polarized [Ru−Cl] bond of the catalyst was tentatively
ascribed to interligand hydrogen bonding in adduct B primarily
formed.11 At the same time, the chloride aligns the incoming
reagent via a hypervalent interaction with the −MR3 group.
The resulting directionality within the ligand sphere of the
loaded complex C explains the observed regioselectivity;
subsequent inner-sphere hydride delivery forms a ruthenacy-
clopropene derivative of type D which accounts for the
stereochemically intriguing trans-addition mode.11,20
INTRODUCTION
■
Shortly after the turn of the millennium, Trost and co-workers
reported that hydrosilylation reactions of internal alkynes
switch from the canonical cis-addition1 to an unorthodox trans-
addition mode when catalyzed by [Cp*Ru(MeCN)3]PF6.2 This
remarkable discovery has opened access to structural motifs
that are difficult to make otherwise. As a consequence, trans-
hydrosilylation was rapidly embraced by the synthesis
community,3,4 despite the fact that nonsymmetrical substrates
usually lead to the formation of regioisomers.5
Prompted by early applications of this methodology,6,7 our
group managed to expand the underlying principle in two
dimensions: On one hand, we were able to impose the trans-
addition mode8 onto hydroboration,9 hydrostannation,10,11
hydrogermylation11,12 and even hydrogenation reactions13−15
of internal alkynes; all of these transformations seemingly
violate reigning paradigms of organometallic chemistry and
have little or no precedent in the literature. On the other hand,
we could solve the regioselectivity problem for unsymmetrical
substrates comprising a protic functional group in vicinity to
the reacting triple bond. To this end, it suffices to replace the
cationic precatalyst [Cp*Ru(MeCN)3]PF6 used by Trost2,16 by
neutral ruthenium complexes featuring a [Cp*Ru−Cl] motif in
This basic mechanistic scenario is in line with various pieces
of indirect evidence, but all attempts at characterizing a
complex of type C met with failure. Even the preceding alkyne
adduct B could not be caught unless the Cp* ring of the active
catalyst was replaced by a much more encumbered and more
Received: December 5, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. Proposed Mechanism of Ruthenium-Catalyzed
trans-Hydrometalation (• = CMe)
Scheme 2. Formation of Alkyne Complexes Suitable for X-
ray Diffraction
assuming that both orthogonal acetylenic π-systems engage in
bonding; if the substrate serves as a four-electron donor,23,24
the resulting complex gains a formal 18-electron count. At the
same time, the −OH proton experiences significant deshielding,
indicative of interligand hydrogen bonding.
Although adduct formation is fast and clean according to
NMR, our early attempts at isolating the resulting π-complexes
had met with failure; as mentioned above, only downstream
[2+2] cycloadducts such as 3 had been obtained.11,25 It was
only after considerable experimentation that our efforts were
finally rewarded with success. Specifically, single crystals
suitable for X-ray diffraction could be grown from the
representative complexes 4−6 comprising a propargyl alcohol
or a propargylic sulfonamide, respectively (Figure 1 and Figures
S2 and S3 in the Supporting Information).
While the gross structural attributes of these complexes are
similar, the differences in details are arguably telling. In all cases
is π-complex formation supported by peripheral hydrogen
bonding between the protic substituent and the chloride atom
as inferred from 1H NMR.26 The strength of this contact seems
correlated with the acidity of the protic group, in that the −X−
(H)···Cl bond length increases from a short 2.975(1) Å in 5,
bearing two electron-withdrawing −CF3 groups, to 3.072(3) Å
in 4, comprising an “ordinary” propargyl alcohol, to 3.153(2) Å
in the sulfonamide adduct 6; yet, even this distance must be
rated “short” and “strong” according to the categories of
organometallic chemistry.27 Whereas the protons engaged in
hydrogen bonding are less well located in the X-ray analyses,
the differences in the −X···Cl distances can be considered
significant.
A closer inspection of the bound alkyne is equally
informative. The substantial elongation of the CC bond28
indicates a high degree of activation which is qualitatively in line
with the massive deshielding observed by 13C NMR.
Interestingly, the CC bond in 5 (1.283(1) Å) is longer
than that in 4 (1.264(4) Å) and the alkyne slightly more bent
away from linearity (C3−C2−C1 144.75(8)° versus
145.3(3)°). This geometric attribute reveals non-negligible
electron back-donation from the filled d-orbitals of the metal
into the π*-orbitals of the bound alkyne, which becomes
stronger upon lowering the LUMO level by virtue of the
electron withdrawing fluorine atoms in 5.
electron-rich η5-1-methoxy-2,4-di-tert-butyl-3-neopentyl-cyclo-
pentadienyl (Cp^) ligand;21 only [2+2] cycloadducts were
obtained as long as the Cp* ring was in place.11 Although the
structural attributes of the model compound 2 and cycloadduct
3 are certainly in accord with our working hypothesis, they do
not take away the need to investigate the real catalyst system in
order to gain a better understanding for the origin and
implications of this powerful steering effect ascribed to
interligand interactions.
Outlined below is a comprehensive study along these lines.
Specifically, we present a series of X-ray structures that reveal
how prototype alkynes behave on coordination to the germane
[Cp*Ru−Cl] fragment. Adducts with ligands of different donor
strength mimic the electronic response of such an alkyne
complex to the incoming H−MR3 reagent (M = Si, Ge, Sn).
Moreover, spectral evidence indicates ready product formation
once the loaded complex has formed and shows that product
release is very facile. The gathered experimental data are
correlated with computational results that provide insights into
the peculiar bonding situation of [Cp*RuCl(alkyne)] com-
plexes formed at the outset and map the reaction coordinate,
leading to the trans-addition product from there on. Moreover,
it is shown that the virtues of interligand hydrogen bonding
reach way beyond regiocontrol over H−MR3 delivery: notably,
the effect allows acetylene derivatives of similar steric and/or
electronic character to be distinguished with meaningful levels
of selectivity; it even proves strong enough to invoke substrates
that are otherwise inert.
RESULTS AND DISCUSSION
■
Hydroxyl-Assisted Alkyne Binding: Structural Data
and Bonding Analysis. Treatment of [Cp*RuCl]4 (1)22 as
the standard catalyst for hydrometalations of all sorts with a
tert-propargyl alcohol (1 equiv) leads to an instant color change
from brown to cherry red (Scheme 2). NMR indicates clean
formation of the corresponding π-complexes such as 4−6,
which are distinguished by massive deshielding of the alkyne C-
atoms; the observed downfield shift Δδ is in the order of ∼50−
70 ppm (Table 1). This immense effect is best explained by
The qualitative conclusion drawn from the experimental data
that strong four-electron donation from the alkyne to the metal
B
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 1. Characteristic Spectral Data (CD2Cl2, ppm) of Representative Free Alkynes and the Corresponding Complexes 4−6
a
Δδ = δ(π-complex) − δ(free alkyne). The numbering scheme for the alkyne C-atoms is as shown in Scheme 2 and the X-ray structures; arbitrary
numbering as indicated.
Figure 1. Structure of complex 4 in the solid state; all H-atoms, except
the one engaged in hydrogen bonding to the chloride ligand, are
omitted for clarity; only one of the two independent molecules in the
unit cell is shown; for the X-ray structures of complexes 5 and 6, see
the Supporting Information.
goes hand in hand with non-negligible back-bonding is
substantiated by the orbital interaction diagram constructed
using orbital perturbation theory at the M06/TZ2P level
(Figure 2).29 The strongest interaction in the model complex
[Cp*Ru(Cl)(CH3CCCH3)] results from the σ-symmetric
electron donation from the [Cp*RuCl] fragment HOMO(a′)
2
of largely dz character. This interaction is strongest with the
alkyne carbon that is distal to the Cl ligand (proximal to the
Cp*) due to the orbital directionality. The [Cp*RuCl]
HOMO−1(a″) is of similar energy but only overlaps slightly
with the alkyne LUMO+1(a″). The [Cp*RuCl] LUMO(a′),
mostly of 5s character, and LUMO−1(a″) experience σ- and π-
symmetric electron donation from both π-clouds (a′ and a″,
respectively). The summation of both π-cloud donations is
slightly greater than the donation from the [Cp*RuCl]
HOMO. The charge transfer from complexation amounts to
a net charge transfer of Δq = 0.133e from the ligated alkyne to
the [Cp*RuCl] fragment. Therefore, it seems justified to
qualitatively rate the alkyne as a four-electron donor.
Additional support for the alkyne four-electron donor
behavior comes from comparison to the analogous alkene
complex [Cp*Ru(Cl)(trans-2-butene)]. The enhanced dis-
tortion in the alkyne relative to the alkene is caused by the
greater orbital interactions within [Cp*Ru(Cl)(CH3C
CCH3)] as compared with [Cp*Ru(Cl)(trans-2-butene)].
Figure 2. Orbital diagram of the model complexes [Cp*Ru(Cl)-
(CH3CCCH3)] (top) and [Cp*Ru(Cl)(trans-2-butene)] (bottom)
computed at the M06/TZ2P level. The binding energy ΔE is the sum
of the interaction energy ΔEint and the distortion energy ΔEdist (given
in kcal mol−1); charge transfers (Δq) are computed from NBO partial
atomic charges, see the Supporting Information.
The alkene has only one π-cloud donor which is nearly offset
by the back-donation from the [Cp*RuCl] HOMO.
C
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Furthermore, the computed complexation induced charge
transfer is computed to be Δq = 0.085e, which is less than
with the alkyne complexed, consistent with the reduced
donation from the ligand π framework.
The orbital schemes for bound propargyl alcohol are basically
identical, as are the computed bond lengths. This result shows
that the presence of a hydroxyl group in the substrate able of
entertaining an interligand hydrogen bond does not change the
orbital interactions and hence the intrinsic reactivity of the
resulting π-complex; however, it is computed to increase its
stability by ∼2.8−4.8 kcal/mol, depending on the particular
substrate (vide infra). As will be shown below, this differential
allows otherwise unreactive substrates to be processed and/or
site selectivity to be imposed on trans-hydrometalation reaction
of various diyne derivatives.
Secondary Propargyl Alcohols: Competing Redox
Isomerization. The preparation of complexes 4 and 5
comprising tert-propargyl alcohols as ligands had proven taxing,
but the isolation of the analogous complex 8 carrying a sec-
propargyl alcohol was even more difficult (Scheme 3). The
Figure 3. Structure of complex 8 in the solid state featuring an
intermolecular hydrogen-bonding pattern between two monomer
units; all H-atoms, except the ones engaged with the chlorides, are
omitted for clarity.
Scheme 3. Coordination Behavior of a Secondary Propargyl
Alcohol
(H)···Cl, 3.22(4) Å). Since 1,2-shift of this particular H-atom
entails redox isomerization, it is tempting to assume that this
step might enjoy assistance by the flanking chloride too. In line
with this notion, H4 in complex 8 (δH = 4.80 ppm) resonates
downfield from the proton in the unbound substrate (δH = 4.45
ppm), which suggests that a weak interaction persists in
solution. An equilibrium with a second conformer 8b is
reasonable; this assumption is based on the characteristic
deshielding of the −OH proton which persists even at low
concentrations at which intermolecular contacts are highly
unlikely. In any case, the downfield shift of H4 speaks against a
developing hydridic character and hence against a mechanism
of redox isomerization via simple 1,2-hydride transfer. In this
context, it should be noted that related interactions between the
[Ru−Cl] unit and the R3M group (M = Si, Ge, Sn) are crucial
for trans-hydrometalations to proceed.11 Therefore, it seems
legitimate to assume that ruthenium catalyzed redox isomer-
ization, trans-hydrometalations and trans-hydrogenationin
their essenceare different manifestations of the same
fundamental reactivity mode.
additional challenge arises from the propensity of such
substrates to undergo redox isomerization with formation of
the corresponding enones even at low temperature.30,31 It was
only after considerable experimentation that single crystals of 8
suitable for X-ray diffraction could be grown, which must be
kept cold at all time to avoid instantaneous formation of hex-3-
en-5-one (9) and regeneration of [Cp*RuCl]4 (1).
The bonding of the alkyne unit to the metal fragment in 8 is
very similar to that in 4 or 5. This is evident from the massive
downfield shift of the ligated C-atoms in the 13C NMR
spectrum (δC = 150.7/135.5 ppm versus 85.4/81.3 ppm in the
free alkyne), the elongation of the triple bond (1.268(3) Å),
and the significant bending of the acetylene unit (C4−C1−C2
147.0(10)°).
The arguably most striking feature of 8 is the now
intermolecular hydrogen-bonding pattern between the −OH
group and the chloride ligand of the next unit at an −O−(H)···
Cl distance of 3.12(3) Å (average of two crystal forms, see the
SI) (Figure 3). As if to compensate, the hydrogen atom H4 at
the foot of the propargylic alcohol unit makes a rather short
intramolecular contact with the neighboring chloride (−C−
If a tertiary propargyl alcohol (amide), when ligated to
[Cp*RuCl], were to entertain an intermolecular hydrogen-
bonding pattern as manifest in 8a, a clash between the chloride
and one of the carbon substituents would ensue (R instead of
the H-atom labeled in red in Scheme 3). To avoid such penalty,
these substrates strongly favor the intramolecular hydrogen-
bonding mode seen in the X-ray structures of 4 and 5, which
corresponds to conformer 8b. This “Thorpe−Ingold effect”
translates into particularly effective regiocontrol in the
subsequent trans-hydrometalation. Experimental results from
our group indeed show the somewhat counterintuitive trend
that delivery of the bulky R3M group to the proximal C-atom
becomes more selective when going from a secondary (or
primary) to an arguably more encumbered tertiary propargyl
alcohol substrate. The examples compiled in Table 2 are
representative.11
Binding of an Unfunctionalized Alkyne. The computa-
tional data outlined above show that a protic site in vicinity to
the triple bond does not change the bonding interactions with
the [Cp*Ru−Cl] fragment per se but markedly enhances the
stability of the resulting complex. To confirm this important
D
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 2. Comparison of the Regioselectivity of trans-
Hydrometalation Reactions of Prototype Secondary and
Tertiary Propargyl Alcohol Substrates Catalyzed by 1 (1.25
mol%)11
Scheme 4. trans-Hydrostannation of 10 at Low
Temperature
a
a
Successful transformation, however, requires slow addition to avoid
decomposition of the precatalyst by excess stannane; product 11
formed at −65 °C was ∼95% isomerically pure.
standard conditions at ambient temperature gave poor results,
the same reaction carried out at −65 °C afforded the desired
product 11 in appreciable yield. The major isomer results from
hydride delivery to the less hindered acetylene C-atom; this
preference leads to the somewhat counterintuitive but
potentially very useful substitution pattern in which the bulky
Me3Sn residue ends up adjacent to rather than distal from the
bulky tert-butyl group.
Binding to Cationic Catalysts. If a peripheral hydrogen
bond with a Ru−Cl unit stabilizes an alkyne π-complex by
∼2.8−4.8 kcal/mol, the halide-free cationic catalyst [Cp*Ru-
(MeCN)3]PF6 originally used by Trost and co-workers for
trans-hydrosilylation2,16 should bind a propargyl alcohol
derivative much less tightly than [Cp*RuCl]4. This is indeed
the case: according to 19F NMR, only 16% of 1,1,1-trifluoro-
2(2-trifluoromethyl)pent-3-yn-2-ol are bound to the cationic
template at −50 °C (Scheme 5).32 Interestingly though, the
conclusion, binding studies with an unfunctionalized substrate
of similar steric demand were carried out.
Attempted crystallization of an adduct between 4,4-dimethyl-
2-pentyne (10) and complex 1 have so far met with failure.
Rather, the single crystals precipitating from a concentrated
solution in CH2Cl2 at −20 °C solely consisted of unchanged
[Cp*RuCl]4 and solute 4,4-dimethyl-2-pentyne (Figure 4). The
triple bond is unperturbed (1.168(5) Å) and the alkyne almost
linear (C3−C1−C2 175.4(4)°).
Scheme 5. Temperature-Dependent Coordination of a
Propargyl Alcohol to a Cationic [Cp*Ru]+ Fragment
Figure 4. Asymmetric unit of the cocrystals formed from 1 and 4,4-
dimethyl-2-pentyne (10).
Further cooling of a more dilute solution in CD2Cl2,
however, led to a characteristic color change from brown to
purple. According to NMR, 70% of 10 is bound to the metal
fragment at −50 °C; cooling/warming cycles proved the
process to be fully reversible. As expected, the C-atoms of
bound 10 show the same characteristic deshielding (δC = 159.4,
131.1 ppm) as those of the propargylic substrates discussed in
the previous sections.
These data prove that substrates with and without protic
functionality bind in the same way to the catalytically active
[Cp*RuCl] fragment and therefore undergo the same trans-
addition reactions; however, they exhibit largely different
affinities. In a preparative context this distinction can be turned
into an advantage as it allows propargyl alcohols to be
addressed without touching an “ordinary” alkyne in vicinity
(see below). Furthermore, hydrometalations of substrates with
low binding affinity have a better chance to proceed when
carried out at low temperatures, provided the R3MH reagent is
added slowly to avoid competing decomposition of the catalyst.
The example shown in Scheme 4 is representative: Whereas the
trans-hydrostannation of the bulky substrate 10 under the
CF3 groups are diastereotopic at this low temperature, which
likely indicates that a chelate structure 13 is formed by
coordination of the −OH group to the metal center (Figure
5).33 Determination of the coalescence temperature and line
shape analysis reveal a low barrier between 12 and 13 (ΔH⧧ ≈
6 kcal/mol) and an associative pathway for chelate formation/
displacement of MeCN (ΔS⧧ = −22.4 cal·mol−1·K−1).
Attempted Characterization of the Loaded Catalyst.
The data outlined above leave no doubt that the polarized
[Ru−Cl] unit is key for the preorganization of the ligand sphere
of the catalyst in that it locks the alkyne substrate via interligand
hydrogen bonding. At the same time, a hypervalent Cl···MR3
interaction has been proposed to position the incoming reagent
H−MR3 in a matching orientation to give a loaded complex of
type C as the key intermediate accountable for regio- and
stereoselective trans-hydrometalation.11
In an attempt to verify this scenario, Me3SnH was quickly
added to a cherry-red solution of the well-defined alkyne adduct
E
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
even at low temperatures implies a favorable reaction
coordinate with only low barriers once the loaded π-complex
has formed; this is well in line with the computational results
outlined below. The argument is reinforced by the fact that the
trans-hydrostannation of 4,4-dimethyl-2-pentyne (10) was also
successful at −65 °C, as long as the stannane was added slowly
(Scheme 4).
Emulation of the Loaded Catalyst. The gathered data
leave no doubt that an alkyne substrate, when bound to
[Cp*RuCl], acts as a four-electron donor in the first place. The
empirical 18-electron rule, however, mandates that it mutates to
a two-electron donor when H−MR3 is taken up and a
conceived intermediate of type C is formed (Scheme 1). If this
switch of the donor character does not occur, the proposed
inner sphere mechanism is obsolete. The fact that C could not
be observed by NMR even at −90 °C might either mean that it
is too short-lived for direct inspection or that it does not form
at all.
Figure 5. Variable-temperature 19F NMR spectrum (CD2Cl2) of a 1:1
mixture of [Cp*Ru(MeCN)3]PF6 and 1,1,1-trifluoro-2(2-
trifluoromethyl)pent-3-yn-2-ol.
To clarify this critical aspect, attempts were made to emulate
the loaded catalyst by formal replacement of H−MR3 by
nonreactive donor ligands (Scheme 7). In a first foray, Me3P
5 in CD2Cl2 at −90 °C, causing an instant color change to
purple (Scheme 6). A single new species is formed according to
Scheme 7. Emulation of the Loaded Complex
Scheme 6. Attempted Characterization of a Loaded Complex
Proves That trans-Hydrostannation Is Facile Even at Low
Temperatures
was added to a solution of complex 5 in CD2Cl2 at −78 °C,
which engenders a pronounced upfield shift of the acetylene
signals. The alkyne resonances in the resulting product 16 are
very close to those of the unbound substrates (Table 3). The
2
NMR; however, it is not the loaded complex C but the
ruthenium adduct of the hydrostannylated alkene product. The
fact that hydrogen and tin are trans to each other in the alkene
ligand of complex 14 even when formed at very low
temperature rules out the possibility that the unorthodox
stereochemical outcome is no more but the net result of a
conventional cis-addition followed by isomerization;1 rather,
this observation constitutes the most rigorous experimental
proof so far that trans-addition is truly an inherent quality.
Even gentle warming of the solution of 14 to −60 °C suffices
to release product 15 and [Cp*RuCl]4 (1); this step is
irreversible within the limits of detection of NMR, thus
showing that the alkene product is a much weaker ligand than
the alkyne substrate. Associated with this is a characteristic
fact that the 13C NMR signals show characteristic JP,C
couplings, however, proves that the alkyne remains bound
and has not been displaced from the metal center by the
phosphine. This aspect is also manifest in the diastereotopic
character of the two CF3 groups in complex 16.
Weaker σ-donor ligands such as pyridine or MeCN entail
similar effects, even though the resulting adducts are more
labile and coordination is (fully) reversible.34 In the case of 17,
the two CF3 groups display a singlet resonance at RT but
resolve to two quartets at −50 °C. For 18, the singlet CF3
resonance begins to split only at −85 °C and full resolution
could not be reached above the freezing point of the solvent
(−96.7 °C).
Adducts 16 (Figure S7, Supporting Information) and 17
(Figure 6) have also been characterized by X-ray diffraction. It
is interesting to note that the switch of the acetylene from a
four- to a two-electron donoras clearly expressed in the
NMR data (Table 3)is not reflected in the CC bond
lengths that remain long (1.256(3) and 1.260(1) Å,
respectively); likewise, a significant bending of the alkyne
away from linearity persists in either case. Obviously, electron
3
change of the JSn,H, which increases from 88 Hz in 14 to 124
Hz in 15: this feature indicates substantial rehybridization
during decomplexation, which in turn implies significant
electron back-donation from the Ru center into the π*-orbital
of the alkene in adduct 14. This finding concurs with the
computational results for trans-2-butene outlined above. The
ready formation of the product and its facile decomplexation
F
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
silane complexes with [Cp*RuCl] fragments.11,35−37 Although
the gathered information is no rigorous proof, it is definitely in
accord with the proposed inner sphere mechanism involving a
loaded complex of type C. Qualitatively, electron donation
from the metal fragment to the ligated R3M−H bond increases
the hydridic character and should hence render addition to the
activated triple bond more facile.
Table 3. Characteristic 13C NMR Data (CD2Cl2, ppm)
Computational Studies. While all experimental data speak
for an inner-sphere rather than an outer-sphere mechanism, this
aspect was deemed sufficiently important to warrant further
scrutiny. Therefore, a detailed computational study of trans-
hydrostannation was carried out at the SMD(DCM)-M06/
def2-TZVP//M06/def2-SVP level of theory (see Supporting
Information for details), which complements earlier inves-
tigations on trans-hydrosilylation20 as well as trans-hydro-
genation.15 The focus on trans-hydrostannation seems legit-
imate given its particularly wide scope and excellent
selectivity.10,11,18,19
In excellent agreement with the experimental data, binding of
a propargyl alcohol to the metal fragment affords a complex
with a peripheral interligand hydrogen bond, which is capable
of binding Me3SnH in the form of the σ-complex INT1 (Figure
7). This starting point corresponds to complex C that we had
proposed in this and our earlier experimental work.11 The
ruthenium center in INT1 cleaves the Sn−H bond while
transferring the hydride to the ligated butynol substrate in a
concerted fashion via TS1-2 (ΔG° = 4.9 kcal/mol). The
resulting metallacyclopropene (η2-vinylruthenium)38 intermedi-
ate INT2 undergoes direct reductive elimination of the Me3Sn-
group to the “carbene-type” α-center to form the trans-addition
product via TS2-Z with a barrier of only 4.4 kcal/mol. This
route is by far the lowest energy pathway available to INT2 and
leads to the experimentally observed trans/α-product (Z-prod,
−21.1 kcal/mol).
a
Not detected because of massive line broadening caused by fast
MeCN exchange even at −85 °C.
Evolution of INT2 into the thermodynamically slightly more
stable cis/α-product (E-prod, −22.9 kcal/mol) is possible by
either a concerted (via TS2-5) or a stepwise pathway (via INT3
and TS3-5) to give the isomeric ruthenacyclopropene INT5,
which forms the product upon reductive coupling. Even the
more favorable of these two possible routes, however, is 9.5
kcal/mol higher in energy than the pathway leading to the
trans/α-product. This differential finds excellent correspond-
ence in the generally very high levels of selectivity observed in
such reactions.11,18,19
Figure 6. Structure of complex 17 in the solid state; all H-atoms,
except the one engaged in hydrogen bonding to the chloride ligand,
are omitted for clarity. For the X-ray structure of the corresponding
phosphine adduct 16, see the Supporting Information.
Unlike the trans-hydrogenation of alkynes,15 where the
evolution of the ruthenacyclopropene into a true carbene
constitutes an important additional reaction channel, the
formation of an analogous carbene INT4 by reductive
elimination of the tin residue and the Ru−Cβ bond proved
highly unfavorable. INT4 could derive from INT3 passed
through en route to the cis-addition product, but the barrier to
its formation is prohibitively high.
Attempts at finding pathways in which Me3Sn delivery
precedes hydrogen transfer met with failure; they invariably
converged to the “hydrogen first” route shown in Figure 7. This
preference distinguishes the hydrostannation from hydro-
silylation, for which Me3Si transfer prior to H transfer is
feasible.20 The distinct behavior manifest in these computa-
tional data, however, is quite intuitive if one considers that σ-
stannane complexes are certainly more hydridic than their σ-
silane counterparts.
back-donation from [Cp*Ru] to the antibonding π*-orbital as
the major determining factor of either structural attribute is
operative independent of whether the alkyne acts as a four- or
two-electron donor ligand. While the Ru1−P1 bond in 16 is
rather short (2.299(9) Å), the Ru1···N1 distance (2.157(1) Å)
in the pyridine adduct 17 is well above the sum of the van der
Waals radii (2.01 Å), indicative of an only weak interaction
even in the solid state. Importantly, the peripheral interligand
−O−(H)···Cl hydrogen bonding persists in either adduct
although it is slightly longer than that in 5 (2.975(1) Å)
[(3.039(3) Å in complex 16; 3.044(1) Å in complex 17].
Nevertheless, the bridges remain strong enough to lock the
propargyl alcohol in place.27
These data showcase that ligands of different donor ability
readily enter into the coordination sphere of an alkyne complex
of type B; reagents of type H−MR3 are expected to do the
same as they are known to form well-defined σ-stannane or σ-
Possible stannation of the distal C-atom was also investigated
to determine if hydrogen bonding could effectively account for
G
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 7. Computed Gibbs free energy profile of a prototype hydrostannation catalyzed by [Cp*RuCl] in CH2Cl2; energies in kcal/mol.
the regioselectivity (Figure 8). Indeed association of the alkyne
leads to an association complex INT0′ that is higher in energy
or a 1,3-diene unit; likewise, 1,3-enynes also usually fail to react.
Compounds 19−24 are representative (Scheme 8).10,11
We reasoned that this shortcoming might be remedied, at
least in part, with the help of appropriately positioned protic
substituents (Scheme 8). While a regular enyne F likely traps
Scheme 8. (Top) Representative Examples of Poorly
Reactive Substrates Comprising Four-/Six-Electron Donor
Sites; (Bottom) The Presence of a Protic Group in the
Vicinity of the Alkyne Might Allow a Reactive Binding Mode
To Be Achieved and Hence the Substrate Scope To Be
Increased
Figure 8. Computed Gibbs free energy profile in CH2Cl2 (in units of
kcal/mol) for distinguishing distal/proximal hydrostannation of a
prototype propargyl alcohol.
than the INT0 containing the hydrogen bond by 4.8 kcal/mol.
The barrier height of their interconversion (TS0-0′) is similar
in magnitude to the highest barrier heights leading to the
stereoiomeric products Z-prod and Z-prod′ suggesting that the
product selectivity depends largely on the energy difference
between INT0 and INT0′. The energy difference between the
HSnMe3 association complexes INT1 and INT1′ is smaller due
to a hydrogen bond with the complexed hydride in INT1′.
Overall, this result is in agreement with the experimentally
observed preference for the proximal product, Z-prod.
Preparative Implications. trans-Hydrometalations as well
as trans-hydrogenation have proven difficult, if not impossible,
for all substrates that contain structural elements competing
with the alkyne unit for binding to the active ruthenium
fragments [Cp*Ru]+ or [Cp*RuCl]; all potential four- or six-
electron donors fall into this category.39−41 That is why
problems are frequently encountered when working with
starting materials that comprise an (electron-rich) arene and/
H
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 4. trans-Hydrostannations of Substrates Containing a Protic Site Next to a 1,3-Enyne or Arene Unit
a
The reactions were carried out at ambient temperature with 1.05−1.2 equiv of R3SnH in CH2Cl2 in the presence of 1 (1−2 mol %); unless stated
b
otherwise, the selectivity for trans-addition was ≥95:5 (major regioisomer). “Proximal” denotes delivery of the R3M group to the alkyne C-atom
next to the protic group, whereas “distal” refers to the more remote C-atom. Z:E = 87:13 (major isomer). Yield of pure trans/α-product after flash
chromatography. Using [Cp*RuCl2]n as the catalyst. Traces of distannylated product were detected. The crude product contains other isomers
c
d
e
f
g
h
i
j
that were not rigorously assigned. At −30 °C. At −10 °C. Ca. 20% distannylated product was formed. For the sake of a better overview, the table
includes some examples from our previous work.11,19
the catalyst in the form of stable adducts of type G, the steering
effect of a propargylic −OH group might allow a compound of
type H to reach a reactive coordination mode J and hence
trans-addition to proceed. Similar arguments can be given for
ordinary arylalkynes and their hydroxylated siblings, respec-
tively.
Scheme 9. Intermolecular Competition Experiments
In line with this notion, the examples compiled in Table 4
(entries 1−6) show that adequately functionalized 1,3-enynes
are indeed amenable to trans-hydrostannation, whereas their
unfunctionalized relatives 19−21 (Scheme 8) basically fail to
react. The presumed steering effect responsible for this
favorable outcome definitely surfaces in the high preference
for proximal R3M delivery in all cases investigated.
The higher affinity of propargyl alcohols to [Cp*RuCl] also
allows meaningful levels of site-selectivity to be harnessed in
reactions of polyunsaturated substrates. This is manifest in the
intermolecular competition experiments shown in Scheme 9, in
which an unprotected propargyl alcohol clearly wins over the
corresponding methyl ether as well as over an unfunctionalized
alkyne. The results obtained with a set of diyne substrates are
more compelling (Table 4, entries 7−14): not only did the
propargylic site react preferentially in all cases investigated, but
the trans-hydrostannation also outperformed conceivable inter-
or intramolecular [2+2] cycloadditions, which are known to be
very facile (see above).11 Moreover, entries 12−14 show that
the discrimination between two different acetylene units even
I
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 5. trans-Hydrometalations of (Hetero)arene Derivatives Endowed with Protic Sites
a
The reactions were carried out at ambient temperature with 1.05−1.2 equiv of R3MH in CH2Cl2 in the presence of 1 (1−2 mol %); unless stated
b
otherwise, the selectivity for trans-addition was ≥95:5 (major regioisomer). “Proximal” denotes delivery of the R3M group to the alkyne C-atom
next to the protic group, whereas “distal” refers to the more remote C-atom. The reaction was performed in the presence of B(C6F5)3 to block the
N-donor site; the additive was then removed on treatment of the crude product with DABCO. Incomplete conversion even after 46 h reaction time.
c
d
pertains to 1,3-diynes in which the triple bonds are conjugated
and hence electronically coupled. The generality of this non-
obvious reactivity pattern is currently investigated in more
detail and its implementation as a strategic design element into
natural product total synthesis is underway.
ing less electron-rich N-tosyl derivative; in fact, its hydro-
silylation never reached full conversion. Most notably though,
the N-unprotected substrate furnished the trans/α-product,
whereas its N-tosylated sibling gave the distal isomer. Such a
mutually inverse pattern was also seen in the pyridone series
(compare entries 16/17); it clearly showcases that silylation at
the distal position predominates as soon as the directing effect
of the NH group is lost. Of note is also the indole derivative
shown in entry 8 that reacted smoothly, whereas a related
indazole (entry 9) mandated addition of B(C6F5)3 to the
mixture to block the high affinity N-donor site (compare the
pyridine adduct 17 discussed above); after hydrosilylation, the
boron-free product was obtained on treatment of the crude
material with DABCO. Collectively, the data compiled in Table
5 substantiate the profound effect that a protic site is able to
exert: it allows trans-hydrometalation of compounds that are
otherwise catalyst poisons to proceed with respectable
outcomes and, in so doing, significantly expands the substrate
scope of this valuable transformation.
As mentioned above, even simple arylalkynes such as 23 and
24 proved problematic in the past,10,11 again most likely
because of the ability of an aromatic ring to act as competing
four- or even six-electron donor (Scheme 8).40,41 Therefore, we
reinvestigated whether or not a suitably placed protic site allows
this severe limitation in substrate scope to be overcome. In line
with our expectation, a set of arylalkynes bearing a hydroxyl
group or a sulfonamide in the chain underwent trans-
hydrometalation with ease (Table 5, entries 1−6); the
compatibility with a labile aldol motif or an aldehyde
underscores the mildness of the method. This favorable
outcome suggested that an unprotected −NH group as part
of a heterocyclic ring might do the same and force a
neighboring triple bond to succumb to trans-hydrometalation.
Gratifyingly, this proved to be the case even for fairly electron-
rich heteroarenes; for reasons of product stability, this study
was focused on hydrosilylation, with the different silanes giving
largely similar results (entries 7−17). The comparison of
entries 10 and 15 is compelling in that the unprotected
imidazole derivative reacted much faster than the correspond-
Limitations: Binding and Reactivity of a TMS-Capped
Propargyl Alcohol. Silylated alkynes make another interesting
case, as Me3Si− groups are likely non-innocent. While all
substrates investigated so far led to single well-defined
complexes on treatment with 1, 3-(trimethylsilyl)prop-2-yn-1-
ol (31) gives rise to three adducts at −20 °C (Scheme 10); one
J
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 10. Binding of a Prototype C-Silylated Propargyl
Alcohol
of them (34), however, is likely off cycle.42 The coordination
mode of the major product 32 (ca. 55%) is similar to the one of
the propargylic alcohols described above as inferred from the
characteristic downfield shift of the alkyne C-atoms (δC
=
143.0, 95.5 ppm).
The NMR data of the other relevant complex 33 in solution
(ca. 35%, δC = 114.6, 92.5 ppm), however, suggest that the
chloride ligand engages with the silyl group which resonates at
δSi = +1.2 ppm. Such a distinct upfield shift relative to the
unbound substrate 31 (δSi = −17.5 ppm) is certainly in line
with a hypervalent Si···Cl interaction, although it cannot be
considered a firm proof.43,44
To clarify this interesting aspect, the coordination behavior
of the TMS capped propargyl alcohol 31 was explored
computationally (Figure 9). The preferred conformations
either contain a bridging OH···Cl hydrogen bond (INT0-Si)
or what appears to be a hypervalent Si···Cl type interaction
(INT0′-Si). There is predicted to be very little difference (0.3
kcal/mol) in energy between the conformers consistent with
the experimental observations. Some degree of hypervalent
interaction in INT0-Si may be supported by the slight
elongation of the pseudoapical Si−Me bond (1.89 Å) relative
to the two pseudoequatorial Si−Me bonds (1.87 Å). The Cl−Si
bond length is within the sum of their van der Waals radii,
suggesting either a steric or bonding interaction. The largest
difference however is between the C−C−Si bond angles. The
C−C−Si bond angle in INT0-Si is significantly contracted by
15.3° relative to INT0-Si′. The compressed angle in INT0-Si
can be attributed to a steric interaction between the methyl
groups of the Cp* and the SiMe3 moiety which causes an
energy increase relative to INT0-Si′ thus counteracting the
stabilizing effect of the hydrogen bond.
Overall, it is likely a combination of a favorable hypervalent
interaction in INT0-Si′ and an unfavorable steric interaction in
INT0-Si that cancels out the hydrogen bonding in INT0-Si,
resulting in nearly equal populations of both conformers. This
is in stark contrast to the significant energy difference of 4.8
kcal/mol between the conformers of the nonsilylated complex
which is readily attributed to the OH···Cl bond as no
hypervalent or steric interaction differences are present. The
OH···Cl distances are similar between the nonsilylated and
silylated cases suggesting that both have similar bond strengths.
In view of these spectroscopic and computational data, it is
not surprising that hydrometalations of compound 31 are
hardly selective. Specifically, hydrostannation gave a ∼1:1
mixture of 35 and 36. Although hydrosilylation with (EtO)3SiH
afforded the trans/α-product 37 as the major compound (ca.
57%, spontaneous condensation of the product primarily
formed leads to the cyclic siloxane structure), substantial
amounts of the cis/α-product 38 are also present in the crude
Figure 9. Computed geometries and energy differences for the
preferred conformations of the complex between [Cp*RuCl] and the
C-silylated propargyl alcohol 31 versus the complex with 2-butyne.
The numbers in red are computed bond distances (Å) and angles (°).
The bond angles are displayed on the ball-and-stick rendered models
for clarity.
mixture; the question as to why cis-adducts are formed at all is
not trivial to answer at this point and has to await further
studies.
CONCLUSION
■
The combined experimental and theoretical approach chosen in
this investigation provides a detailed picture of the course of
ruthenium catalyzed trans-hydrometalation reactions of internal
alkynes. Primarily, such substrates behave as four-electron
donor ligands to the catalytically active metal fragment
[Cp*RuCl], which entails massive activation of the bound
triple bond. Once the reagent R3MH (M = Si, Ge, Sn) enters
the ligand sphere, the coordinated alkyne switches to assume a
two-electron donor character; the resulting loaded complex
then evolves via a ruthenacyclopropene as the key intermediate
into the final trans-addition product. In accordance with the low
computed barriers, spectral and preparative data show that the
reaction can proceed even at low temperatures. While this
explanation for the unorthodox trans-selectivity is reminiscent
of previous studies on trans-hydrosilylation and trans-hydro-
genation, the current investigation was also able to uncover why
unsymmetrical alkynes bearing protic substituents are partic-
ularly privileged substrates: a nascent hydrogen bond between
the −OH (−NHR) group and the polarized [Ru−Cl] unit
leads to tight binding and, at the same time, imposes
directionality onto the ligand sphere of the resulting complex,
which ultimately accounts for the faithful delivery of the R3M−
K
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(2) (a) Trost, B. M.; Ball, Z. T. J. Am. Chem. Soc. 2001, 123, 12726−
group to the alkyne C-atom proximal to the steering
substituent. Importantly, the cooperativity between protic
substituents and catalyst also allows the substrate scope to be
significantly expanded: while regular 1,3-enynes or arylalkynes
tend to be inert, such compounds react smoothly when carrying
an −OH group in vicinity to the triple bond. Even notoriously
challenging electron-rich heterocycles prove amenable to trans-
hydrosilylation, provided they contain an unprotected − NH
group as part of the ring system. Finally, the synergy between
the [Ru−Cl] unit and a protic group allows site selectivity to be
exerted, so that a propargylic triple bond reacts preferentially
over a normal alkyne. This roster of enabling features holds
significant promise for further applications of alkyne trans-
addition chemistry.
12727. (b) Trost, B. M.; Ball, Z. T.; Joge, T. J. Am. Chem. Soc. 2002,
̈
124, 7922−7923. (c) Trost, B. M.; Machacek, M. R.; Ball, Z. T. Org.
Lett. 2003, 5, 1895−1898. (d) Trost, B. M.; Ball, Z. T. J. Am. Chem.
Soc. 2005, 127, 17644−17655.
(3) Trost, B. M.; Ball, Z. T. Synthesis 2005, 2005, 853−887.
(4) Frihed, T. G.; Furstner, A. Bull. Chem. Soc. Jpn. 2016, 89, 135−
̈
160.
(5) Intramolecularity is arguably the most effective countermeasure;
for instructive cases of intramolecular trans-hydrosilylations, see:
(a) Trost, B. M.; Ball, Z. T. J. Am. Chem. Soc. 2003, 125, 30−31.
(b) Trost, B. M.; Ball, Z. T.; Laemmerhold, K. M. J. Am. Chem. Soc.
2005, 127, 10028−10038. (c) Denmark, S. E.; Pan, W. Org. Lett. 2002,
4, 4163−4166. (d) Denmark, S. E.; Pan, W. Org. Lett. 2003, 5, 1119−
1122.
(6) (a) Furstner, A.; Radkowski, K. Chem. Commun. 2002, 2182−
̈
2183. (b) Lacombe, F.; Radkowski, K.; Seidel, G.; Furstner, A.
̈
ASSOCIATED CONTENT
■
Tetrahedron 2004, 60, 7315−7324.
(7) (a) Micoine, K.; Furstner, A. J. Am. Chem. Soc. 2010, 132,
14064−14066. (b) Lehr, K.; Mariz, R.; Leseurre, L.; Gabor, B.;
Furstner, A. Angew. Chem., Int. Ed. 2011, 50, 11373−11377.
S
* Supporting Information
̈
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.6b12517.
̈
(8) The term trans-hydrostannation (silylation, germylation) of an
alkyne as used herein denotes a reaction in which the H and the MR3
unit (M = Sn, Si, Ge) end up trans to each other. It is pointed out that
the resulting product is correctly termed Z-configured for the
formalism of nomenclature, whereas the product of an analogous
trans-hydroboration is E-configured.
Experimental part including characterization data, NMR
spectra of new compounds, and supporting crystallo-
graphic information (PDF)
Computational part including methodological details,
orbital interaction data, free energy profile for a
regioisomeric pathway, energy tables, and optimized
Cartesian coordinates (PDF)
(9) Sundararaju, B.; Furstner, A. Angew. Chem., Int. Ed. 2013, 52,
̈
14050−14054.
X-ray crystallographic data (CIF)
(10) Rummelt, S. M.; Furstner, A. Angew. Chem., Int. Ed. 2014, 53,
̈
3626−3630.
(11) Rummelt, S. M.; Radkowski, K.; Rosca, D.-A.; Furstner, A. J. Am.
̈
AUTHOR INFORMATION
Corresponding Author
*fuerstner@kofo.mpg.de
ORCID
■
Chem. Soc. 2015, 137, 5506−5519.
(12) (a) Matsuda, T.; Kadowaki, S.; Yamaguchi, Y.; Murakami, M.
Org. Lett. 2010, 12, 1056−1058. (b) Matsuda, T.; Kadowaki, S.;
Murakami, M. Chem. Commun. 2007, 2627−2629.
(13) Radkowski, K.; Sundararaju, B.; Furstner, A. Angew. Chem., Int.
̈
Walter Thiel: 0000-0001-6780-0350
Ed. 2013, 52, 355−360.
Alois Furstner: 0000-0003-0098-3417
̈
(14) Fuchs, M.; Furstner, A. Angew. Chem., Int. Ed. 2015, 54, 3978−
̈
Present Addresses
3982.
‡L.M.W.: Department of Chemistry, Univesity of Massachu-
setts Lowell, 1 University Avenue, Lowell, MA 01854, USA
§M.W.: Indian Institute of Technology Bombay, Mumbai-
400076, India
(15) Leutzsch, M.; Wolf, L. M.; Gupta, P.; Fuchs, M.; Thiel, W.;
Fares
̀
, C.; Furstner, A. Angew. Chem., Int. Ed. 2015, 54, 12431−12436.
̈
(16) For the first application of the cationic catalyst to propargylic
alcohols, see: Trost, B. M.; Ball, Z. T.; Joge, T. Angew. Chem., Int. Ed.
̈
2003, 42, 3415−3418.
Notes
(17) Similarly directed trans-hydroborations of propargyl alcohols
have not been achieved because pinacol borane reacts faster with the
−OH group than with the triple bond.
The authors declare the following competing financial
interest(s): Patent application filed.
(18) Rummelt, S. M.; Preindl, J.; Sommer, H.; Furstner, A. Angew.
̈
ACKNOWLEDGMENTS
Chem., Int. Ed. 2015, 54, 6241−6245.
■
(19) (a) Sommer, H.; Furstner, A. Org. Lett. 2016, 18, 3210−3213.
(b) Schaubach, S.; Furstner, A.; Michigami, K. Synthesis 2016, 49,
202−208. (c) Sommer, H.; Furstner, A. Chem. - Eur. J. 2017, 23, 558−
̈
Generous financial support by the Alexander-von-Humboldt
Foundation (fellowship to D.A.R.) and the MPG is gratefully
acknowledged. We thank the analytical departments of the MPI
̈
̈
562.
for excellent support; Dr. J. Flasz, M. Ilg, Dr. M. Muller, Dr. S.
̈
(20) (a) Chung, L. W.; Wu, Y.-D.; Trost, B. M.; Ball, Z. T. J. Am.
Chem. Soc. 2003, 125, 11578−11582. (b) Ding, S.; Song, L.-J.; Chung,
L. W.; Zhang, X.; Sun, J.; Wu, Y.-D. J. Am. Chem. Soc. 2013, 135,
13835−13842. (c) Zhang, X.; Chung, L. W.; Wu, Y.-D. Acc. Chem. Res.
2016, 49, 1302−1310.
(21) Dutta, B.; Curchod, B. F. E.; Campomanes, P.; Solari, E.;
Scopelliti, R.; Rothlisberger, U.; Severin, K. Chem. - Eur. J. 2010, 16,
8400−8409.
(22) (a) Fagan, P. J.; Mahoney, W. S.; Calabrese, J. C.; Williams, I. D.
Organometallics 1990, 9, 1843−1852. (b) Fagan, P. J.; Ward, M. D.;
Calabrese, J. C. J. Am. Chem. Soc. 1989, 111, 1698−1719.
(23) For pioneering studies, see: (a) Templeton, J. L.; Ward, B. C. J.
Am. Chem. Soc. 1980, 102, 3288−3290. (b) Ward, B. C.; Templeton, J.
L. J. Am. Chem. Soc. 1980, 102, 1532−1538.
Rummelt, Dr. S. Schaubach, Dr. H. Sommer, Dr. F. Ungeheuer
for some of the examples shown in the tables; and Umicore AG
& Co KG, Hanau, for a generous gift of noble metal salts.
REFERENCES
■
(1) (a) Dobbs, A. P.; Chio, F. K. I. In Comprehensive Organic Synthesis
II, 2nd ed.; Knochel, P., Molander, G. A., Eds.; Elsevier: Amsterdam,
2014; Vol. 8; pp 964−998. (b) Smith, N. D.; Mancuso, J.; Lautens, M.
Chem. Rev. 2000, 100, 3257−3282. (c) Hydrosilylation. A Compre-
hensive Review on Recent Advances; Marciniec, B., Ed.; Advances in
Silicon Science 1; Springer: Amsterdam, 2009. (d) Barbeyron, R.;
Benedetti, E.; Cossy, J.; Vasseur, J.-J.; Arseniyadis, S.; Smietana, M.
Tetrahedron 2014, 70, 8431−8452.
L
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(24) For propargylic alcohols acting as four-electron donors in the
(38) Frohnapfel, D. S.; Templeton, J. L. Coord. Chem. Rev. 2000,
206−207, 199−235.
osmium series, see: (a) Carbo,
Jean, Y.; Lledos, A.; Lopez, A. M.; Onate, E. Organometallics 2002, 21,
305−314. (b) Esteruelas, M. A.; Lop
Organometallics 1997, 16, 4657−4667.
(25) Compare: (a) Le Paih, J.; Derien, S.; Bruneau, C.; Demerseman,
B.; Toupet, L.; Dixneuf, P. H. Angew. Chem., Int. Ed. 2001, 40, 2912−
2915. (b) Le Paih, J.; Derien, S.; Demerseman, B.; Bruneau, C.;
́
J. J.; Crochet, P.; Esteruelas, M. A.;
(39) See ref 17 and the following: (b) Fagan, P. J.; Ward, M. D.;
Calabrese, J. C. J. Am. Chem. Soc. 1989, 111, 1698−1719. (c) Steines,
́
́
̃
́
ez, A. M.; Ruiz, N.; Tolosa, J. I.
S.; Englert, U.; Drießen-Holscher, B. Chem. Commun. 2000, 217−218.
̈
(40) O’Connor, J. M.; Friese, S. J.; Rodgers, B. L.; Rheingold, A. L.;
Zakharov, L. J. Am. Chem. Soc. 2005, 127, 9346−9347.
́
(41) (a) Gill, T. P.; Mann, K. R. Organometallics 1982, 1, 485−488.
(b) Schmid, A.; Piotrowski, H.; Lindel, T. Eur. J. Inorg. Chem. 2003,
2003, 2255−2263.
́
Dixneuf, P. H.; Toupet, L.; Dazinger, G.; Kirchner, K. Chem. - Eur. J.
2005, 11, 1312−1324.
(42) The third discrete species in solution (ca. 10%) was identified as
the vinylruthenium hydride species 34, which seems to be a byproduct
of the stoichiometric binding experiments and does not play a role in
catalysis. The resonance at δH = −9.28 ppm reveals its hydridic nature,
whereas the diastereotopicity of the CH2−OH protons indicates a
chelate structure. The shielded 29Si resonance (δSi = +7.6 ppm)
indicates a hypervalent interaction, probably with the hydride ligand.
Deuterium labeling experiments suggest that the −OH group of the
substrate is the source of the Ru−H unit as well as the alkenyl proton.
(43) Lachaize, S.; Sabo-Etienne, S. Eur. J. Inorg. Chem. 2006, 2006,
2115−2127.
(26) Deuterium bond energies may be lower than hydrogen bond
energies, but the effect has previously been estimated to be in the
order of only 1−2 kJ/mol (∼0.24−0.48 kcal/mol): Rao, C. N. R. J.
Chem. Soc., Faraday Trans. 1 1975, 71, 980−983. Therefore one
cannot expect any significant impact on the observed selectivities. In
fact, the π-complex formed from a deuterated propargyl alcohol
showed the same spectral characteristics (except for the absence of the
1
−OH shift in the H NMR spectrum) as its protio analogue 4 and
identical reactivity in trans-hydrostannation and trans-hydrosilylation.
(27) (a) Aullon
A. G. Chem. Commun. 1998, 653−654. (b) Kovac
Coord. Chem. Rev. 2006, 250, 710−727.
́
, G.; Bellamy, D.; Brammer, L.; Bruton, E. A.; Orpen,
́
s, A.; Varga, Z.
(44) See ref 35 and the following: Mai, V. H.; Korobkov, I.; Nikonov,
G. I. Organometallics 2016, 35, 936−942.
(28) (a) The average CC bond length of an alkyne is 1.18 Å, see:
Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen, A. G.;
Taylor, R. J. Chem. Soc., Perkin Trans. 2 1987, S1. (b) The alkyne
bond length in a related propynylcyclohexanol derivative is 1.179 Å,
see: Robinson, R. P.; Buckbinder, L.; Haugeto, A. I.; McNiff, P. A.;
Millham, M. L.; Reese, M. R.; Schaefer, J. F.; Abramov, Y. A.; Bordner,
J.; Chantigny, Y. A.; Kleinman, E. F.; Laird, E. R.; Morgan, B. P.;
Murray, J. C.; Salter, E. D.; Wessel, M. D.; Yocum, S. A. J. Med. Chem.
2009, 52, 1731−1743.
(29) See the Supporting Information for the methods used for
constructing the orbital interaction diagram.
(30) (a) Trost, B. M.; Livingston, R. C. J. Am. Chem. Soc. 1995, 117,
9586−9587. (b) Trost, B. M.; Livingston, R. C. J. Am. Chem. Soc.
2008, 130, 11970−11978.
(31) For an improved catalyst for redox isomerization, see:
Schaubach, S.; Gebauer, K.; Ungeheuer, F.; Hoffmeister, L.; Ilg, M.
K.; Wirtz, C.; Furstner, A. Chem. - Eur. J. 2016, 22, 8494−8507.
̈
(32) Despite this rather low affinity, the exchange between bound
and unbound substrate is slow; it mandates heating to +50 °C to
ensure fast exchange on the NMR time scale.
(33) Similar interactions have been proposed to serve as
regiochemical control elements. For pertinent discussion, see the
following and the references cited therein: (a) Trost, B. M.; Ryan, M.
C.; Rao, M.; Markovic, T. Z. J. Am. Chem. Soc. 2014, 136, 17422−
17425. (b) Trost, B. M.; Cregg, J. J. J. Am. Chem. Soc. 2015, 137, 620−
623.
(34) At room temperature, the solution of 5 and MeCN shows the
characteristic cherry red color of the alkyne-π-complex, whereas it is
dark green at −78 °C, which indicates formation of the adduct 5·
MeCN; this process is fully reversible. In the case of pyridine, the line
broadening in the NMR spectra of the brown solution at room
temperature suggests fast exchange, which is largely frozen out at −78
°C; see the Supporting Information.
(35) (a) Gutsulyak, D. V.; Churakov, A. V.; Kuzmina, L. G.; Howard,
J. A. K.; Nikonov, G. I. Organometallics 2009, 28, 2655−2657.
(b) Osipov, A. L.; Vyboishchikov, S. F.; Dorogov, K. Y.; Kuzmina, L.
G.; Howard, J. A. K.; Lemenovskii, D. A.; Nikonov, G. I. Chem.
Commun. 2005, 3349−3351.
(36) Kubas, G. J. Metal Dihydrogen and σ-Bond Complexes; Kluwer
Academic/Plenum Publishers: Dordrecht, The Netherlands, 2001.
(37) While stannanes form fairly stable adducts even at ambient
temperature (see ref 11), the temperature dependence of the
̀
coordination behavior of (EtO)3SiH vis-a-vis [Cp*RuCl(PiPr3]
resembles that of MeCN: according to NMR, it is fully coordinated
at − 80 °C but unbound at room temperature. Associated with this is a
characteristic and reversible color change from blue to brown.
M
DOI: 10.1021/jacs.6b12517
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX