A highly active and air-stable ruthenium complex for the ambient temperature anti-markovnikov reductive hydration of terminal alkynes

Article abstract of DOI:10.1021/ja501738a

The conversion of terminal alkynes to functionalized products by the direct addition of heteroatom-based nucleophiles is an important aim in catalysis. We report the design, synthesis, and mechanistic studies of the half-sandwich ruthenium complex 12, which is a highly active catalyst for the anti-Markovnikov reductive hydration of alkynes. The key design element of 12 involves a tridentate nitrogen-based ligand that contains a hemilabile 3-(dimethylamino) propyl substituent. Under neutral conditions, the dimethylamino substituent coordinates to the ruthenium center to generate an air-stable, 18-electron, κ³-complex. Mechanistic studies show that the dimethylamino substituent is partially dissociated from the ruthenium center (by protonation) in the reaction media, thereby generating a vacant coordination site for catalysis. These studies also show that this substituent increases hydrogenation activity by promoting activation of the reductant. At least three catalytic cycles, involving the decarboxylation of formic acid, hydration of the alkyne, and hydrogenation of the intermediate aldehyde, operate concurrently in reactions mediated by 12. A wide array of terminal alkynes are efficiently processed to linear alcohols using as little as 2 mol % of 12 at ambient temperature, and the complex 12 is stable for at least two weeks under air. The studies outlined herein establish 12 as the most active and practical catalyst for anti-Markovnikov reductive hydration discovered to date, define the structural parameters of 12 underlying its activity and stability, and delineate design strategies for synthesis of other multifunctional catalysts.

Full text of DOI:10.1021/ja501738a

Article
pubs.acs.org/JACS
A Highly Active and Air-Stable Ruthenium Complex for the
Ambient Temperature Anti-Markovnikov Reductive Hydration
of Terminal Alkynes
Mingshuo Zeng, Le Li, and Seth B. Herzon*
Department of Chemistry, Yale University, New Haven, Connecticut 06520, United States
S
* Supporting Information
ABSTRACT: The conversion of terminal alkynes to function-
alized products by the direct addition of heteroatom-based
nucleophiles is an important aim in catalysis. We report the
design, synthesis, and mechanistic studies of the half-sandwich
ruthenium complex 12, which is a highly active catalyst for
the anti-Markovnikov reductive hydration of alkynes. The key
design element of 12 involves a tridentate nitrogen-based li-
gand that contains a hemilabile 3-(dimethylamino)propyl sub-
stituent. Under neutral conditions, the dimethylamino substituent coordinates to the ruthenium center to generate an air-stable,
18-electron, κ3-complex. Mechanistic studies show that the dimethylamino substituent is partially dissociated from the ruthenium
center (by protonation) in the reaction media, thereby generating a vacant coordination site for catalysis. These studies also show
that this substituent increases hydrogenation activity by promoting activation of the reductant. At least three catalytic cycles,
involving the decarboxylation of formic acid, hydration of the alkyne, and hydrogenation of the intermediate aldehyde, operate
concurrently in reactions mediated by 12. A wide array of terminal alkynes are efficiently processed to linear alcohols using as
little as 2 mol % of 12 at ambient temperature, and the complex 12 is stable for at least two weeks under air. The studies outlined
herein establish 12 as the most active and practical catalyst for anti-Markovnikov reductive hydration discovered to date, define
the structural parameters of 12 underlying its activity and stability, and delineate design strategies for synthesis of other
multifunctional catalysts.
INTRODUCTION
■
Linear alcohols are used throughout the commodity and fine
chemical industries and this large demand renders the
production of alcohols from hydrocarbon feedstocks an
important goal in catalysis.1 We have focused on developing
methods to form linear alcohols from alkynes by anti-
Markovnikov reductive hydration (Figure 1A). Such reactions
have the potential to replace stoichiometric pathways, such as
hydroboration−oxidation−reduction, which are mainstays of
synthetic chemistry. Related and contemporaneous efforts to
convert unsaturated hydrocarbons to linear alcohols have focused
primarily on terminal alkenes and comprise the dual-catalytic anti-
Markovnikov hydration of styrenes,2 photoredox approaches to
alkene hydroacyloxyation and hydroalkoxylation,3 hydroformyla-
tion−reduction,4 metal-catalyzed transfer hydrogenation−carbonyl
addition,5 and allylic oxidation−hydrogenation.6
In the reductive hydration reaction we have developed, the
alkyne is converted to an aldehyde by anti-Markovnikov
hydration, and the aldehyde is reduced in situ to provide the
alcohol (Figure 1A). We first reported a tandem two-catalyst
system7 that employed the highly efficient hydration catalyst
acetonitrile bis(2-diphenylphosphino-6-t-butylpyridine)
(η5-cyclopentadienyl)ruthenium hexafluorophosphate, devel-
oped by Grotjahn and co-workers,8 and Shvo’s catalyst9 to re-
duce the aldehyde intermediate (using 2-propanol as reductant).
By utilizing a cationic gold−N-heterocyclic carbene complex10
Figure 1. (A) Anti-Markovnikov reductive hydration. (B) Presumed
active catalyst 2 in the anti-Markovnikov reductive hydration mediated
by 1 and 2,2′-bipyridine (bipy).
in place of Grotjahn's catalyst, branched alcohols were also
accessible.7
Received: February 24, 2014
Published: April 4, 2014
© 2014 American Chemical Society
7058
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
We recently reported that a single catalyst can support both
the hydration and hydrogenation activities. A complex formed in
situ from (η5-cyclopentadienyl)(η6-naphthalene)ruthenium hex-
afluorophosphate (1) and 2,2′-bipyridine (bipy) transforms
terminal alkynes to linear alcohols in aqueous tetrahydrofuran
using formic acid as reductant (Figure 1B).11 In this system,
preheating a mixture of the precursor 1 and an excess of bipy was
necessary to generate the active catalyst. Preparative-scale experi-
ments and X-ray crystallographic analysis suggested generation
of structure 2, in which the half-sandwich ruthenium fragment
is ligated by bipy and a labile solvent molecule (S; acetonitrile in
the X-ray analysis). A time-course analysis revealed that the
hydration and hydrogenation activities of the catalyst were
temporally separated, an effect we attributed to the production of
stable catalyst−alkyne intermediates. As a consequence of its
coordination environment, the complex 2 is air-sensitive and is
difficult to isolate in pure form. Although this catalyst efficiently
promoted the hydration step, the hydrogenation activity was
modest, and elevated temperatures were required to achieve
useful rates of reduction.
We have continued mechanistic studies of this reaction, with
the goal of improving the catalyst activity, scope, and utility of the
transformation. Herein we report the full evolution of our
catalyst design studies, which have culminated in the identi-
fication of an air-stable, single-component catalyst that promotes
the anti-Markovnikov reductive hydration at ambient temper-
ature. This catalyst contains a hemilabile 3-(dimethylamino)-
propyl substituent that creates an 18-electron κ3-structure,
rendering it amenable to use on a benchtop. Mechanistic studies
show that the 3-(dimethylamino)propyl substituent partially
dissociates under the reaction conditions to facilitate catalysis
and that this functional group also increases hydrogenation
activity. These studies also reveal a complex competition
between at least three catalytic cycles involving the decarbox-
ylation of the formic acid, alkyne hydration, and aldehyde
hydrogenation.
Condensation of pyridine 2-carboxaldehyde with various amines
formed the ligands 4−8 in 84−94% yield (Table 1). These
ligands were conveniently associated with ruthenium by stirring
with (η5-cyclopentadienyl)tris(acetonitrile)ruthenium hexafluoro-
phosphate (3) in methylene chloride-d2 at 25 °C. 1H NMR ana-
lysis revealed nearly quantitative conversion to the iminopyridine
complexes for all ligands except 5.
By comparing the chemical shifts of the hydrogen atoms
closest to the terminal functional group (Ha and Hb in 4−8) in
the free ligand and the complex, and the detection of coordinated
acetonitrile, the mode of binding could be inferred.17 Thus,
the 3-(methoxy)propylimine ligand 4 coordinated to ruthenium
in the κ2-mode, providing the complex 9 (Δδ Ha, Hb, Hb = 0.04,
0.02, −0.02, respectively; δ CH3CN = 2.13), whereas the
alkynylimine ligand 6 formed the κ3-complex 10 (Δδ Ha, Hb,
Hb = 1.87, 0.85, 0.43, respectively). Observation of a nominal
change in Ha and Hb chemical shifts (Δδ Ha, Hb, Hb = 0.03, 0.22,
0.32, respectively) and the detection of bound acetonitrile
(δ 2.13) in the complex formed from the 2-(dimethylamino)-
ethyl imine ligand 7 suggested generation of the κ2-structure
11. On the other hand, the 3-(dimethylamino)propylimine 8
provided the κ3-complex 12, as evidenced by a large shift in
ligand resonances (Δδ Ha, Ha, Hb, Hb = −0.92, 0.98, 0.18,
and 1.06, respectively) and the absence of coordinated ace-
tonitrile (see Supporting Information for additional spectro-
scopic data).
The structure of the complex 12 was confirmed by X-ray
crystallography (Figure 3). The catalyst adopts a piano stool
geometry, which is typical of cationic, half-sandwich ruthenium
complexes.18 The dimethylamino−ruthenium (N-3−Ru) bond
length in 12 is 2.227(2) Å, whereas the pyridine−ruthenium
(N-1−Ru) and the imine−ruthenium bonds (N-2−Ru) are sub-
stantially shorter [2.083(3) and 2.048(2) Å, respectively]. The κ3
binding mode may be favored by adoption of a well-defined
chairlike geometry, as evidenced by the X-ray analysis and
analysis of vicinal 1H coupling constants.
The carbonyl complexes 13 and 14 were prepared to compare
the electronic nature of the iminopyridine and bipyridine ligands
(Table 2). The bipyridine complex 13 (1962 cm−1)18b possesses
a carbonyl stretch that is comparable to the iminopyridine com-
plex 14 (1972 cm−1), indicating that the two ligands possess
similar electronic properties.
To benchmark the efficiencies of these catalysts, the reductive
hydration of phenylacetylene (15a) was studied using 4.5 mol %
of 2, 10, or 12 and 4.0 equiv of formic acid in aqueous N-methyl-
2-pyrrolidinone (NMP) at 25 °C (Table 3).19 Under these
conditions, the catalyst 2 efficiently converted 15a to phenyl-
acetaldehyde (16a), but consistent with the modest hydro-
genation activity of the complex, none of the reduction product
2-phenethanol (17a) was detected (entry 1). We hypothesized
that the alkyne function of the κ3-complex 10 may undergo
conversion to an aldehyde or terminal alcohol, allowing catalysis,
but this complex did not display activity (entry 2). This may
reflect unfavorable geometries in the intramolecular func-
tionalization of the alkyne. The κ3-complex 12, however, was
CATALYST DESIGN, SYNTHESIS, AND ACTIVITY
■
In accord with literature reports,12 we attributed the air-
sensitivity of 2 to the presence of a labile monodentate ligand.
To address this, we targeted for synthesis a hemilabile13 complex
that could interconvert between a coordinatively saturated
κ3-isomer and a reactive unsaturated κ2-isomer (Figure 2). We
reasoned that the 18-electron κ3-state should be stable toward
oxidative decomposition pathways. Upon activation and isomer-
ization to κ2-binding, an open coordination site is generated,
allowing catalysis to proceed. On the basis of studies of related
half-sandwich ruthenium hydride complexes,14 it seemed likely
that a monohydride [e.g., CpRu(bipy)H] is the active reducing
agent in reactions catalyzed by 2. We posited that a basic
hemilabile functional group, such as an amine (X in Figure 2),
could promote formation of these monohydride intermediates
by mediating heterolytic cleavage of the reductant.15
Our studies focused on iminopyridine derivatives due to the
modular nature of their synthesis and their similarity to bipy.16
Figure 2. Catalyst design strategy.
7059
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
a
Table 1. Syntheses and Structures of Complexes Derived from [CpRu(CH3CN)3]PF6 (3) and the Iminopyridine Ligands 4−8
a
1
The structures of the complexes 9−12 were elucidated by H NMR spectroscopy of the unpurified product mixtures in methylene chloride-d2.
b
See Supporting Information for full spectroscopic data. n/d = not detected.
remarkably active, and 4.5 mol % of 12 afforded a 98% yield of
17a after 24 h at 25 °C (entry 3). The yield of product was
unchanged after storing the catalyst under air in a desiccator for
two weeks (entry 4).
Given the superior activity of the complex 12, a preparative-
scale synthesis was developed. Thus, stirring an equimolar
mixture of the ligand 8 and the commercially available, air-stable
ruthenium precursor 1 in acetonitrile at 60 °C formed 12 (eq 1).
The product was isolated in 95% yield and >95% analytical purity
(1H NMR analysis) by extraction of the acetonitrile layer with
reagent-grade hexanes (in a separatory funnel, open to air) and
concentration. The complex 12 is stable for at least 2 weeks
under air (vide supra) and can be stored in a vial under an inert
gas (argon or nitrogen) indefinitely. Complex 12 will soon be
available from Aldrich.
Figure 3. Molecular structure of the κ3-complex 12 (50% probability
level). Hydrogen atoms and counterion are omitted for clarity.
Table 2. IR Stretching Frequencies of the Ruthenium
Carbonyl Complexes 13 and 14
CATALYST ACTIVATION
■
Studies were conducted to probe the mechanism of the reductive
hydration mediated by 12 and to determine the origins of its
superior activity. A partially annotated 1H NMR spectrum of 12
in methylene chloride-d2 is shown in Figure 4A. Upon addition of
7060
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
a
Table 3. Comparative Analysis of the Reductive Hydration Activity of 2, 10, and 12
a
Reaction conditions: 15a (150 μmol), 10 or 12 (6.75 μmol), formic acid (600 μmol), NMP−H2O (4:1 v/v, [15a]0 = 0.3 M), 25 °C, 24 h.
b
c
Determined by 1H NMR spectroscopy using mesitylene as an internal standard. Catalyst prepared in situ from 3 (6.75 μmol) and bipy
d
e
(6.75 μmol). None was detected under conditions where 1% could be observed. Employing 12 after storage under air for two weeks in a
desiccator.
Figure 4. (A) 1H NMR spectra of 12. (B) 1H NMR spectra of 12 and [2-13C]phenylacetylene (5 equiv). (C) 1H NMR spectra of 12,
[2-13C]phenylacetylene (5 equiv), and trifluoroacetic acid (4 equiv). Spectra were recorded in methylene chloride-d2 at 25 °C.
[2-13C]phenylacetylene (5 equiv), the resonances corresponding
to 12 were essentially unchanged (Figure 4B). Addition of
trifluoroacetic acid (4 equiv) resulted in formation of a new
complex in 95% yield (Figure 4C). We identified this species as
the η2-alkyne complex 18. The alkyne C−H (Hd′) was observed
as a doublet centered at δ 5.46 (1JC−H = 246 Hz), downfield of
7061
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
free [2-13C]phenylacetylene (Hd, doublet, δ 3.12, JC−H = 251
Hz). The free and coordinated terminal acetylene carbon atoms
were observed at 137.2 and 122.9, respectively (13C NMR and
HMQC analysis). In addition, the iminopyridine chemical shifts
of 12 and 18 are distinct. The methylene protons adjacent to the
dimethylamino substituent (Hb) were observed at δ 2.50 and δ
3.38 in the κ3-complex 12, and coalesced into one peak centered
at δ 3.27 in 18 (Hb′). The methyl substituents of 12 (Ha) were
well-resolved at δ 1.27 and δ 3.17, and coalesced into two
partially overlapping doublets centered at δ 2.95 in the η2-alkyne
complex 18 (Ha′). A COSY correlation was observed between
the putative ammonium proton (He′, δ 9.45, br singlet) and
Ha′ and Hb′ in the complex 18. Addition of triethylamine
(4 equiv) led to regeneration of 12 (18%) along with unidenti-
fied decomposition products. These results demonstrate that
acidic conditions are needed to promote coordination of the
alkyne.
1
Figure 6. Reaction progress profile for the reductive hydration of
(2-fluorophenyl)acetylene (19). Conditions: [19]0 = 0.300 M, [12]0 =
0.0135 M, [H2O]0 = 11.1 M, [HCO2H]0 = 1.20 M, NMP−H2O (4:1 v/v),
23 °C.
To probe for this equilibrium in a setting relevant to catalysis, a
related study was conducted in a mixture of N,N-dimethylforma-
mide-d7−D2O and in the presence of formic acid (Figure 5A, B).
To obtain information about the hydration step, we monitored
the disappearance of (2-fluorophenyl)acetylene (19) and
analyzed these data by reaction progress kinetic analysis.20 The
instantaneous rate of alkyne consumption was obtained from
differentiation of plots of the fraction conversion of 19 versus
Figure 5. (A) 1H NMR spectrum of 12 in N, N-dimethylformamide-d7−
D2O (4:1, v/v) at 0 °C. (B) 1H NMR spectrum of 12, [2-13C]-
phenylacetylene (6.7 equiv), and formic acid (10.5 equiv), in N, N-
dimethylformamide-d7−D2O (4:1, v/v) at 0 °C, immediately after
mixing. See Figure 4 for resonance assignments.
Figure 7. Reaction progress profile for the reductive hydration of
(2-fluorophenyl)acetylene (19) monitored by 19F NMR spectroscopy.
Blue series: fraction conversion of 19 versus time. Red series:
instantaneous rate versus time. Conditions: [19]0 = 0.300 M, [12]0 =
0.0135 M, [H2O]0 = 11.1 M, [HCO2H]0 = 1.20 M, NMP−H2O (4:1 v/v),
23 °C.
An NMR spectrum acquired immediately after mixing at 0 °C
revealed formation of 18 in 20% yield, and the remainder of the
ruthenium-containing complexes were accounted for by 12
(Figure 5B). Consistent with a reversible equilibrium between 12
and 18, we observed a gradual decrease in the concentration of
12 as the reaction was warmed to ambient temperature and the
[2-13C]phenylacetylene was consumed.
time (Figure 7). Plots of the instantaneous rate versus time
(red series in Figure 7) indicate that the rate gradually increases
before decreasing in the manner expected for a reaction
exhibiting positive-order kinetics in substrate. The observed
induction period prevents analysis by the method of initial
rates.
To establish the order in catalyst, two reactions were carried out
using the same initial concentration of (2-fluorophenyl)acetylene
(19), water, and formic acid, but with varying initial concentrations
of 12 (13.5 and 20.4 mM). Plots of rate/[12]0 versus [19]
overlaid, establishing the hydration as first order in 12 (Figure 8).
Figure 9 shows that the relationship between reaction rate
and formic acid equivalents is complex. The rate of reaction at
[HCO2H]0 = 0.600 or 1.50 M (2.0 and 5.0 equiv with respect to
19, green and orange curves, respectively) is less than that at
MECHANISTIC STUDIES
■
The reductive hydration of (2-fluorophenyl)acetylene (19, eq 2)
was monitored by 19F NMR spectroscopy at 10 min intervals
(Figure 6). Temporal separation of the hydration and hydro-
genation activities11 was not observed, likely because the
formation of catalyst−substrate complexes is incomplete using
12 (Figure 5). This behavior complicates a kinetics analysis
because the rate of hydrogenation is coupled to the rate of
hydration. Accordingly, we investigated each step separately to
gain insights into the behavior of the catalyst.
7062
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
deactivation or product inhibition.20 To remove complications
arising from the induction period, we disregard the early data
points in the red series (indicated by the solid red points) from
the analysis. The concentrations of (2-fluorophenyl)acetylene
(19) and water in both runs are identical at the time point
indicated by the arrows in the figure. By shifting the first time
point of the red series to overlay with the blue series, the rates of
these two reactions can be compared. In a normal time-adjusted
plot, the experiment with a higher concentration of starting
material is typically as fast as or slower than that beginning with
a lower concentration of starting material, and a slower rate is an
indication of product inhibition or catalyst deactivation.20 As is
apparent from this analysis, the time-adjusted plots diverge and
the experiment starting from a higher concentration of 19 (blue
series) displays a faster rate than that beginning from a lower
concentration of 19 (red series). The faster rate for the reaction
starting from a higher concentration of 19 may arise from a
decrease in the ratio of formic acid to 19, which shifts the
distribution of catalyst toward the hydration cycle (see Figure 11
below). However, this increase in rate may mask smaller effects
arising from product inhibition or catalyst decomposition.
To gain insight into the reduction step,21 we studied the
hydrogenation of phenylacetaldehyde (16a) using DCO2H as
reductant (eq 3). The experiment was conducted in DMF to avoid
the potential for H/D exchange with NMP. The reduction
proceeded in nearly quantitative yield and with 10% deuterium
atom incorporation at the carbinol center of 17a, as determined by
1H and 2H NMR analysis. This result suggests loss of the identity of
the formyl hydrogen (deuterium) and oxygen hydrogen atoms by
decarboxylation of formic acid22 and formation of HD. Reduction
may occur via a ruthenium hydride, and the selective incorporation of
hydrogen into the aldehyde may reflect primary and equilibrium
kinetic isotope effects for formation of the ruthenium hydride and
addition to the carbonyl.14d Repeating the experiment in NMP
allowed for quantification of the formic acid isotopologs at the end of
the experiment. In this experiment, we obtained a 200% yield of
DCO2(H/D) (theoretical: 300%) and a 27% yield of HCO2(H/D).
The most reasonable explanation for production of HCO2(H/D)
involves exchange of the deuterium and proton positions by
reduction of carbon dioxide with HD (e.g., reversal of the
decarboxylation pathway).
Figure 8. A. Plot of rate/[12]0 versus [19] reveals a first order
dependence on 12. Conditions: [19]0 = 0.300 M, [12]0 = 0.0135 or
0.0204 M, [H2O]0 = 11.1M, [HCO2H]0 = 1.20M, NMP−H2O (4:1 v/v),
23 °C.
Figure 9. Plot [19] versus time at varying initial concentrations of
formic acid. Conditions: [19]0 = 0.300 M, [H2O]0 = 11.1 M, [12]0 =
0.0135 M, NMP−H2O (4:1 v/v), 23 °C.
[HCO2H]0 = 1.20 M (4.0 equiv with respect to 19, blue curve),
and the reaction beginning with [HCO2H]0 = 0.6 M does not
proceed to completion. The slower rate at a higher concentration
of formic acid is attributed to inhibition of the hydration step by
partitioning of the catalyst toward a nonproductive, decarbox-
ylation pathway (see below). At lower concentrations of formic
acid, the slower rate may arise from gradual neutralization of the
reaction media by this decarboxylation pathway and deactivation
of the catalyst by formation of the κ3-isomer.
Figure 10 shows [19] versus time in a pair of “same excess”
experiments, which are frequently used to probe for catalyst
This latter result suggested that a ruthenium hydride may be
accessible from 12 and dihydrogen. To test this, we monitored
the reduction of phenylacetaldehyde (16a) using dihydrogen
(introduced by bubbling dihydrogen through the reaction
solution for 1 min), catalyst 12, and trifluoroacetic acid (TFA,
1 equiv with respect to ruthenium) as activator (eq 4). Under
these conditions, 2-phenethanol (17a) was formed in 30% yield,
with the remainder of material in this experiment accounted for
as unreacted 16a. The difference in yields of 17a using
dihydrogen and formic acid may arise from variations in the
acidity of the reaction media and the low pressure of dihydrogen
Figure 10. Plot of [19] versus time in a pair of “same-excess” experiments;
the experiments have different values of [19]0 and [H2O]0, with the dif-
ference between the two (defined as the “excess” = [19]0 − [H2O]0) held
the same. Common conditions: [12]0 = 0.0135 M, [HCO2H]0 = 1.20 M,
NMP−H2O (4:1 v/v), 23 °C. Solid red points denote the induction period.
7063
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
a
Table 4. Reduction of Phenylacetaldehyde (16a) Using Complexes Formed In Situ from Ligand 7, 8, or 22 and the Precursor 3
a
Reaction conditions: 16a (150 μmol), 3 (6.75 μmol), 7, 8, or 22 (6.75 μmol), formic acid (600 μmol), NMP−H2O (4:1 v/v, 500 μL, [16a]0 =
b
c
1
0.300 M), 25 °C, 24 h. Determined by H NMR spectroscopy using mesitylene as an internal standard. Yield of formic acid based on starting
material.
Figure 11. Proposed catalytic cycles that are operative during reductive hydration reactions mediated by 12.
7064
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
a
Table 5. Substrate Scope for Reductive Hydration Employing Complex 12
a
b
Reactions were performed on a 500 μmol scale at 25 °C in 4:1 (v/v) NMP−H2O, for 48 h; [alkyne]0 = 300 mM. Isolated yield after purification
c
d
e
f
by flash-column chromatography. Employing 1.28 g (6.00 mmol) of the alkyne 15c. 60 h. 2.5 mol % of trifluoroacetic acid was added. 5 mol %
of trifluoroacetic acid was added. [15x]0 = 204 mM, 8 mol % of trifluoroacetic acid was employed.
g
7065
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
used in the experiment in eq 4.23 In the absence of TFA, the
reduction did not occur, in accord with our NMR studies above.
hydride complexes from dihydrogen,25 and the lower (9%) yield of
reduction product employing the alkylimine ligand 22 (Table 4) is
consistent with participation of the amine in the hydrogenation
step. We postulate that reduction occurs via outer sphere delivery
of hydride from 24, potentially assisted by hydrogen bonding of
the carbonyl oxygen to the ammonium ion.15
SUBSTRATE SCOPE
■
To test for a rate enhancement in the activation of formic acid
arising from the dimethylamino substituent of 12, we evaluated
the reduction of phenylacetaldehyde (16a) by catalysts formed
in situ from the acetonitrile precursor 3 and the ligands 7, 8,
and 22 (Table 4). In these experiments, four equivalents of
formic acid were employed. The complexes derived from the
2-(dimethylamino)ethyl ligand 7 and the 3-(dimethylamino)-
propyl ligand 8 efficiently reduced the starting material 16a
(entries 1 and 2). The complex derived from the N-butylimino
ligand 22, however, was less effective for the reduction (9% yield of
17a, entry 3), even though it cannot access an unreactive κ3-form.
In addition, we observed that nearly all of the formic acid was
depleted in the reaction employing ligand 7 (entry 1), whereas
only 75% of the formic acid employed was depleted in reac-
tions employing 8 (entry 2). These experiments indicate that
the decarboxylation of formic acid occurs at a rate that exceeds
hydrogenation of 16a.
The scope and limitations of the catalyst 12 are shown in
Table 5. For all of the substrates examined, the reductive hydra-
tion proceeded at ambient temperature, and for many substrates
(entries 1−11), 2 mol % of 12 was sufficient to achieve complete
conversion. Both aromatic and aliphatic alkynes (15a−k) gave
high isolated yields at 2 mol % of catalyst loading. The catalyst
and reaction conditions are compatible with many functional
groups including phthalimides (15c), primary alkyl chlorides
(15d), esters (15e), and carboxylic acids (15f). Substrates
containing free alcohols (15m), di- or trisubstituted alkenes
(15n, 15r, 15s), amides (15q, 15s), and tertiary amines (15t)
required 4.5 mol % of 12 to achieve full conversion. Of particular
note, propargylic alcohols such as 15v and 15x are converted to
product at ambient temperature (69 and 49% yield for 15v and
15x, respectively). Substrates such as 15v do not provide high
yields of product with earlier reductive hydration catalysts.11
A
reaction conducted on a 6 mmol scale proceeded in comparable
yield to 500 μmol scale (89% and 94% for 6 mmol and 500 μmol
scale, respectively; entry 3). A limitation of this catalyst can
be observed with sterically-encumbered substrates such as 15l.
These substrates require higher catalyst loading (9 mol %) and
additional trifluoroacetic acid to obtain high conversions.
Presumably, the rates of hydration and/or hydrogenation are
decreased due to unfavorable nonbonded interactions with the
catalyst.
MECHANISTIC MODEL
■
Based on the experiments presented above, we propose that
three catalytic cycles are operating concurrently in reductive
hydration reactions mediated by 12. These include a hydration
cycle, a decarboxylation cycle that interconverts formic acid with
dihydrogen and carbon dioxide, and a hydrogenation cycle that
reduces the aldehyde to an alcohol (Figure 11).
In the absence of acid, the catalyst exists in the κ3-form. Under
the acidic conditions of the reductive hydration reaction,
protonation, solvolysis, and substrate association occur, to lead
to the η2-alkyne complex 26. The complex 26 lies outside of the
hydrogenation and decarboxylation cycles, and the observed
induction period in the hydration step (Figure 7) can be
explained by the partitioning of 12 between these cycles and the
hydration cycle. At the initial stages of the reaction, a larger
proportion of the catalyst is contained within the decarboxylation
cycle, which leads to a slower initial rate of alkyne consumption.
As the concentration of formic acid decreases, the distribution of
catalyst shifts toward the hydration cycle, increasing the rate of
consumption of alkyne. Significant decreases in the formic acid
concentration induce catalyst deactivation by neutralization and
formation of the κ3-isomer, leading to incomplete conversion of
substrate (see Figure 9). The hydration step may proceed by
addition of water to a metal vinylidene intermediate (derived from
the η2-alkyne complex 26), as has been proposed for ruthenium-
based anti-Markovnikov alkyne hydration catalysts.8c−e,24
Our studies of the hydrogenation step conclusively establish
that the pendant amine accelerates the rate of reduction. The
amine may serve as a proton shuttle, facilitating the forma-
tion of the ruthenium formate 23, which may decarboxylate
to generate the ruthenium hydride 24.14d Elimination of
dihydrogen would regenerate 12. The observed deuterium
scrambling in reactions employing DCO2H and the successful
reduction of 16a with dihydrogen indicate this pathway is
reversible. Prior studies have established the activating role
of pendant amines in the generation of ruthenium and iridium
CONCLUSION
■
In summary, we have reported the design, synthesis, and
mechanistic studies of a highly active catalyst for the anti-
Markonvikov reductive hydration of alkynes. The catalyst builds
on our preliminary studies of the bipy complex 2. The catalyst 12
is easily prepared on a multigram scale, is stable for at least two
weeks under air (and can be stored indefinitely under argon), and
can be employed without recourse to a glovebox. Mechanistic
studies suggest that the stability of the catalyst arises from
coordination of a hemilabile amine to the ruthenium center.
These studies also showed that the pendant amine increases the
hydrogenation activity of the catalyst. These investigations
revealed a complex interplay between hydration, hydrogenation,
and decarboxylation cycles. The catalyst displays the highest
activity of any reductive hydration system reported to date,
allowing reactions to be conducted at ambient temperature for
the first time. The mechanistic studies outlined herein will inform
the development of other anti-Markovnikov reductive function-
alization reactions.
ASSOCIATED CONTENT
■
S
* Supporting Information
Detailed experimental procedures and characterization data for
all new compounds. This material is available free of charge via
the Internet at http://pubs.acs.org.
7066
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067

Journal of the American Chemical Society
Article
J. Organomet. Chem. 1996, 523, 187. (f) Philippopoulos, A. I.;
Hadjiliadis, N.; Hart, C. E.; Donnadieu, B.; Mc Gowan, P. C.;
Poilblanc, R. Inorg. Chem. 1997, 36, 1842.
(18) (a) Albers, M. O.; Liles, D. C.; Robinson, D. J.; Singleton, E.
Organometallics 1987, 6, 2179. (b) Balavoine, G. G. A.; Boyer, T.; Livage,
C. Organometallics 1992, 11, 456.
(19) We have found that polar aprotic solvents such as N-methyl-2-
pyrrolidinone or N,N-dimethylformamide are superior to tetrahydrofur-
an. Le Li and Seth B. Herzon, unpublished results.
(20) (a) Blackmond, D. G. Angew. Chem., Int. Ed. 2005, 44, 4302.
(b) Mathew, J. S.; Klussmann, M.; Iwamura, H.; Valera, F.; Futran, A.;
Emanuelsson, E. A. C.; Blackmond, D. G. J. Org. Chem. 2006, 71, 4711.
(21) For a review of metal-catalyzed transfer hydrogenation, see
AUTHOR INFORMATION
Corresponding Author
seth.herzon@yale.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support from the David and Lucile Packard Foundation
is gratefully acknowledged.
REFERENCES
■
Samec, J. S. M.; Backvall, J.-E.; Andersson, P. G.; Brandt, P. Chem. Soc.
̈
(1) Arpe, H.-J.; Weissermel, K. Industrial Organic Chemistry; Wiley-
VCH: Weinheim, 2010.
Rev. 2006, 35, 237.
(22) For selected examples of transition metal-mediated decarbox-
ylation of formic acid, see (a) Gao, Y.; Kuncheria, J. K.; Jenkins, H. A.;
Puddephatt, R. J.; Yap, G. P. A. J. Chem. Soc., Dalton Trans. 2000, 3212.
(b) Junge, H.; Boddien, A.; Capitta, F.; Loges, B.; Noyes, J. R.; Gladiali,
S.; Beller, M. Tetrahedron Lett. 2009, 50, 1603. (c) Fukuzumi, S.;
Kobayashi, T.; Suenobu, T. J. Am. Chem. Soc. 2010, 132, 1496.
(d) Majewski, A.; Morris, D. J.; Kendall, K.; Wills, M. ChemSusChem
2010, 3, 431. (e) Boddien, A.; Gartner, F.; Federsel, C.; Sponholz, P.;
Mellmann, D.; Jackstell, R.; Junge, H.; Beller, M. Angew. Chem., Int. Ed.
2011, 50, 6411. (f) Boddien, A.; Mellmann, D.; Gartner, F.; Jackstell, R.;
Junge, H.; Dyson, P. J.; Laurenczy, G.; Ludwig, R.; Beller, M. Science
2011, 333, 1733. (g) Himeda, Y.; Miyazawa, S.; Hirose, T.
ChemSusChem 2011, 4, 487. (h) Papp, G.; Csorba, J.; Laurenczy, G.;
Joo, F. Angew. Chem., Int. Ed. 2011, 50, 10433. (i) Hull, J. F.; Himeda, Y.;
Wang, W.-H.; Hashiguchi, B.; Periana, R.; Szalda, D. J.; Muckerman, J.
T.; Fujita, E. Nat. Chem. 2012, 4, 383. (j) Barnard, J. H.; Wang, C.; Berry,
N. G.; Xiao, J. Chem. Sci. 2013, 4, 1234.
(23) Effervescence is observed upon venting reactions employing
formic acid as reductant, suggesting the pressure of the vessel is higher
than that of equation 4 (1 atm). This increase in pressure may increase
the hydrogenation activity of the catalyst.
(24) Tokunaga, M.; Suzuki, T.; Koga, N.; Fukushima, T.; Horiuchi, A.;
Wakatsuki, Y. J. Am. Chem. Soc. 2001, 123, 11917.
(25) (a) Chu, H. S.; Lau, C. P.; Wong, K. Y.; Wong, W. T.
Organometallics 1998, 17, 2768. (b) Lee, D.-H.; P. Patel, B.; H. Crabtree,
R.; Clot, E.; Eisenstein, O. Chem. Commun. 1999, 297. (c) Gruet, K.;
Crabtree, R. H.; Lee, D.-H.; Liable-Sands, L.; Rheingold, A. L.
Organometallics 2000, 19, 2228. (d) Gruet, K.; Clot, E.; Eisenstein,
O.; Lee, D. H.; Patel, B.; Macchioni, A.; Crabtree, R. H. New J. Chem.
2003, 27, 80. (e) DiMondo, D.; Thibault, M. E.; Britten, J.; Schlaf, M.
Organometallics 2013, 32, 6541.
(2) Dong, G. B.; Teo, P. L.; Wickens, Z. K.; Grubbs, R. H. Science 2011,
333, 1609.
(3) (a) Hamilton, D. S.; Nicewicz, D. A. J. Am. Chem. Soc. 2012, 134,
18577. (b) Perkowski, A. J.; Nicewicz, D. A. J. Am. Chem. Soc. 2013, 135,
10334.
(4) (a) Chevallier, F.; Breit, B. Angew. Chem., Int. Ed. 2006, 45, 1599.
(b) Takahashi, K.; Yamashita, M.; Ichihara, T.; Nakano, K.; Nozaki, K.
Angew. Chem., Int. Ed. 2010, 49, 4488.
(5) Smejkal, T.; Han, H.; Breit, B.; Krische, M. J. J. Am. Chem. Soc.
2009, 131, 10366.
(6) Campbell, A. N.; White, P. B.; Guzei, I. A.; Stahl, S. S. J. Am. Chem.
Soc. 2010, 132, 15116.
(7) Li, L.; Herzon, S. B. J. Am. Chem. Soc. 2012, 134, 17376.
(8) (a) Grotjahn, D. B.; Incarvito, C. D.; Rheingold, A. L. Angew.
Chem., Int. Ed. 2001, 40, 3884. (b) Grotjahn, D. B.; Lev, D. A. J. Am.
Chem. Soc. 2004, 126, 12232. (c) Grotjahn, D. B.; Miranda-Soto, V.;
Kragulj, E. J.; Lev, D. A.; Erdogan, G.; Zeng, X.; Cooksy, A. L. J. Am.
Chem. Soc. 2007, 130, 20. (d) Grotjahn, D. B.; Kragulj, E. J.; Zeinalipour-
Yazdi, C. D.; Miranda-Soto, V.; Lev, D. A.; Cooksy, A. L. J. Am. Chem.
Soc. 2008, 130, 10860. For a review, see (e) Grotjahn, D. B. Pure Appl.
Chem. 2010, 82, 635.
(9) (a) Blum, Y.; Shvo, Y. Isr. J. Chem. 1984, 24, 144. (b) Shvo, Y.;
Czarkie, D.; Rahamim, Y.; Chodosh, D. F. J. Am. Chem. Soc. 1986, 108,
7400. For a review, see (c) Warner, M. C.; Casey, C. P.; Backvall, J.-E.
̈
Top. Organomet. Chem. 2011, 37, 85.
(10) Marion, N.; Ramon
131, 448.
(11) Li, L.; Herzon, S. B. Nat. Chem. 2014, 6, 22.
(12) (a) Palacios, M. D.; Puerta, M. C.; Valerga, P.; Lledos
́
, R. S.; Nolan, S. P. J. Am. Chem. Soc. 2008,
́
, A.; Veilly, E.
Inorg. Chem. 2007, 46, 6958. (b) Kim, K.; Kishima, T.; Matsumoto, T.;
Nakai, H.; Ogo, S. Organometallics 2012, 32, 79.
(13) For a review of hemilabile ligands in catalysis, see Braunstein, P.;
Naud, F. Angew. Chem., Int. Ed. 2001, 40, 680.
(14) For selected examples, see (a) Jia, G.; Morris, R. H. J. Am. Chem.
Soc. 1991, 113, 875. (b) Jia, G.; Lough, A. J.; Morris, R. H.
Organometallics 1992, 11, 161. (c) Magee, M. P.; Norton, J. R. J. Am.
Chem. Soc. 2001, 123, 1778. (d) Guan, H.; Iimura, M.; Magee, M. P.;
Norton, J. R.; Zhu, G. J. Am. Chem. Soc. 2005, 127, 7805. (e) Guan, H.;
Saddoughi, S. A.; Shaw, A. P.; Norton, J. R. Organometallics 2005, 24,
6358. (f) O, W. W. N.; Lough, A. J.; Morris, R. H. Organometallics 2012,
31, 2137. (g) Silantyev, G. A.; Filippov, O. A.; Tolstoy, P. M.; Belkova,
N. V.; Epstein, L. M.; Weisz, K.; Shubina, E. S. Inorg. Chem. 2013, 52,
1787.
(15) For reviews of functional ligands in catalysis, see (a) Morris, R. H.
Chem. Soc. Rev. 2009, 38, 2282. (b) Crabtree, R. H. New J. Chem. 2011,
35, 18. (c) Eisenstein, O.; Crabtree, R. H. New J. Chem. 2013, 37, 21.
(16) For a review of iminopyridine ligands, see Chelucci, G. Coord.
Chem. Rev. 2013, 257, 1887.
(17) For examples of related approaches to ligand binding analysis, see
(a) Tein-Fu, W.; Tsung-Yi, L.; Jim-Wen, C.; Chi-Wi, O. J. Organomet.
Chem. 1992, 423, 31. (b) Wang, T.-F.; Wen, Y.-S. J. Organomet. Chem.
1992, 439, 155. (c) Flores, J. C.; Chien, J. C. W.; Rausch, M. D.
Organometallics 1994, 13, 4140. (d) van der Zeijden, A. A. H. J.
Organomet. Chem. 1996, 518, 147. (e) Wang, T.-F.; Lai, C.-Y.; Wen, Y.-S.
7067
dx.doi.org/10.1021/ja501738a | J. Am. Chem. Soc. 2014, 136, 7058−7067