Magnesium-catalyzed mild reduction of tertiary and secondary amides to amines

Article abstract of DOI:10.1021/acscatal.5b01038

The first example of a catalytic hydroboration of amides for their deoxygenation to amines is reported. This transformation employs an earth-abundant magnesium-based catalyst. Tertiary and secondary amides are reduced to amines at room temperature in the presence of pinacolborane (HBpin) and catalytic amounts of To^MMgMe (To^M = tris(4,4-dimethyl-2-oxazolinyl)phenylborate). Catalyst initiation and speciation is complex in this system, as revealed by the effects of concentration and order of addition of the substrate and HBpin in the catalytic experiments. To^MMgH₂Bpin, formed from To^MMgMe and HBpin, is ruled out as a possible catalytically relevant species by its reaction with N,N-dimethylbenzamide, which gives Me₂NBpin and PhBpin through C-N and C-C bond cleavage pathways, respectively. In that reaction, the catalytic product benzyldimethylamine is formed in only low yield. Alternatively, the reaction of To^MMgMe and N,N-dimethylbenzamide slowly gives decomposition of To^MMgMe over 24 h, and this interaction is also ruled out as a catalytically relevant step. Together, these data suggest that catalytic activation of To^MMgMe requires both HBpin and amide, and To^MMgH₂Bpin is not a catalytic intermediate. With information on catalyst activation in hand, tertiary amides are selectively reduced to amines in good yield when catalytic amounts of To^MMgMe are added to a mixture of amide and excess HBpin. In addition, secondary amides are reduced in the presence of 10 mol % To^MMgMe and 4 equiv of HBpin. Functional groups such as cyano, nitro, and azo remain intact under the mild reaction conditions. In addition, kinetic experiments and competition experiments indicate that B-H addition to amide C-O is fast, even faster than addition to ester C=O, and requires participation of the catalyst, whereas the turnover-limiting step of the catalyst is deoxygenation.

Full text of DOI:10.1021/acscatal.5b01038

Research Article
pubs.acs.org/acscatalysis
Magnesium-Catalyzed Mild Reduction of Tertiary and Secondary
Amides to Amines
Nicole L. Lampland, Megan Hovey,† Debabrata Mukherjee,‡ and Aaron D. Sadow*
Department of Chemistry, 1605 Gilman Hall, Iowa State University, Ames, Iowa 50011, United States
S
* Supporting Information
ABSTRACT: The first example of a catalytic hydroboration
of amides for their deoxygenation to amines is reported. This
transformation employs an earth-abundant magnesium-based
catalyst. Tertiary and secondary amides are reduced to amines
at room temperature in the presence of pinacolborane
(HBpin) and catalytic amounts of ToMMgMe (ToM
=
tris(4,4-dimethyl-2-oxazolinyl)phenylborate). Catalyst initia-
tion and speciation is complex in this system, as revealed by
the effects of concentration and order of addition of the
substrate and HBpin in the catalytic experiments.
ToMMgH2Bpin, formed from ToMMgMe and HBpin, is ruled out as a possible catalytically relevant species by its reaction
with N,N-dimethylbenzamide, which gives Me2NBpin and PhBpin through C−N and C−C bond cleavage pathways, respectively.
In that reaction, the catalytic product benzyldimethylamine is formed in only low yield. Alternatively, the reaction of ToMMgMe
and N,N-dimethylbenzamide slowly gives decomposition of ToMMgMe over 24 h, and this interaction is also ruled out as a
catalytically relevant step. Together, these data suggest that catalytic activation of ToMMgMe requires both HBpin and amide,
and ToMMgH2Bpin is not a catalytic intermediate. With information on catalyst activation in hand, tertiary amides are selectively
reduced to amines in good yield when catalytic amounts of ToMMgMe are added to a mixture of amide and excess HBpin. In
addition, secondary amides are reduced in the presence of 10 mol % ToMMgMe and 4 equiv of HBpin. Functional groups such as
cyano, nitro, and azo remain intact under the mild reaction conditions. In addition, kinetic experiments and competition
experiments indicate that B−H addition to amide CO is fast, even faster than addition to ester CO, and requires
participation of the catalyst, whereas the turnover-limiting step of the catalyst is deoxygenation.
KEYWORDS: hydroboration, amide reduction, magnesium, deoxygenation, catalysis
Lewis acid-coordinated CO intermediate are similarly
INTRODUCTION
■
invoked for ester and amide reductions.5 Although this idea is
accepted for stoichiometric reductions, minimal experimental
evidence is available for amide reductions.3 Well-defined main
group reductants, either as stoichiometric reagents or as part of
the catalytic systems, may also contribute experimental support
for elementary steps in LiAlH4 or NaBH4 reductions.
The demand for efficient syntheses of amines is ever increasing
because of the need to produce chemicals through sustainable
processes and amines’ continued importance in pharmaceutical,
agrochemical, and materials chemistry applications.1,2 Amides,
which are naturally prevalent among biological molecules or are
readily synthetically accessed, provide attractive starting
materials for amine preparations through reductive trans-
formations.3−5 However, selective reduction of the amide
functional group is challenging for thermodynamic and kinetic
reasons and often requires strongly reducing metal hydride
reagents, such as LiAlH4, NaBH4, or B2H6, that also react with a
number of functional groups. For example, nitrile and nitro
groups are readily reduced by LiAlH4, nitriles are reduced by
B2H6, and olefins are readily hydroborated by B2H6 or BH3·
THF.6 Amide reductions that avoid LiAlH4 and BH3 were
identified as key challenging conversions by the ACS Green
Chemistry Institute Pharmaceutical Roundtable,2 and this need
continues, even with remarkable progress in the last five years.
Notably, these stoichiometric reagents contain both reducing
hydrides and Lewis acid sites (Li, Na, B, or Al), presumably to
activate amides (as well as esters and other carbonyls) for
reduction. Thus, pathways involving hydride attack upon a
A number of pathways have been reported for reductive
interactions of metal compounds and amides, including
deoxygenation to amines,7−12 deoxygenation and alkyla-
tion,13−15 dehydration to nitriles,16 and C−N bond cleavage
to amines and aldehydes.17,18 Although catalytic hydro-
genations are appealing on the basis of their atom economy,
most amide deoxygenations via hydrogenation require forcing
conditions, including high pressure and elevated temperature
(>150 °C). In addition, hydrodeoxygenation of primary and
secondary amines often yields mixtures resulting from alkyl
group disproportionation.3,13,17,18 Although a range of late
metal complexes have been reported to efficiently reduce
amides through hydrosilylation,8,9,19 new catalytic processes are
Received: May 18, 2015
© XXXX American Chemical Society
4219
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226

ACS Catalysis
Research Article
Scheme 1. Chemical Routes for Amide Reduction
RESULTS AND DISCUSSION
■
N,N-Dimethylbenzamide reacts with 2 equiv of pinacolborane
(HBpin) upon addition of ToMMgMe (10 mol %) to form
benzyldimethylamine in 54% yield (eq 1). Control experiments
reveal that N,N-dimethylbenzamide and 2 equiv of pinacolbor-
ane are unchanged after 24 h at temperatures up to 120 °C in
the absence of ToMMgMe.
still needed to address the challenges facing amide deoxyge-
nations (Scheme 1). Few catalytic systems are able to effectively
reduce primary, secondary, and tertiary amides,11 and few
catalysts are based on metal complexes other than non-
oxophilic late metals.20,21 Moreover, many catalysts require
elevated temperatures for effective operation, and disproportio-
nation of primary and secondary amides into a mixture of
tertiary, secondary, and primary amines also hinders the
application of catalytic hydrosilylation for deoxygenation. An
exception is the [Ir(η2-C8H14)2Cl]2/Et2SiH2 system, which
reduces secondary amides at room temperature (but also
catalyzes silane redistribution).8 This catalyst, as well as many
of the first-row transition-metal hydrosilylation catalysts such as
Zn(OAc)2,22 are appealing for their simplicity, but the tuning
effects of ancillary ligands may be difficult to introduce for
conversions in which advanced activity and selectivity is
needed. Other mild systems are applicable only for tertiary
amides and employ main group metal catalysts.21−23
In contrast, the reduction of amides by catalytic hydro-
boration with HBpin is not reported.24 Magnesium compounds
have been shown to be good catalysts for the hydroboration of
a number of carbonyl compounds and unsaturated substrates,
such as pyridine.25−28 Our group has recently found that
ToMMgMe (ToM = tris(4,4-dimethyl-2-oxazolinyl)-
phenylborate) catalytically reduces and cleaves esters with 2
equiv of pinacolborane (HBpin) to give alkoxyboronic pinacol
esters (ROBpin).29 Amides are slightly stabilized relative to
esters, which can be shown by the ΔHrxn for the metathesis
reaction of methyl acetate and dimethylamine to N,N-
dimethylacetamide and methanol, which is −7.5 kcal/mol.30
On this basis, the reduction of amides might be predicted to be
slower than the reduction of esters.
Thus, the feasibility and reaction pathway(s) of a
magnesium-catalyzed amide reduction is of interest in relation
to the ester reduction. In addition, the classic studies of Brown
on hydroboration and deoxygenation of amides with B2H6
reveal good selectivity, even with this highly reactive
reagent,31,32 which nonetheless is limited by its reactivity
toward olefins and alkynes. Herein, we report the catalytic
reduction of tertiary and secondary amides to amines using
pinacolborane and the precatalyst ToMMgMe under mild, room
temperature conditions. Pinacolborane does not react at room
temperature with secondary and tertiary amides, which allows
for functional group tolerance and potential selectivities not
known with B2H6 or LiAlH4. The isolated magnesium
dihydridopinacolborate adduct ToMMgH2Bpin, which is a
precatalyst in the aforementioned ester cleavage, is not effective
as a catalytic species in the reduction of amides. To the best of
our knowledge, this report describes the first example of the
catalytic hydroboration of amides.
The ToMMgMe-catalyzed amide deoxygenation reaction of
eq 1 contrasts the hydroboration/reductive ester cleavage
catalyzed by ToMMgMe, instead following the typically
observed conversion of amides to amines in the presence of
strong metal hydrides, such as LiAlH4 or NaBH4.6 The
byproduct of the catalytic hydroboration/deoxygenation is
pinBOBpin, which is characterized by 1H and 11B NMR signals
at 1.02 and 21.7 ppm, respectively (in benzene-d6).33
These catalytic experiments were initially performed by
dissolving N,N-dimethylbenzamide and HBpin in benzene,
followed by addition of ToMMgMe. In contrast, benzyldime-
thylamine is not efficiently produced (18% yield) in experi-
ments in which N,N-dimethylbenzamide is added to a solution
of ToMMgMe and HBpin. Instead, the magnesium compound
and HBpin react instantaneously to give MeBpin and a black,
intractable and catalytically inactive precipitate. Although this
decomposition may be avoided (in the absence of amide) by
adding ToMMgMe to excess HBpin (>10 equiv) to form
ToMMgH2Bpin, experiments in which N,N-dimethylbenzamide
is added to a 20:1 mixture of HBpin/ToMMgMe do not afford
greater than ∼20% benzyldimethylamine. In those experiments,
the characteristic black precipitate is still observed. Further-
more, isolated ToMMgH2Bpin is less efficient as a catalyst
precursor for amide reduction than ToMMgMe (Table 1).
Finally, the deoxygenation yield is poorer and C−N cleavage
products are higher in reactions performed in methylene
chloride (see below and Scheme 2 for discussion of the
observed pathways in this catalytic system). That solvent is
effective for the formation of ToMMgH2Bpin, and it is also
effective as a solvent for the ToMMgMe-catalyzed cleavage of
esters. Thus, these empirical observations for amide deoxyge-
nation contrast those of the ToMMg-catalyzed ester hydro-
boration,29 which is proposed to involve ToMMg{RO(H)Bpin}
as a catalytic intermediate formed from ToMMgH2Bpin and
esters.
Other oxazoline-based magnesium compounds also show
catalytic activity for hydroboration and deoxygenation of
amides (Table 1), and the catalytic efficacy varies with substrate
and ancillary oxazolinylborate-based ligand. The chiral C3-
symmetric ToPMgMe (ToP = tris((4S)-isopropyl-2-oxazolinyl)-
phenylborate) gives slightly greater conversion to benzyldime-
thylamine (67% NMR yield) than ToMMgMe after 12 h, and it
gives 97% yield after 24 h. This product is obtained with 54%
yield after 12 h when ToMMgMe is the catalyst, and the same
yield is measured after 24 h, indicating that the product does
not decompose under the reaction conditions, and the ToPMg-
derived catalyst is longer-lived than the ToMMg-derived species.
4220
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226

ACS Catalysis
Research Article
Table 1. Effects of Ancillary Ligand and Reaction Conditions on Catalytic Deoxygenation of N,N-Dimethylbenzamide with
Pinacol Borane
With ToMMgMe, yields are improved with even 4 equiv of
HBpin, whereas amine yield is decreased with a greater
concentration of HBpin when ToPMgMe is the catalyst. The
less efficient C1-symmetric ToMPMgMe (ToMP = bis(4,4-
dimethyl-2-oxazolinyl)((4S)-isopropyl-2-oxazolinyl)-
phenylborate) requires >24 h for <50% yield, and amine yields
are not improved with longer reaction times. Similarly, a
bis(oxazolinyl)boratomagnesium methyl gives an even lower
yield with both short and long reaction times, implying that the
catalyst quickly deactivates.
Thus, the interaction of ToMMgH2Bpin and amides could
provide insight into pathways available during catalytic amide
reduction to guide further optimizations. ToMMgH2Bpin and 1
equiv of N,N-dimethylbenzamide give complete consumption
of the amide and formation of benzyldimethylamine in 11%
yield in a process that affords a mixture of species. In Scheme 2,
pathways associated with C−O, C−N, and C−C cleavage are
identified on the basis of the assigned products. Two sets of
new ToM-containing signals were observed. The major species
was assigned to ToMMgOCH2Ph on the basis of a comparison
with an authentic sample generated from ToMMgMe and
HOCH2Ph. In addition, a signal at 2.15 ppm was assigned to
HNMe2 on the basis of an identical chemical shift of an
authentic dilute sample of dimethylamine in benzene-d6.
The reactant ToMMgH2Bpin, as well as ToMMgNMe2, are
ruled out as the other ToM-containing species on the basis of a
comparison with authentic samples. PhCH2OBpin and
Me2NBpin are formed; PhCH2OBpin was assigned on the
Although ToPMgMe gives the highest yield in Table 1, most
other substrates below in Table 2 are reduced in equal or higher
yield with ToMMgMe than with the C3-symmetric ToPMgMe
complex. In fact, for most tertiary and secondary amides,
ToMMgMe is a superior precatalyst in terms of reaction times,
NMR yields, and selectivity. Moreover, 10, 5, or even 2 mol %
ToMMgMe as the catalyst provides benzyldimethylamine in
>95% yield under optimized conditions with excess HBpin.
There appears to be a trade-off between ToMMgMe and HBpin
loadings: high catalyst loading and low HBpin concentration
give >90% yield of the amine, and low catalyst loading requires
a larger excess of HBpin. This is undoubtably a kinetic effect
and suggests that the intermediate formed from CO
hydroboration is catalytically deoxygenated through reactions
that involve both ToMMgX and HBpin species.
1
basis of H and 11B NMR spectroscopy and GC/MS (δH 1.04,
4.96, 7.05, 7.10, 7.30; δB 22.7; and m/z 234.1),34,35 and
1
Me2NBpin was assigned on the basis of H and 11B NMR
spectroscopy (δH 1.12, 2.64; δB 23.9).36 A small signal in the
GC/MS was assigned to PhBpin on the basis of its identical
retention time as an authentic sample, its parent ion peak of
204.1 m/z in the MS, and the overall similarity of daughter ion
4221
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226

ACS Catalysis
Research Article
observations suggest that the formation of the active catalytic
species requires all three reaction components (ToMMgMe,
HBpin, and amide) and may involve an unusual dual substrate-
catalyst initiation process.
Scheme 2. Multiple Pathways Observed in Reactions of
ToMMgH2Bpin and N,N-Dimethylbenzamide
Unfortunately, attempts to determine a quantitative catalytic
rate law for tertiary amide hydroboration/reduction were
hindered by precipitation that occurs during the reaction.
Qualitatively, an increased concentration of HBpin (with all
other variables kept constant) results in a faster disappearance
of amide and faster appearance of the amine product. A higher
catalyst concentration also provides faster conversions.
Valuable, albeit nonquantitative, mechanistic information is
provided by in situ UV−vis spectroscopy. A transient
absorption at 330 nm, assigned to a reaction intermediate,
quickly increases in intensity in the early stages of the reaction
and then slowly decays (Figure 1). This signal is attributed to a
species that is formed from the combination of ToMMgMe,
HBpin, and Ph(Me2N)CO. Independent experiments
indicate that bimolecular combinations (ToMMgMe and
HBpin or ToMMgMe and Ph(Me2N)CO) do not provide
this absorbance; the catalytic products pinBOBpin and
PhCH2NMe2 also do not produce this signal.
peaks in the MS. In addition, a signal in the GC/MS was
assigned to Me2NCH2OBpin. Similarly, two equiv of N,N-
dimethylbenzamide rapidly give benzyldimethylamine in 16%
yield (vs ToMMgH2Bpin as the limiting reagent) after 15 min,
and no changes are observed after that point. In that
experiment, N,N-dimethylbenzamide is partly consumed (37%
N,N-dimethylbenzamide is unreacted), and several unidentified
ToM-containing compounds are observed.
The dominance of the C−N bond cleavage pathway from the
reaction of ToMMgH2Bpin contrasts the catalytic deoxygena-
tion pathway observed with ToMMgMe and excess HBpin.
Amide deoxygenation requires two reducing equivalents, both
of which could, in principle, be provided by ToMMgH2Bpin.
The experiments above, however, indicate that the second
reducing equivalent in ToMMgH2Bpin goes to C−N bond
cleavage if there is no excess HBpin present. On the basis of
this idea, the effect of excess HBpin on deoxygenation yields
and reaction times was investigated. Neat HBpin results in
nearly quantitative yields and fast reaction rates, and the
optimal yield is obtained with excess pinacolborane (20 equiv;
see Table 1). Moreover, the C−N cleavage products Me2NBpin
and PhCH2OBpin are not detected when a high concentration
of HBpin is used. Thus, the role of excess HBpin under
catalytic conditions is not to give ToMMgH2Bpin in high yield
because catalyst activation does not involve that compound as
an intermediate. Instead, HBpin plays an important role in
controlling selectivity toward deoxygenation vs C−N cleavage.
An alternative pathway for catalyst activation might instead
involve the interaction of ToMMgMe and an amide.
Interestingly, 1:1 or 1:2 mixtures of ToMMgMe and N,N-
dimethylbenzamide are unchanged after 2 h at room temper-
ature in benzene. After 24 h, all ToMMgMe signals disappeared,
and multiple broad signals associated with unidentified ToM-
containing species appeared. Neither H[ToM] nor methane
signals were detected, ruling out adventitious hydrolysis.
However, ∼95% of the initial N,N-dimethylbenzamide was
unreacted. This process is much slower than the catalytic
reaction and likely unrelated to amide deoxygenation. These
observations also contrast the reaction of ToMMgMe and
EtOAc, which react instantaneously at room temperature to
give acetone and ToMMgOEt.11 The catalytic yields of
benzyldimethylamine (51−54%), however, are similar either
when 2 equiv of HBpin are added to a mixture of ToMMgMe
and N,N-dimethylbenzamide or when catalyst is added to the
mixture of HBpin and N,N-dimethylbenzamide. In total, these
Figure 1. Plot of absorbance vs time for the transient signal at 330 nm
assigned to the catalytically active species.
To analyze the formation of the intermediate, we
independently and systematically varied [ToMMgMe],
[HBpin], and [N,N-dimethylbenzamide] and measured the
signal intensity and slope of curves of absorbance at 330 nm vs
time for the initial portion of the reaction. Variation in the
catalyst concentration from 3.95 to 7.91 mM provided a linear
plot of initial rate (d(absorbance)/dt) vs [ToMMgMe] (kobs
≈
0.02 s−1; [HBpin] = 0.791 M and [Ph(Me2N)CO] = 0.040
M). Moreover, the absorbance increases as [ToMMgMe]
increases, suggesting that the signal is associated with a
magnesium species. Variation of [HBpin] from 0.395 to 1.79 M
shows a generally increasing slope of d(absorbance)/dt with
increasing [HBpin], although the data are not sufficient for a
quantitative linear least-squares analysis. Interestingly, at low
concentration of N,N-dimethylbenzamide, the initial rate shows
little change at low concentration until 40 mM, but shows a
sharp increase in rate above ∼45 mM. These rate dependences
4222
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226

ACS Catalysis
Research Article
parallel our general observations regarding the effects of catalyst
initiation on the catalytic conversion. On the basis of these
similarities, we suggest that the 330 nm signal is due to a
catalytically relevant species that contains ToMMg, amide, and
HBpin derived moieties.
Under catalytic hydroboration conditions, N,N-dibenzyl-4-
cyanophenylamide gives the amide C−N cleavage product
rather than the deoxygenated product (eq 2); however, the
This system gives reduction of amides under mild conditions
at room temperature with good yields (Table 2) and
advantageously tolerates nitro and azo moieties typically
reduced by common stoichiometric metal hydride reagents
such as LiAlH4. Aryl bromide is tolerated, and benzyl groups on
the amide are not cleaved under reaction conditions, as might
be expected under late-metal catalyzed hydrogenations.
Notably, the conversion works with both aliphatic, aromatic,
and formamide-based substrates. However, N,N-dimethyl
acrylamide reacts instantaneously with ToMMgMe and HBpin
to give an unidentified precipitate.
cyano moiety is not reduced under the reaction conditions.
Although the boronate-protected p-cyanobenzyl alcohol was
not isolated, its NMR yield is equal to the NMR yield of
dibenzylamine (89%), and dibenzylammonium chloride is
isolated in 77% yield. The starting material contained
13C{1H} NMR signals at 114.1 and 118.5 ppm assigned to
ipso-C6H4CN and CN signals, and similar resonances in
pinBOCH2C6H4CN were measured at 111.7 and 119.3 ppm.
Deoxygenation of the related electron-poor para-nitrophenyl
amide is straightforward under the standard catalytic con-
ditions, in contrast to LiAlH4 reductions.
Table 2. ToMMgMe-Catalyzed Hydroboration and
Reduction of Tertiary and Secondary Amides
Secondary amides are also deoxygenated to secondary
amines, although increased catalyst loading (Table 2) is
required for high yield. Interestingly, the highest yields are
obtained with 4 equiv of HBpin with respect to the amide
rather than the larger excess preferred for tertiary amides. The
fastest reactions and the highest yields are obtained with
substrates containing small groups adjacent to the carbonyl,
most notably the formamides. These observations, along with
the significant variation of yield with a series of similar
oxazoline-based ancillary ligands, suggest that steric effects
greatly affect the reaction. Significantly, formation of tertiary
amines (e.g., PhNMe2) via imine alkylation is not observed.
This amine alkylation is a common pathway under hydro-
silylation and hydrogenation conditions.3
A closer investigation of the secondary amide hydroboration
reaction through in situ NMR spectroscopy is revealing, even
though the reaction pathway appears complex. First, N-
phenylformamide and HBpin are unchanged at room temper-
ature in the absence of catalyst. That is, the amide NH and the
pinacolborane BH do not eliminate H2 and form a B−N bond
at room temperature. Upon addition of ToMMgMe as the
catalyst, N-phenylformamide and HBpin react rapidly to
consume all of the formamide starting material within 5 min.
Effervescence of hydrogen gas is observed, as is a mixture of
species in solution that includes a formimidate boronic ester
intermediate (A) and reduced hydroboration intermediate C
(Scheme 3).
1
A H NMR spectrum of the reaction mixture contained a
new singlet at 8.10 ppm assigned to a formimidate boronic
ester (A in Scheme 3), and this peak is downfield of the signal
at 7.87 ppm for the N-phenylformamide starting material. The
formimidate signal correlates to a phenyl signal at 6.85 ppm in a
NOESY experiment, which allows assignment of the ortho
phenyl group. From a COSY experiment, the para and meta
phenyl protons are identified at 6.76 and 7.00 ppm. The
13C{1H} NMR spectrum contained a signal at 168.7 ppm that
1
1
correlated with the H NMR signal at 8.10 ppm in a H−13C
HMQC experiment. For comparison, the imine carbon of N-
phenyl iminoether is 154.2 ppm.37 The formimidate proton
also correlated to a 15N NMR signal at −202 ppm in a 1H−15N
HMBC experiment. This 15N NMR chemical shift is similar to
the three-coordinate imidate nitrogen in a protonated
oxazolinylborate.38 The 1H NMR signal at 8.10 ppm also
a
Isolated yield in parentheses.
4223
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226

ACS Catalysis
Research Article
Over 48 h, signals assigned to the product MePhN−Bpin
appear in the NMR spectra of the reaction mixture. The signals
in the 1H NMR spectrum for A at 8.10 ppm and C at 5.65 ppm
disappear as MePhN−Bpin resonances increase in intensity.
Thus, the slow steps in the catalysis are the conversion of A and
C to product, and these steps are faster with greater ToMMgMe
loading.
Scheme 3. Pathways for Catalytic Conversion Based on the
Spectroscopically Detected Composition of the Reaction
Mixture
To further probe the catalytic additions, we turned to
competition experiments between amides and esters. Previous
studies of ToMMg-catalyzed ester reductive cleavage showed
very fast conversions,29 with reactions completing in less time
than the above amide deoxygenation pathway. Our kinetic
studies in that system implicated a catalyst-mediated reversible
ester cleavage prior to hydroboration with ToMMg{RO(H)-
Bpin}. This mechanism is based on a half-order rate
dependence on ester concentration and zero-order dependence
on HBpin concentration. Likely, the initial ester cleavage
involves addition to the carbonyl, whereas the studies of amide
hydroboration suggest concurrent catalytic dehydrocoupling
and hydroboration reactions. Moreover, studies of secondary
and tertiary amides suggest rapid consumption of starting
materials and rapid formation of intermediate(s), followed by
slow formation of the amine products. Thus, competition
experiments probe the relative rates of HBpin addition to
amide vs ester.
1
correlated in a H−11B HMBC experiment to a 11B NMR
resonance at 5.1 ppm. The intensity of that 11B NMR signal is
much larger than the ToM-derived 11B NMR signals of the
catalyst. In addition, the 11B NMR chemical shift appeared in
the range consistent with a neutral four-coordinate boron
center.39 Finally, a high resolution mass spectrum contained the
parent ion signal at 247.1480 m/z, and the isotopic pattern in
the mass spectrum indicated that this species contained only
one boron atom. These data strongly implicate A as a
formimidate species, likely formed from catalytic dehydrogen-
ative borylation of the amide proton.
A second intermediate, tentatively assigned as the boronic
ester resulting from hydroboration of the CO and nitrogen
borylation (C), was characterized by a methylene signal at 5.65
ppm in the 1H NMR spectrum. A 1H−13C HMQC experiment
revealed a correlation between this signal and a 13C NMR signal
at 55.1 ppm. A DEPT-135 experiment indicated that the signal
is from a CH2 group. Unfortunately, crosspeaks to this signal
were not detected in 1H−11B and 1H−15N HMBC experiments.
However, two signals in the 11B NMR spectrum are assigned to
O−Bpin (22.5 ppm) and N−Bpin (24.7 ppm) moieties. The
mass spectrum of the reaction mixture did not contain the
parent ion peak of 248 or 375 m/z for such a species or its
borylated derivative, possibly because of HOBpin or
pinBOBpin elimination and hydrolysis during the mass
spectrometry experiment. Instead, a signal at 106 m/z assigned
to the protonated iminium [Ph(H)N=CH2]+ is detected. This
The initial rates of consumption of phenyl formate and
phenylformamide reactants, chosen for their similar steric and
structural features, were compared. Phenyl formate and
phenylformamide react with HBpin in the presence of 2.5
mol % ToMMgMe to give phenylmethylamidopinacolborane
and the boryl ether products MeOBpin and PhOBpin (eq 3).
Under the conditions of excess carbonyl vs HBpin
(formamide:formate:HBpin =1:1:1), the competition experi-
ment reveals that the rate of amide consumption is faster than
the rate of ester consumption. The initial concentrations of
phenylformamide and phenyl formate decrease by 0.14 and
0.04 M, respectively, over the first 5 min of the reaction. In
contrast, when the concentration of HBpin is increased
(formamide:formate:HBpin =1:1:2 with all other variables
held constant), the rate of amide consumption is slower than
the rate of ester consumption. In this case, the initial
concentrations of phenylformamide and phenyl formate
decreased by 0.09 and 0.13 M, respectively, over the first 5
min of the reaction. In both cases, these experiments show that
the presence of formamide substantially inhibits the rate of
catalytic ester cleavage. However, with a larger amount of
HBpin (formamide:formate:HBpin =1:1:4), all of the ester
substrate is consumed within 5 min, while ca. 50% phenyl-
formamide is unreacted at that time. Because the ester cleavage
is zero-order in [HBpin], these observations suggest that the
active catalytic species are not the same for amide and ester
conversion.
1
intermediate C associated with the methylene H NMR signal
at 5.65 ppm is formed preferentially under reaction conditions
with lower amounts of HBpin (2 equiv), whereas the
formimidate boronic ester is favored at higher HBpin loading
(4 equiv relative to formamide).
Intermediate C contains 2 Bpin groups, whereas A is only
monoborylated; this is the opposite of the selectivity that might
be expected on the basis of the reaction conditions. The
contrast between the stoichiometry of the intermediates and
their apparent kinetic preference can be rationalized as shown
in Scheme 3. Intermediates A and C form concurrently, and
study of the reaction timecourse suggests that A is not on the
pathway to C. From this data, we suggest that the rate
constants for formation of A and B have different [HBpin]
dependences and a fast uncatalyzed NH/BH dehydrocoupling
occurs from intermediate B. Even though intermediates A and
C are formally related by a hydroboration event, the
hydroboration of A is not observed in this catalytic system.
There are no further changes to the NMR spectra of these
reaction mixtures after 15 min. In the experiments with 1:1:1
formamide:formate:HBpin, substantially more of the inter-
mediates A and C are formed (21% and 29% NMR yield) from
the catalytic addition of HBpin and N-phenylformamide than
4224
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226

ACS Catalysis
Research Article
MeOBpin (8%) from ester cleavage. Thus, the formamide
reacts approximately 5× faster than the formate, but the
addition of HBpin to formate leads directly to the product
whereas the catalyzed interaction of HBpin and N-phenyl-
formamide provides intermediate species that are further
reduced at longer reaction times.
distinct for the two transformations. Further work to clarify the
pathway(s) and identify the reactive species for deoxygenation
and C−N cleavage in amides is currently underway.
Organosilanes are not effective reductants of amides in this
oxazolinylborato magnesium system; neither are silanes
effective in the related magnesium-catalyzed reductive cleavage
of esters. Although silanes reduce amides to amines in many
transition-metal-based catalytic systems, these catalysts are
typically less oxophilic (e.g., iron group or later). In the present
reduction employing a highly oxophilic magnesium center, the
HBpin is likely important because of its ability to act as a
hydride donor and as a Lewis acid. This principle may guide
future developments of catalytic amide reductions to improve
efficiency, yield, and selectivity for mild conversion methods.
Our current efforts are directed toward this goal.
An oxazolidinone substrate provides an alternative competi-
tion between ester cleavage and deoxygenation (eq 4). This
ASSOCIATED CONTENT
competition experiment compares product formation from
either pathway, rather than initial consumption of ester or
amide under conditions of low HBpin concentration needed in
the above experiments.
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b01038.
Under conditions with a large excess of HBpin (20 equiv),
deoxygenation is favored 2:1 over ester cleavage as determined
from the ratio of products in the 11B NMR spectrum.
Decreasing the amount of HBpin to 2 equiv. reduces the
product ratio to 1.3:1, but deoxygenation is still favored. These
reactions require 1 day for full conversion, which is similar to
the rate of amide deoxygenation. The change in product ratio
with lower [HBpin] in this case likely reflects the direct
dependence of the amide deoxygenation reaction on the
pinacolborane concentration observed in synthetic experiments.
Experimental description of the synthesis and character-
ization of magnesium catalysts and catalytic products,
and representative spectra of isolated products and in situ
experiments (PDF)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: sadow@iastate.edu.
Present Addresses
†(M.H.) Oxbow Carbon LLC
CONCLUSION
■
‡(D.M.) Institut fur Anorganische Chemie-RWTH Aachen
̈
A number of important general observations are revealed from
our study of the first example of catalytic hydroboration for
amide deoxygenation. The ancillary ligands, based on the
oxazolinylborate motif, show a range of catalytic activity and
selectivity, and it is clear that ancillary ligand effects are
important in these magnesium-catalyzed amide deoxygenations.
ToMMgMe generally outperforms the other tested oxazolinyl-
borate-based magnesium complexes as an effective precatalyst
for the hydroboration of amides, although the sensitivity of the
reaction to conditions suggests that other catalysts based on
oxophilic early metal centers should be explored in the future.
Moreover, comparisons between tertiary and secondary amides
reveals that in general, catalytic hydroboration is fast, and
reductive deoxygenation is slow. However, the results of this
study clearly show that the pathway of tertiary amide reduction
is tuned for C−O vs C−N bond cleavage by HBpin
concentration. Interestingly, this C−N cleavage pathway occurs
at low HBpin concentration, and amide deoxygenation shows a
significant HBpin dependence. In contrast, secondary amides
undergo reductive deoxygenation in preference to C−N
cleavage with only a slight excess of HBpin, which is needed
in the case of the reaction with the amide NH in a catalyzed
dehydrocoupling reaction. Thus, the apparent turnover-limiting
steps in the catalysis involve deoxygenation (C−O bond
cleavage), whereas the catalytic addition of HBpin to the amide
is apparently fast. This appears to be the case for both
secondary and tertiary amides. In contrast, the related ester
reductive cleavage pathway is zero-order in HBpin. Overall, the
magnesium catalyst activation and speciation for amide
deoxygenation and ester cleavage appear to be inequivalent,
and the interaction of HBpin with catalytic intermediates are
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Prof. Andreja Bakac is gratefully thanked for help with kinetic
measurements and analysis. The authors gratefully thank the
National Science Foundation (CHE-0955635) for financial
support.
REFERENCES
■
(1) Lawrence, S. A. Amines: synthesis, properties, and applications;
Cambridge University Press: New York, 2004.
(2) Constable, D. J. C.; Dunn, P. J.; Hayler, J. D.; Humphrey, G. R.;
Leazer, J. L., Jr.; Linderman, R. J.; Lorenz, K.; Manley, J.; Pearlman, B.
A.; Wells, A.; Zaks, A.; Zhang, T. Y. Green Chem. 2007, 9, 411−420.
(3) Smith, A. M.; Whyman, R. Chem. Rev. 2014, 114, 5477−5510.
(4) Dodds, D. L.; Cole-Hamilton, D. J., Catalytic Reduction of
Amides Avoiding LiAlH4 or B2H6. In Sustainable Catalysis: Challenges
and Practices for the Pharmaceutical and Fine Chemical Industries; Dunn,
P. J., Hii, K. K., Krische, M. J., Williams, M. T., Eds.; Wiley: Hoboken,
NJ, 2013; pp 1−36.
(5) Gaylord, N. G. Reduction with Complex Metal Hydrides;
Interscience: New York, 1956.
(6) Brown, H. C. Hydroboration. W. A. Benjamin: New York, 1962.
(7) Igarashi, M.; Fuchikami, T. Tetrahedron Lett. 2001, 42, 1945−
1947.
(8) Cheng, C.; Brookhart, M. J. Am. Chem. Soc. 2012, 134, 11304−
11307.
(9) Park, S.; Brookhart, M. J. Am. Chem. Soc. 2012, 134, 640−653.
(10) Das, S.; Addis, D.; Junge, K.; Beller, M. Chem. - Eur. J. 2011, 17,
12186−12192.
4225
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226

ACS Catalysis
Research Article
(11) Li, Y.; Molina de La Torre, J. A.; Grabow, K.; Bentrup, U.;
Junge, K.; Zhou, S.; Bruckner, A.; Beller, M. Angew. Chem., Int. Ed.
̈
2013, 52, 11577−11580.
(12) Reeves, J. T.; Tan, Z.; Marsini, M. A.; Han, Z. S.; Xu, Y.; Reeves,
D. C.; Lee, H.; Lu, B. Z.; Senanayake, C. H. Adv. Synth. Catal. 2013,
355, 47−52.
(13) Stein, M.; Breit, B. Angew. Chem., Int. Ed. 2013, 52, 2231−2234.
(14) Li, B.; Sortais, J.-B.; Darcel, C. Chem. Commun. 2013, 49, 3691−
3693.
(15) Hanada, S.; Ishida, T.; Motoyama, Y.; Nagashima, H. J. Org.
Chem. 2007, 72, 7551−7559.
(16) Zhou, S.; Junge, K.; Addis, D.; Das, S.; Beller, M. Org. Lett.
2009, 11, 2461−2464.
(17) Balaraman, E.; Gnanaprakasam, B.; Shimon, L. J. W.; Milstein,
D. J. Am. Chem. Soc. 2010, 132, 16756−16758.
(18) Barrios-Francisco, R.; Balaraman, E.; Diskin-Posner, Y.; Leitus,
G.; Shimon, L. J. W.; Milstein, D. Organometallics 2013, 32, 2973−
2982.
(19) Hanada, S.; Tsutsumi, E.; Motoyama, Y.; Nagashima, H. J. Am.
Chem. Soc. 2009, 131, 15032−15040.
(20) Fernandes, A. C.; Romao, C. C. J. Mol. Catal. A: Chem. 2007,
̃
272, 60−63.
(21) Xie, W.; Zhao, M.; Cui, C. Organometallics 2013, 32, 7440−
7444.
(22) Das, S.; Addis, D.; Zhou, S.; Junge, K.; Beller, M. J. Am. Chem.
Soc. 2010, 132, 1770−1771.
(23) Sakai, N.; Fujii, K.; Konakahara, T. Tetrahedron Lett. 2008, 49,
6873−6875.
(24) Chong, C. C.; Kinjo, R. ACS Catal. 2015, 5, 3238−3259.
(25) Arrowsmith, M.; Hadlington, T. J.; Hill, M. S.; Kociok-Kohn, G.
̈
Chem. Commun. 2012, 48, 4567−4569.
(26) Arrowsmith, M.; Hill, M. S.; Hadlington, T.; Kociok-Kohn, G.;
̈
Weetman, C. Organometallics 2011, 30, 5556−5559.
(27) Anker, M. D.; Arrowsmith, M.; Bellham, P.; Hill, M. S.; Kociok-
Kohn, G.; Liptrot, D. J.; Mahon, M. F.; Weetman, C. Chem. Sci. 2014,
̈
5, 2826−2830.
(28) Arrowsmith, M.; Hill, M. S.; Kociok-Kohn, G. Chem. - Eur. J.
̈
2013, 19, 2776−2783.
(29) Mukherjee, D.; Ellern, A.; Sadow, A. D. Chem. Sci. 2014, 5,
959−964.
(30) Guthrie, J. P. J. Am. Chem. Soc. 1974, 96, 3608−3615.
(31) Brown, H. C.; Heim, P. J. Am. Chem. Soc. 1964, 86, 3566−3567.
(32) Brown, H. C.; Heim, P. J. Org. Chem. 1973, 38, 912−916.
(33) Bontemps, S.; Sabo-Etienne, S. Angew. Chem., Int. Ed. 2013, 52,
10253−10255.
(34) Oshima, K.; Ohmura, T.; Suginome, M. J. Am. Chem. Soc. 2011,
133, 7324−7327.
(35) Koren-Selfridge, L.; Londino, H. N.; Vellucci, J. K.; Simmons, B.
J.; Casey, C. P.; Clark, T. B. Organometallics 2009, 28, 2085−2090.
(36) Ohmura, T.; Masuda, K.; Suginome, M. J. Am. Chem. Soc. 2008,
130, 1526−1527.
(37) Naulet, N.; Filleux, M. L.; Martin, G. J.; Pornet, J. Org. Magn.
Reson. 1975, 7, 326−30.
(38) Pawlikowski, A. V.; Gray, T. S.; Schoendorff, G.; Baird, B.;
Ellern, A.; Windus, T. L.; Sadow, A. D. Inorg. Chim. Acta 2009, 362,
4517−4525.
(39) Kidd, R. G. Boron-11. Academic Press: New York, 1983; Vol. 2,
p 49−77.
4226
DOI: 10.1021/acscatal.5b01038
ACS Catal. 2015, 5, 4219−4226