Experimental and computational studies of borohydride catalyzed hydrosilylation of a variety of C=O and C=N functionalities including esters, amides and heteroarenes

Article abstract of DOI:10.1039/c3nj01485a

Sodium borohydride and a series of related borohydrides catalyze a transition metal-free hydrosilylation of a variety of C=O and C=N functionalities under mild conditions. Importantly, many of these reactions are possible using the cheap and environmentally benign hydrosilane polymethylhydrosiloxane. A mechanism is proposed based on experimental and computational results.

Full text of DOI:10.1039/c3nj01485a

NJC
View Article Online
View Journal | View Issue
PAPER
Experimental and computational studies of
borohydride catalyzed hydrosilylation of a variety
of CQO and CQN functionalities including esters,
amides and heteroarenes†
Cite this: New J. Chem., 2014,
3
8, 1694
a
a
b
Michael G. Manas,‡ Liam S. Sharninghausen,‡ David Balcells* and
a
Robert H. Crabtree*
Received (in Porto Alegre, Brazil)
6th November 2013,
2
Accepted 13th February 2014
Sodium borohydride and a series of related borohydrides catalyze a transition metal-free hydrosilylation
of a variety of CQO and CQN functionalities under mild conditions. Importantly, many of these
reactions are possible using the cheap and environmentally benign hydrosilane polymethylhydrosiloxane.
A mechanism is proposed based on experimental and computational results.
DOI: 10.1039/c3nj01485a
www.rsc.org/njc
1
3–15
center, facilitating attack at silicon by a CQO oxygen
followed by hydride attack at the carbonyl carbon by a borohydride
1
. Introduction
1
6
The catalytic hydrosilylation of carbon–heteroatom multiple bonds intermediate. Although transition metal-free catalysis has
7
is of continuing interest and has become a popular topic in been seen for hydrosilylation of a range of substrates including
1–5
9,14
11,14
12,17
18,19
the recent literature. Hydrosilylation of carbonyls and imines aldehydes and ketones,
esters,
amides,
imines
1
2
20
generally produces the corresponding silyl ethers and silyl amines and nitriles, as well as ether cleavage, there is room for
that are easily converted to alcohols or amines through acidic improvement in terms of substrate scope, catalyst loading and
cleavage. This provides an alternative to the well-established cost of both catalyst and silane.
stoichiometric reduction using metal-hydride or borohydride
We now find that a catalytic amount of a borohydride brings
6
reagents. The mild and selective nature of the reaction, as well about hydrosilylation of aldehydes, ketones, esters, amides,
as the easy availability of inexpensive and environmentally benign imines, and N-heterocycles under mild conditions. In comparison
hydrosilanes, such as polymethylhydrosiloxane (PMHS), provides a to these borohydrides, previously reported Lewis acid borane
1
useful procedure for these reductions. Hydrosilylation has catalysts are either expensive, as for B(C F ) , or difficult to handle,
6
5 3
13,14
another potential advantage in allowing isolation of the silylated as for BF3.
intermediates that can act as protected alcohols or amines. lithium triethylborohydride has the broadest substrate scope,
A wide variety of transition metal hydrosilylation catalysts while sodium borohydride is notable for its low cost and ease of
have been shown to mediate reduction of CQC, CQO and use. Importantly, most of the LiHBEt -catalyzed transformations
but metal-free hydrosilylation is an attractive described here can be achieved with PMHS, a cheap and
Of the borohydride catalysts screened in this study,
3
1
–5,7,8
CQN bonds,
7
and effective alternative. Most of the known transition metal- environmentally benign hydrosilane polymer, as reductant.
free catalysts operate either through Lewis base or Lewis acid We therefore have a simple hydrosilylation procedure, using a
activation of the hydrosilane. Nucleophilic catalysts such as commercially available hydride catalyst and cheap reagents
9
10–12
CsF or TBAF
Si intermediates capable of hydride donation to the substrate. in order to understand this unusual reaction.
Lewis acid catalysts such as B(C form a three center–two
are thought to form reactive, pentacoordinate under mild conditions. Mechanistic studies were undertaken
6 5 3
F )
electron bond between the silyl-hydride and the Lewis acidic
2
. Results and discussion
a
Department of Chemistry, Yale University, P.O. Box 208107, New Haven,
2.1 Catalytic screening
CT 06520-8107, USA. E-mail: robert.crabtree@yale.edu
b
Centre of Excellence for Theoretical and Computational Chemistry (CTCC),
These studies began as an effort to develop first row transition
metal hydrosilylation catalysts, of which relatively few are known.
In control reactions, we found clean and efficient conversion of
benzaldehyde to the benzylsilyl ether without the ostensible Ni
Dept. of Chemistry, University of Oslo, Oslo, 0315 Norway.
E-mail: david.balcells@kjemi.uio.no
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3nj01485a
‡
These authors contributed equally to this research.
3 3
catalyst. With 1 equivalent of SiPhH , 5 mol% of LiHBEt was an
1694 | New J. Chem., 2014, 38, 1694--1700
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
Paper
NJC
effective catalyst in THF at 20 1C over 1 hour. The following lower than with the generally less reactive PMHS it is possible
discussion therefore refers to the transition metal-free system. that this lower activity can be attributed to steric effects. Extension
The hydrosilylation of various carbonyl substrates was catalyzed to reduction of esters and tertiary amides proved possible by
by 5 mol% LiHBEt with SiPhH or PMHS in THF under inert heating to 601 and using 1.05 equivalents of SiPhH3 or three
3
3
1
atmosphere (eqn (1) and (2)). The resulting silyl ethers were equivalents Si–H in PMHS (Table 1, entries 6–9). H NMR analysis
cleaved by basic work-up to form the respective alcohols. For showed that, prior to work-up or cleavage, tertiary amide hydro-
aldehydes and ketones (Table 1, entries 1–4), the reaction silylation directly yields the tertiary amine as well as a mixture of
proceeds to completion in 1 hour with a B35 mol% loading of siloxanes (eqn (2)). Successful amide and ester reduction is notable
SiPhH
3
with respect to carbonyl substrate, indicating that the as this is a much more difficult reaction. The silylation of both
cinnamaldehyde and ethylcinnamate with excess silane at 601
silane is able to deliver all three of its hydride equivalents.
The silyl ether intermediates were identified in GC-MS spectra of selectively yields cinnamyl alcohol with no trace of the fully reduced
the crude reaction mixture. Trisubstituted SiPh(OR) was the major 3-phenyl-1-propanol (Table 1, entries 5 and 7). Our system therefore
3
product while disubstituted SiPhH(OR) was a minor product selectively reduces a,b-unsaturated carbonyls to the unsaturated
2
(
see ESI,† page S6). Similar yields were obtained for hydrosilyla- silyl-ethers, a desirable selectivity pattern.
tion of carbonyl substrates with PMHS as silane (Table 1, entries
, 3, 4, 6 and 8). Diphenyl- and triphenylsilane in the appropriate
stoichiometric ratios were not as reactive as SiPhH (see ESI,†
Table S1). As the yields with diphenyl- and triphenylsilane were
1
3
(1)
Table 1 Hydrosilylation of carbonyl groups catalyzed by LiHBEt
3
c
Silane Time Temp. Yield
(2)
Entry Substrate
Product
(equiv.) (h)
(1C)
(%)
The LiHBEt system is also active in the hydrosilylation of
3
a
a
0
1.5
.35
495
495
b
1
1
1
RT
b
N-heterocycles and imines, reactions for which there are few
1
2
2
,21,22
23
prior examples either with
or without a transition metal.
catalyzes the hydrosilylation
We find that 5 mol% of LiHBEt
3
a
a
0.35
RT
RT
495
of a range of heterocycles at 601 in 6–36 h (eqn (3); Table 2,
entries 1–7). Quinolines and quinoxaline yield tetrahydro-products,
a
a
0
1
.35
495b together with o7% of the 1,2- and 1,4-dihydro-products. Reduction
3
4
b
495
.5
of 1,10-phenanthroline gives 1,2,3,4,7,8,9,10-octahydro- and 1,2,3,4-
tetrahydro-products without forming any dihydro-, hexahydro-, or
a
a
0
1
.35
.5
495
other species (Table 2, entry 7). Acridine (Table 2, entry 6) is
converted to the 1,4-dihydro species, consistent with a 1,4 addition.
Transition metal catalyzed hydrogenation of heterocycles often
proceeds via initial 1,4 hydrogenation followed by a 1,3 hydride
b
1
RT
b
80
a
a
a
5
6
1.05
4
4
60
60
93
24
shift to give an easily reduced imine. A similar mechanism for
hydrosilylation seems less feasible owing to the difficulty of a 1,3
a
1
3.0
.05
495
495
23
silane shift, but the alternative would require H–Si addition
b
b
across a CQC bond which we have not seen in any other substrate.
Imine reduction requires 2 hours at 451 with 0.35 equiv. of SiPhH3,
affording near-quantitative yields of the respective amines after
basic work-up (Table 2, entries 8 and 9). Selected heterocycle and
imine substrates were also reduced using PMHS in generally lower
but still useful yields (Table 2, entries 2, 5 and 8).
a
a
a
7
8
1.05
4
2
60
60
495
a
b
1
.05
495
82
b
3.0
(3)
a
a
9
1.05
2
60
94
Experiments run with LiHBEt
3
1 M in THF (0.06 mmol, 5 mol%), carbonyl
at RT or 60 degrees.
(0b.42 mmol, 0.35 equivalents–1.26 mmol, 1.05 equivalents) used silylation under similar conditions but they have the advantage
as silane. Polymethylhydrosiloxane (0.108 ml, 1.5 equivalents Si–H or
Both LiBH and NaBH are less active catalysts for hydro-
substrate (1.20 mmol) in THF (4 ml) under 1 atm N
2
4
4
a
SiPhH
3
that the reactions proceed without exclusion of air. NaBH
4
c
0
.216 ml, 3.0 equivalents Si–H) used as silane. Yield determined by
1
proved to be a more effective catalyst than the Li or K salts
(see ESI,† Table S2, entries 1–3). The Li salt is highly
H NMR spectroscopy using 1,3,5-trimethoxybenzene as an internal
standard.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 1694--1700 | 1695

View Article Online
NJC
Paper
Table 2 Hydrosilylation of heterocycles and imines catalyzed by LiHBEt
3
Table 3 Hydrosilylation of aldehydes and ketones catalyzed by NaBH
4
d
Silane Time Temp. Yield
Catalyst
c
Entry Substrate
1
Product
(equiv.) (h) (1C)
(%)
Entry loading
Substrate
Time (h) Yield (%)
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
5 mol%
2 mol%
Sat’d sol’n Benzaldehyde’
Benzaldehyde
Benzaldehyde
1
1
1
1
1
1
1
1
1
495
92
89
495
495
495
495
495
48
a
a,c
1
.05
24
24
60
60
71
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
5 mol%
4-Nitrobenzaldehyde
a
a
b
1
3.0
.05
75
50
4-Chlorobenzaldehyde
4-tertbutylbenzaldehyde
4-Methoxybenzaldehyde
Hexanal
Cinnamaldehyde
Acetophenone
4-Methoxyacetophenone
4-Chloroacetophenone
4-Nitroacetophenone
Cyclopentanone
Benzophenone
Benzaldehyde
Acetophenone
Cyclopentanone
b
2
a
a
a
1
.05
24
24
60
60
52
69
3
4
a,b
b
0
1
2
3
4
5
2/2 /24 30/495/495
2
2
2
2
2
2
2
2
10
76
80
a
1.05
a,b
a,b
a,b
a,b
a,b
b
/24
/24
/20
/20
/20
70/65
81/495
46/495
28/495
44/57
b
a
a
b
a
b
d
d
d
a,b
a,b
a,b
1
3.0
.05
6
24
90
55
16
60
b
5
6
1
1
7
8
a
a
a
a
Experiments run with NaBH
2.64 mmol) SiPhH (0.881 mmol, 0.35 equivalents), entries 1–15, or
polymethylhydrosiloxane (0.216 ml, 1.5 equivalents Si–H) entries 16–18
4
(0.052 or 0.13 mmol), substrate
0.70
24
36
60
60
495
(
3
a
b
in THF (4 ml) at RT or 45 degrees. Reactions run at 45 1C. Reactions
Yield determined by H NMR spectroscopy using
,3,5-trimethoxybenzene as an internal standard. Reaction run with
5a
c
1
5
run under 1 atm N
1
PMHS as silane source.
2
.
d
7
2.10
a
26
This suggests that only the dissolved NaBH is effective, in contrast
to the case for some other metal-free systems. Despite this,
5 mol% was kept as the catalyst loading for our work to ensure
having a conveniently weighable quantity and correspondingly
4 3
reproducible results. NaBH showed activity comparable to LiHBEt
for the reduction of benzylic aldehydes, but was less efficient
for the reduction of ketones (Table 3, entry 10). For comparable
conversion, ketones needed either longer reaction times or slightly
elevated temperatures (Table 3, entries 10–15). As expected,
electron-withdrawing groups such as nitro at the 4-position
accelerated the reduction of acetophenones (Table 3, entries 10–13).
Esters, amides and imines gave poor yields under these mild
4
9
a
a
0
1
.35
.5
2
495b
8
9
b
45
45
b
4
85
0.35a
2
495
a
Experiments run with LiHBEt
3
1 M in THF (0.066 mmol, 5 mol%) and
substrate (1.32 mmol) in THF (4 ml) at 60 degrees under 1 atm N
2
.
a
SiPhH (0.46 mmol, 0.35 equivalents to 2.772 mmol, 2.10 equivalents)
3
b
used as silane source. Polymethylhydrosiloxane (0.119 ml, 1.5 equiva-
lents Si–H or 0.238 ml, 3.0 equivalents Si–H) used as silane source.
c
d
o7% of dihydro-products also observed. Yield determined by conditions. With the milder PMHS as silane source, high yields
1
H NMR spectroscopy using 1,3,5-trimethoxybenzene as an internal
were obtained at elevated temperatures and extended reaction
standard.
times (Table 3, entries 16–18).
Thus, the combination of NaBH4 and PMHS provides a
cheap and benign procedure for hydrosilylation of aldehydes
and ketones.
hygroscopic, making it more difficult to handle than the Na or
K salts; this could also explain its lower activity if some of the
catalyst was degraded by ambient humidity. As a result of its
2.2 Mechanism
ease of handling, low cost and good activity, NaBH
4
was
selected for further study. A variety of aldehydes and ketones The mechanism was studied by a combination of experimental
were screened for hydrosilylation using B35 mol% loading of and computational methods. In past transition metal-free
3
SiPhH and variable mol% loading of borohydride in dry THF and systems, a nucleophile was proposed to attack the hydrosilane
processed with a simpler acidic work-up for more rapid screening. to give a transient 5-coordinate, hypervalent silane species
7
,9,16
16
The NaBH -catalyzed hydrosilylation of benzylic aldehydes capable of hydride donation.
Oestereich, et al., showed
4
proceeded efficiently and tolerated a variety of substituents at that nucleophilic attack on Si is also the likely mode of action
1
5
the 4-position of the arene ring (Table 3, entries 4–7).
Since NaBH was not fully dissolved in 4 ml of THF at Lewis acid acts as a hydride abstractor and forms a coordination
mol% loading, we compared the rate of hydrosilylation after complex between the silane and the borane. Accordingly, we
under B(C F ) catalysis. Piers, et al., also suggested that the
6 5 3
4
5
1
2
5
hour using 5 mol%, 2 mol% and a saturated solution and propose that a nucleophile present in our system activates the
found that the rates were nearly the same (Table 3, entries 1–3). silane substrate, forming a hypervalent Si intermediate that
1696 | New J. Chem., 2014, 38, 1694--1700
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
Paper
NJC
promotes hydride transfer from Si. A Lewis acid derived from
4
We propose that the nucleophile in the NaBH system is the
NaBH
4
may also play a role in stabilizing the hypervalent alkoxide or amide anion formed by substrate reduction. The
resulting borane may act as a Lewis acid, removing a hydride
intermediate and encouraging hydride transfer.
One potential nucleophile that could be present under the from the hypervalent intermediate to yield the silyl-ether products
reaction conditions of the hydrosilylation of benzaldehyde is and regenerating the borohydride catalyst. It is unclear if the
the alkali metal alkoxide salt formed from the stoichiometric borane leaves the reaction sphere during the catalysis or if it
reduction of benzaldehyde with NaBH
4
. A number of experi- remains associated with the nucleophile after the initial hydride
mental results are consistent with nucleophilic attack of this donation.
alkoxide on the silane. Lithium benzyloxide was independently
Alternative mechanistic paths were considered. The exclusion of
synthesized and isolated through stoichiometric reduction of light had no effect, ruling out photochemical mechanisms. The
benzaldehyde. This alkoxide salt in a 10 mol% loading could addition of stable organic radical inhibitors such as pyrene gave no
replace the NaBH4 catalyst under the conditions of Table 1, change in reactivity, supporting a nonradical mechanism (see ESI,†
entry 1. After 1 hour, the reaction gave a 26% conversion of Table S2, entry 4).
benzaldehyde to benzyl alcohol, discounting the protonated
alkoxide added as catalyst. By comparison, the same reaction change in enthalpy for a series of elementary steps in the
carried out with the addition of 10 mol% of BH gave a 495% addition of SiPhH to PhCHO (see Table 4). These DH values
The mechanism was further explored by computing the
3
3
conversion showing a great improvement over the system were calculated at the DFT(M06)/6-311+G** level and included
lacking borane. This shows that (i) alkoxide formation and the solvent effects of THF (see Computational details). The
subsequent attack on the silane are potential steps in the catalyst, NaBH , starts by transferring a hydride to either
4
mechanism and (ii) the presence of additional Lewis acid SiPhH or PhCHO. Hydride transfer to PhCHO is endothermic
3
considerably increases the rate of reaction. While this supports but more favourable than hydride transfer to SiPhH , as shown
3
the potential formation of a 5-coordinate hypervalent silane by the DH values found for entries 2 and 3 in Table 4, 58.0 and
ꢀ
1
intermediate, we were not able to isolate or observe any such 25.0 kcal mol , respectively. In addition, the resulting ion pair,
species. PhCH ONa, is largely stabilized by the addition of SiPhH , DH =
We also varied the alkali metal cation in MBH
2
3
ꢀ1
4
. KBH
4
(5 mol%) ꢀ13.2 kcal mol (Table 4, entry 4), which yields the formation
gave only poor yields of the reduced product, presumably of a Si–O bond between these two species, d(Si–O) = 1.93 Å.
attributable to the role of M, but addition of excess alkali metal The resulting intermediate, PhCH O(Na)SiPhH (Fig. 1), lies
2
3
ꢀ
1
ions gave no increase in activity. Using 5 mol% of KBH4 in 11.8 kcal mol above the starting reactants and contains a five-
addition to 20 mol% of NaBF or LiBF gave nearly identical coordinated hypervalent Si center with the Ph and PhCH O
4
4
2
+
yields after 1 hour; 10% yield versus 11% and 9% respectively groups in equatorial and axial positions, respectively. The Na
see ESI,† Table S2, entries 8 and 9). In addition, it was found cation is seemingly stabilized by both O and the Ph rings,
that sequestration of the alkali metal through the addition of as suggested by the close contacts between these pairs,
5-crown-5 ether to a standard hydrosilylation using 5 mol% of d(Naꢁ ꢁ ꢁO) = 2.19 Å, d(Naꢁ ꢁ ꢁPh centroid) = 2.86 and 3.54 Å.
NaBH gave no change in reactivity (see ESI,† Table S2, entry 10). Further stabilization of the hypervalent Si intermediate,
This does not fully rule out a potential role for the sodium ions, PhCH O(Na)SiPhH , by the addition of BH was also explored
(
1
4
2
3
3
since the sequestration by the crown ether only occurs in an by computing the associated reaction enthalpy. Two different
equatorial belt around the cation, leaving axial sites on the products are possible depending on which of the two axial faces
cation available for coordination. However, crown ether binding holds BH
would cause a steric hindrance which could reasonably affect the substituent, the geometry optimization converges to the inter-
reaction rate if the alkali metal is an essential component of an mediate formulated as PhCH O(Na)SiPhH (Fig. 1). The
ꢁBH
intermediary species. Based on this evidence, we tentatively formation of this species is exothermic by 7.9 kcal mol
3 3 2
. If BH is introduced on the face bearing the PhCH O
2
3
3
ꢀ
1
suggest that the alkali metal does not play an essential role (Table 4, entry 5) and involves a Lewis acid–base interaction
and that the borane is the more likely active Lewis acid; the low between BH and the Ph ring attached to Si, d(B–C) = 2.01 Å. In
3
reactivity of KBH may result from poor solubility in THF.
contrast, if BH3 is introduced on the axial face bearing the
4
ꢀ1
Table 4 Reaction enthalpies in kcal mol
a
0
DH
b
Entry
Reaction
DH
1
2
3
NaBH
NaBH
NaBH
PhCH
PhCH
PhCH
PhCH
4
4
4
2
2
2
2
+ PhCHO + SiPhH
+ SiPhH - SiPhH
+ PhCHO - PhCH
ONa + SiPhH - PhCH
3
- NaBH
Na + BH
ONa + BH
O(Na)SiPhH
4
+ PhCH
2
OSiPhH
2
ꢀ25.5
58.0
25.0
ꢀ13.2
ꢀ7.9
ꢀ37.3
0.0
ꢀ25.5
58.0
25.0
11.8
3.9
3
4
3
2
3
c
4
3
2
3
c
5
O(Na)SiPhH
O(Na)SiPhH
O(Na)SiPhH
3
3
2
+ BH
+ BH
ꢁBH
3
- PhCH
- PhCH
- NaBH
2
O(Na)SiPhH
O(Na)SiPhH
3
ꢁBH
3
4
c
6
3
2
2
ꢁBH
ꢀ25.5
7
4
4
+ PhCH
2
OSiPhH
2
ꢀ25.5
a
b
c
Reaction enthalpy for each reaction. Reaction enthalpy with respect to initial reactants (NaBH
4
3
+ PhCHO + SiPhH ). See Fig. 1 for a 3D-
representation of the optimized geometries of PhCH
2
O(Na)SiPhH
3
, PhCH
2
O(Na)SiPhH
3
ꢁBH
3
and PhCH
2
O(Na)SiPhH
2 4
ꢁBH .
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 1694--1700 | 1697

View Article Online
NJC
Paper
Fig. 2 Relaxed energy scans on the B–H (top) and Si–O (bottom) bonds.
2 3 2
Fig. 1 Optimized structures of PhCH O(Na)SiPhH (top-left) and PhCH
O- a steep energy increase when the B–H bond is cleaved from a
(
Na)SiPhH
structures in the geometry optimization of PhCH
tomic distances are given in Å.
3
ꢁBH (top-right). Initial (bottom-left) and final (bottom-right) distance of 1.25 Å (equilibrium distance = 1.21 Å) to 1.50 Å.
3
2 2 4
O(Na)SiPhH ꢁBH . Intera-
From this point, the rise in energy is eased by the formation of
the C–H bond. Fig. 2(b) shows a similar scenario for the
formation of the Si–O bond but with the opposite energy change,
which is constantly reduced as the Si–O bond is shortened from
hydride, the geometry optimization converges to the intermediate
formulated as PhCH O(Na)SiPhH ; i.e. BH abstracts the
ꢁBH
hydride yielding the reaction product along with regenerated
NaBH
3
.35 Å to 1.95 Å (equilibrium distance = 1.93 Å). These results are
2
2
4
3
‡
consistent with DE E DE in both reactions.
On the basis of this experimental and theoretical evidence
we propose the potential reaction mechanism illustrated in
Fig. 3. As discussed above, the reaction is initiated by reduction
4
, d(B–H) = 1.24 Å and d(HꢁꢁꢁSi) = 2.36 Å.
Interestingly, this also happens if the geometry optimization is
initiated starting from widely separated reactants, d(BH ꢁꢁꢁHSi) =
3
3
.00 Å. This suggests that once BH transfers from the axial
3
PhCH
without any potential energy barrier. In the formation of
PhCH O(Na)SiPhH , which is strongly exothermic (DH =
ꢁBH
37.3 kcal mol , Table 4, entry 6), the Si center becomes
2
O-face to the opposite one, hydride abstraction occurs
2
2
4
ꢀ
1
ꢀ
four-coordinate tetrahedral, thus losing its hypervalent character.
From this intermediate, the dissociation of the final reaction
product, PhCH OSiPhH , regenerates the catalyst, NaBH , in a
2
2
4
ꢀ
1
thermoneutral process, DH = 0.0 kcal mol . The overall reaction
is exothermic by 25.5 kcal mol
ꢀ
1
.
The key steps involved in the postulated mechanism, i.e.
hydride transfer (Table 4, entry 3) and alkoxide addition
(Table 4, entry 4), were further explored computationally by
focusing on their kinetics. Despite numerous attempts, the
transition states of these two steps could not be located. This
suggests that these reactions either proceed through extremely
low and flat barriers with ill-defined transition states having
very small imaginary frequencies or have no barrier on the
potential energy surface. In line with this, the relaxed scans on
the elongation of the B–H bond and the shortening of the Si–O
4
Fig. 3 Reaction mechanism postulated for the NaBH -catalysed hydro-
bond did not yield any energy maximum (Fig. 2). Fig. 2(a) shows silylation of benzaldehyde.
1698 | New J. Chem., 2014, 38, 1694--1700
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
Paper
of the substrate at the electrophilic carbon center by the resulting liquid was taken up in CDCl
NJC
3
. The yield was deter-
borohydride. The resulting nucleophile then attacks the silane mined by comparative integration of the product and starting
1
to form a hypervalent, 5-coordinate silyl ether. Subsequent material peaks to the internal standard in the H NMR spectrum.
addition of BH results in either the stabilization of this species Product identity was determined by comparing to commercially
3
or hydride abstraction, depending on which axial face of the available materials.
hypervalent intermediate is involved. Hydride abstraction
yields the reaction product and regenerates the catalyst.
Optimized procedure for room temperature hydrosilylation
screening with NaBH in air
4
To a disposable 4 dram glass vial was added a stir bar,
trimethoxybenzene (56.0 mg, internal H NMR standard) and
3. Conclusions
1
NaBH
.881 mmol) and substrate (2.64 mmol) were then added in
order, as listed. The reaction was allowed to stir at RT for
4 3
(5.0 mg, 0.13 mmol). THF (4 ml), SiPhH (108 ml,
Transition metal-free hydrosilylation of carbon–heteroatom
double bonds occurs with catalytic amounts of LiHBEt and
NaBH4 under mild conditions in good yields. DFT evidence
suggests that hypervalent Si intermediates are involved.
0
3
1
0
hour. After 1 hour, the reaction was stopped by the addition
.25 M p-toluenesulfonic acid in methanol (2 ml). The reaction
was allowed to stir in air for 18 additional hours at RT. The
solvent was then removed in vacuo. CDCl (2 ml) was added to
the resulting slurry and was filtered to remove insoluble excess
NaBH and silane polymer byproducts. The yield was determined
4
General
. Experimental
3
4
All reagents were purchased from Sigma-Aldrich, Alfa-Aesar and
Fisher Chemicals and used as received unless otherwise indicated.
by comparative integration of the product and starting material
1
peaks to the internal standard in the H NMR spectrum. Product
1-Benzoylpiperidine was synthesized by a previously reported
identity was determined by comparing to commercially available
materials.
2
6
protocol. Lithium benzyloxide was synthesized by LiBH
reduction of benzaldehyde (see ESI†). All air-free reactions were
4
run under 1 atm of dry nitrogen using standard Schlenk techniques Computational details
with dried and sparged solvents. All reusable glassware and stir
Structures were fully optimized without any geometry or symmetry
bars were extensively soaked in concentrated base bath to ensure a
minimal risk of reaction contamination by reactive transition
metals. NMR spectra were recorded at room temperature on a
Bruker Avance 400 MHz NMR spectrometer running Topspin
version 1.3 with a 5 mm gradient BBI probe and a BACS-60
automatic sample changer and referenced to the residual solvent
peak (d in ppm and J in Hz). GC-MS spectra recorded on Agilent
constraints. These optimizations were done in-solvent (THF),
modelled with the continuum solvation model of Truhlar and
27
Cramer, SMD. The calculations were carried out at the DFT
28
level by using the dispersion functional of Truhlar, M06, as
29
implemented in Gaussian09. The triple-z quality 6-311+G**
30,31
basis set
was used to represent all elements. Analytic frequencies
were computed in order to confirm the minimum-energy nature of
all stationary points. These calculations were also used to determine
the zero-point and thermal energies needed to compute the reaction
enthalpies.
6890N GC with a 5973N MSD and a 7683 series injector using a
15 meter, 0.25 mm ID Restek Rtx-1 column.
Optimized procedure for hydrosilylation screening under N
2
To a flame dried 15 ml Schlenk tube under nitrogen was added
a stir bar and 1,3,5-trimethoxybenzene (26.5 mg, internal Acknowledgements
1
H NMR standard). Dried and N
added to the sealed Schlenk tube via syringe. Substrate
1.32 mmol), LiHBEt , 1 M in THF (66 ml, 0.066 mmol) and
2
sparged THF (4 ml) was
This research was supported by the U.S. Department of Energy
DOE), Office of Science, Basic Energy Sciences (BES), under
(
(
3
catalysis award DE-FG02-84ER13297 (M.G.M, L.S.). We would
like to acknowledge the aid of Prof. Ulrich Hintermair and
Dr Jesus Campos-Manzano in advising and copy-editing this
manuscript. D.B. acknowledges the Research Council of Norway for
funding (Grant 179568/V30) and the Norwegian Supercomputing
Program for computational resources (Grant nn4654k).
Si–H source SiPhH (54.0 ml, 0.44 mmol–324.0 ml, 2.64 mmol) or
PMHS (0.119 ml, 1.5 equivalents Si–H – 0.238 ml, 3.0 equivalents
Si–H) were then added by syringe in the order listed. Elevated
3
temperature NaBH
same way, with the initial addition of solid NaBH
.066 mmol) to the Schlenk tube instead of LiHBEt
4
catalyzed reactions were prepared in the
(2.5 mg,
. The
4
0
3
reaction was allowed to stir at RT ꢀ60 1C for 1–24 hours
depending on the reaction. After the allotted time, the reaction
was stopped by the addition of base and methanol (4 ml of
Notes and references
1
: 1 2 M aqueous KOH + MeOH). The reaction was allowed to stir
1 D. Addis, S. Das, K. Junge and M. Beller, Angew. Chem., Int.
Ed., 2011, 50, 6004–6011.
2 S.-H. Lee, D. V. Gutsulyak and G. I. Nikonov, Organometallics,
2013, 32, 4457–4464.
in air for 2 additional hours at RT. Additional water was added to
the solution (4 ml) and then the mixture was extracted with
diethyl ether. The organic layers were collected and dried over
sodium sulfate. The solvent was removed in vacuo and the
3 V. K. Chidara and G. Du, Organometallics, 2013, 32, 5034–5037.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 1694--1700 | 1699

View Article Online
NJC
Paper
4
5
A. J. Ruddy, C. M. Kelly, S. M. Crawford, C. A. Wheaton, 14 D. J. Parks and W. E. Piers, J. Am. Chem. Soc., 1996, 118,
O. L. Sydora, B. L. Small, M. Stradiotto and L. Turculet,
Organometallics, 2013, 32, 5581–5588.
9440–9441.
15 D. J. Parks, J. M. Blackwell and W. E. Piers, J. Org. Chem.,
2000, 65, 3090–3098.
(a) J. Zheng, S. Chevance, C. Darcel and J.-B. Sortais, Chem.
Commun., 2013, 49, 10010; (b) L. Gonz ´a lez-Sebasti ´a n, 16 S. Rendler and M. Oestreich, Angew. Chem., Int. Ed., 2008,
M. Flores-Alamo and J. J. Garc ´ı a, Organometallics, 2013,
2, 7186–7194.
47, 5997–6000.
3
17 (a) S. Zhou, K. Junge, D. Addis, S. Das and M. Beller, Org.
Lett., 2009, 11, 2461–2464; (b) Y. Li, J. A. M. de La Torre,
K. Grabow, U. Bentrup, K. Junge, S. Zhou, A. Bruckner and
M. Beller, Angew. Chem., Int. Ed., 2013, 52, 11577–11580;
(c) J. Fernandez-Salas, S. Manzini and S. P. Nolan, Chem.
Commun., 2013, 49, 9758.
6
7
L. Banfi, E. Narisano, R. Riva, N. Stiasni and M. Hiersemann,
Sodium Borohydride, in Encyclopedia of Reagents for Organic
Synthesis, ed. L. Paquette, J. Wiley & Sons, New York, 2004.
G. L. Larson and J. L. Fry, Ionic and Organometallic-
Catalyzed Hydrosilane Reductions, in Organic Reactions,
ed. S. E. Denmark, J. Wiley & Sons, New York, 2008, 18 J. M. Blackwell, E. R. Sonmor, T. Scoccitti and W. E. Piers,
vol. 71. Org. Lett., 2000, 2, 3921–3923.
(a) I. Cabrita and A. C. Fernandes, Tetrahedron, 2011, 67, 19 K. Muther, J. Mohr and M. Oestreich, Organometallics, 2013,
183–8186; (b) R. J. P. Corriu, J. J. E. Moreau and M. Pataud- 32, 6643–6646.
Sat, J. Organomet. Chem., 1982, 228, 301–308; (c) D. V. Gutsulyak 20 A. Fedorov, A. A. Toutov, N. A. Swisher and R. H. Grubbs,
and G. I. Nikonov, Angew. Chem., Int. Ed., 2010, 49, 7553–7556; Chem. Sci., 2013, 4, 1640.
d) T. Murai, T. Sakane and S. Kato, J. Org. Chem., 1990, 55, 21 A. M. Voutchkova, D. Gnanamgari, C. E. Jakobsche,
8
8
(
449–453; (e) C. Cheng and M. Brookhart, J. Am. Chem. Soc.,
C. Butler, S. J. Miller, J. Parr and R. H. Crabtree,
2012, 134, 11304–11307; ( f ) D. Bezier, G. T. Venkanna,
J. Organomet. Chem., 2008, 693, 1815–1821.
L. C. Misal Castro, J. Zheng, T. Roisnel, J.-B. Sortais and 22 L. Hao, J. F. Harrod, A.-M. Lebuis, Y. Mu, R. Shu, E. Samuel
C. Darcel, Adv. Synth. Catal., 2012, 354, 1879–1884; and H.-G. Woo, Angew. Chem., Int. Ed., 1998, 37, 3126–3129.
g) K. Junge, K. Schroder and M. Beller, Chem. Commun., 23 S. Geier, P. A. Chase and D. W. Stephan, Chem. Commun.,
011, 47, 4849–4859; (h) A. M. Tondreau, C. C. H. Atienza, 2010, 46, 4884–4886.
K. J. Weller, S. A. Nye, K. M. Lewis, J. G. P. Delis and P. J. Chirik, 24 D.-W. Wang, X.-B. Wang, D.-S. Wang, S.-M. Lu, Y.-G. Zhou
Science, 2012, 335, 567–570; (i) J. Campos, A. C. Esqueda, and Y.-X. Li, J. Org. Chem., 2009, 74, 2780–2787.
(
2
J. Lopez-Serrano, L. Sanchez, F. P. Cossio, A. de Coszar, 25 F. Schubert and K. Lang, Angew. Chem., 1960, 24, 994–1000.
E. A. Ivarez, C. Maya and E. Carmona, J. Am. Chem. Soc., 26 S. Zhou, K. Junge, D. Addis, S. Das and M. Beller, Angew.
2
010, 132, 16765–16767.
R. J. P. Corriu, R. Perz and C. Reye, Tetrahedron, 1983, 39, 27 A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys.
99–1009. Chem. B, 2009, 113, 6378–6396.
0 M. Fujita and T. Hiyama, J. Am. Chem. Soc., 1984, 106, 28 Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2007, 120,
629–4630. 215–241.
1 M. D. Drew, N. J. Lawrence, D. Fontaine, L. Sehkri, 29 M. J. Frisch, et al., Gaussian 09 (Revision. D.01), Gaussian,
W. Watson and S. A. Bowles, Synlett, 1997, 989–991. Inc., Wallingford CT, 2009.
2 C. Bornschein, S. Werkmeister, K. Junge and M. Beller, 30 R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem.
New J. Chem., 2013, 37, 2061–2065. Phys., 1980, 72, 650–654.
3 J. Fry, M. Orfanopoulo, M. G. Adlington, W. R. Dittman and 31 A. D. McLean and G. S. Chandler, J. Chem. Phys., 1980, 72,
Chem., Int. Ed., 2009, 48, 9507–9510.
9
9
1
1
1
1
4
S. B. Silverman, J. Org. Chem., 1978, 43, 374–375.
5639–5648.
1700 | New J. Chem., 2014, 38, 1694--1700
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014