Implications of CO2 Activation by Frustrated Lewis Pairs in the Catalytic Hydroboration of CO2: A View Using N/Si+ Frustrated Lewis Pairs

Article abstract of DOI:10.1021/acscatal.6b00421

A series of base-stabilized silylium species were synthesized and their reactivity toward CO₂ explored, yielding the characterization of a novel N/Si⁺ FLP-CO₂ adduct. These silicon species are active catalysts in the hydroboration of CO₂ to the methoxide level with 9-BBN, catecholborane (catBH), and pinacolborane (pinBH). Both experiments and DFT calculations highlight the role of the FLP-CO₂ adduct in the catalysis. Depending on the nature of the hydroborane reductant, two distinct mechanisms have been unveiled. While 9-BBN and catBH are able to reduce an intermediate FLP-CO₂ adduct, the hydroboration of CO₂ with pinBH follows a different and novel path where the B-H bond is activated by the silicon-based Lewis acid catalyst. In these mechanisms, the formation of a highly stabilized FLP-CO₂ adduct is found detrimental to the kinetics of the reaction.

Full text of DOI:10.1021/acscatal.6b00421

Research Article
pubs.acs.org/acscatalysis
Implications of CO2 Activation by Frustrated Lewis Pairs in the
Catalytic Hydroboration of CO2: A View Using N/Si+ Frustrated Lewis
Pairs
N. von Wolff, G. Lefevre, J.-C. Berthet, P. Thuery, and T. Cantat*
̀
́
NIMBE, CEA, CNRS, Universite
́
Paris-Saclay, Gif-sur-Yvette, France
S
* Supporting Information
ABSTRACT: A series of base-stabilized silylium species were synthesized and
their reactivity toward CO2 explored, yielding the characterization of a novel N/
Si+ FLP-CO2 adduct. These silicon species are active catalysts in the
hydroboration of CO2 to the methoxide level with 9-BBN, catecholborane
(catBH), and pinacolborane (pinBH). Both experiments and DFT calculations
highlight the role of the FLP-CO2 adduct in the catalysis. Depending on the
nature of the hydroborane reductant, two distinct mechanisms have been unveiled.
While 9-BBN and catBH are able to reduce an intermediate FLP-CO2 adduct, the
hydroboration of CO2 with pinBH follows a different and novel path where the
B−H bond is activated by the silicon-based Lewis acid catalyst. In these
mechanisms, the formation of a highly stabilized FLP-CO2 adduct is found detrimental to the kinetics of the reaction.
KEYWORDS: organocatalysis, CO2, hydroboration, mechanisms, DFT calculations
the catalysis remain yet unclear.29−34Experimentally, FLPs that
INTRODUCTION
■
catalyze the hydroelementation of CO2 do not form observable
adducts with CO2, and conversely, only a single N/B FLP-CO2
adduct has been reported so far to promote the catalytic
hydroboration of CO2 by our group.27,35 Mechanistically, CO2
adducts may have a decreased reactivity compared to that of
free CO2 in the presence of a reductant, and reactive CO2
adducts that are not thermodynamically favored under catalytic
conditions may form without being detected.
In order to better address the role and influence of CO2
adducts in the metal-free catalytic reduction of CO2, we have
thus sought stable CO2 adducts able to perform as catalysts in
the hydroboration of CO2. Whereas FLP-CO2 adducts mostly
feature borane or alane Lewis acids, we reasoned that an
isolobal and isoelectronic silylium cation (R3Si+) would exhibit
a stronger Lewis acidity and a greater affinity for the O-center
in CO2. In addition, the versatile synthesis of organosilicon
compounds may facilitate the access to structures exhibiting
different affinities for CO2. In fact, a few FLP systems involving
highly electrophilic silyl cations in combination with
phosphines as Lewis base were reported to activate small
Carbon dioxide is an abundant, inexpensive, low toxic, and
renewable carbon feedstock and therefore an appealing C1-
building block in organic chemistry.1−3 The efficient reduction
of CO2 to important raw materials for the chemical industry or
to value-added chemicals is, however, hampered by its high
thermodynamic stability and kinetic inertness toward reduction.
This drawback calls for efficient reduction catalysts able to
synchronize C−O bond cleavage with C−H and C−C bond
formation, at a high kinetic rate and with a low energy demand.
While CO2 activation and reduction at a metal center has been
thoroughly investigated over the last four decades,2,4−8 organic
systems able to perform this task, based on main group
elements (e.g., B, N, O, P, and Si), remain sporadic and
underexplored. Many different coordination modes have been
unveiled for the interaction of CO2 with a transition metal or an
f-element, based on the nature of the metal ion and the redox
state of the CO2 ligand.9,10 In organic chemistry, CO2 activation
is quite monotonous and relies on the formation of stable
adducts with organic Lewis bases, such as guanidines or N-
heterocyclic carbenes (NHCs).11−13 Nevertheless, most pro-
gress has been gained via the synergistic action of Lewis pairs,
able to coordinate both the C and O centers of CO2. Within
the concept of frustrated Lewis pairs (FLPs), introduced by
Stephan and Erker,14−17 several CO2 adducts have been
successfully isolated over the last 6 years, in which the C
atom is attached to a C, N, or P donor, while one or two O
atoms coordinate a group XIII Lewis acid (B or Al).18−28
Although FLP-CO2 adducts are commonly proposed as
catalytic intermediates in the reduction of CO2 with hydro-
silanes or hydroboranes, their possible involvement and role in
molecules. In 2011, Muller and co-workers demonstrated that
̈
triarylsilylium cations in combination with bulky phosphines
could activate H2, and the same system was shown recently to
activate CO2 in the formation of an adduct (I) with the
intermolecular P/Si+ FLP (Scheme 1).36,37 Recently, Stephan
and co-workers characterized P/Si+ FLP-CO2 adducts (II)
Received: February 12, 2016
Revised: April 13, 2016
© XXXX American Chemical Society
4526
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
Scheme 1. Silylium-Based FLP Systems for the Activation of
Small Molecules (CO2 and H2)
Scheme 3. Synthesis of Hypervalent Silanes and Their
Reaction with CO2
starting from basic phosphines (namely, PEt3 and Pt-Bu3) and
silyl triflates.38 Further, Ashley et al. reported on a phosphine
stabilized silylium species able to cleave H2.39 We report herein
the formation of novel intramolecular FLP-CO2 adducts by
reaction of CO2 with nitrogen stabilized silyl cations. These
systems exhibit excellent catalytic activity in the hydroboration
of CO2, and the catalytic involvement of FLP-CO2 adducts is
addressed, based on mechanistic experimental, and DFT
investigations (Scheme 1).
RESULTS AND DISCUSSION
■
and Ph2SiCl2 led to the isolation of the analogous compounds
iPr1 and Ph1, respectively. Crystals of 1, grown from a pentane
solution, revealed a monomeric structure with a hypervalent
silicon center bound to two methyl groups, the two nitrogen
donors of TBD, and one chloride (Figure 1).44 The silicon
center adopts a distorted trigonal bipyramidal geometry, the
Cl1 and N2 atoms occupying the axial positions with a Cl−Si−
N2 bond angle of 159.20(4)°. The Si−N2 bond length is
Synthesis of Novel FLP-CO2 Adducts. We recently
reported that both 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD)
and guanidine substituted borane III are able to bind CO2 to
afford adducts IV and V, respectively (Scheme 2).27
Scheme 2. TBD and TBD-Substituted Borane II Form CO2-
Adducts III and IV
Importantly, these species are highly active hydroboration
catalysts, and they promote the reduction of CO2 to the
methanol level with 9-borabicyclo[3.3.1]nonane 9-BBN, at
room temperature. In order to increase the activation of the C−
O bond in CO2, the replacement of the organoborane Lewis
acid with a more oxophilic silyl cation was envisioned.
Unstabilized silyl cations are strong σ and π acceptors40 and
can undergo unwanted side reactions with the solvent.41,42
The formation of base stabilized silyl species with modulated
reactivity was thus attempted by reacting the guanidinate
[K+,TBD−] with various dichlorosilanes (Scheme 3a).
−
Figure 1. ORTEP view of 1 (left) and cation (1+)2 in [1+,B(C6F5)4 ]2·
C4H8O2 (right). The solvent molecules and hydrogen atoms are
omitted for clarity. Selected bond lengths (Å) and angles (deg) in 1:
Si1−Cl1 2.2783(5), Si1−N1 1.7914(12), Si1−N2 1.9843(12), N1−
C1 1.3578(17), N2−C1 1.3087(18), and N1−Si1−N2 69.09(5). In
(1+)2: Si1−N1 1.7714(13), Si1−N3 1.7796(13), N1−C1 1.3628(16),
and N3−C5 1.3598(17).
Following the procedure developed by Andell and co-
workers,43 the hypervalent chlorosilane 1 was isolated from
[K+,TBD−] and Me2SiCl2. Replacing Me2SiCl2 with i-Pr2SiCl2
4527
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
significantly longer than the Si−N1 bond length (1.9843(12) vs
1.7914(12) Å), and the Si−Cl bond is elongated in comparison
to Si−Cl bonds in tetravalent chlorosilanes (2.2783(5) vs,
2.01(2) Å in SiCl4). Taken together, these data indicate a 3c/4e
ω-interaction for the axial bonding.45,46 A natural bond orbital
(NBO) analysis indeed confirms a notable acidity at the Si
atom with a positive charge of +1.89. A similar structure was
found for iPr1 (see SI). Treating 1 with a Lewis acid (i.e.,
AgNTf2, [Ph3C][B(C6F5)4], or Me3SiI) readily led to the
formation of dimeric dicationic species in agreement with the
reported ease of halogen−silicon bond ionization of base-
stabilized halosilanes (Scheme 3b).47 Crystals suitable for X-ray
reported by Muller et al. and Stephan et al., respectively, were
̈
found unstable and could not be characterized structurally,38
[2+,Cl−] and [2+,B(C6F5)4 ] (Scheme 3d) represent the only
−
examples of N/Si+ FLP-CO2 adducts, which are characterized
structurally in the solid state. The bonding of CO2 in adduct 2+
resembles that of V and IV, with a more pronounced activation
of the C−O bonds. Indeed, the activation of CO2 with FLPs
bends the linear CO2 molecule, and the O−C−O bond angle is
smaller in 2+ (121.4(3) °) compared to V (128.6(2) °) and IV
(129.9(2) °). In parallel, the C−O bond activated by the Lewis
acid is somewhat longer in 2+ (1.322(4) Å), compared to V
(1.299(3) Å) and IV (1.257(3) Å), while the CO bond is
shorter (1.206(4) vs 1.222(2) in V and 1.229(2) Å in IV).
These data confirm the stronger Lewis acidity of the silyl cation
in the activation of CO2, compared to the N/B system V.
Spectroscopically, the 13C resonance for CO2 is shifted upfield
in 2+ (145.3 ppm, see SI Figures S25 or S23), compared to V
and IV (153.1 and 154.4 ppm, respectively).
−
analysis were obtained for both [1+,B(C6F5)4 ]2 and
[1+,NTf2 ]2 (Figure 1 and SI). A dicationic dimeric structure
−
is also proposed for [1+,I−]2 due to its very poor solubility in
common polar solvents, such as THF, DMSO, acetonitrile, and
pyridine.48
Having in hand a series of base-stabilized chlorosilanes and
silyl cations, the activation of CO2 was explored. Exposing a d8-
THF solution of 1 to an atmosphere of CO2 readily produces a
novel species, namely, [2+,Cl−], in 30% conversion, according
It is notable that the formation of CO2 adducts with the
intramolecular N/Si+ FLPs seems sensitive to the steric
crowding around the silylium center. In fact, no adduct is
observed when the isopropyl or phenyl derivatives iPr1 or Ph1
are exposed to an atmosphere of CO2 (Scheme 3e). This
behavior provides FLP systems for which activation of CO2 is
disfavored, and hence, a platform to study the occurrence and
catalytic importance of CO2 activation by FLPs (vide infra).
Catalytic Hydroboration of CO2. The catalytic hydro-
boration of CO2 is an efficient method to reduce CO2 to the
methanol level under mild reaction conditions, to produce
methoxyboranes or methylamines (in the presence of an amine
source). This catalytic reaction, first unveiled by the Guan
group in 2010 with nickel(II) catalysts,49,50 has attracted much
attention, and a variety of organometallic catalysts have been
proposed.51−53 Interestingly, in 2013, our group27 has shown in
parallel with the work of Fontaine and co-workers31 that simple
organic catalysts such as TBD, Me-TBD, or N/B, and P/B
FLPs (e.g., V) could also serve as efficient catalysts in this
reductive process. As shown with the synthesis and chemical
behavior of 1, 1+, iPr1, Ph1, and 2+, the introduction of a silyl
cation on the guanidine backbone of TBD enables a fine-tuning
in the activation of CO2. Whereas a strong coordination of CO2
is observed in 2+, no reaction with CO2 is observed in the case
of iPr1 and Ph1 showing that the corresponding adducts are not
thermodynamically favored (although they could exist as
metastable intermediates). These compounds therefore con-
stitute a suitable platform to investigate the effect of CO2-
adduct formation in the catalytic reduction of CO2.
1
to the H and 13C NMR spectra of the crude solution (see SI,
Figure S22). Crystals suitable for X-ray diffraction were
successfully obtained by slow diffusion of pentane into a
concentrated pyridine solution of 1 under a CO2 atmosphere.
As depicted in Figure 2, [2+,Cl−] is a novel CO2 adduct where
the carbon atom coordinates to a N-donor on the TBD
backbone, while one oxygen atom binds to the silylium center.
Figure 2. ORTEP view of cation 2+ in [2+,Cl−]·0.5 C5H5N. The
solvent molecules and hydrogen atoms are omitted for clarity. Selected
bond lengths (Å) and angles (deg): Si1−O1 1.671(2), Si1−N1
1.756(2), O1−C8 1.322(4), O2−C8 1.206(4), N1−C1 1.331(4),
N2−C1 1.392(4), Si1−O1−C8 125.7(2), and O1−C8−O2 121.4(3).
In the presence of 9-BBN and 1 atm of CO2 in THF, 2.5 mol
% of 1 catalyzes the hydroboration of CO2 to the
corresponding methoxyborane MeOBBN in >99% yield, after
20 h at 20 °C (Table 1, entry 1, TOF = 1.3 h−1).54 THF is the
best solvent under these conditions, as both more polar
(CH3CN, pyridine) and less polar solvents (toluene, CH2Cl2)
gave lower TOFs (Table 1, entries 2−5, TOFs up to 0.3 h−1).
In fact, as shown in the work of Kummer and co-workers,47 the
solvent (Gutmann donor and acceptor number) can have an
important influence on the ionization of the Si−Cl bond and
hence on the Lewis acidity of the silicon center. The observed
solvent effects could, however, not be correlated to the
respective solvent acceptor (nor donor) numbers and might
therefore be due to differences in CO2 and borane solubility.
Increasing the steric bulk around silicon leads to a lower
reactivity for the hydroboration of CO2 using 9-BBN. After 24
The formation of 2+ also reveals that 1 can act as an
intramolecular FLP. The addition of CO2 is reversible, and 1 is
recovered after evaporating a THF solution of [2+,Cl−] under
high vacuum (10−2 mbar) overnight (Scheme 3c). Interestingly,
the formation of [2+,Cl−] is more favored in CH2Cl2 compared
to THF, and it reaches a conversion of 80% (see SI, Figure
S22). This is in agreement with the observed ease of ionization
of Si−Cl bonds by polar solvents reported by Kummer and co-
workers,47 and the more polar CH2Cl2 solvent might thus shift
the equilibrium toward the ionized CO2-adduct. Starting from
[1+,B(C6F5)4 ]2, CO2-adduct [2+,B(C6F5)4 ] is obtained
quantitatively under a CO2 atmosphere (see SI, Figure S26).
This adduct shows a much greater stability as 7 days under
reduced pressure (10−2 mbar) are needed to promote the
−
−
−
decarboxylation of [2+,B(C6F5)4 ]. While adducts I and II,
4528
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
a
Table 1. Catalyst Screening for the Hydroboration of CO2 to Methoxyboranes
b
c
entry
1
catalyst
borane
T [°C]
t [h]
yield [%]
TOF (h−1
)
1
9-BBN
9-BBN
9-BBN
9-BBN
9-BBN
9-BBN
9-BBN
9-BBN
9-BBN
catBH
catBH
catBH
catBH
catBH
pinBH
pinBH
pinBH
pinBH
pinBH
catBH
catBH
20
23(6.5)
78
100(77)
17.5
0
1.3 (3.7)
0.1
d
2
1
20
e
3
1
20
78
0.0
f
4
1
20
78
83
0.3
g
5
1
20
78
43
0.2
6
iPr1
Ph1
20
24
60
0.7
7
20
24
56
0.8
0.5[1+,B(C6F5)4
0.5[1+,I−]2
]
2
]
2
]
2
20
72(3)
24
30(6)
100
74
0.1(0.9)
1.3
−
−
−
8
9
20
10
11
12
13
14
15
16
17
18
19
20
1
iPr1
Ph1
70
120
67
0.5
70
31
0.3
70
96
44
0.3
0.5[1+,B(C6F5)4
[1+,I−]2
70
42
18
0.2
70
96
49
0.3
1
iPr1
Ph1
90(70)
90(70)
90
28(28)
16(28)
28
75(21)
100(78)
82
0.8(0.2)
2(0.8)
0.9
0.5[1+,B(C6F5)4
0.5[1+,I−]2
90
90
50
0.6
90
96
40
0.5
Ph5
70
96
40
0.3
h
21
7
90
24
10
0.1
a
b
Values in parentheses correspond to shorter reaction times with corresponding yields. Reaction conditions: catalyst (2.5 mol %) in 400 μL of
c
THF, (0.48 mmol of B−H) and 1 atm CO2. Yields determined by 1H NMR vs internal standard (mesitylene). TOFs were calculated under steady-
state regime as the derivative of TON (B−H bond used) vs reaction time. In CH3CN. In pyridine. In DCM. In toluene. In benzene.
d
e
f
g
h
h at 20 °C, MeOBBN was obtained in 60% yield using
isopropyl compound iPr1 (TOF = 0.7 h−1) and in 56% yield
with the phenyl analogue Ph1 (TOF = 0.8 h−1) (Table 1, entries
6 and 7). The nature of the anion X− also has a drastic influence
on the catalytic performances, and replacing the chloride ligand
in Table 1). All together, these results suggest that the
formation of a CO2-adduct (e.g., 2) facilitates the catalytic
hydroboration of CO2 with 9-BBN and catBH. Nevertheless,
−
catalysts binding strongly to CO2 such as [1+,B(C6F5)4 ]2) are
sluggish.
−
in 1 with the noncoordinating B(C6F5)4 anion inhibits the
Although pinacolborane (pinBH) presents a much lower
hydridic character and a weaker Lewis acidity at boron than 9-
BBN or catBH,56 it was found to be an efficient hydride source
in the hydroboration of CO2 with 1. MeOBpin (75%) was
obtained selectively after 28 h at 90 °C under 1 atm of CO2,
from pinBH and 2.5 mol % 1 (entry 15 in Table 1, TOF = 0.8
h−1). Whereas metal-based catalysts have been shown to
effectively catalyze the hydroboration of CO2 using a variety of
borane sources (e.g., (9-BBN)2, BH3SMe2, and catBH), few
catalysts exist that utilize pinacolborane. For example, Sabo-
Etienne and Bontemps et al. have reported a ruthenium hydride
complex able to afford MeOBpin in a low 40% yield after 5 h at
70 °C (TOF = 5 h−1).52 Only three metal-free systems have
been described so far as catalysts in the hydroboration of CO2
with pinBH: Stephan et al. designed a ring expanded-carbene
providing a mixture of two-, four-, and six-electron reduction
products,30 whereas Fontaine et al. prepared a phosphino-
borane FLP producing 60% pinBOMe at 70 °C (TOF = 21
h−1).29,31 Recently, the group of Wegner reported a bidentate
borane catalyst able to reduce CO2 to the methanol level in
catalytic hydroboration of CO2 (entry 8, Table 1, TOFs up to
0.1 h−1).55 Poorly soluble in THF, [1+,I−]2 promotes the
reduction of CO2 to MeOBBN, after an induction period of
about 3 h, in 100% yield, after 24 h (TOF = 1.3 h−1 after
−
induction period). [1+,NTf2 ]2 does not react with CO2 and
shows no catalytic activity, in the presence of 9-BBN,
presumably because of its very poor solubility in THF.
When 9-BBN is replaced with catechol borane (catBH), the
selective reduction of CO2 to methoxycatecholborane (MeOB-
cat) proved less efficient requiring 120 h at 70 °C in 74% yield
(TOF = 0.5 h−1, entry 10, Table 1). Although the catalytic
performances are modest, these results are interesting as neither
the parent guanidine base TBD nor the dimeric boron analogue
III, nor the N/B FLP-CO2 adduct V (Scheme 2) show any
catalytic activity with catBH and pinBH.27 The same trends in
reactivity are observed for catBH and 9-BBN when catalyst 1 is
−
replaced with iPr1, Ph1, [1+,I−]2, or [1+,B(C6F5)4 ]2, and the
catalytic activity plummets in the presence of poorly
coordinating anions or with bulky silyl cations (entries 11−14
4529
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
a
Scheme 4. Computed Pathway for the Catalytic Hydroboration of CO2 via CO2 Activation (M05-2X/PCM THF)
a
Values given are the Gibbs Free Energies using 1 as a reference (G = 0.0 kcal/mol).
89% with pinBH at 70 °C with high reactivity (TOF = 93
h−1).33 As the catalyst itself possesses six active B−H bonds,
this TOF-value must be considered with caution, and a TOF of
around 16 h−1 is calculated per active B−H bond, facilitating its
comparison with other catalysts bearing only one active site.
Strikingly, catalysts iPr1 and Ph1 perform better than 1 in the
hydroboration of CO2 with pinBH, while the opposite effect
was noted with 9-BBN and catBH. MeOBpin is formed in 82
and 100% yield within 28 h at 90 °C with Ph1 and iPr1,
respectively, corresponding to TOFs of 0.9 and 2 h−1 (entries
16 an 17 in Table 1). Although the amino-borane catalyst of
Fontaine and co-workers showed higher catalytic activity (TOF
= 21 h−1, at 70 °C),31 iPr1 is able to promote the formation of
MeOBpin quantitatively without the formation of other
reduction side products within only 16 h. In contrast to
reactions with 9-BBN and catBH, the influence of the anion is
less marked in the case of pinBH. Both compounds
products over time reveals that the reduction of CO2 to
methoxyborane is kinetically controlled by the first hydro-
boration step, yielding HCOOBR2 (see SI page 6). The
mechanism of the catalytic hydroboration of CO2 to
formoxyboranes has thus been investigated using both DFT
calculations and stoichiometric reactions.
Mechanism Involving the Activation of CO2. CO2-
adducts are often described as key-intermediates in the catalytic
cycle of the CO2 reduction processes,25,57 and both adducts IV
and V were shown to react with an electrophilic borane (9-
BBN) to yield formate species.27 In light of these data and the
reaction chemistry of 1 described above, activation of CO2 by 1
is a reasonable starting point for the hydroboration of CO2 with
1 (Scheme 4). DFT calculations, performed at the M05-2X/
PCM = THF/6-31+G* (6-311++G** for the active hydride)
level of theory confirm that the formation of a CO2 adduct
[2+,Cl−] from 1 is indeed exergonic by −6.1 kcal/mol with a
low activation barrier of only 10.6 kcal/mol (Scheme 4).
Interestingly, the coordination of the chloride anion to the
silylium center is also possible, and the corresponding CO2
adduct 2 lies at −8.6 kcal/mol. These data are consistent with
the experimental results showing a reversible CO2 activation
with 1 and the possible formation of hypervalent guanidinate
chlorosilanes. Reduction of the activated CO2 molecule in 2
with 9-BBN was found to involve the addition of the B−H
bond to the unbound CO functionality with a low lying
transition state (TS2−3) at 13.9 kcal/mol. The resulting
intermediate 3 features a tetrahedral carbon. Elimination of
the product HCOOBBN (4) from 3 is essentially barrier-less
and favored by the thermodynamics of the overall process
(ΔG= −13.9 kcal/mol). The hydroboration of CO2 with 9-
BBN catalyzed by 1 thus involves three successive steps,
namely, the activation of CO2 by the intramolecular FLP, the
addition of the hydroborane, and finally, the elimination of
HCOOBBN to regenerate catalyst 1 (Scheme 4). Reduction of
3 before release by a second equivalent of hydride might
therefore also be likely, although experimentally the first
hydride transfer is rate determining (see SI). As depicted in
Scheme 4, CO2 adduct 2 and TS2−3 are the rate determining
[1+,B(C6F5)4 ]2 and [1+,I−]2 exhibit a catalytic activity
−
comparable to that of 1, affording pinBOMe in 40 and 50%
yield, respectively, after 24 h (TOFs of 0.5 and 0.6 h−1
respectively, Table 1, entries 18 and 19). It is remarkable that
pinBH imposes different demands on the nature of the catalyst,
and unlike 9-BBN and catBH, catalysts showing no CO2
activation (e.g., iPr1 and Ph1) are more efficient for the
hydroboration of CO2 with pinBH.
Reaction Mechanism and DFT Calculations. The
catalytic hydroboration of CO2 with 1, 1+, iPr1, and Ph1
constitutes an appropriate system to explore the role and
impact of FLP-CO2 adducts in the reduction of CO2 because
these four active catalysts interact differently with CO2.
Whereas 1 and 1+ form stable adducts with CO2, replacing
the methyl groups on the silicon atom in 1 with i-Pr or Ph
substituents provides iPr1 and Ph1, which do not react with
CO2. In addition, different behaviors and trends were noticed
for 9-BBN and pinBH suggesting that different mechanisms
might take place in this important transformation. Regardless of
the nature of the hydroborane R2BH, the formoxyborane
HCOOBR2 and the acetal H2C(OBR2)2 are intermediates in
the formation of MeOBR2. In addition, the distribution of the
4530
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
states, and they control the kinetics of the catalysis.58−61 The
computed energy span of 22.5 kcal/mol is much lower than the
one required for the uncatalyzed addition of 9-BBN to CO2
(30.8 kcal/mol)27 and is consistent with a fast reaction at room
temperature (computed TOF of 0.7 h−1).57 Notably, as the
CO2 adduct 2 constitutes a TOF determining intermediate
(TDI), increasing its stability is detrimental to the catalytic
performances. Computationally, removing the chloride anion in
the energy surface depicted in Scheme 4, an energy span of 24.7
kcal/mol (1+ as 0.0 kcal/mol reference; see Figure S38) is
obtained, that is, 2.2 kcal/mol higher than that with catalyst 1.
This result is in excellent agreement with the experimental
findings pointing to a significant diminution of the catalytic
boron atom that is less Lewis acidic than 9-BBN because the π
lone pairs on the oxygen substituents are delocalized into the
boron vacant p orbital. As such, coordination of the
hydroborane to the CO group of the activated CO2
cat
molecule is disfavored, and the TOF determining TS2−3
pin
and TS2−3 are higher in energy than 2 (spans of 34.0 and
37.9 kcal/mol for catBH and pinBH, respectively). The
computed order of reactivity 9-BBN > catBH > pinBH is
thus in agreement with the experimental observations.
Nonetheless, a span of 39.7 kcal/mol is relatively high for a
catalytic reaction proceeding at 70 °C, thereby questioning the
validity of a mechanism relying on CO2 activation for pinBH. In
addition, while iPr1 presents a better catalytic performance than
1 with pinBH, its computed energy span is 1.8 kcal/mol higher
in energy (37.9 kcal/mol). Experimentally, no reaction is
−
performance when 1 is replaced with [1+,B(C6F5)4 ]2.
Moreover, the addition of one equivalent of 9-BBN
(monomer) to a 70:30 equilibrium mixture of 1/[2+,Cl−]
results in the fast formation of the acetal CH2(OBBN)2 after
only 2.5 h at RT, while a reaction time of 17 h is required for
−
observed upon mixing [2+,B(C6F5)4 ]2 with pinBH for >30 h at
−
90 °C, although 2.5 mol % [1+,B(C6F5)4 ]2 efficiently catalyzes
the reduction of CO2 to MeOBpin at 90 °C (Scheme 5 and
Table 1, entry 18). These facts definitively rule out a
mechanism relying on CO2 activation for the hydroboration
of CO2 with pinBH, and an alternative pathway was sought.
Mechanism Involving the Electrophilic Activation of
the Hydroborane. In order to account for the catalytic
[1+,B(C6F5)4 ]2 (Scheme 5a,b). The formation of CO2 adducts
−
Scheme 5. Stoichiometric and Catalytic Reactions between
1, 2+, CO2, and Hydroboranes
−
activity of 1 and [1+,B(C6F5)4 ]2 in the reduction of CO2 with
pinBH, an alternative mechanism was explored based on the
potential electrophilic activation of the B−H bond with the
catalyst (Scheme 6). Indeed, the ionization of the silicon-
chloride bond is readily accessible in hypervalent chlorosi-
lanes.47 Starting from 1, such dissociation would be also favored
by the presence of an electrophile such as a hydroborane to
yield masked silylium ion 1+. The hydride transfer from
R2BHCl− to 1+ was computed for the three hydroborane
reagents, to form hydrosilane 5 (Scheme 6). The formation of
5 and chloroborane R2BCl (6) byproduct is slightly endergonic
for the three hydroborane sources (+7.0 kcal/mol for 9-BBN, +
10.9 kcal/mol for catBH, and +7.8 kcal/mol for pinBH, Scheme
6). Nevertheless, the respective transition state energies are all
accessible, suggesting that it is a viable intermediate on the
hypersurface (17.0 kcal/mol for 9-BBN, 21.8 kcal/mol for
catBH, and 24.8 kcal/mol for pinBH, Scheme 6). Species 5
appears to be a potent reductant, and outersphere hydride
transfer from the Si−H group in 5 to an incoming CO2
molecule yields silylium 1+ and the formate anion with a low
energy barrier of 18.9 kcal/mol, in agreement with the reported
reactivity of other hypervalent hydrosilanes toward CO2.62
Subsequent salt metathesis among HCOO−, R2BCl, and 1+
finally regenerates catalyst 1 and releases formoxyborane
HCOOBR2, without any significant energy barrier (<6 kcal/
mol).
from iPr1 and Ph1 is around 4 kcal/mol higher in energy
compared to that of 2 (see SI, Table S1). Although this
destabilization could have a positive effect on the catalysis, it
also results in a lower degree of activation of the CO2 molecule
and, hence, in a greater activation barrier of 25.2 and 22.6 kcal/
mol for the rate determining C−H bond formation, for the iPr
and Ph substituted sytems, respectively (see SI, page 9 for
detailed span values). Again, these results are in good
qualitative agreement with the catalytic data showing a lower
catalytic activity for iPr1 and Ph1 compared to that of 1, and they
emphasize the necessity for a fine balance in the activation of
CO2.
TS5−4 associated with the reduction of CO2 with 5
constitutes the highest energy rate determining state in this
pathway.63 Interestingly, the structure of TS5−4 does not
depend on the nature of the hydroborane, although its energy
level is indirectly dictated by the formation of the chloroborane
coproduct R2BCl (6). Although catalyst 1 is the lowest energy
state in this pathway, CO2 adduct 2 easily forms under the
1
applied conditions, as revealed by H NMR spectroscopy, and
it should therefore be included on the energy surface. 2 is thus
an off-cycle intermediate that further enlarges the surface
landscape yielding energetic spans of 34.5 kcal/mol (9-BBN),
35.3 kcal/mol (pinBH), and 38.3 kcal/mol (catBH) for the
different hydroboranes.
Obviously, changing 9-BBN with catBH or pinBH does not
modify the energy level of the CO2 adduct 2. Nevertheless, it
affects the position of the corresponding TS for the hydride
transfer to CO2 (TS2−3). Both catBH and pinBH possess a
4531
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
Scheme 6. Computed Pathway for the Electrophilic Activation of Hydroboranes in the Reduction of CO2 to Formoxyboranes
a
(M05-2X/PCM THF)
a
Values given are Gibbs Free Energies with respect to 1 (0.0).
This mechanistic route hence highlights two important
experimental observations. First, an electrophilic pathway
significantly lowers the energetic span (35.3 kcal/mol) by
about 3 kcal/mol for the hydroboration of CO2 with pinBH
compared to a route where CO2 is activated via the formation
of an adduct (37.9 kcal/mol). Second, a stable CO2-adduct
hampers reaction kinetics in CO2 hydroboration under
electrophilic activation, as witnessed by the lower reactivity of
Scheme 7. Reduction of CO2 with Ph5 to a Formate Species
[1+,B(C6F5)4 ]2 compared to that of 1. In fact, such a
−
mechanism also explains the differences in reactivity among
1, iPr1, and Ph1, but care has to be taken to avoid quantitative
conclusions, as the calculated spans are close in energy.
Computationally, the energy span for the catalytic hydro-
boration of CO2 with pinBH decreases from 35.3 kcal/mol with
1 to 34.2 and 32.4 kcal/mol with Ph1 and iPr1, respectively. The
DFT model is thus consistent with the experimental data
providing the following order of reactivity: iPr1 > Ph1 > 1.
Although the order of reactivity is reproduced correctly, these
results have to be taken as a qualitative measure, given the
accuracy of the DFT models. It is notable that, although the
TOF-limiting transition state with iPr1 corresponds to the
hydride abstraction from pinBH (TS5−4), the main factor
governing its higher catalytic activity is the destabilization of
CO2-adduct iPr2 (+3.7 kcal/mol with respect to 2) (see SI,
Table S1).
To further emphasize the feasibility of CO2 reduction by a
base-substituted silane 5 (Scheme 7), the synthesis of such a
species was attempted. Guanidine substituted hydrosilane Ph5 is
easily accessible through dehydrogenative coupling of TBD
with diphenylsilane and shows a characteristic hydride peak at
5.03 ppm in the 1H NMR (Scheme 7 and Figure S28). Crystals
suitable for X-ray analysis were obtained by slowly cooling
down a concentrated acetonitrile solution of Ph5 from 90 °C to
room temperature (Figure 3). With bond angles of 97.4,
114.7(2)° and 116.8(2) for N1−Si−H1, N1−Si−C8a, and
N1−Si−C14a, respectively, the silicon geometry is slightly
distorted from a tetrahedron, although a ω-bond as in 1 or iPr1
Figure 3. ORTEP view of Ph5. Only one position of the disordered
groups is represented (see SI for details). Selected bond lengths (Å):
Si1−N1 1.722(3), Si1−H1 1.427, N1−C1 1.341(4), and N2−C1
1.323(4).
is less likely due to a small bond angle of 153.6° for N2−Si−H1
and a long N2−Si distance of 2.648(5) Å. Under 1 atm of CO2,
a THF solution of Ph5 evolves within minutes at room
4532
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
temperature to yield a formate species (Scheme 7 and Figure
S30). In the absence of an electrophilic scavenger, such as a
chloroborane, this species then converts to formylated TBD by
deoxygenation of the formate anion (Figure S30). This
chemical behavior demonstrates the strong reductive capability
of Ph5 in the presence of CO2,64 in agreement with the DFT
model depicted in Scheme 6.
bond to yield a boron electrophile and convert the catalyst to a
strong hydride donor, as observed for iPr1 with pinBH (Scheme
8B). In these two mechanisms, the formation of stable CO2
adducts hampers the catalysis by stabilizing the catalyst’s resting
state. Finally, the organocatalyst can also activate the CO2
molecule, directly. As CO2 is a weak Lewis base, this mode of
action usually involves a bifunctional activation of CO2 with the
cooperative effect of a Lewis base and a Lewis acid (Scheme
8C). Catalysts 1, 2+, and IV were shown to follow this path,
with 9-BBN. Importantly, the degree of activation of the CO2
molecule must be finely tuned to open this reduction path.
Indeed, the “activation” of CO2 in the form of an adduct is both
deleterious and necessary to the catalytic activity, as it stabilizes
the lowest intermediate in the potential energy surface yet
prepares CO2 for the subsequent reduction step by removing
electron density from the carbon by coordination to the Lewis
acid. It is noteworthy that future catalytic systems based on this
approach should thus target compounds that show a low
affinity for CO2, based on thermodynamics, yet increase the
electrophilicity of the carbon center with strong Lewis acids.
The catalytic implication of Ph5 was further assessed in the
reduction of CO2 with catBH. Ph5 shows comparable activity as
the related phenyl catalyst Ph1 (TOF = 0.3 h−1 in both cases)
providing CH3OBcat in 40% yield after 96 h at 70 °C (44% for
Ph1, Table 1, entries 20 and 12).
Overall, the mechanistic route depicted in Scheme 6 relies on
the abstraction of a hydride from the hydroborane reagent with
the Lewis acid catalyst and its subsequent transfer to CO2 for
the reduction of a CO bond. This route is hence reminiscent
of the work of Piers and Gevorgyan groups on the
hydrosilylation of CO and C−O bonds, where the
electrophilic borane (B(C6F5)3) abstracts the hydride from
the hydrosilane.64−70 Here, DFT calculations suggest that the
inverse reaction should also be possible. With the work of
Oestreich and co-workers71,72 on the transient formation of
such Si−H−B species during the hydrosilylation of carbonyl
functional groups and the recent isolation and characterization
of such a species by Piers and co-workers,73 the availability of a
transition state (TS1−5) for the hydride transfer from a
chloroborate (e.g., [pinBH−Cl]−) to the masked silylium 1+
is further accredited. Interestingly, such a mechanism relying on
an electrophilic activation of a hydroborane reductant is
unprecedented in hydroboration chemistry, although it
resembles the mode of action of Lewis acid catalysts in
hydrosilylation chemistry. It is found to be favored for electron
rich boranes such as pinBH (and possibly catBH, although
DFT calculation does not allow us to draw a final conclusion),
with a boron center stabilized by electron donors.
CONCLUSIONS
■
A novel class of intramolecular nitrogen−silylium FLPs, e.g.,
TBDSiR2Cl (R = Me, iPr, and Ph) and their cationic derivatives
have been reported, and their reactivity with CO2 has been
investigated. Formation and stability of the CO2-adducts can be
modulated with the steric and electronic environment at the
silicon center, and only the methylated CO2−FLP have been
observed and characterized. All of these novel silicon species
were shown to promote the catalytic hydroboration of CO2 to
the methoxide level, in high yield under mild reaction
conditions. DFT calculations were carried out, in support of
experimental results, to better explore the role of FLP-CO2
adducts in the catalytic reduction of CO2 with 3 different
boranes (9-BBN, catBH, and pinBH). To date, CO2-adducts 2+
and V, where CO2 is activated by a N/Si+ or N/B
intramolecular FLP, remain the only two examples of stable
FLP-CO2 adducts exhibiting a catalytic activity in the hydro-
elementation of CO2.
Altogether, these results provide a clearer perspective of the
role of FLP-CO2 adducts in the catalytic reduction of CO2 and
of the different mechanisms at play in the organocatalytic
hydroboration of CO2. Organic catalysts have been shown to
promote the hydroboration of CO2 via three different modes
(Scheme 8). First, strong nucleophiles favor the hydride
From a mechanistic point of view, the reduction of CO2 by 9-
BBN with 1 relies on a hydride transfer from boron to carbon
through a [2 + 2] mechanism, where the B−H bond adds to
the CO bond of the activated CO2 molecule in the FLP-CO2
adduct. However, using pinBH with iPr1 as catalyst a novel
electrophilic pathway is at play. For catBH and other catalysts, a
clear distinction between both pathways cannot be made.
Nevertheless, a strong stabilization of CO2 in the form of an
adduct is deleterious to the catalytic activity in all cases.
Scheme 8. Different Modes of Action of an Organocatalyst
in the Hydroboration of CO2
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.6b00421.
Experimental procedures, characterization of com-
pounds, kinetic profile of catalytic CO2 hydroboration,
crystallographic data, Cartesian coordinates of all
optimized stationary points and transition states with
the corresponding energies and lowest frequencies, and
overview of the energetic spans of the different proposed
reaction mechanisms including different catalysts and
hydroboranes sources, as well as an alternative reaction
pathway (PDF)
transfer to CO2 by increasing the hydridicity of the B−H
linkage by coordination to the boron center (Scheme 8A). As
an example, catalyst Me-TBD or pro-azaphosphatranes operate
through this process.27,74 In a recent study by Bourissou and
Fontaine,32 the importance of an intermediate formaldehyde
adduct was highlighted, which can also serve as a Lewis base. In
contrast, Lewis acids may abstract the hydride of the B−H
4533
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
X-ray data (CIF)
Research Article
(29) Courtemanche, M.-A.; Leg
J. Am. Chem. Soc. 2014, 136, 10708−10717.
́
are,
́
M.-A.; Maron, L.; Fontaine, F.-G.
(30) Wang, T.; Stephan, D. W. Chem. - Eur. J. 2014, 20, 3036−3039.
(31) Courtemanche, M.-A.; Legare, M.-A.; Maron, L.; Fontaine, F.-G.
́ ́
AUTHOR INFORMATION
Corresponding Author
*Fax: +33-1-6908-6640. E-mail: thibault.cantat@cea.fr.
■
J. Am. Chem. Soc. 2013, 135, 9326−9329.
(32) Declercq, R.; Bouhadir, G.; Bourissou, D.; Leg
́
́
are, M.-A.;
Courtemanche, M.-A.; Nahi, K. S.; Bouchard, N.; Fontaine, F.-G.;
Maron, L. ACS Catal. 2015, 5, 2513−2520.
Notes
The authors declare no competing financial interest.
(33) Lu, Z.; Hausmann, H.; Becker, S.; Wegner, H. A. J. Am. Chem.
Soc. 2015, 137, 5332−5335.
ACKNOWLEDGMENTS
(34) Berkefeld, A.; Piers, W. E.; Parvez, M. J. Am. Chem. Soc. 2010,
132, 10660−10661.
■
For financial support of this work, we acknowledge CEA,
CNRS, the University Paris-Saclay, the CHARMMMAT
Laboratory of Excellence, and the European Research Council
(ERC Starting Grant Agreement no. 336467). T.C. thanks the
Fondation Louis D. − Institut de France for its great support.
N.v.W. thanks S. Kozuch for providing the Autotof software.
(35) Cantat, T.; Gomes, C.; Blondiaux, E.; Jacquet, O. Method for
preparing oxyborane compounds. C07F 5/02 (2006.01), C07B 59/00
(2006.01), G01N 33/60 (2006.01). WO/2014/162266, 2014.
(36) Reißmann, M.; Schafer, A.; Jung, S.; Muller, T. Organometallics
̈
̈
2013, 32, 6736−6744.
(37) Schafer, A.; Reißmann, M.; Schafer, A.; Saak, W.; Haase, D.;
̈
̈
Muller, T. Angew. Chem., Int. Ed. 2011, 50, 12636−12638.
(38) Weicker, S. A.; Stephan, D. W. Chem. - Eur. J. 2015, 21, 13027−
13034.
̈
REFERENCES
■
(1) Fiorani, G.; Guo, W.; Kleij, A. W. Green Chem. 2015, 17, 1375−
1389.
(39) Herrington, T. J.; Ward, B. J.; Doyle, L. R.; McDermott, J.;
White, A. J. P.; Hunt, P. A.; Ashley, A. E. Chem. Commun. 2014, 50,
12753−12756.
(2) Aresta, M.; Nobile, C. F.; Albano, V. G.; Forni, E.; Manassero, M.
J. Chem. Soc., Chem. Commun. 1975, 0, 636−637.
(3) Tlili, A.; Blondiaux, E.; Frogneux, X.; Cantat, T. Green Chem.
2015, 17, 157−168.
(4) Leitner, W. Coord. Chem. Rev. 1996, 153, 257−284.
(5) Walther, D. Nachr. Chem., Tech. Lab. 1992, 40, 1214−1227.
(6) Walther, D.; Ruben, M.; Rau, S. Coord. Chem. Rev. 1999, 182,
67−100.
(7) Fernandez-Alvarez, F. J.; Aitani, A. M.; Oro, L. A. Catal. Sci.
Technol. 2014, 4, 611−624.
(8) Vaska, L. J. Mol. Catal. 1988, 47, 381−388.
(9) Gibson, D. H. Coord. Chem. Rev. 1999, 185−186, 335−355.
(10) Gibson, D. H. Chem. Rev. 1996, 96, 2063−2096.
(11) Weyers, G.; Steimann, M.; Kuhn, N. Z. Naturforsch. 1999, 54b,
427.
(12) Villiers, C.; Dognon, J.-P.; Pollet, R.; Thuer
M. Angew. Chem., Int. Ed. 2010, 49, 3465−3468.
(40) Douvris, C.; Ozerov, O. V. Science 2008, 321, 1188−1190.
(41) Lambert, J. B.; Zhang, S.; Stern, C. L.; Huffman, J. C. Science
1993, 260, 1917−1918.
(42) Reed, C. A. Acc. Chem. Res. 1998, 31, 325−332.
(43) Maaranen, J.; Andell, O. S.; Vanne, T.; Mutikainen, I. J.
Organomet. Chem. 2006, 691, 240−246.
(44) Andell and co-workers solved a dimeric structure of 1 by cooling
down a refluxing pentane solution to −30 °C. This might indicate a
higher thermodynamic stability of the dimeric structure.
(45) Pounds, A. J. J. Chem. Educ. 2007, 84, 43.
(46) Weinhold, F.; Landis, C. R. Valency and Bonding: A Natural
Bond Orbital Donor-Acceptor Perspective; Cambridge University Press:
Cambridge, U.K., 2005.
(47) Kummer, D.; Abdel Halim, S. H.; Kuhs, W.; Mattern, G. J.
́
y, P.; Ephritikhine,
Organomet. Chem. 1993, 446, 51−65.
(48) Crystals suitable for X-ray analysis were obtained by slow
diffusion of pentane into a concentrated solution of [1+,I−]2 in DMF.
However, the solved structure shows the decomposition of [1+,I−]2 in
DMF with complete deoxygenation of several DMF molecules and
four TBD residues, as well as four iodides in the unit cell.
(49) Chakraborty, S.; Zhang, J.; Krause, J. A.; Guan, H. J. Am. Chem.
Soc. 2010, 132, 8872−8873.
(13) Murphy, L. J.; Robertson, K. N.; Kemp, R. A.; Tuononen, H.
M.; Clyburne, J. A. C. Chem. Commun. 2015, 51, 3942−3956.
(14) Stephan, D. W. Dalton Transactions 2009, 3129−3136.
(15) Stephan, D. W. Org. Biomol. Chem. 2008, 6, 1535−1539.
(16) Erker, G. C. R. Chim. 2011, 14, 831−841.
(17) Stephan, D. W. Acc. Chem. Res. 2015, 48, 306−316.
́ ́
(18) Fontaine, F.-G.; Courtemanche, M.-A.; Legare, M.-A. Chem. -
(50) Huang, F.; Zhang, C.; Jiang, J.; Wang, Z.-X.; Guan, H. Inorg.
Chem. 2011, 50, 3816−3825.
Eur. J. 2014, 20, 2990−2996.
(19) Momming, C. M.; Otten, E.; Kehr, G.; Frohlich, R.; Grimme, S.;
̈
̈
(51) Shintani, R.; Nozaki, K. Organometallics 2013, 32, 2459−2462.
(52) Bontemps, S.; Vendier, L.; Sabo-Etienne, S. Angew. Chem., Int.
Ed. 2012, 51, 1671−1674.
Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2009, 48, 6643−6646.
(20) Theuergarten, E.; Bannenberg, T.; Walter, M. D.;
Holschumacher, D.; Freytag, M.; Daniliuc, C. G.; Jones, P. G.;
Tamm, M. Dalton Transactions 2014, 43, 1651−1662.
(53) Sgro, M. J.; Stephan, D. W. Angew. Chem., Int. Ed. 2012, 51,
11343−11345.
(21) Stephan, D. W. J. Am. Chem. Soc. 2015, 137, 10018−10032.
(22) Menard, G.; Gilbert, T. M.; Hatnean, J. A.; Kraft, A.; Krossing,
́
(54) We chose to calculate the TOF following the proposition of
Kozuch and Martin (see ref 62), i.e., as the derivative of the number of
turnovers (number of B−H bonds used) per catalyst with respect to
the reaction time under a steady state regime.
I.; Stephan, D. W. Organometallics 2013, 32, 4416−4422.
(23) Travis, A. L.; Binding, S. C.; Zaher, H.; Arnold, T. A. Q.; Buffet,
J.-C.; O’Hare, D. Dalton Transactions 2013, 42, 2431−2437.
(24) Appelt, C.; Westenberg, H.; Bertini, F.; Ehlers, A. W.; Slootweg,
J. C.; Lammertsma, K.; Uhl, W. Angew. Chem., Int. Ed. 2011, 50, 3925−
3928.
−
(55) [1+,NTf2 ]2 is almost insoluble in THF and showed no catalytic
activity under the employed reaction conditions.
(56) Plumley, J. A.; Evanseck, J. D. J. Phys. Chem. A 2009, 113, 5985−
5992.
(25) Men
1797.
́
ard, G.; Stephan, D. W. J. Am. Chem. Soc. 2010, 132, 1796−
(57) Lu, Z.; Wang, Y.; Liu, J.; Lin, Y.-j.; Li, Z. H.; Wang, H.
Organometallics 2013, 32, 6753−6758.
(26) Bertini, F.; Lyaskovskyy, V.; Timmer, B. J. J.; de Kanter, F. J. J.;
Lutz, M.; Ehlers, A. W.; Slootweg, J. C.; Lammertsma, K. J. Am. Chem.
Soc. 2012, 134, 201−204.
(58) Kozuch, S.; Shaik, S. Acc. Chem. Res. 2011, 44, 101−110.
(59) Uhe, A.; Kozuch, S.; Shaik, S. J. Comput. Chem. 2011, 32, 978−
985.
́
(27) Das Neves Gomes, C.; Blondiaux, E.; Thuery, P.; Cantat, T.
(60) Stephan, D. W.; Erker, G. Chemical Science 2014, 5, 2625−2641.
(61) Kozuch, S. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2012, 2,
795−815.
Chem. - Eur. J. 2014, 20, 7098−7106.
(28) Barnett, B. R.; Moore, C. E.; Rheingold, A. L.; Figueroa, J. S.
Chem. Commun. 2015, 51, 541−544.
4534
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535

ACS Catalysis
Research Article
(62) Kozuch, S.; Martin, J. M. L. ACS Catal. 2012, 2, 2787−2794.
(63) The formation of hypervalent hydrosilanes has also been
proposed as a competent catalytic pathway in the hydrosilylation of
CO2, based on DFT calculation. See: Zhou, Q.; Li, Y. J. Am. Chem. Soc.
2015, 137, 10182−10189.
(64) Arya, P.; Boyer, J.; Corriu, R. J. P.; Lanneau, G. F.; Perrot, M. J.
Organomet. Chem. 1988, 346, C11−C14.
(65) Gevorgyan, V.; Liu, J.-X.; Rubin, M.; Benson, S.; Yamamoto, Y.
Tetrahedron Lett. 1999, 40, 8919−8922.
(66) Gevorgyan, V.; Rubin, M.; Benson, S.; Liu, J.-X.; Yamamoto, Y.
J. Org. Chem. 2000, 65, 6179−6186.
(67) Gevorgyan, V.; Rubin, M.; Liu, J.-X.; Yamamoto, Y. J. Org. Chem.
2001, 66, 1672−1675.
(68) Blackwell, J. M.; Foster, K. L.; Beck, V. H.; Piers, W. E. J. Org.
Chem. 1999, 64, 4887−4892.
(69) Blackwell, J. M.; Sonmor, E. R.; Scoccitti, T.; Piers, W. E. Org.
Lett. 2000, 2, 3921−3923.
(70) Parks, D. J.; Blackwell, J. M.; Piers, W. E. J. Org. Chem. 2000, 65,
3090−3098.
(71) Hog, D. T.; Oestreich, M. Eur. J. Org. Chem. 2009, 2009, 5047−
5056.
(72) Rendler, S.; Oestreich, M. Angew. Chem., Int. Ed. 2008, 47,
5997−6000.
(73) Houghton, A. Y.; Hurmalainen, J.; Mansikkamaki, A.; Piers, W.
̈
E.; Tuononen, H. M. Nat. Chem. 2014, 6, 983−988.
(74) Blondiaux, E.; Pouessel, J.; Cantat, T. Angew. Chem., Int. Ed.
2014, 53, 12186−12190.
4535
DOI: 10.1021/acscatal.6b00421
ACS Catal. 2016, 6, 4526−4535