Hydroboration of Carbon Dioxide Using Ambiphilic Phosphine-Borane Catalysts: On the Role of the Formaldehyde Adduct

Article abstract of DOI:10.1021/acscatal.5b00189

Ambiphilic phosphine-borane derivatives 1-B(OR)₂-2-PR′₂-C₆H₄ (R′ = Ph (1), iPr (2); (OR)₂ = (OMe)₂ (1a, 2a); catechol (1b, 2b) pinacol (1c, 2c), -OCH₂C-(CH₃)₂CH₂O- (1d)) were tested as catalysts for the hydroboration of CO₂ using HBcat or BH₃·SMe₂ to generate methoxyboranes. It was shown that the most active species were the catechol derivatives 1b and 2b. In the presence of HBcat, without CO₂, ambiphilic species 1a, 1c, and 1d were shown to transform to 1b, whereas 2a and 2c were shown to transform to 2b. The formaldehyde adducts 1b·CH₂O and 2b·CH₂O are postulated to be the active catalysts in the reduction of CO₂ rather than being simple resting states. Isotope labeling experiments and density functional theory (DFT) studies show that once the formaldehyde adduct is generated, the CH₂O moiety remains on the ambiphilic system through catalysis. Species 2b·CH₂O was shown to exhibit turnover frequencies for the CO₂ reduction using BH₃·SMe₂ up to 228 h^-1 at ambient temperature and up to 873 h^-1 at 70 °C, mirroring the catalytic activity of 1b. (Figure Presented)

Full text of DOI:10.1021/acscatal.5b00189

Research Article
pubs.acs.org/acscatalysis
Hydroboration of Carbon Dioxide Using Ambiphilic Phosphine−
Borane Catalysts: On the Role of the Formaldehyde Adduct
Richard Declercq,†,⊥,∥ Ghenwa Bouhadir,*,†,⊥ Didier Bourissou,*,†,⊥ Marc-Andre Legare,‡,∥
́
́
́
Marc-Andre Courtemanche,‡ Karine Syrine Nahi,‡ Nicolas Bouchard,‡ Frederic-Georges Fontaine,*,‡
́
́
́
and Laurent Maron§
†Universite
Toulouse, France
⊥CNRS, LHFA, UMR 5069, 31062 Toulouse, France
́ ́ ́ ́
de Toulouse, UPS, Laboratoire Heterochimie Fondamentale et Applique (LHFA), 118 route de Narbonne, 31062
‡Dep
́
artement de Chimie, Universite
́
Laval, 1045 Avenue de la Med
́ ́ ́
ecine, Quebec (Quebec), Canada, G1V 0A6
§Universite
France
́
de Toulouse, INSA, UPS, LCPNO, CNRS, UMR 5215 CNRS-UPS-INSA, 135 avenue de Rangueil, 31400 Toulouse,
S
* Supporting Information
ABSTRACT: Ambiphilic phosphine−borane derivatives 1-
B(OR)2-2-PR′2−C6H4 (R′ = Ph (1), iPr (2); (OR)2 = (OMe)2
(1a, 2a); catechol (1b, 2b) pinacol (1c, 2c), −OCH2C-
(CH3)2CH2O− (1d)) were tested as catalysts for the
hydroboration of CO2 using HBcat or BH3·SMe2 to generate
methoxyboranes. It was shown that the most active species
were the catechol derivatives 1b and 2b. In the presence of
HBcat, without CO2, ambiphilic species 1a, 1c, and 1d were
shown to transform to 1b, whereas 2a and 2c were shown to
transform to 2b. The formaldehyde adducts 1b·CH2O and 2b·
CH2O are postulated to be the active catalysts in the reduction
of CO2 rather than being simple resting states. Isotope labeling
experiments and density functional theory (DFT) studies
show that once the formaldehyde adduct is generated, the CH2O moiety remains on the ambiphilic system through catalysis.
Species 2b·CH2O was shown to exhibit turnover frequencies for the CO2 reduction using BH3·SMe2 up to 228 h−1 at ambient
temperature and up to 873 h−1 at 70 °C, mirroring the catalytic activity of 1b.
KEYWORDS: organocatalysis, carbon dioxide reduction, methanol, frustrated Lewis pairs, hydroboranes, ambiphilic systems
the promise for more active and selective processes. In that
regard, some transition metal-based systems have been used for
the reduction of carbon dioxide to formic acid,4 formate,5
formaldehyde,6 methanol,6d,7 methane,8 acetals,9 and carbon
monoxide.10
In the past few years, there have been many important
developments in the design of metal-free catalytic systems for
the reduction of carbon dioxide. Indeed, some highly reactive
species such as aluminum11 and silylium cations12 have been
shown to reduce carbon dioxide with low selectivity to
methane, methanol, and other alkylation side-products. FLP
(frustrated Lewis pair) systems, even if they are known to bind
carbon dioxide,13 have demonstrated very limited efficiency in
its reduction.14,15 Since the seminal report by Ying and co-
workers that N-heterocyclic carbenes can reduce carbon dioxide
in the presence of hydrosilanes to methoxysilanes, which upon
INTRODUCTION
■
The general consensus of the scientific community on the role
of greenhouse gases on global climate changes has led to several
initiatives to limit their emissions. As a consequence, a large
number of contributions on the search for economical and
efficient ways to sequester and valorize carbon dioxide, the
principal gas which causes climate change, have appeared.
Nowadays, several technologies are used in order to capture
carbon dioxide from flue exhausts of major producers but
sequestration remains a costly solution.1 One way to make the
capture of CO2 more economically viable is in its valorization
by using this molecule as a C-1 building block for the synthesis
of valuable chemicals.2 One transformation that has attracted
much attention is the reduction of carbon dioxide to methanol,
which is at the core of the methanol economy, as promoted by
Nobel laureate George A. Olah.3 Several heterogeneous
systems are known to catalytically reduce carbon dioxide to
methanol, and some of these technologies are now
commercialized.3 Nevertheless, the search for novel catalysts
is still ongoing, notably using homogeneous systems which give
Received: January 29, 2015
Revised: March 11, 2015
© XXXX American Chemical Society
2513
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520

ACS Catalysis
Research Article
hydrolysis yield methanol,16 there have been few other
organocatalytic systems reported for carbon dioxide reduction.
Among these figures the report by Stephan that some
phosphine−borane systems derived from 9-BBN catalyze the
hydroboration of CO2.17 For their part, Cantat and co-workers
demonstrated that strong nitrogen bases, such as guanidines
and amidines can be used as organocatalysts for the reduction
of CO2 to formamides using hydrosilanes18a or to methox-
yboranes using 9-borabicyclo[3.3.1]nonane (9-BBN) and,
although with very low turnover frequency, HBcat (cat =
catechol).18b It was recently reported that bidentate Lewis
bases, such as proton sponge, can also act as a catalyst for the
reduction of carbon dioxide using BH3·SMe2 as a reductant18c,d
and that even NaBH4 can catalyze the reduction of carbon
dioxide using BH3 adducts.18e Moreover, Cantat and co-
workers reported the methylation of amines with carbon
dioxide and 9-BBN mediated by proazaphosphatrane super-
bases.18f However, these last systems, although active, exhibit
either low turnover numbers or low turnover frequencies.
One of the most active metal-free systems for the reduction
of carbon dioxide to date is 1-Bcat-2-PPh2−C6H4 (1b), which
can be generated in situ from the addition of HBcat to
precatalyst Al(2-PPh2−C6H4)3.19 1b can generate methoxybor-
anes of general formula CH3OBR2 from several hydroboranes
(HBR2), such as catecholborane, pinacolborane, and BH3·
SMe2, which in turn can be hydrolyzed to methanol.20
Although such a process has limited commercial potential
because of the cost of boranes, it serves as an excellent model to
understand the mechanism of FLP-derivatives in the catalytic
functionalization of unsaturated substrates such as carbon
dioxide. Indeed, the very high turnover frequency (TOF),
which can reach 973 h−1, and the turnover numbers (TONs)
obtained, which exceed 2950 at 70 °C, surpass that of the most
active transition metal catalysts. An in-depth density functional
theory (DFT) study of this system using HBcat as reductant
allowed postulating that the activation of both the catecholbor-
ane and CO2 in a concerted fashion led to much lower
transition state energies than classical reduction pathways, a
process that does not occur without catalyst.21 In order to
optimize this class of catalyst, we were interested in looking at
the influence of the substituents on the phosphine and borane
parts. Herein, we report that the influence of the phosphine and
the borane moieties on the catalyst is of limited importance
because of substitutions that occur prior to the beginning of the
catalytic activity. Notably, the formaldehyde generated forms an
adduct with the phosphine−borane that turns out to be an
active catalyst for the reduction of carbon dioxide rather than
only a resting state as previously postulated. DFT calculations
suggest that the dual activation of the borane and carbon
dioxide plays a key role in the reduction process.
the synthesis of 1-B(OR)2-2-PR′2−C6H4 derivatives with
diphenyl- and di-iso-propyl-phosphine moieties (R′ = Ph and
iPr) was envisioned (Figure 1).
Figure 1. Synthesis of the various derivatives of general formula 1-
B(OR)2-2-R′2P−C6H4 tested in the course of this study.
The 1-B(OR)2-2-PPh2−C6H4 derivatives (R2 = catechol
(1b), pinacol (1c), and −OCH2C(CH3)2CH2O− (1d)) were
easily synthesized from the addition of the corresponding diols
to in situ generated 1-B(OMe)2-2-PPh2−C6H4 (1a).23b The
isolation of the latter compound was also possible on the gram
scale but proved to be quite sensitive to hydrolysis. On the
other hand, the reaction of catechol with 1-B(OMe)2-2-
P(iPr)2−C6H4 (2a) surprisingly did not yield the expected 1-
Bcat-2-P(iPr)2−C6H4 (2b), although the generation of the
pinacol derivative 1-Bpin-2-P(iPr)2−C6H4 (2c) was successful
using a similar reaction pathway. It was however possible to
synthesize the desired analogue 2b by the electrophilic trapping
of 1-Li-2-P(iPr)2−C6H4 with ClBcat. Unfortunately, species 2b
proved to be particularly unstable in solution (Figure S18 in the
Supporting Information (SI)), which might be caused by
redistribution of the catechol moiety promoted by the
nucleophilic phosphine,24 and had to be protected by
coordination to borane, generating species 1-Bcat-2-[P(iPr)2·
BH3]−C6H4 (2b·BH3).
The catalytic activity of catalysts 1a−d, 2a, 2b·BH3, and 2c
toward the hydroboration of CO2 was evaluated first using
HBcat as a reducing agent under approximately 2 atm of CO2
in C6D6.25 As can be seen in Table 1 (entries 1−3, 6, and 9),
species 1a−c, 2a, and 2c exhibit similar catalytic activity for the
formation of methoxyboranes, sustaining turnover frequencies
between 46 to 56 h−1 range in the first 105 min of reaction. A
stoichiometric amount of catBOBcat was also formed during
the process, as previously characterized for this system.20a The
exception is 2b·BH3, presumably because of the BH3 protection
of the phosphine moiety (entry 7). However, this problem is
circumvented after some time since a TON of 74 was observed
after 12 h of reaction (entry 8). Species 1d is very slow to start
(entry 4), showing insignificant production of methoxyboranes
after 105 min of reaction, but with increasing activity as time
goes, reaching TON of 48 after 4 h (entry 9).
RESULTS AND DISCUSSION
Spectroscopic Monitoring of the Reduction Process.
In order to rationalize the similarities in catalytic activity of the
different phosphine−boranes species and explain the difference
of reactivity for the neopentyl glycol derivative 1d, the CO2
■
Efficiency of 1-B(OR)2-2-PR′2−C6H4 Derivatives As
Catalysts for the HBcat Hydroboration of CO2.
Phosphine−boranes of the general structure 1-BR2-2-PR′2−
C6H4 have been reported before and used extensively as ligands
for transition metals.22 Some of these metal-free species have
shown a large range of activities, notably in the fixation,
splitting, and transfer of singlet dioxygen,23a in the bifunctional
organo-catalyzed Michael addition23b and in the trapping of
reactive intermediates of organic transformations.23c,d In order
to better understand the importance of structural parameters of
these phosphine−boranes derivatives on the catalytic activity,
1
reduction using HBcat was monitored using H and 31P{1H}
NMR spectroscopy. As previously reported, one species at −1.0
ppm was observed by 31P{1H} NMR spectroscopy when 1b
was used as catalyst, which was attributed to the formaldehyde
adduct 1-Bcat-2-PPh2−C6H4·CH2O (1b·CH2O). Interestingly,
it was observed that 1b·CH2O was also present at the end of
the CO2 reduction catalytic experiments when using any of the
1a−d derivatives as catalyst. In order to account for possible
2514
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520

ACS Catalysis
Research Article
Table 1. Catalytic Hydroboration of 2 atm of CO2 Using
HBcat
Scheme 1. Synthesis of Formaldehyde Adducts of Phosphine
Boranes
a
a
b
b
entry
catalyst
borane
time (min)
TON
TOF (h−1
)
1
2
3
4
5
6
7
8
9
1a
HBcat
HBcat
HBcat
HBcat
HBcat
HBcat
HBcat
HBcat
HBcat
105
98
92
98
81
1
53
56
46
<1
12
48
0
1b
1c
105
105
240
105
105
720
105
1d
1d
48
84
0
2a
2b·BH3
2b·BH3
2c
a
1b·CH2O: path A using 1b, 80 °C, 16 h. 1b·13CH2O: path B using
74
75
6
1b, 13CO2, 5 HBcat, 80 °C, 16 h. 1c·CH2O: path A, using 1c, 70 °C,
15 min. 2b·CH2O: path B using 2c, 3 HBcat, 80 °C, 12 h. 2c·CH2O:
path A, using 2c, 70 °C, 15 min.
43
a
Reaction conditions: In a J-Young NMR tube, 530 μmol of
catecholborane (56.4 μL) was added to a mixture of 5.3 μmol of
catalyst (∼0.6 mL of stock solution at ∼8.8 mmol/L) and an internal
standard of hexamethylbenzene (2 mg, 3.3 g/L) in C6D6. The final
concentration of catalyst is 8.1 mM. The J-young NMR tube was
placed under ∼2 atm of CO2 and was heated at 70 °C. The catalysis
were observed by 1H NMR spectroscopy at 5.37 ppm for both
1b·CH2O and 1c·CH2O, which is consistent with the change in
hybridization of the carbon atom and with the loss of planarity
of the aldehyde. In the 31P{1H} NMR spectra, 1c·CH2O was
characterized by a singlet at −5.5 ppm, while 1b·CH2O
resonates at δ = −3.5. The 11B NMR spectra show signals
consistent with tetravalent boron atoms at 5.7 and 8.8 ppm for
1c·CH2O and 1b·CH2O, respectively. The formaldehyde
adducts of the PiPr2 derivatives appeared to be more stable
than their PPh2 counterparts. Species 2b·CH2O was isolated in
quantitative yield by the addition of 3 equiv of HBcat to 2c
under 1 atm of CO2 in C6D6 or by the addition of 30 equiv of
b
was followed by 1H NMR spectroscopy. Turnover numbers (TONs)
and turnover frequencies for the formation of CH3OBcat according to
the number of hydrogen atoms transferred to CO2 based on NMR
integration of the corresponding resonance with internal standard.
transformations of the catalyst during the reduction process, the
reactions of species 1a−d with HBcat were carried out. It was
found that 1a, c, and d readily convert to species 1b when 5
equiv of catecholborane was added to these species in the
absence of CO2. However, species 1d was shown to undergo
such replacement more slowly than the other ambiphilic
species. Monitoring the catalytic reduction of CO2 with 1d as a
catalyst showed that catalysis did not begin until it transforms
to 1b. Similarly, the CO2 reduction reactions using 2a, 2b·BH3,
or 2c as catalysts allowed the observation of a single
phosphorus containing species by 31P{1H} NMR at δ = 14.8
as the resting state of the reaction. The addition of HBcat to
2b·CH2O (vide infra) in solution without the presence of CO2
gave the same resonance observed under catalytic conditions at
≈14 ppm, suggesting that the latter species consist in a Lewis
adduct between 2b·CH2O and HBcat, presumably by an
interaction between the HBcat and one O adjacent to B,
although the exact mode of interaction could not be
determined experimentally. This transformation implies the
reduction of CO2 to formaldehyde, as previously reported for
the PPh2 derivatives, but also the substitution of the pinacol
moiety on boron by a catechol moiety. Similarly to the case of
the −PPh2 catalysts, the formaldehyde adduct 2b·CH2O is the
resting state of the CO2 reduction reaction.
Synthesis of the Formaldehyde Adducts and Labeling
Experiments. As observed in Scheme 1, the formaldehyde
adduct of 1b can be prepared and isolated by heating the
phosphine−boranes 1b in the presence of excess paraformalde-
hyde. Unfortunately, other phosphine−boranes formaldehyde
adducts based on the PPh3 framework could not be isolated. In
fact, 1a decomposed when heated with paraformaldehyde.
While 1c is stable in the presence of paraformaldehyde, its
binding of formaldehyde was found to be reversible and the
adduct 1c·CH2O could not be isolated in the solid form.
Consequently, 1c·CH2O could be prepared in situ and
characterized by NMR spectroscopy, but its catalytic activity
could not be quantified. In CDCl3, the methylene resonances
HBcat to species 2c·CH2O (Scheme 1). This species has a 31
P
NMR chemical shift of δ = 12.2 and the CH2 resonance was
observed at δ = 4.96 in the H NMR spectrum. 2c·CH2O was
1
isolated by the addition of paraformaldehyde to species 2c in
toluene at 70 °C and was characterized by 31P NMR (δ = 7.7).
In order to verify if the aldehyde adduct is only a resting state
or a significant part of the active catalyst, labeled 1b·13CH2O
and 2b·13CH2O were prepared directly from 13CO2.
1b·13CH2O was synthesized by exposing 1b to ca. 2 atm of
13CO2 in a closed Schlenk vessel in the presence of 5 equiv of
catecholborane. The resulting white precipitate was isolated
and washed several times with toluene to yield pure
1b·13CH2O. The 31P{1H} NMR spectrum for 1b·13CH2O
1
shows a characteristic doublet at −3.5 ppm with a JC−P of 57
Hz. In the 1H NMR spectrum, the CH2O resonance was found
as a doublet at 5.37 ppm with a 1JC−H of 151 Hz. Interestingly,
when this species was used as catalyst for the hydroboration of
CO2 using HBcat (Scheme 2, eq 1), no 13C incorporation was
observed in the methoxyborane formed and the 13C−31P
coupling within the formaldehyde adduct could be observed by
NMR spectroscopy throughout the catalytic process. Con-
sistently, when 1b·CH2O and 2b·CH2O were used to reduce
13CO2, no 13C incorporation was observed in the formaldehyde
adducts. These experiments demonstrate that the CH2O
fragment bridging P and B is not reduced but stays coordinated
on the ambiphilic molecule. Thus, the formaldehyde adduct is
more than a resting state and is in fact the active species
responsible for the fast reduction of CO2 in the presence of
ambiphilic phosphine−boraness 1 and 2.
DFT Studies of CO2 Reduction Using Formaldehyde
Adducts. In order to determine whether formaldehyde adduct
1b·CH2O and 2b·CH2O in the catalytic transformation, the
possible interactions between HBcat and these species were
2515
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520

ACS Catalysis
Research Article
find three new transition states which all involve simultaneous
activation of HBcat and CO2. In the first transition state, TS1a
(−1.1 (28.6)), the H(2)−O(1) distance of 2.15 Å suggests that
the Lewis acidic activation is achieved through hydrogen
bonding with one of the hydrogen atoms of the bound-
formaldehyde (Figure 3). Although unusual, the activation of
CO2 by hydrogen bonding was suggested before in the
Ni(II)(cyclam) electrochemical reduction of CO2 to CO by
Sauvage and co-workers26 and in the reduction of CO2 to
sodium formate by an Ir(III) pincer complex by Hazari and co-
workers.27 Another interesting aspect of this transition state is
the difference in B−O bond lengths. While the B(1)−O(4)
bond of 1.38 Å is in the expected range, the B(1)−O(3) bond
of 2.27 Å is significantly elongated, showing that the catechol
group is disconnected and that the B-catechol ring is opened.
The second transition state also involved the formation of an
oxygenate species with HBcat, but this time from the opening
of the formaldehyde adduct (IM3) (TS1b; 1.9 (29.6) kcal·
mol−1). The oxygen atom of CO2 is perfectly aligned with the
Lewis acidic boron center. The B−O bond distance of 2.44 Å
and the planar geometry around the boron center (sum of
angles = 358.7°) suggest weak interaction with the carbon
dioxide molecule. The third transition state with similar energy
values than the two other transition states was also located
starting from intermediate IM5, (TS1c −1.8 (26.3) kcal·
mol−1). It was found to occur by a simultaneous delivery of a
boronium/hydridoborate ion pair which would be formed from
the transfer of the hydride from one B-catechol moiety to the
other. In all these intermediates, the formaldehyde adduct plays
an important role, either by activating the borane or the carbon
dioxide. Once the hydride and the Bcat moieties are transferred
to generate the catBOC(O)H species, the B−O bond that was
broken in the transition state (from the Bcat in pathway A and
from B−OCH2 in pathways B and C) is regenerated to reform
species 1b·CH2O.
Scheme 2. Labeling Experiments Carried out with the
Phosphine−Boranes Formaldehyde Adducts 1b·CH2O and
2b·CH2O
studied using computational chemistry. In order to compare the
energies with our previous computational studies, calculations
were performed at the same level of theory (B97D/6-31G**)
with the solvation effects (benzene) accounted for by the SMD
model. As shown in Figure 2, several possible adducts could be
optimized where the borane interacts with one of the oxygen
on the phosphine−boranes species, notably from the catechol
and the aldehyde moieties. Although the previously reported
formaldehyde adduct remains the most thermodynamically
stable adduct, many other minima could be located on the
energy surface. Unsurprisingly, all the isomers have very similar
energy values, making difficult the identification of one ground-
state structure, especially when keeping in account the
uncertainty of the method ( 5 kcal·mol−1), and the possible
involvement of fluxional processes and conformational changes.
Intrigued by the important number of isomers identified and
by the possibility for the aldehyde adduct to be an active
catalyst rather than a resting state, pathways for CO2 reduction
were investigated with 1b·CH2O acting as a catalyst. In the
study of the reduction of CO2 by 1b, the lowest energy barriers
were found upon simultaneous activation of the catecholborane
by the Lewis base and of carbon dioxide by the Lewis acid.
Direct reduction of carbon dioxide, once bound in a classical
fashion by the phosphine and the borane of the catalyst, was
found to require a much higher barrier (Scheme 3).21
Comparing these transition states with the direct reduction
of formaldehyde (TS 1d, 22.3 (38.8 kcal·mol−1)) supports the
experimental observation that formaldehyde remains bound to
the ambiphilic framework during catalysis. The transition states
TS1a−TS1c were found to have free energies that are
respectively 9.7, 8.7, and 12.0 kcal·mol−1 lower than the
activation by the phosphine that was originally reported.
Although it is likely that the initial formation of the aldehyde
occurs through the phosphine−borane mechanism, the results
reported herein show that the oxygen atoms are more potent
than phosphorus in the activation of hydroboranes.
After a thorough investigation of the possibilities for 1b·
CH2O to act as a catalyst for CO2 reduction, it was possible to
Figure 2. Some of the calculated adducts between the boranes ([B]H = HBcat) and species 1b·CH2O with calculated free enthalpy and free energy
(kcal·mol−1) computed at B97D/6-31G** with the experimental solvent (benzene) accounted for the SMD model. The * indicates that two minima
were observed for two rotamers (see the SI).
2516
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520

ACS Catalysis
Research Article
a
Scheme 3. Previously reported transition states (and Energies in Kilocalories Per Mole) for the Hydroboration of CO2 by 1b
a
H[B] = catecholborane.21
Figure 3. Transition state structures for the hydroboration of CO2 by 1b·CH2O and HBcat. Calculated free enthalpy and free energy (kcal·mol−1)
computed at B97D/6-31G** with the experimental solvent (benzene) accounted for the SMD model.
Catalytic Efficiency of Aldehyde Adducts in the
Reduction of CO2 using BH3·SMe2. The efficiency of the
aldehyde adducts was then evaluated in the reduction of CO2
using the dimethylsulfide−borane adduct as a reducing agent.
BH3·SMe2 is a reagent of choice for the hydroboration of CO2
since it is less reactive and dangerous than the diborane reagent
B2H6 while possessing high hydrogen content by having three
transferable hydrides per boron atom. Interestingly, no
transition metal system has been reported to use BH3 adducts
for the reduction of carbon dioxide and only a limited number
of organocatalysts can do such reaction, including 1b that
proved the most active to date, with TOF of 973 h−1 at 70
°C.20 To follow up on known literature, the values of TON and
TOF were calculated according to the number of hydrogen
atoms transferred on CO2.
The catalytic reduction of CO2 was carried out using 1a−1d
as catalysts (Table 2, entries 1−4). In these experiments, about
2 atm of CO2 was added to a 8 mM solution of the catalyst in
C6D6 containing 100 equiv of BH3·SMe2 in a J-Young NMR
tube. As can be observed, both species 1a and 1b showed good
activities, reaching TOF of 257 and 242 h−1 in the first hour of
reaction (Table 2, entries 1 and 2, respectively). There is a
significant decrease in activity when the pinacol derivative was
used, since only a TOF of 43 h−1 was observed (Table 2, entry
3), whereas catalyst 1d did not show significant activity (Table
2, entry 4). While the reduction of CO2 using BH3·SMe2 and
1b as catalyst proceeds rapidly at 70 °C, it was shown to be
much slower at room temperature. As it was reported, the
spectroscopic monitoring of the latter reaction reveals that it
suffers from a long induction period before reaching its peak
2517
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520

ACS Catalysis
Research Article
Table 2. Catalytic Hydroboration of 2 atm of CO2 Using
a
BH3·SMe2
b
b
entry
catalyst
1a
time (min)
T (°C)
TON
TOF (h−1
)
1
60
67
60
60
35
30
90
30
30
90
70
70
70
70
25
20
20
70
70
70
257
271
43
257
242
43
220a
3
1b
1c
4
1d
<1
<1
5
1b·CH2O
2b·CH2O
2b·CH2O
2b·CH2O
2b·CH2O
2b·CH2O
84
144
228
136
594
873
670
6
114
204
297
435
1005
Figure 4. Number of turnovers for the reduction of CO2 in
[B(OMe)O]n in the presence of 100 equiv of BH3·SMe2 as a
7
8
■
●
c
reductant using 9 mM solution of ( ) 1b·CH2O and ( ) 1b in C6D6.
The reaction was carried out at room temperature using about 1−2
atm of CO2 in a J-Young NMR tube. The exact pressure could not be
measured, but under the loading present it is expected that CO2 will be
the limiting reagent since 1.5 atm is needed to obtain 100% yield (300
TON), explaining the lower rate at high conversion.
9
c
10
a
Reaction conditions: In a J-Young NMR tube, 530 μmol of BH3·
SMe2 (56.4 μL) was added to a mixture of 5.3 μmol of catalyst (∼0.6
mL of stock solution at ∼8.8 mmol/L) and an internal standard of
hexamethylbenzene (2 mg, 3.3 g/L) in C6D6 for entries 1−5, but in
CDCl3 for entries 6−10. The final concentration of catalyst is ∼8.1
mM. The J-young NMR tube was placed under 2 atm of CO2 and was
heated at 70 °C. The catalysis was followed by 1H NMR spectroscopy.
b
tube conditions surpass the catalytic activity of catalyst 1b.
Decreasing catalyst loading to 1.8 mM but keeping all the other
parameters identical, it was possible to observe that catalyst 2b·
CH2O could give a TON of 435 in the first 30 min (TOF of
873 h−1), while running the reaction for 150 min increased the
TON to 1005 (Table 2, entries 9 and 10). These results are of
significant interest since some of the formaldehyde adducts
exhibit higher stability than their phosphine−borane precur-
sors, notably in the case of 2b. Also, the high activity at ambient
temperature and the absence of induction period contrasts
drastically with all systems reported to date to reduce CO2
using BH3 derivatives which usually need induction periods and
higher temperatures to proceed efficiently.
Turnover numbers (TONs) and turnover frequencies for the
formation of [B(OMe)O]n according to the number of hydrogen
atoms transferred to CO2 based on NMR integration of the
corresponding resonance with internal standard. The reaction was
performed under the same conditions but with a catalyst concentration
of ∼1.8 mM.
c
turnover frequency. The intermediates of the reaction were
monitored using H and 31P NMR spectroscopy. As long as
1
species 1b·BH3 is the only one in solution, no significant
reduction process is taking place. However, as soon as the
presence of the formaldehyde adduct was observed, the
catalytic reduction rate was shown to increase significantly,
confirming that 1b·CH2O is playing a significant role on the
catalytic reduction.
When carrying out the reduction of CO2 in benzene at room
temperature with 1b·CH2O as the catalyst, the rapid reduction
of CO2 was observed in the first minutes of reaction. In fact,
within the 5 min needed for the acquisition of the first NMR
spectrum, it was possible to ascertain a nominal TON value of
26 for the conversion of CO2 to [MeOBO]n, which can be
associated with an initial TOF of >520 h−1 (Figure 4). This
reaction rate is significantly superior to previously observed
catalytic activities at ambient temperature when using 1b as the
starting catalyst, especially when considering that a significant
amount of the catalyst was not dissolved at this time. After 35
min, a TON of 84 was observed, corresponding to a TOF of
144 h−1 (Table 2, entry 5).
The formaldehyde adduct 2b·CH2O was also evaluated with
BH3·SMe2 as a reductant. The catalytic tests were performed in
CDCl3 to ensure good solubility. Accordingly, it was found that
a 1 mol % of species 2b·CH2O could give over 99% of the
conversion expected within 30 min, giving TON of 114 and
TOF over 228 h−1 at ambient temperature, which is slightly
higher that the activity of 1b·CH2O (Table 2, entry 6). Letting
the reaction run for 90 min increased the TON value to 204,
while reducing the TOF to 136 h−1 (Table 2, entry 7). Looking
at the reactivity at 70 °C it was possible to observe that catalyst
2b·CH2O converted almost quantitatively BH3·SMe2 to
methoxyboranes in 30 min, giving TON and TOF of 297
and 594 h−1 (Table 2, entry 8), respectively, which under NMR
It should be noted that the increase in activity of the
formaldehyde adducts compared to the ambiphilic phosphine−
boranes is really counterintuitive, especially when comparing
with the usual chemistry of Frustrated Lewis Pairs. Indeed, in
latter systems, the activation of substrates such as molecular
hydrogen and carbon dioxide is possible by preventing the
formation of Lewis adducts between the two components of
the FLP. Although such process is particularly important in
hydrogen activation, it does not seem to play a similar role in
the catalytic reduction of carbon dioxide. Indeed, the most
important factor seems to be the activation of the reducing
agent, which is done in the current system by interaction of the
hydroborane with O atoms of the formaldehyde adduct;
something that can be related to the activity of several borates
−
in the reduction of carbon dioxide, including notably the BH4
salts that are well-known to reduce CO2 in a stoichiometric
fashion. It seems however that the simultaneous activation of
the carbon dioxide by either hydrogen bonding or by a weak
Lewis acid such as the boroncatecholate moiety is helping to
lower significantly the transition state energy. As it was stated
before, such weak Lewis acid can also prevent the formation of
very stable formate adducts which might make difficult the
release of the reduced species and high catalytic turnovers. The
formaldehyde adducts 1b·CH2O and 2b·CH2O presented in
this system have also the advantage to be quite stable to both
air and water, which are significant advantages compared to
some ambiphilic species, notably 2b.
2518
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520

ACS Catalysis
Research Article
Climate Change (IPCC), Cambridge University Press: New York,
2005.
(2) Olah, G. A.; Goeppert, A.; Surya Prakash, G. K. J. Org. Chem.
CONCLUSION
■
Ambiphilic species reported in this study all show the ability to
hydroborate catalytically CO2 to CH3OBR2 derivatives using
catecholborane and BH3·SMe2. However, in the presence of
catecholborane, all species undergo transformations to generate
respectively 1b·CH2O or 2b·CH2O, which are believed to be
the active catalysts in the latter reaction. Although the exact
mechanism for such transformation is unknown, derivatives of
1b and 2b seem to be the favored species in such system. The
latter reaction also puts in evidence the importance of possible
redistribution occurring in such systems with the B−O bonds
being quite kinetically labile. As it was demonstrated by using
1b·13CH2O and 2b·13CH2O as catalysts, the interaction
between the formaldehyde and the phosphine−borane moiety
seems to remain intact throughout catalysis. DFT modeling of
possible transition states show that the ambiphilic activation
remains possible, as it was proposed for species 1b, but this
time one oxygen atom, either from the formaldehyde or the
catechol moieties, acts as Lewis base to activate the reducing
agent, whereas the CO2 is activated either by one borane or by
hydrogen bonding with one hydrogen of the formaldehyde
adduct. Species 2b·CH2O was shown to be the best catalyst at
room temperature reported for this system, obtaining TOF of
228 h−1. Although it represents a much harder challenge, we are
currently interested in using such concept for the hydro-
genation of carbon dioxide.
2009, 74, 487−498.
(3) Beyond Oil and Gas: The Methanol Economy; Olah, G. A., Alain
Goeppert, G. K.; Prakash, S., Eds.; Wiley-VCH: Weinheim, 2006.
(4) (a) Tanaka, R.; Yamashita, M.; Nozaki, K. J. Am. Chem. Soc. 2009,
131, 14168−14169. (b) Federsel, C.; Jackstell, R.; Beller, M. Angew.
Chem., Int. Ed. 2010, 49, 6254−6257. (c) Schaub, T.; Paciello, R. A.
Angew. Chem., Int. Ed. 2011, 50, 7278−7282.
(5) (a) Matsuo, T.; Kawaguchi, H. J. Am. Chem. Soc. 2006, 128,
12362−12363. (b) Federsel, C.; Boddien, A.; Jackstell, R.; Jennerjahn,
R.; Dyson, P. J.; Scopelliti, R.; Laurenczy, G.; Beller, M. Angew. Chem.,
Int. Ed. 2010, 49, 9777−9780. (c) Langer, R.; Diskin-Posner, Y.;
Leitus, G. W.; Shimon, L. J.; Ben-David, Y.; Milstein, D. Angew. Chem.,
Int. Ed. 2011, 50, 9948−9952. (d) Schmeier, T. J.; Dobereiner, G. E.;
Crabtree, R. H.; Hazari, N. J. Am. Chem. Soc. 2011, 133, 9274−9277.
́
(e) Park, S.; Bezier, D.; Brookhart, M. J. Am. Chem. Soc. 2012, 134,
11404−11407. (f) Huff, C. A.; Sanford, M. S. ACS Catal. 2013, 3,
2412−2416. (g) Jeletic, M. S.; Mock, M. T.; Appel, A. M.; Linehan, J.
C. J. Am. Chem. Soc. 2013, 135, 11533−11536. (h) Zhang, L.; Cheng,
J.; Hou, Z. Chem. Commun. 2013, 49, 4782−4784.
(6) (a) Bontemps, S.; Vendier, L.; Sabo-Etienne, S. Angew. Chem., Int.
Ed. 2012, 51, 1671−1674. (b) Bontemps, S.; Sabo-Etienne, S. Angew.
Chem., Int. Ed. 2013, 52, 10253−10255. (c) Bontemps, S.; Vendier, L.;
Sabo-Etienne, S. J. Am. Chem. Soc. 2014, 136, 4419−4425.
(d) Metsanen, T. T.; Oestreich, M. Organometallics 2015, 34, 543−
̈
546.
(7) (a) Chakraborty, S.; Zhang, J.; Krause, J. A.; Guan, H. J. Am.
Chem. Soc. 2010, 132, 8872−8873. (b) Balaraman, E.; Gunanathan, C.;
Zhang, J.; Shimon, L. J. W.; Milstein, D. Nature Chem 2011, 3, 609 −
614. (c) Huff, C. A.; Sanford, M. S. J. Am. Chem. Soc. 2011, 133,
18122−18125. (d) Wesselbaum, S.; von Stein, T.; Klankermayer, J.;
Leitner, W. Angew. Chem., Int. Ed. 2012, 51, 7499−7502.
(8) (a) Matsuo, T.; Kawaguchi, H. J. Am. Chem. Soc. 2006, 128,
ASSOCIATED CONTENT
■
S
* Supporting Information
The following file is available free of charge on the ACS
Publications website at DOI: 10.1021/acscatal.5b00189.
́
12362−12363. (b) Park, S.; Bezier, D.; Brookhart, M. J. Am. Chem. Soc.
Detailed experimental procedures and additional DFT
information (other calculated reaction pathways, Carte-
sian coordinates, free enthalpies, and energies) (PDF)
2012, 134, 11404−11407. (c) Mitton, S. J.; Turculet, L. Chem. - Eur. J.
2012, 48, 15258−15262. (d) Berkefeld, A.; Piers, W. E.; Parvez, M.;
Castro, L.; Maron, L.; Eisenstein, O. Chem. Sci. 2013, 4, 2152−2162.
(9) LeBlanc, F. A.; Piers, W. E.; Parvez, M. Angew. Chem. 2014, 126,
808−811.
AUTHOR INFORMATION
■
(10) (a) Laitar, D. S.; Muller, P.; Sadighi, J. P. J. Am. Chem. Soc. 2005,
̈
Corresponding Authors
127, 17196−17197. (b) Kleeberg, C.; Cheung, M. S.; Lin, Z.; Marder,
T. B. J. Am. Chem. Soc. 2011, 133, 19060−19063. (c) Zhao, H.; Lin, Z.;
Marder, T. B. J. Am. Chem. Soc. 2006, 126, 15637−15643.
(11) (a) Khandelwal, M.; Wehmschulte, R. J. Angew. Chem., Int. Ed.
2012, 51, 7323−7326. (b) Wehmschulte, R. J.; Saleh, M.; Powell, D. R.
Organometallics 2013, 32, 6812−6819.
*E-mail: frederic.fontaine@chm.ulaval.ca (F.-G.F.).
*E-mail: bouhadir@chimie.ups-tlse.fr. Fax: 33 (0)5 61 55 82
04. Tel.: 33 (0)5 61 55 68 03 (G.B.).
*E-mail: dbouriss@chimie.ups-tlse.fr (D.B.).
Author Contributions
(12) Schafer, A.; Saak, W.; Haase, D.; Muller, T. Angew. Chem., Int.
∥R.D. and M.-A.L. contributed equally to this work.
̈
̈
Ed. 2012, 51, 2981−2984.
Notes
(13) (a) Momming, C.; Otten, M.; Kehr, E. G.; Frohlich, R.;
̈
̈
The authors declare no competing financial interest.
Grimme, S.; Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2009, 48,
6643−6646. (b) Appelt, C.; Westenberg, H.; Bertini, F.; Ehlers, A. W.;
Slootweg, J. C.; Lammertsma, K.; Uhl, W. Angew. Chem., Int. Ed. 2011,
50, 3925−3928. (c) Boudreau, J.; Courtemanche, M.-A.; Fontaine, F.-
G. Chem. Commun. 2011, 47, 11131−11133. (d) Peuser, I.; Neu, R. C.;
ACKNOWLEDGMENTS
■
This work was supported by the National Sciences and
Engineering Research Council of Canada (NSERC, Canada)
and the Centre de Catalyse et Chimie Verte (Quebec). M.-A.
C., M.-A. L, and N.B. would like to thank NSERC and FQRNT
for scholarships. The Centre National de la Recherche
Zhao, X.; Ulrich, M.; Schirmer, B.; Tannert, J. A.; Kehr, G.; Frohlich,
̈
R.; Grimme, S.; Erker, G.; Stephan, D. W. Chem. - Eur. J. 2011, 17,
9640−9650. (e) Hounjet, L. J.; Caputo, C. B.; Stephan, D. W. Angew.
Chem., Int. Ed. 2012, 51, 4714 −4717. (f) Takeuchi, K.; Stephan, D. W.
Chem. Commun. 2012, 48, 11304−11306. (g) Theuergarten, E.;
́
Scientifique (CNRS), the Universite Paul Sabatier (UPS),
Schlosser, J.; Schluns, D.; Freytag, M.; Daniliuc, C. G.; Jones, P. G.;
̈
̈
and French MESR (Ph.D. grant to R.D.) are acknowledged for
financial support of this work. We would like to thank R.
Nadeau for the design of the TOC graphic.
Tamm, M. Dalton Trans. 2012, 41, 9101−9110. (h) Lu, Z.; Liu, Y. J.;
Wang, Y.; Lin, J.; Li, Z. H.; Wang, H. Organometallics 2013, 32, 6753−
6758. (i) Barry, B. M.; Dickie, D. A.; Murphy, L. J.; Clyburne, J. A. C.;
Kemp, R. A. Inorg. Chem. 2013, 52, 8312−8314.
REFERENCES
(14) (a) Ashley, A. E.; Thompson, A. L.; O’Hare, D. Angew. Chem.,
■
(1) Carbon Dioxide Capture and Storage; Metz, B., Davidson, O., de
Coninck, H., Loos, M., Meyer, L., Eds.; Intergovernmental Panel on
Int. Ed. 2009, 48, 9839−9843. (b) Men
́
ard, G.; Stephan, D. W. J. Am.
Chem. Soc. 2010, 132, 1796−1797. (c) Men
́
ard, G.; Stephan, D. W. J.
2519
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520

ACS Catalysis
Research Article
Am. Chem. Soc. 2010, 132, 1796−1797. (d) Roy, L.; Zimmerman, P.
M.; Paul, A. Chem. - Eur. J. 2011, 17, 435−439. (e) Men
́
ard, G.;
Stephan, D. W. Dalton Trans. 2013, 42, 5447−5453. (f) Men
́
ard, G.;
Gilbert, T. M.; Hatnean, J. A.; Kraft, A.; Krossing, I.; Stephan, D. W.
Organometallics 2013, 32, 4416−4422.
(15) (a) Berkefeld, A.; Piers, W. E.; Parvez, M. J. Am. Chem. Soc.
2010, 132, 10660−10661. (b) Wen, M.; Huang, F.; Lu, G.; Wang, Z.-
X. Inorg. Chem. 2013, 52, 12098 −12107.
(16) Riduan, S. N.; Zhang, Y.; Ying, J. Y. Angew. Chem., Int. Ed. 2009,
48, 3322−3325.
(17) Wang, T.; Stephan, D. W. Chem. - Eur. J. 2014, 20, 3036−3039.
́
(18) (a) Das Neves Gomes, C.; Jacquet, O.; Villiers, C.; Thuery, P.;
Ephritikhine, M.; Cantat, T. Angew. Chem., Int. Ed. 2012, 51, 187−190.
(b) Das Neves Gomes, C.; Blondiaux, E.; Thuery, P.; Cantat, T. Chem.
are, M.-A.; Courtemanche, M.-
́
- Eur. J. 2014, 20, 7098−7106. (c) Leg
́
́
A.; Fontaine, F.-G. Chem. Commun. 2014, 50, 11362−11365. (d) For
a related system deriving from the [ClC(PPh2NMes)Li] carbenoid,
́
see: Ho, S. Y.-F.; So, C.-W.; Saffon-Merceron, N.; Mezailles. Chem.
Commun. 2015, 51, 2107−2110. (e) Fujiwara, K.; Yasuda, S.; Mizuta,
T. Organometallics 2014, 33, 6692−6695. (f) Blondiaux, E.; Pouessel,
J.; Cantat, T. Angew. Chem., Int. Ed. 2014, 53, 12186−12190.
(19) Courtemanche, M.-A.; Larouche, J.; Leg
Maron, L.; Fontaine, F.-G. Organometallics 2013, 32, 6804−6811.
(20) (a) Courtemanche, M.-A.; Legare, M.-A.; Maron, L.; Fontaine,
F.-G. J. Am. Chem. Soc. 2013, 135, 9326−9329. (b) Fontaine, F.-G.;
Courtemanche, M.-A.; Legare, M.-A. Chem. - Eur. J. 2014, 20, 2990−
2996.
(21) Courtemanche, M.-A.; Legare, M.-A.; Maron, L.; Fontaine, F.-G.
́ ́
are, M.-A.; Wenhua, B.;
́
́
́
́
́
́
J. Am. Chem. Soc. 2014, 136, 10708−10717.
(22) (a) Bontemps, S.; Bouhadir, G.; Miqueu, K.; Bourissou, D. J.
Am. Chem. Soc. 2006, 128, 12056−12057. (b) Bontemps, S.; Bouhadir,
G. D.; Apperley, C.; Dyer, P. W.; Miqueu, K.; Bourissou, D. Chem. -
Asian J. 2009, 428−435. (c) Bouhadir, G.; Amgoune, A.; Bourissou, D.
Phosphine-Boranes and Related Ambiphilic Compounds: Synthesis,
Structure, and Coordination to Transition Metals. Advances in
Organometalic Chemistry; Elsevier, 2010; Vol. 58, pp 1−107.
(23) (a) Porcel, S.; Bouhadir, G.; Saffon, N.; Maron, L.; Bourissou, D.
́
Angew. Chem., Int. Ed. 2010, 49, 6186−6189. (b) Basle, O.; Porcel, S.;
Ladeira, S.; Bouhadir, G.; Bourissou, D. Chem. Commun. 2012, 48,
4495−4497. (c) Moebs-Sanchez, S.; Bouhadir, G.; Saffon, N.; Maron,
L.; Bourissou, D. Chem. Commun. 2008, 3435−3437. (d) Bebbington,
M. W. P.; Bontemps, S.; Bouhadir, G.; Bourissou, D. Angew. Chem., Int.
Ed. 2007, 46, 3333−3336.
(24) Westcott, S. A.; Blom, H. P.; Marder, T. B.; Baker, R. T.;
Calabrese, J. C. Inorg. Chem. 1993, 32, 2175−2182.
(25) The exact protocol described in the experimental section was
carried out for all catalytic reactions. However, since the volume of the
NMR tubes differ slightly and some small difference in temperature or
pressure can occur when the CO2 is transferred in the J-Young tube,
the exact pressure cannot be known. In a typical experiment with
HBcat the CO2 will be in significant excess, but in the case of the
reactions with BH3·SMe2, about 1.5 atm is needed for full conversion,
which might alter the activity after more than 200 TON.
(26) Beley, M.; Collin, J.-P.; Ruppert, R.; Sauvage, J.-P. J. Am. Chem.
Soc. 1986, 108, 7461−7467.
(27) Schmeier, T. J.; Dobereiner, G. E.; Crabtree, R. H.; Hazari, N. J.
Am. Chem. Soc. 2011, 133, 9274−9277.
2520
DOI: 10.1021/acscatal.5b00189
ACS Catal. 2015, 5, 2513−2520