Decamethylscandocinium-hydrido-(perfluorophenyl)borate: Fixation and tandem tris(perfluorophenyl)borane catalysed deoxygenative hydrosilation of carbon dioxide

Article abstract of DOI:10.1039/c3sc50145k

This work was directed at studying the capability of structurally defined, strongly Lewis-acidic metal centers to effect catalytic reductive fixation of the small molecule substrate CO₂. Exposing solutions or solid samples of the ion pair [Cp*₂Sc][HB(C₆F₅) ₃] 1_CIP, in which the highly electrophilic decamethyl-scandocene cation and [HB(C₆F₅)]^- as a potentially reactive source of hydride equivalents are associated, to CO ₂ selectively produces ion pair [Cp*₂Sc][HCO ₂B(C₆F₅)₃] 2_CIP. The results of solution and solid state structural analysis of 2_CIP imply ionic association of [Cp*₂Sc]⁺ and [HCO ₂B(C₆F₅)₃]^- rather than B(C₆F₅)₃-adduct formation to neutral Cp*₂Sc-formate. In the presence of B(C₆F ₅)₃ co-catalyst and excess triethylsilane, the formation of 2_CIP from 1_CIP initiates the catalytic deoxygenative hydrosilation of CO₂ to CH₄. The roles of ion pairs 1 and 2, borane co-catalyst, and silane in the catalytic reaction were studied mechanistically by NMR spectroscopy. Intermediately formed 3,3,7,7-tetraethyl-3, 7-disila-4,6-dioxanonane product was found to exert an accelerating effect on the overall reaction rate by promoting [HCO₂B(C₆F ₅)₃]^- dissociation to give 2_SIP through formation of separated ion pairs [Cp*₂Sc(κ ²-(Et₃SiO)₂CH₂)][HCO ₂{B(C₆F₅)₃}_n], n = 1, 2. DFT calculations show that the formation 2_CIP from the reaction of 1_CIP with CO₂ is exoergic and without significant energy barriers. This work lays the basis for future studies of reactive ion pairs of this kind in the context of small molecule chemistry.

Full text of DOI:10.1039/c3sc50145k

Chemical Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Decamethylscandocinium-hydrido-(perfluorophenyl)-
borate: fixation and tandem tris(perfluorophenyl)-
borane catalysed deoxygenative hydrosilation of
carbon dioxide†
Cite this: Chem. Sci., 2013, 4, 2152
a
a
Andreas Berkefeld,‡ Warren E. Piers, Masood Parvez,a Ludovic Castro,bc
Laurent Maron
*
bc
d
*
*
and Odile Eisenstein
This work was directed at studying the capability of structurally defined, strongly Lewis-acidic metal centers
to effect catalytic reductive fixation of the small molecule substrate CO2. Exposing solutions or solid
samples of the ion pair [Cp*2Sc][HB(C6F5)3] 1CIP, in which the highly electrophilic decamethyl-
scandocene cation and [HB(C6F5)]ꢀ as a potentially reactive source of hydride equivalents are associated,
to CO2 selectively produces ion pair [Cp*2Sc][HCO2B(C6F5)3] 2CIP. The results of solution and solid state
structural analysis of 2CIP imply ionic association of [Cp*2Sc]+ and [HCO2B(C6F5)3]ꢀ rather than B(C6F5)3-
adduct formation to neutral Cp*2Sc-formate. In the presence of B(C6F5)3 co-catalyst and excess
triethylsilane, the formation of 2CIP from 1CIP initiates the catalytic deoxygenative hydrosilation of CO2
to CH4. The roles of ion pairs 1 and 2, borane co-catalyst, and silane in the catalytic reaction were
studied mechanistically by NMR spectroscopy. Intermediately formed 3,3,7,7-tetraethyl-3,7-disila-4,6-
dioxanonane product was found to exert an accelerating effect on the overall reaction rate by
promoting [HCO2B(C6F5)3]ꢀ dissociation to give 2SIP through formation of separated ion pairs
[Cp*2Sc(k2-(Et3SiO)2CH2)][HCO2{B(C6F5)3}n], n ¼ 1, 2. DFT calculations show that the formation 2CIP from
the reaction of 1CIP with CO2 is exoergic and without significant energy barriers. This work lays the basis
for future studies of reactive ion pairs of this kind in the context of small molecule chemistry.
Received 16th January 2013
Accepted 24th February 2013
DOI: 10.1039/c3sc50145k
www.rsc.org/chemicalscience
critical mechanistic commonality that emerges is the need to
overcome thermodynamically strong element-oxygen bonds at
Introduction
The development of catalytic processes that allow for use of CO2 as
a C1 synthon for the production of value added chemicals or
recycling of CO2 into usable fuels on an industrial scale has
challenged the ingenuity of chemists for several decades.1–6 In this
context, a plethora of thermodynamically feasible schemes for the
up-conversion of CO2 into products such as formic acid, meth-
anol, or CO have been outlined and discussed in the literature. A
some point in a catalytic cycle, a step that in practice can be also
be associated with substantial kinetic barriers. As a consequence,
achieving substrate turnover at appreciable overall rates oen
necessitates that the reactions be run at elevated temperatures
and pressures,7–10 or in the presence of highly energetic, oxophilic
transmetallating co-reactants11,12 such as aluminum13,14 or
zinc13,15–20 alkyls, (di)boranes and stannanes,21–28 or silanes.29–33
A widely pursued strategy to introduce CO2 into a catalytic
process takes advantage of its generally rapid insertion into
reactive metal–hydrogen or –alkyl bonds.34–37 Late transition
metals have been predominantly studied in this context in light
of their comparatively low oxophilicity, which manifests in
metal–oxygen linkages amenable to catalytic turnover under
less forcing conditions. Recent reports on the catalytic hydro-
genation of CO2 to formic acid using molecular Ir,38,39 Ru,9,10,40–44
Fe,45,46 and Co47,48 based catalysts have illustrated the potential
of this strategy.35,49,50 A notable feature of these catalyst systems
is their ability to form stable monomeric metal-hydrides
capable of activating the CO2 substrate.
aDepartment of Chemistry, University of Calgary, 2500 University Drive NW, Calgary,
Alberta, Canada T2N 1N4. E-mail: wpiers@ucalgary.ca; Tel: +1 403-220-5746
b
´
LPCNO, Universite de Toulouse, INSA, UPS, LPCNO, 135 avenue de Rangueil,
F- 31077 Toulouse, France
cCNRS, LPCNO, F-31077 Toulouse, France
d
´
Institut Charles Gerhardt, UMR 5253 CNRS Universite Montpellier, cc 1501, Place
E. Bataillon, F-34095, Montpellier, France
† Electronic supplementary information (ESI) available: Crystallographic data
les for 2CIP and 3, additional experimental and spectroscopic details and
tables of coordinates, energies E and Gibbs energies G in a.u. for all calculated
structures and the full list of authors for ref. 1 and 103. CCDC 914770 and
914771. For ESI and crystallographic data in CIF or other electronic format see
DOI: 10.1039/c3sc50145k
In contrast to their late metal counterparts, monomeric,
base-free early metal-hydrides are very reactive. They tend to be
¨
¨
‡ Present address: Technische Universitat Berlin, Institut fur Chemie Sekr. C 2,
Straße des 17. Juni 13, 10623 Berlin, Germany.
2152 | Chem. Sci., 2013, 4, 2152–2162
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
transient species that, if not kinetically stabilised through bulky the form of ion pair 1CIP that rapidly reacts with CO2 to
ancillary ligands, show a high propensity to form less reactive replenish 2CIP. Overall, tandem 2CIP–B(C6F5)3 catalyzes the
dimeric or polyhydride cluster compounds.51–55 Reports on the deoxygenative hydrosilation of CO2 to CH4 in the presence of
catalytic, or pseudo catalytic conversion of CO2 into reduced Et3SiH as sacricial reductant and oxygen acceptor. Mecha-
products by early metal compounds are thus relatively nistic details of the reaction are presented that shed light on the
scarce.56–60
Kawaguchi zirconium system63 and show similarities to late
A notable example is the deoxygenative hydrosilation of CO2 transition metal-based systems very recently reported by Broo-
to CH4 catalyzed by a tandem61,62 cationic Zr complex–tris(per- khart et al.71 and Mitton and Turculet,72 illustrating the versa-
uorophenyl)borane system, I, reported in 2006 by Matsuo and tility of this approach.73
Kawaguchi.63 Here, silane serves to act as the sacricial reduc-
tant and oxygen acceptor, rendering the reaction thermody-
Results and discussion
Activation of CO2 by 1CIP
namically favorable. Although the observed activity is moderate,
with a reported 150 TO hꢀ1 (moles Si–H per mole of Zr) at
ambient temperatures, this work demonstrated that breaking
Contact ion pair 1CIP was prepared and fully characterised as
described previously73 from the reaction of deca-
methylscandocene chloride74 and B(C6F5)3 in the presence of a
slight excess of triethylsilane. Solution NMR spectroscopy and
X-ray diffraction data established the association of the ionic
components via dynamic k2-F bonds between the ortho and
meta uorine atoms of the hydridoborate counterion and the
scandocene cation.73 The B–H moiety is not involved in the ion
pair association (Scheme 1). This bonding motif most likely
results from two factors, (i), the small radius of the Sc3+ ion,
thermodynamically strong early transition metal–oxygen bonds
is kinetically feasible and sustains catalytic turnover. Besides
facilitating the abstraction of a benzyl ligand to form the reac-
tive bis-phenoxide-aryl ether ligand-stabilised cationic Zr
monobenzyl complex I, the B(C6F5)3 co-catalyst also mediates
the catalytic activation of silane Si–H bonds.64–66 On the basis of
the limited mechanistic data available, the authors speculated
that coordination to the cationic Zr center activates CO2 toward
reduction by the aforementioned reactive borane–silane
species. Given that formation of cationic Zr-carboxylates via
alkyl migration to coordinated CO2 is both thermodynamically
and kinetically highly favorable,67 the ability of the B(C6F5)3–
HSiR3 system to convert these presumed intermediate species
to fully reduced products is signicant. Furthermore, the
inability of B(C6F5)3–HSiR3 to hydrosilate CO2 in the absence of
a cationic co-catalyst underscores the need for a tandem cata-
lytic system for this transformation.
˚
0.732 A, and, (ii), the thermodynamic preference of a Sc–F bond,
141 kcal molꢀ1, over a Sc–H bond, ꢂ43 kcal molꢀ1
.
In the
75
compositionally related ion pair [Cp*2Sm][HB(C6F5)3], the dec-
amethylsamarocene cation binds [HB(C6F5)3]ꢀ via two o-F–C
contacts and the B–H moiety.76 This presumably reects the
signicantly larger radius of the Sm3+ ion, 0.964 A, as compared
˚
to Sc3+, given that the metal–F bond strengths of Sm, 135 kcal
molꢀ1 75 and Sc are comparable.
,
As we have convincingly shown previously,73 the reactivity of
1CIP contrasts with that of related ion pairs in that highly reactive
Cp*2ScH74,77 is not present in chemically meaningful quantities in
solutions of 1CIP. For comparison, Marks and co-workers repor-
ted on the high propensity of the ion pair [Cp*2ZrH][HB(C6F5)3]
to undergo H–D isotope scrambling into both the Zr–H and the
B–H moiety in d6-benzene solution, which involves reversible
intramolecular hydride-migration between zirconium and
boron.78,79 In another example, Mindiola and co-workers reported
on the competence of the ion pair [k2-(ArNC(Me)CHC(Me)NAr)
Sc(NHAr)][HBEt3], Ar ¼ 2,6-(iPr)2C6H3, to stoichiometrically react
with the C–O bonds of ethers and benzophenone.80 Based on
experimental observations, the authors inferred that the highly
Here we report on the reactivity of the contact ion pair (CIP)
[Cp*2Sc][HB(C6F5)3] 1CIP, in which potentially reactive hydrido-
tris(pentauorophenyl)borate is partnered with the strongly
Lewis-acidic decamethylscandocene cation, for catalytic CO2
hydrosilation at ambient temperatures. Initial substrate xation
occurs within the spatially conned reactive pocket created by
the ionic components of 1CIP and results in the exclusive
formation of the ion pair [Cp*2Sc][HCO2B(C6F5)3] 2CIP. The high
reactivity of 1CIP toward CO2 contrasts that of [HB(C6F5)3]ꢀ
partnered with non Lewis-acidic [TMPH]+,68 (TMP ¼ 2,2,6,6-
tetramethylpiperidine), that requires heating to temperatures
ꢁ
>50 C in order to facilitate the formal B–H bond addition to
CO2 to form [TMPH]+[HCO2B(C6F5)3]ꢀ at an appreciable rate.69
In the presence of catalytic quantities of reactive borane–Et3SiH,
borato-formate is subject to rapid B–O bond hydrosilation to
form triethylsilylformate and regenerate [HB(C6F5)3]ꢀ,70 here in Scheme 1 Synthesis and properties of contact ion pair 2CIP
.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2152–2162 | 2153

View Article Online
Chemical Science
Edge Article
reactive neutral Sc-hydride parent complex plays a role in these
The bond distances in the Sc–O1–C21–O2–B fragment
reactions. Several experiments of this nature show conclusively suggest that the structure is closer to a zwitterionic species as
73
that such hydride migration is absent in the ion pair 1CIP
.
depicted in Scheme 1, as opposed to an alternate neutral reso-
The absence of reversible intramolecular hydride ligand nance structure in which the C]O bond is more localised
transfer from boron to scandium in 1CIP is important in light of between C21 and O2 (inset, Scheme 1). For example, the C21–O1
˚
˚
its observed high reactivity towards CO2. Exposure of a 45 mM bond (1.236(5) A) is shorter than the C21–O2 bond (1.272(5) A).
d5-bromobenzene solution of 1CIP to 1 atm of 13C-labeled CO2 at Further, the Sc–O1 bond of 2.088(3) A is slightly longer than a
ambient temperatures instantaneously and selectively afforded typical Sc-alkoxide distance,82–84 while the B–O2 length of
˚
˚
the ion pair decamethylscandocinium formato-tris(per- 1.560(5) A is somewhat shorter than a typical distance in
uorophenyl)borate, [Cp*2Sc][HCO2B(C6F5)3] 2CIP, as shown in datively bonded carbonyl adducts of B(C6F5)3.85 This contrasts
Scheme 1. Ion pair 1CIP also underwent reaction with CO2 in the with the delocalised bonding found in the more symmetrical
solid state, albeit more slowly. The IR spectrum of the recovered anion [HCO2{B(C6F5)3}2]ꢀ in which the observed bond distances
˚
solid (taken in a KBr pellet) is identical to that of 2CIP prepared are 1.586(3) and 1.583(3) A for the B–O and 1.255(3) and 1.268(5)
70
˚
independently in solution (Fig. S1†). A comparison of the IR A for the C–O bonds. The asymmetry in the bonding that
spectra of 2CIP with that of [TMPH][HCO2B(C6F5)3]69 (KBr) emphasises the zwitterionic structure is likely key to the
revealed a shi of the carbonyl band observed for 2CIP, 1622 observed activity of 2CIP as a CO2 hydrosilation catalyst
cmꢀ1, to longer wavenumbers by DnO]C ¼ 22 cmꢀ1. This indi- (vide infra).
cates a tighter bonding of formato-borate to the Lewis-acidic
Further scrutiny of the solution NMR data indicates that 2CIP
decamethylscandocene cation as compared to [TMPH]+. Reac- is in equilibrium with free B(C6F5)3 and neutral species
tion of d1-1CIP with CO2 shows that the hydrogen in the bridging {Cp*2ScOC(O)H}n. The nuclearity of the neutral Sc-formate
formato group derives from the hydridoborate anion.
species is unknown; its low concentration in solution favors a
1H and 13C NMR solution spectra of 2CIP showed resonances monomer, but the high electrophilicity of the metal center
at 8.77 and 172.9 ppm, respectively, characteristic of the renders monomeric Cp*2ScOC(O)H highly susceptible to
1
formate moiety. A JCH coupling constant of 216.6 Hz was dimerisation.86–89 Based on the degree of B(C6F5)3 dissociation
observed for this C–H moiety (Fig. S2†). A broad resonance determined from 19F NMR data an equilibrium constant Keq
ꢂ
1
(n ꢂ 630 Hz) at ꢀ0.7 ppm in the H coupled 11B NMR spec- 10ꢀ4 M can be estimated (Scheme 1). Observation of this equi-
½
trum, and a set of three sharp resonances in the 19F NMR librium suggests that the neutral resonance structure depicted
spectrum with a DdFm–p value of 6.8 ppm81 establish the pres- in the inset of Scheme 1 is also chemically relevant in this
ence of the [HCO2B(C6F5)3]ꢀ ion.69,70 The assignment of 2CIP to a compound.
contact ion pair in solution is supported by 1H–19F HOESY data
The reaction with 1CIP and CO2 was too fast to study by NMR
that revealed dipolar couplings between the formate proton and spectroscopy. Therefore, we probed its mechanism using DFT
the ortho- and meta-F nuclei, as well as between the Cp* protons computations; the derived Gibbs energy prole is shown in
and the ortho-F nuclei on the borate (Fig. S3†). The solid state Fig. 2. The reaction begins by the coordination of the CO2
molecular structure of 2CIP determined by single crystal X-ray molecule to the scandium center by one oxygen atom,90 with
diffraction is in agreement with the solution structure derived concomitant loss of one Sc–F interaction within the ion-pair. A
from the NMR data and is shown in Fig. 1.
similar slippage from k2-F to k1-F bonding was proposed in the
reaction of 1CIP with CO.73 Here, this step is endoergic by +8.5
kcal molꢀ1, slightly below the expected loss of entropy at room
temperature (10–15 kcal molꢀ1
)
indicating that CO2 can
91
coordinate to the scandium cation. The geometry of the k1-CO2
adduct is such that a transition state for hydride transfer from
boron to the electrophilic CO2 carbon is easily reached. At the
transition state (TS), the CO2 molecule is activated and adopts a
bent geometry with an O–C–O angle of 151ꢁ in order to develop
an empty p orbital at the carbon atom in order to prepare for
formation of the C–H bond. The B–H bond is elongated by
˚
0.08 A with respect to the free ion-pair but the C–H is not yet
˚
fully formed; the C–H distance is still quite long at 1.60 A. The
Gibbs energy barrier is only +12.7 kcal molꢀ1 above the reac-
tants, consistent with the observed kinetic facility of reaction.
From the TS, the system evolves to the formation of a neutral k2-
O formate intermediate and free borane, which remains in the
vicinity of the complex. This intermediate is very stable
(ꢀ15.5 kcal molꢀ1 with respect to the separated reactants)
because of the close proximity between the positively charged
Fig. 1 ORTEP representation of the molecular structure of 2CIP (thermal ellip-
soids at 50% probability level). Hydrogen atoms, except for H21, and co-crystal-
˚
lised C6H6 are not shown for clarity. Selected bond distances, in A: Sc–O1,
ꢀ
2.088(3); O1–C21, 1.236(5); C21–O2, 1.272(5); O2–B, 1.560(5). Bond angles, in ꢁ
:
scandium atom and the negatively charged HCO2 ligand.
Sc–O1–C21, 151.2(3); C21–O2–B, 129.2(3).
However, these reactants easily pass through
a second
2154 | Chem. Sci., 2013, 4, 2152–2162
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
Fig. 2 Gibbs energy profile of the reaction of CO2 with the ion-pair 1CIP to form 2CIP. Sc, dark green; F, light green; O, red; C grey; B pink; H white.
transition state leading exergonically to the observed product spectroscopy and by X-ray diffraction (see Fig. S4†). These data
2CIP, which is even more stable at ꢀ21.9 kcal molꢀ1. This step revealed similar k2-F binding of the methyl-borate counterion in
corresponds to abstraction of the formato ligand from scan- 3 as was observed for hydrido-borate in 1CIP; full details are
dium by interaction of the electrophilic B(C6F5)3 with a nega- given in the ESI.† As can be seen in Scheme 2, no reaction
tively charged oxygen atom of the formate. At the TS, one Sc–O between 3 and CO2 was observed at room temperature over
˚
bond is broken with an Sc–O distance of 2.80 A compared to extended time periods, suggesting that CO2 is not able to bind
˚
2.23 A in the formate intermediate. The Gibbs energy barrier is to the Sc center in 3 and/or that transfer of a methide group to
only 9.8 kcal molꢀ1 with respect to the intermediate and is thus coordinated CO2 is not energetically feasible.
an energetically facile step. The computed structure of the nal
Interestingly, exposing a C6D5Br solution containing 3, 1.0
product 2CIP is in excellent agreement with the experimentally equivalent of Et3SiH, and B(C6F5)3 (<1 equivalent) to CO2
determined structure. Although 2CIP is the most stable resulted in the gradual formation of 2CIP. As shown in Scheme 2,
stationary point on the potential energy surface with respect to the reaction involves the intermediate formation of 1CIP from 3,
the reactants, the energy barrier back over the last TS is not followed by rapid reaction of 2CIP with CO2 as detailed above.
high. Furthermore, the computed difference in Gibbs energies Monitoring the reaction by 1H NMR spectroscopy corroborated
between the last intermediate and 2CIP of 6.4 kcal molꢀ1 the formation of the Et3SiMe by-product along with 2CIP
represents the maximum energy difference between
.
{Cp*2ScOC(O)H}n and free borane, so it is not surprising that
borane dissociation is observable experimentally (Scheme 1).
Catalytic hydrosilation of CO2 using 1CIP
The facility with which ion pair 1CIP reacts with CO2 led us to Given that formato borate ion pair 2CIP is in equilibrium with
attempt the analogous reaction which would lead to C–C bond {Cp*2ScOC(O)H}n and free B(C6F5)3 the potential for borane
formation using the related ion pair 3, which contains a methyl catalyzed hydrosilation of the C]O functions64,65 in the system
borate counteranion rather than a hydridoborate. Compound 3 is present. Accordingly, we examined the efficacy of ion pair 1CIP
forms cleanly by abstraction of the methyl group of Cp*2ScMe as a catalyst precursor for the hydrosilation of CO2 under a
by B(C6F5)3 and was fully characterised by 11B and 19F NMR variety of conditions. Initially, ion pair 1CIP (25 mmol; 47 mM Sc)
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2152–2162 | 2155

View Article Online
Chemical Science
Edge Article
Sc-formate species, with the ratio between the ion paired and
neutral species being 3 : 2. Following complete silane depletion,
the neutral species eventually diminished entirely in favor of the
two ion paired species that fully emerged (along with their
corresponding Cp* signals) as signals of identical line widths at
half height as a sharp singlet at 8.52 (xa ¼ 0.36; n ꢂ 1 Hz; Cp*,
½
1.90), and a broadened singlet at 8.53 ppm (xb ¼ 0.64; n ꢂ 7 Hz;
½
Cp*, 1.66). Consistent with the 1H NMR data, 19F NMR spectra
revealed the formation of two distinct formato-borate species
with DdFm–p values of 6.5 and 7.0 ppm.
The rather slow progression of the reaction and the incom-
plete reduction of the CO2 is likely due to the low concentrations
of free B(C6F5)3, which is necessary to effect hydrosilation at
reasonable rates. Monitoring the reaction in the presence of 0.5
and 4 equivalents of additional B(C6F5)3 under otherwise iden-
tical reaction conditions provided further details on the nature
of the species observed and their role in catalytic CO2 reduction.
The key ndings are summarised as follows.
Scheme 2 Synthesis of 3 and reaction with CO2.
was exposed to excess CO2 (PCO $ 1 atm) in the presence of
2
First, increasing the concentration of additional B(C6F5)3 co-
catalyst in the reaction mixture favors the formation of CH4, as
excess Et3SiH (23 equivalents) in d6-benzene, and the reaction
was monitored by 1H and 19F NMR spectroscopy at T ¼ 298 K.92
Spectra taken immediately aer CO2 addition showed the
expected formation of 2CIP (1H, 8.51 ppm; DdFm–p ¼ 7.0 ppm)
and free B(C6F5)3, (ꢂ6%, see Scheme 1), along with emerging
resonances of the organic products of CO2 reduction: 3,3,7,7-
tetraethyl-3,7-disila-4,6-dioxanonane, (Et3SiO)2CH2, at 5.06,
triethyl(methoxy)silane, Et3SiOCH3, at 3.32, and CH4 at 0.17
ppm.70 The reaction's speciation prole vs. time is shown in
Fig. 3. The rate of silane consumption followed an overall 1st
order dependence on silane concentration, with silane decaying
expected. The molar fraction of CH4, xCH , in the nal product
4
mixture increased from z0.5, with no added B(C6F5)3 (Fig. 3), to
0.97 in the presence of 4 equivalents of co-catalyst (see Fig. S6
and S7†). Second, resonances assignable to neutral Sc-formate
species {Cp*2ScOC(O)H}n were not detected in the presence of
additional B(C6F5)3 co-catalyst at any time during the reaction.
This is expected, since additional borane should suppress the
equilibrium production of {Cp*2ScOC(O)H}n as depicted in
Scheme 1b. Aer complete silane consumption, three distinct
formato-borate species were observed at 8.54, ꢂ8.51, assigned
to contact ion pair 2CIP, and 8.30 ppm, as can be seen from
Fig. S8.† The molar fraction of the species detected at 8.30 ppm
increased from 0.1 to 0.3 in the presence of 0.5 and 4.0 equiv-
alents of added B(C6F5)3. 19F NMR spectra corroborated the
formation of three distinct decamethylscandocinium formato-
borate ion pairs, DdFm–p values being 7.0, 6.9, and 7.4 ppm
(Fig. S9†). As described below, we assign these other two for-
matoborate ions to belong to separated ion pairs in which the
formato anion is coordinated by either one or two equivalents of
borane.
to one-half of its initial concentration in ꢂ32 min (s
)
½,SiH
according to a pseudo rst-order rate constant kobs,SiH ¼ (3.8 ꢃ
0.1) ꢄ 10ꢀ4 sꢀ1; see also Fig. S5, ESI.† However, careful exami-
nation of the traces in Fig. 3, suggests that there is a short
induction period of z20 minutes where silane consumption is
slower.
Accompanying silane depletion, the initially observed
formate proton signal at 8.51 ppm assigned to 2CIP, (Cp*, 1.57)
ppm, diminished in favor of a set of two new formato-borate
resonances along with another set of at least two neutral
A third observation is that, contrasting the overall 1st order
dependence of the reaction rate on silane concentration in the
absence of added B(C6F5)3 co-catalyst, the presence of extra
borane resulted in a kinetic prole that featured a more
pronounced initiation period of slow silane consumption fol-
lowed by a period of constant, but more rapid silane conver-
sion; cf. Fig. S5–S7, ESI.† This induction period is longer as the
amount of added borane is increased. Furthermore, as shown
in Fig. 4, the overall reaction rate aer the initiation phase
appears to be of overall zeroth order with regard to silane
concentration and independent of added B(C6F5)3 within
experimental accuracy, the averaged kobs,SiH being ꢂ0.17 mM
sꢀ1. Notably, the change of the overall order in silane and the
higher rate of silane consumption aer the initiation period
appears to coincide with the rise in molar fraction of the
(Et3SiO)2CH2, xacetal, product to $0.06 (see below).93 Once
silane consumption was complete, no changes in the
Fig. 3 Plot of normalised reactant and product concentrations, [Y]t/[total-Et3Si]t,
vs. time as monitored by 400 MHz 1H NMR spectroscopy in d6-benzene solution;
47 mM 1CIP, T ¼ 298 K, PCO $ 1 atm. No additional B(C6F5)3 co-catalyst added.
2
2156 | Chem. Sci., 2013, 4, 2152–2162
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
We interpret these observations within the framework of the
proposed mechanism depicted in Scheme 3. As in our original
tandem catalysis system for the hydrosilation of CO2,70 the
availability of free B(C6F5)3 co-catalyst is critical for the overall
activity of the system, since it is the catalyst responsible for
hydrosilating the C]O and C–O bonds en route to the nal
reduction product CH4 aer the scandium-mediated initial
addition of Et3SiH to CO2. The active species for these trans-
formations is the borane–silane complex64,65 4, which is depic-
ted in Scheme 3.
The rst hydrosilation event emanates from the scandium
contact ion pair 2CIP, which forms rapidly from 1 as discussed
above. Two paths for this step appear to be operative depending
on whether or not sufficient quantities of the reduction product
(Et3SiO)2CH2 are present. At the beginning of a reaction, when
there is no (Et3SiO)2CH2 available, the reaction must follow Path
A, shown on the bottom of Scheme 3, in which the neutral
decamethylscandocene formate {Cp*2ScOC(H)O}n, is hydro-
silated by 4 to produce the unobserved HC(O)OSiEt3 product
and ion pair 1. Once the former species is generated, it is rapidly
converted to (Et3SiO)2CH2. This nding is completely consistent
with the observations concerning the stoichiometric reaction of
[TMPH][HCO2B(C6F5)3] and silane catalysed by borane, which
instantly produced a mixture of unreacted starting material
(42%), small amounts of Et3SiOCH3 and methane (ꢂ1%), and
(Et3SiO)2CH2 (57%).70 The latter compound is thus the slowest
substrate for hydrosilation by 4 and builds up to sufficient
concentrations to convert 2CIP into the separated ion pair (SIP)
Fig. 4 Plot of normalised silane concentration, [Et3SiH]t/[Et3Sitotal]t, vs. time
showing the commonly observed initiation period followed by constant silane
depletion. Rate constants were determined by linear regression: red lines, R ¼
0.998 each; average kobs,SiH ꢂ 0.17 mM sꢀ1
.
composition of the reaction mixtures were observed for at least
24 hours. However, catalytic activity ensued immediately aer
the addition of further silane ($20 equivalents) and the re-
pressurizing of the system with CO2 at a rate consistent with
the maximum rate observed aer the initiation period. The use
of 13C-enriched CO2 corroborated that all observed ion paired
species participate in the reaction, since the resonances for
both the proton and the carbon nuclei of all the formato
species immediately exhibited the expected doublet coupling
1
1
in the H and H-coupled 13C NMR spectra.
Scheme 3 Proposed mechanism for the hydrosilation of CO2 by the tandem catalyst system of 2CIP and B(C6F5)3.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2152–2162 | 2157

View Article Online
Chemical Science
Edge Article
2SIP in which (Et3SiO)2CH2 serves as a chelating ligand to the
[Cp*2Sc]+ cation. The formatoborate ion in 2SIP is much more
susceptible to hydrosilation by 4, which constitutes Path B (top
of Scheme 3). The induction periods observed are thus related
to the time taken to produce enough (Et3SiO)2CH2 via Path A to
“turn on” Path B. Since added B(C6F5)3 suppresses the equi-
librium formation of {Cp*2ScOC(H)O}n, the induction period is
longer as more borane is added.
However, it is notable that while the induction periods are
longer with added borane, the maximum rate of silane
conversion achievable is higher with even a small excess of
borane present, since this results in higher concentrations of
active hydrosilation catalyst 4. Furthermore, the observed
changes in behavior of the system with respect to order in
[Et3SiH] can be rationalised via this model. When no extra
borane is added, the observation of a 1st order dependence of
the overall reaction rate on silane concentration indicates that
the equilibrium quantity of B(C6F5)3 co-catalyst available from
2CIP is too low to generate sufficient amounts of reactive silane–
borane adduct species 4, and a dependence on [Et3SiH] is
observed. On the other hand, when sufficient quantities of
B(C6F5)3 are present, and Path B is operative, it is likely that the
rate limiting step is the formation of acetal separated ion pair
2SIP from 2CIP, and thus a 0th order dependence on silane
concentration is observed (Fig. 4). This mechanistic hypothesis
is also consistent with the observed formatoborate speciation in
the reaction. The observation of up to three ion paired formato-
borate species upon complete silane consumption can thus be
related to the equilibrium formation of 2CIP, and acetal-sepa-
rated species 2SIP in which the counteranions are the
[HCO2B(C6F5)3]ꢀ and the known B(C6F5)3 adduct of this species,
[HCO2{B(C6F5)3}2]ꢀ.70 With the formation of 2SIP likely repre-
senting the rate limiting step of the overall catalytic process, a
turnover frequency of 325 TO SiH per Sc per h can be estimated
for the conversion of 2CIP into 1CIP in the presence of excess
Scheme 4 Properties of separated ion pairs.
conrmed the presence of an outer sphere [HB(C6F5)3]ꢀ anion,
observed DdFm–p ¼ 2.9 ppm being typical for this anion in
bromobenzene solution.
The chelating acetal is labile since exposure of the solution of
ion pair 1SIP to excess CO2 (1 atm) at r.t. afforded contact ion pair
2CIP instantaneously. This reaction likely proceeds through the
highly reactive contact ion pair, i.e. 1CIP as described above. There
was no indication of acetal coordination in the NMR spectra of
2CIP under these conditions, suggesting that Keq(1) strongly favors
2CIP over species 2SIP under these conditions. However, a distinct
set of resonances for [Cp*2Sc(k2-(iPrO)2CH2)]+ of ion pairs
[Cp*2Sc(k2-(iPrO)2CH2)][HCO2{B(C6F5)3}n] 2SIP (n ¼ 1, 2) emerged
1
in the H NMR spectrum upon addition of further (iPrO)2CH2
silane and 4 equiv. of added borane co-catalyst (xCH ¼ 0.97). It
4
equals one fourth of overall activity of 1300 TO SiH hꢀ1 based on
the observed zeroth order rate constant of overall silane
conversion of ꢂ0.17 mM sꢀ1 (47 mM 1CIP) which includes all
subsequent steps preceding methane formation. The latter
overall catalytic activity can be compared to the overall activity
(150 TO hꢀ1) reported for the cationic zirconium complex/
borane system by Matsuo and Kawaguchi.63
(total 8.6 equivalents) and B(C6F5)3 (4.0 equivalents), the total
molar fraction of species 2SIP being 0.22. Characteristic 1H NMR
resonances
of
outer
sphere
formato-borate
ions
[HCO2{B(C6F5)3}n]ꢀ (n ¼ 1, 2) were observed at 8.42 (n ¼ 1; DdFm–p
¼ 6.3), and 8.24 (n ¼ 2, DdFm–p ¼ 7.2) ppm. Equilibrium
constants Keq(1) and Keq(2) as dened in Scheme 4 were esti-
mated from the NMR spectra to be ꢂ0.7 Mꢀ1 and ꢂ3.4 Mꢀ1
,
respectively, under these conditions.
An obvious test of this mechanistic picture would be the
purposeful addition of (Et3SiO)2CH2 to the reaction mixture and
observe the effect on the initiation period. Unfortunately, this
silylacetal was not readily accessible, so we used diisopropoxy-
methane, (iPrO)2CH2, as a model acetal for this purpose. As
summarised in Scheme 4, the addition of (iPrO)2CH2 (4.5
equivalents) to a 40 mM solution of 1 in d5-bromobenzene
afforded the acetal separated ion pair [Cp*2Sc(k2-(iPrO)2CH2)]
[HB(C6F5)3] 1SIP exclusively; no residual 1CIP was detected. A set
of three 1H NMR resonances of k2-O bound (iPrO)2CH2 to
[Cp*2Sc]+ was observed (NMR data is also summarised in
Scheme 4), showing that the exchange of bound with free acetal
is slow on the NMR time scale in this solvent. A set of three
sharp JF–F coupled resonances in the 19F NMR spectrum
Similar observations were made when a solution of 1CIP
(32 mM Sc) was exposed to 1 atm CO2 in the presence of 10
equivalents of (iPrO)2CH2 in d6-benzene. Contact ion pair 2CIP
(1H, 8.50; DdFp–m ¼ 7.1 ppm) was cleanly afforded. However,
when 4 equivalents of B(C6F5)3 were added characteristic reso-
nances of ion pair species 2SIP in equilibrium with 2CIP emerged
at 8.39 (n ¼ 1; DdFp–m ¼ 6.7), and 8.16 (n ¼ 2, DdFp–m ¼ 7.6) ppm.
Consistent with the discussion concerning Scheme 3 above,
hydrosilation catalysis ensued immediately upon addition of
$20 equivalents silane and pressurizing with CO2, with no
induction period observed. This suggests that the system is able
to access the more rapid Path B mechanism immediately in the
presence of (iPrO)2CH2, which serves as a surrogate for
2158 | Chem. Sci., 2013, 4, 2152–2162
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
(Et3SiO)2CH2. This catalyst formulation was, however, less lacking the aryl ether functionality gave catalysts that were less
robust than that of Scheme 3 because the borane–silane active, suggesting that the pendant aryl ether plays a role in
complex was also reactive towards (iPrO)2CH2, causing it to separating the formatoborate anions that are likely involved from
degrade over time. Nonetheless, its effect on the reaction prole the cationic zirconium center. Thus, the details concerning the
provided good evidence for the operation of Path B in the overall structural properties, dynamic behavior and thermodynamics of
mechanistic scheme.
equilibria of these ion pairs are of great interest in the develop-
ment of further catalyst systems that utilise this tandem strategy.
Conclusions
Experimental
The hydrosilation of CO2 to methane using a tandem catalysis
strategy in which the rst addition of R3SiH to CO2 is mediated
by an electrophilic species partnered with the [HB(C6F5)3]ꢀ
anion and subsequent reductive steps carried out with the
versatile B(C6F5)3/silane catalyst system 464,65 is an emerging
methodology.63,71,72 In our initial contribution,70 the activation
of CO2 by [TMPH]+[HB(C6F5)3]ꢀ was rate limiting; here, we have
described a catalyst in which the cation is the much more
General procedures and other experimental details are
described in the ESI.†
Preparation of [Cp*2Sc][HCO2B(C6F5)3] 2CIP
Solution synthesis. In a glove box, 305 mg (0.35 mmol) of
1
CIP$0.5C6H6 was weighed out into a 100 mL thick walled glass
reactive [Cp*2Sc]+ Lewis acid, which signicantly activates CO2 vessel equipped with a Kontes valve. The vessel was sealed,
towards acceptance of hydride from the anion. In the presence removed from the glove box and evacuated on the high vacuum
ꢁ
of excess borane, this system comprises a highly active and line. Benzene (10 mL) was condensed onto the solid at ꢀ78 C
robust catalyst formulation for the hydrosilation of CO2 to (acetone/dry ice), and the vessel was lled with (13C-labeled)
methane. Due to the dual catalysts involved, we have not yet CO2 and sealed, and the suspension allowed to warm to r.t.
quantitatively evaluated the intrinsic TOF of the critical turn- Compound 1CIP immediately dissolved upon reaction with CO2
over step of 2CIP into 1CIP, but macroscopically, the current affording a colourless solution. Aer stirring for another 30 min
ꢁ
system performs at competitive efficiency and longevity in at r.t., the solution was frozen at ꢀ78 C, excess CO2 was care-
comparison to other catalysts.
fully removed, and benzene was nally sublimed off at r.t. under
The reaction of 1CIP with CO2 is remarkably facile, also dynamic vacuum to leave a straw colored solid that was recov-
corroborated by DFT calculations that show that the conversion ered in the glove box. Yield: 189 mg (0.22 mmol), 62%. Single
to 2CIP occurs with no large energy barriers. The ionic compo- crystalline samples 2CIP$1/3 C6H6 suitable for X-ray diffraction
nents of 1CIP are associated through multiple dynamic Sc–F were obtained from layering hexanes on top of a saturated
contacts. The ability of hydrido-borate to bind the Sc center in benzene solution of 2CIP. 1H NMR (400 MHz, C6D5Br, 298 K): d
1
both a k2-F2 and a k1-F fashion is critical to allow small molecule 8.77 (d, 1JH–C ¼ 216.6 Hz, 1H, [HCO2B(C6F5)3]ꢀ); 7.20 (s, 2H, /3
substrates, such as CO2 or CO,73 to enter the polarised reaction C6H6); 1.67 (s, 30H, C5Me5) ppm. 13C, and 13C{1H} NMR (101
1
“pocket” between cation and anion.94 This highly reactive, MHz, C6D5Br, 298 K):
d
172.9 (d, JH–C
140.0,
¼
216.6 Hz,
and 118.5
spatially restricted environment resembles that which is [HCO2B(C6F5)3]ꢀ);
148.2,
137.3,
proposed to be operative in “frustrated Lewis pairs”.94,95 Here, the ([HCO2B(C6F5)3]ꢀ); 128.3 (1/3 C6H6); 124.2 (C5Me5); 10.4 (C5Me5)
binding of CO2 to the cationic metal center in a k-O fashion90 ppm. 19F (376 MHz, C6D5Br, 298 K): d ꢀ132.8 (m, 6F, o-F);
perturbs the energy levels of the frontier orbitals of CO2 ꢀ156.4 (t, 3JF–F ¼ 21 Hz, 3F, p-F); ꢀ163.1 (m, 6F, m-F) ppm. 11
B
rendering the central carbon atom highly electrophilic. The (128 MHz, C6D5Br, 298 K): d ꢀ0.7 (br s, n ¼ ꢂ630 Hz) ppm. IR
½
spatial proximity of the reactive B–H moiety in the anion, caused (KBr): 1622 cmꢀ1 (nO]C); elemental analysis (calcd, found for
by the k1-F interaction, contributes to the very low barrier for
C
41H33BF15O2Sc, 2*1/3 C6H6): C (54.81, 54.96), H (3.70, 3.67).
Solid state synthesis. In a glove box, a 100 mL thick-walled
conversion of CO2 into HCO2ꢀ, which occurs in a concerted
fashion, selectively forming 2CIP. The exoergicity of the reaction, glass vessel was charged with 40 mg (46 mmol) of bright yellow
demonstrated by the DFT calculations, is driven by the affinity of prisms of 1CIP. The vessel was sealed, removed from the glove
the Sc cation for negatively oxygen atom(s) in the formato or box and evacuated on the high vacuum line. Aer cooling to
ꢁ
formatoborate anions. This strong Sc–O affinity is determining ꢀ78 C (acetone/dry ice) the vacuum was equalised with CO2,
for the equilibrium between the formatoborate ion pair 2CIP and and the vessel sealed and warmed to r.t. The surface of the
the k2-O2CH formato scandium complex and free borane.
prisms gradually turned orange-red (reaction monitored under
Mechanistic studies reveal a complex kinetic prole for the a microscope) within ꢂ1 week at r.t. Excess CO2 was removed
process and show that ion-pairing effects play a signicant role in under vacuum and the crystalline material recovered in the
dictating the overall rate of the conversion. Specically, as the glove box. The IR spectrum of this material taken in KBr is
reaction proceeds and the product of double hydrosilation of CO2 shown in Fig. S1, ESI.† NMR spectra acquired in d5-bromo-
is produced, it can aid in the separation of formatoborate ion benzene matched those given above. Viewed under a micro-
pairs, which results in more rapid hydrosilation of the anion by 4 scope, the crystal's habit appeared to be unchanged but an X-ray
and overall higher turnover rates. In this context, it is interesting diffraction experiment revealed the complete loss of the original
to note that in Kawaguchi's original system,63 the most active lattice order, consistent with the reaction of 1CIP to afford 2CIP in
catalyst utilised the specic ligand set depicted in I; ligands the solid state.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2152–2162 | 2159

View Article Online
Chemical Science
Edge Article
(m, 6F, o-F); ꢀ157.6 (t, 3JF–F ¼ 20.7 Hz, 3F, p-F); ꢀ167.2 (m, 6F,
Representative procedure for catalytic CO2 hydrosilation
3
m-F); 2SIP, ꢀ134.3 (m, 6F, o-F); ꢀ156.1 (t, JF–F ¼ 20.7 Hz, 3F,
In a glove box, a sealable NMR tube was charged with 21 mg (24
mmol) of 1CIP$0.5C6H6, and a solution of 65 mg (559 mmol)
Et3SiH, 49 mg (96 mmol) B(C6F5)3, and 7.0 mL (66 mmol) h8-
toluene in 690 mg d6-benzene was added by pipet. The tube was
sealed, removed from the glove box and the solution degassed
by three freeze–pump–thaw cycles on a high vacuum line. Aer
freezing the solution at ꢀ78 ꢁC (acetone/dry ice), CO2 was
condensed into the tube by cooling the upper part of the tube
with a plug of cotton chilled with liquid N2 (ꢂ2 to 3 atm CO2,
r.t.). The tube was sealed, carefully warmed to r.t., and thor-
oughly agitated to dissolve CO2 and initiate the reaction.
Insoluble 1CIP quickly dissolved upon reaction with CO2 to
afford of 2CIP. Initial 1H (400 MHz) and 19F (376 MHz) NMR
spectra were acquired immediately aer sample preparation,
showing the formation of 2CIP and trace amounts of the reaction
products. Concentrations of reactants and products were
determined by comparison of absolute peak areas to that of the
methyl singlet resonance of h8-toluene, and were normalised to
the concentration of total Et3Si, [total-Et3Si]t. Reactions were
monitored by regularly acquiring 1H (400 MHz) NMR spectra
(recording time of FIDs, 4.31 s; number of scans, 16; delay
p-F); ꢀ163.3 (m, 6F, m-F) ppm.
Computational details
The scandium atom was treated with the small core Stuttgart–
Dresden ECP, associated with its adapted basis set.96 The
hydrogen, boron, carbon and oxygen atoms were described with
the all-electrons 6-31G(d,p) basis set.97 Fluorine atoms were
treated with the Stuttgart–Dresden ECP, associated with its
adapted basis set98 and completed with a set of d polarisation
functions.99 Calculations were carried out at the DFT level using
the hybrid functional B3PW91.100–102 Geometry optimisations
were performed without any symmetry restrictions and the
nature of the extrema was veried with analytical frequency
calculations, which showed that the transition state had only
one imaginary frequency and the minimum had none. The
connectivity between minima and transition states was veried
with IRC calculations. Gibbs free energies were obtained at T ¼
298.15 K and P ¼ 1 atm within the harmonic approximation of
frequencies. DFT calculations were carried out with the
Gaussian 09 suite of programs.103
between scans, 30 s) at intervals of 375 s at T ¼ 298 K. 1H NMR
dip
(400 MHz, C6D6, 298 K): d 8.54 (m,
J
ꢂ 1.7 Hz,
H–F
dip
[HCO2B(C6F5)3]ꢀ, 2SIP); 8.30 (m,
J
ꢂ
1.7 Hz,
Acknowledgements
H–F
[HCO2{B(C6F5)3}2]ꢀ, 2SIP); 1.69 (s, [Cp*2Sc(k2-(Et3SiO)2CH2)]+)
ppm. No distinct resonances of k2-O bound (Et3SiO)2CH2 were
observed, indicating a fast exchange with non-bound acetal on
Funding for the experimental work described was provided by
the Natural Sciences and Engineering Research Council of
Canada in the form of a Discovery Grant to W. E. P. W. E. P. also
thanks the Canada Council for the Arts for a Killam Research
Fellowship (2012–2014). A. B. thanks the Deutsche For-
schungsgemeinscha for nancial support in the form of a
1
the H NMR chemical shi time scale in d6-benzene solution.
19F NMR (376 MHz, C6D6, 298 K): d ꢀ134.1 (o-F, 2SIP); ꢀ134.3
(o-F, 2SIP); ꢀ156.1 (p-F, 2SIP); ꢀ157.2 (p-F, 2SIP); ꢀ163.5 (m-F,
2SIP); ꢀ164.1 (m-F, 2SIP) ppm.
`
Postdoctoral Fellowship. L. C. thanks the Ministere de l'En-
´
seignement Superieur et de la Recherche for a Ph. D. fellowship.
Representative procedure for studies of acetal separated
species 1SIP and 2SIP
`
L. M. and O. E. thank the CNRS and the Ministere de l'En-
´
seignement Superieur et de la Recherche for funding. The
CALMIP (Toulouse) and CINES (French National centre) are
gratefully acknowledged for a generous donation of computa-
tional time.
In a glove box, 15 mg (24 mmol) of 1CIP$0.5C6H6 were dissolved
in a solution of 2.0 mL (19 mmol; int. standard) h8-toluene, and
13 mL (81 mmol) (iPrO)2CH2 in 800 mg d5-bromobenzene. The
solution was transferred into a sealable NMR tube by pipet. 1SIP
:
1H{11B} NMR (400 MHz, C6D5Br, 298 K): d 4.71 (s, 2H,
Notes and references
[Cp*2Sc(k2-(iPrO)2CH2)]+); 4.16 (br s,
n
ꢂ
19 Hz, 1H,
½
[HB(C6F6)3]ꢀ); 4.13 (heptet, JH–H ¼ 6.15 Hz, PrO-CH); 1.64 (s,
1 H. Arakawa, et al., Chem. Rev., 2001, 101, 953–996. See ESI†
for full citation.
2 M. Aresta and A. Dibenedetto, Dalton Trans., 2007, 2975–
2992.
3 G. A. Olah, A. Goeppert and G. K. S. Prakash, J. Org. Chem.,
2009, 74, 487–498.
4 G. A. Olah, G. K. S. Prakash and A. Goeppert, J. Am. Chem.
Soc., 2011, 133, 12881–12898.
5 T. Sakakura, J.-C. Choi and H. Yasuda, Chem. Rev., 2007,
107, 2365–2387.
6 I. Omae, Coord. Chem. Rev., 2012, 256, 1384–1405.
7 Y. Inoue, Y. Itoh and H. Hashimoto, Chem. Lett., 1977, 6,
855–856.
3
i
30H, [Cp*2Sc(k2-(iPrO)2CH2)]+); 1.01 (d, JH–H ¼ 6.19 Hz, PrO-
CH3) ppm. 19F NMR (376 MHz, C6D5Br, 298 K): d ꢀ132.3 (m, 6F,
o-F); ꢀ163.2 (t, 3JF–F ¼ 21 Hz, 3F, p-F); ꢀ166.1 (m, 6F, m-F) ppm.
11B{1H} (128 MHz, C6D5Br, 298 K): d ꢀ24.9 ppm. The solution
was degassed by three-pump–thaw cycles, frozen at ꢀ78 ꢁC
(acetone/dry ice), and exposed to CO2 (1 atm). The tube was
warmed to r.t. and thoroughly agitated. 1H and 19F NMR spectra
showed the formation of 2SIP. In a glove box, 35 mg (68 mmol)
B(C6F5)3, and 13 mL (81 mmol) (iPrO)2CH2 were added. 1H NMR
(400 MHz, C6D5Br, 298 K): d 8.43 (s, [HCO2B(C6F5)3]ꢀ, 2SIP); 8.43
(s, [HCO2{B(C6F5)3}2]ꢀ, 2SIP); 4.59 (s, 2H, [Cp*2Sc(k2-
3
i
3
i
(iPrO)2CH2)]+); 4.04 (heptet, JH–H ¼ 6.2 Hz, PrO-CH); 1.64 (s,
30H, [Cp*2Sc(k2-(iPrO)2CH2)]+); 0.99 (d, JH–H ¼ 6.2 Hz, PrO-
8 J. Louie, J. E. Gibby, M. V. Farnworth and T. N. Tekavec, J.
Am. Chem. Soc., 2002, 124, 15188–15189.
3
i
CH3) ppm. 19F NMR (376 MHz, C6D5Br, 298 K): 2SIP, d ꢀ133.3
2160 | Chem. Sci., 2013, 4, 2152–2162
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
´
9 S. Sanz, M. Benıtez and E. Peris, Organometallics, 2010, 29,
275–277.
38 R. Tanaka, M. Yamashita and K. Nozaki, J. Am. Chem. Soc.,
2009, 131, 14168–14169.
10 S. Sanz, A. Azua and E. Peris, Dalton Trans., 2010, 39, 6339–
39 T. J. Schmeier, G. E. Dobereiner, R. H. Crabtree and
N. Hazari, J. Am. Chem. Soc., 2011, 133, 9274–9277.
40 A. D. Getty, C.-C. Tai, J. C. Linehan, P. G. Jessop,
M. M. Olmstead and A. L. Rheingold, Organometallics,
2009, 28, 5466–5477.
6343.
´
11 A. Correa and R. Martın, Angew. Chem., Int. Ed., 2009, 48,
6201–6204.
12 Y. Zhang and S. N. Riduan, Angew. Chem., Int. Ed., 2011, 50,
6210–6212.
13 J. Takaya and N. Iwasawa, J. Am. Chem. Soc., 2008, 130,
15254–15255.
41 E. Balaraman, C. Gunanathan, J. Zhang, L. J. W. Shimon
and D. Milstein, Nat. Chem., 2011, 3, 609–614.
¨
42 A. Boddien, F. Gartner, C. Federsel, P. Sponholz,
14 N. P. Mankad, T. G. Gray, D. S. Laitar and J. P. Sadighi,
Organometallics, 2004, 23, 1191–1193.
15 M. Takimoto and M. Mori, J. Am. Chem. Soc., 2001, 123,
2895–2896.
D. Mellmann, R. Jackstell, H. Junge and M. Beller, Angew.
Chem., Int. Ed., 2011, 50, 6411–6414.
43 C. A. Huff and M. S. Sanford, J. Am. Chem. Soc., 2011, 133,
18122–18125.
16 M. Takimoto and M. Mori, J. Am. Chem. Soc., 2002, 124,
10008–10009.
44 T. Schaub and R. A. Paciello, Angew. Chem., Int. Ed., 2011,
50, 7278–7282.
¨
17 C. M. Williams, J. B. Johnson and T. Rovis, J. Am. Chem.
Soc., 2008, 130, 14936–14937.
18 C. S. Yeung and V. M. Dong, J. Am. Chem. Soc., 2008, 130,
7826–7827.
19 R. Johansson, M. Jarenmark and O. F. Wendt,
Organometallics, 2005, 24, 4500–4502.
20 K. Shimizu, M. Takimoto, Y. Sato and M. Mori, Org. Lett.,
2005, 7, 195–197.
21 T. Ohishi, L. Zhang, M. Nishiura and Z. Hou, Angew. Chem.,
Int. Ed., 2011, 50, 8114–8117.
45 A. Boddien, D. Mellmann, F. Gartner, R. Jackstell, H. Junge,
P. J. Dyson, G. Laurenczy, R. Ludwig and M. Beller, Science,
2011, 333, 1733–1736.
46 R. Langer, Y. Diskin-Posner, G. Leitus, L. J. W. Shimon,
Y. Ben-David and D. Milstein, Angew. Chem., Int. Ed.,
2011, 50, 9948–9952.
47 C. Federsel, C. Ziebart, R. Jackstell, W. Baumann and
M. Beller, Chem.–Eur. J., 2012, 18, 72–75.
48 C.-C. Tai, T. Chang, B. Roller and P. G. Jessop, Inorg. Chem.,
2003, 42, 7340–7341.
¨
22 D. S. Laitar, P. Muller and J. P. Sadighi, J. Am. Chem. Soc.,
49 W. Wang, S. Wang, X. Ma and J. Gong, Chem. Soc. Rev.,
2011, 40, 3703–3727.
2005, 127, 17196–17197.
´
23 S. Chakraborty, J. Zhang, J. A. Krause and H. Guan, J. Am.
Chem. Soc., 2010, 132, 8872–8873.
50 P. G. Jessop, F. Joo and C.-C. Tai, Coord. Chem. Rev., 2004,
248, 2425–2442.
24 A. J. M. Miller, J. A. Labinger and J. E. Bercaw,
Organometallics, 2011, 30, 4308–4314.
51 A. J. Hoskin and D. W. Stephan, Coord. Chem. Rev., 2002,
233–234, 107–129.
25 K. Ukai, M. Aoki, J. Takaya and N. Iwasawa, J. Am. Chem.
Soc., 2006, 128, 8706–8707.
52 M. Konkol and J. Okuda, Coord. Chem. Rev., 2008, 252,
1577–1591.
26 D. P. Hruszkewycz, J. Wu, N. Hazari and C. D. Incarvito, J.
Am. Chem. Soc., 2011, 133, 3280–3283.
27 J. Wu and N. Hazari, Chem. Commun., 2011, 47, 1069–1071.
28 R. Johansson and O. F. Wendt, Dalton Trans., 2007, 488–
492.
53 S. Arndt and J. Okuda, Chem. Rev., 2002, 102, 1953–1976.
54 W. E. Piers and D. J. H. Emslie, Coord. Chem. Rev., 2002,
233–234, 131–155.
55 P. M. Zeimentz, S. Arndt, B. R. Elvidge and J. Okuda, Chem.
Rev., 2006, 106, 2404–2433.
29 T. C. Eisenschmid and R. Eisenberg, Organometallics, 1989,
8, 1822–1824.
56 J. S. Silvia and C. C. Cummins, J. Am. Chem. Soc., 2010, 132,
2169–2171.
30 T. Fujihara, T. Xu, K. Semba, J. Terao and Y. Tsuji, Angew.
Chem., Int. Ed., 2011, 50, 523–527.
57 J. S. Silvia and C. C. Cummins, Chem. Sci., 2011, 2, 1474–
1479.
31 P. Deglmann, E. Ember, P. Hofmann, S. Pitter and
O. Walter, Chem.–Eur. J., 2007, 13, 2864–2879.
32 W. Sattler and G. Parkin, J. Am. Chem. Soc., 2011, 133, 9708–
9711.
58 A. M. Chapman, M. F. Haddow and D. F. Wass, J. Am. Chem.
Soc., 2011, 133, 18463–18478.
59 M. A. Rankin and C. C. Cummins, J. Am. Chem. Soc., 2010,
132, 10021–10023.
33 M. Khandelwal and R. J. Wehmschulte, Angew. Chem., Int.
Ed., 2012, 51, 7323–7326.
34 M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann and
60 V. A. Williams, D. R. Manke, P. T. Wolczanski and
T. R. Cundari, Inorg. Chim. Acta, 2011, 369, 203–214.
61 D. E. Fogg and E. N. dos Santos, Coord. Chem. Rev., 2004,
248, 2365–2379.
¨
F. E. Kuhn, Angew. Chem., Int. Ed., 2011, 50, 8510–8537.
35 W. Leitner, Angew. Chem., Int. Ed. Engl., 1995, 34, 2207–
62 J.-C. Wasilke, S. J. Obrey, R. T. Baker and G. C. Bazan, Chem.
Rev., 2005, 105, 1001–1020.
2221.
36 K. Huang, C.-L. Sun and Z.-J. Shi, Chem. Soc. Rev., 2011, 40,
63 T. Matsuo and H. Kawaguchi, J. Am. Chem. Soc., 2006, 128,
12362–12363.
2435–2452.
37 S. N. Riduan and Y. Zhang, Dalton Trans., 2010, 39, 3347–
64 D. J. Parks and W. E. Piers, J. Am. Chem. Soc., 1996, 118,
9440–9441.
3357.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2152–2162 | 2161

View Article Online
Chemical Science
Edge Article
65 D. J. Parks, J. M. Blackwell and W. E. Piers, J. Org. Chem.,
2000, 65, 3090–3098.
toluene, to 1 atm of CO2. A precipitate immediately
formed, suggesting a dimeric structure for the putative
66 G. I. Nikonov, S. F. Vyboishchikov and O. G. Shirobokov, J.
Am. Chem. Soc., 2012, 134, 5488–5491.
67 M. Hill and O. F. Wendt, Organometallics, 2005, 24, 5772–
5775.
neutral formate. Prolonged stirring of
a
C6D5Br
suspension of this material in the presence of B(C6F5)3 (1
equiv.) resulted in partial formation of contact ion pair
2CIP, as determined by NMR spectroscopy.
87 B. K. Campion, R. H. Heyn and T. D. Tilley, Inorg. Chem.,
1990, 29, 4355–4356.
¨
68 V. Sumerin, F. Schulz, M. Nieger, M. Leskela, T. Repo and
B. Rieger, Angew. Chem., Int. Ed., 2008, 47, 6001–6003.
69 A. E. Ashley, A. L. Thompson and D. O'Hare, Angew. Chem.,
Int. Ed., 2009, 48, 9839–9843.
88 R. S. P. Coutts and P. C. Wailes, J. Organomet. Chem., 1970,
25, 117–122.
70 A. Berkefeld, W. E. Piers and M. Parvez, J. Am. Chem. Soc.,
2010, 132, 10660–10661.
89 F. A. LeBlanc, A. Berkefeld, W. E. Piers and M. Parvez,
Organometallics, 2012, 31, 810–818.
´
71 S. Park, D. Bezier and M. Brookhart, J. Am. Chem. Soc., 2012,
90 I. Castro-Rodriguez, H. Nakai, L. N. Zakharov,
A. L. Rheingold and K. Meyer, Science, 2004, 305, 1757–
1759.
91 L. A. Watson and O. Eisenstein, J. Chem. Educ., 2002, 79,
1269–1277.
134, 11404–11407.
72 S. J. Mitton and L. Turculet, Chem.–Eur. J., 2012, 18, 15258–
15262.
73 A. Berkefeld, W. E. Piers, M. Parvez, L. Castro, L. Maron and
O. Eisenstein, J. Am. Chem. Soc., 2012, 134, 10843–10851.
74 M. E. Thompson, S. M. Baxter, A. R. Bulls, B. J. Burger,
M. C. Nolan, B. D. Santarsiero, W. P. Schaefer and
J. E. Bercaw, J. Am. Chem. Soc., 1987, 109, 203–219.
75 71st Handbook of Chemistry and Physics, ed. D. R. Lide, CRC
Press, Boca Raton, 1990.
76 W. J. Evans, K. J. Forrestal, M. A. Ansari and J. W. Ziller, J.
Am. Chem. Soc., 1998, 120, 2180–2181.
77 W. E. Piers, P. J. Shapiro, E. E. Bunel and J. E. Bercaw,
Synlett, 1990, 74–84.
78 X. Yang, C. L. Stern and T. J. Marks, Angew. Chem.,Int. Ed.
Engl., 1992, 31, 1375–1377.
92 No reaction between 1 and Et3SiH is observed under these
conditions.
93 The zeroth order rate constant for silane consumption is
slightly slower in the case where more borane is added.
This is likely related to the faster rate of consumption of
the silyl acetal when the concentration of borane is
higher (see also Fig. S5†).
94 D. W. Stephan and G. Erker, Angew. Chem., Int. Ed., 2010,
49, 46–76.
¨
¨
95 C. M. Momming, E. Otten, G. Kehr, R. Frohlich, S. Grimme,
D. W. Stephan and G. Erker, Angew. Chem., Int. Ed., 2009,
48, 6643–6646.
79 X. Yang, C. L. Stern and T. J. Marks, J. Am. Chem. Soc., 1994,
116, 10015–10031.
96 M. Dolg, U. Wedig, H. Stoll and H. Preuss, J. Chem. Phys.,
1987, 86, 866–872.
80 F. Basuli, J. Tomaszewski, J. C. Huffman and D. J. Mindiola,
Organometallics, 2003, 22, 4705–4714.
97 W. Hehre, R. Ditcheld and J. A. Pople, J. Chem. Phys., 1972,
56, 2257–2258.
¨
81 W. E. Piers, Adv. Organomet. Chem., 2005, 52, 1–76.
82 P. B. Hitchcock, M. F. Lappert and A. Singh, J. Chem. Soc.,
Chem. Commun., 1983, 1499–1501.
98 A. Bergner, M. Dolg, W. Kuchle, H. Stoll and H. Preuss, Mol.
Phys., 1993, 80, 1431–1441.
99 L. Maron and C. Teichteil, Chem. Phys., 1998, 237, 105–
83 G. B. Deacon, P. E. Fanwick, A. Gitlits, I. P. Rothwell,
122.
B. W. Skelton and A. H. White, Eur. J. Inorg. Chem., 2001, 100 J. P. Perdew and Y. Wang, Phys. Rev. B: Condens. Matter,
1505–1514. 1992, 45, 13244–13249.
84 P. L. Arnold, Z. R. Turner, R. Bellabarba and R. P. Tooze, J. 101 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
Am. Chem. Soc., 2011, 133, 11744–11756.
85 D. J. Parks, W. E. Piers, M. Parvez, R. Atencio and
M. J. Zaworotko, Organometallics, 1998, 17, 1369–1377.
102 K. Burke, J. P. Perdew and Y. Wang, Electronic Density
Functional Theory: Recent Progress and New Directions,
Plenum, 1998.
86 Neutral Cp*2ScOC(O)H was generated by exposing 103 M. J. Frisch, et al., Gaussian 09 Revision A.1, Gaussian Inc.,
Cp*2ScH, generated in situ from Cp*2ScCH3 and H2 in Wallingford CT, 2009, see ESI† for full citation.
2162 | Chem. Sci., 2013, 4, 2152–2162
This journal is ª The Royal Society of Chemistry 2013