Metal-free carbon dioxide reduction and acidic C-H activations using a frustrated Lewis pair

Article abstract of DOI:10.1016/j.ica.2010.12.022

Activation of CO₂ and acidic C-H bonds by the lutidine-tris(pentafluorophenyl)borane [Lut/B(C₆F₅) ₃] frustrated Lewis pair (FLP) are described (lutidine = 2,6-dimethylpyridine). Lut/B(C₆F₅)

Full text of DOI:10.1016/j.ica.2010.12.022

Inorganica Chimica Acta 369 (2011) 126–132
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Metal-free carbon dioxide reduction and acidic C–H activations using a
frustrated Lewis pair
1
⇑
Sophia D. Tran, Tristan A. Tronic, Werner Kaminsky , D. Michael Heinekey, James M. Mayer
Department of Chemistry, University of Washington, P.O. Box 351700, Seattle, WA 98195-1700, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 16 December 2010
Activation of CO2 and acidic C–H bonds by the lutidine-tris(pentafluorophenyl)borane [Lut/B(C6F5)3] frus-
trated Lewis pair (FLP) are described (lutidine = 2,6-dimethylpyridine). Lut/B(C6F5)3 reacts with CO2 and
H2 at ambient temperature and 4 atm of pressure to form the lutidinium boro-formate salt
[LutH+][HC(@O)OB(C6F5)3ꢀ]. This salt has been fully characterized including an X-ray crystal structure
and independent synthesis from formic acid and Lut/B(C6F5)3. Attempts to activate a C–H bond in meth-
ane by Lut/B(C6F5)3, analogous to its heterolytic cleavage of H2, were unsuccessful, which are consistent
with published calculations showing significant barriers to this reaction. Lut/B(C6F5)3 does react with
more acidic C–H bonds, including acetone and nitroalkanes. With nitromethane, the boro-nitrone anion
Dedicated to Robert G. Bergman
Keywords:
Frustrated Lewis pair
Carbon dioxide
Methane
ꢀ
H2C@NO2B(C6F5)3 is formed, as indicated by NMR and mass spectral analyses.
Acidic C–H activation
Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction
FLPs have been reported to accomplish a variety of small molecule
activations such as the heterolytic cleavage of H2 [13,14], N2O and
Chemical transformations of carbon dioxide (CO2) and methane
(CH4) to produce useful chemical feedstocks are very attractive
goals due to current concerns about climate change and decline
of petroleum reserves [1]. The activation of these and related small
molecules in solution has typically been accomplished by transi-
tion metal complexes [2]. CO2 could be an attractive carbon feed-
stock [1,3], however current methods of CO2 reduction to formic
acid, formaldehyde, methanol, and their derivatives often involve
expensive metal catalysts and extreme conditions [4]. Methane is
the primary component of natural gas and is expensive to trans-
port. It is also a potent and substantial greenhouse gas (CH4 is 62
times more efficient than CO2 as a greenhouse gas [5]). Because
of interest in converting methane to a liquid such as methanol,
and for related conversions of more complex organic molecules,
catalytic C–H activation and functionalization has long been an
attractive goal [6–8].
CO2 complexation [15,16], THF ring-opening [9], and more re-
cently, the conversion of CO2 to CH4 via a FLP with triethylsilane
as the reductant [17].
We set out to determine whether pyridine-based FLPs can re-
duce CO2 using molecular hydrogen. The reported methods for me-
tal-free CO2 hydrogenation do not use hydrogen (H2) as the
hydrogen source [18]. As described below, facile formation of a
boro-formate salt from an FLP, CO2, and H2 under mild conditions
was observed [19]. While this work was in progress, Ashley,
Thompson, and O’Hare reported similar reactivity using a bulky
piperidine base, albeit at higher temperatures [20]. They also
showed that methanol can be produced under harsher conditions
and with destruction of their FLP. Our focus then shifted towards
C–H activation reactions, starting with methane to compare with
the calculations by Wang and co-workers [21], and concluding
with FLP reactions of acidic C–H bonds.
Recently, Stephan and coworkers have shown that frustrated
Lewis pairs (FLPs) can accomplish a range of interesting bond acti-
vation and catalytic reactions [9–12]. FLPs involve a Lewis acid and
base, typically fluorinated organoborane acids combined with
bulky phosphine or pyridine bases, whose steric bulk prevents
strong dative bond formation. The weak Lewis acid/base adduct
thus retains much of the reactivity of the individual acid and base.
2. Results and discussion
2.1. CO2 reduction by H2 using 2,6-lutidine/B(C6F5)3
The FLP formed from 2,6-lutidine (Lut = 2,6-dimethylpyridine)
and B(C6F5)3 heterolytically cleaves H2 to yield the borohydride salt
[LutH+][HB(C6F5)3ꢀ] as a white solid, as reported by Geier and Ste-
phan [9]. This hydrogen product was treated with 4 atm of CO2 in
toluene-d8 in a flame-sealed NMR tube. Over the course of hours at
room temperature, monitoring by 1H NMR spectroscopy revealed
⇑
Corresponding author. Tel.: +1 206 543 2083; fax: +1 206 685 8665.
E-mail address: mayer@chem.washington.edu (J.M. Mayer).
1
UW Chemistry crystallographic facility.
0020-1693/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2010.12.022

S.D. Tran et al. / Inorganica Chimica Acta 369 (2011) 126–132
127
Scheme 1.
that the initial signals characteristic of the lutidinium borohydride
decrease, including the broad singlet for the N–H proton at d
9.82 ppm2 and the borohydride four line pattern at d 3.85 ppm
(JBH = 83 Hz). These resonances are replaced by a sharp singlet at d
8.31 ppm, assigned as the formate proton HC(@O)O of a boroformate
anion (Eq. (1), Fig. S1). The chemical shift for the formyl hydrogen in
free formic acid is d 7.78 ppm in toluene-d8. When 13C-labeled CO2 is
used, the singlet is replaced by a doublet with JCH = 216 Hz at the
same chemical shift, and a resonance is observed at d 153 ppm in
the 13C{1H} NMR spectrum (Fig. S3). These data confirm that a for-
mate C–H bond has been formed.
(Fig. 1) has two independent lutidinium–boroformate ion pairs in
the asymmetric unit with the unit cell containing in total eight
pairs. In each pair, the acidic lutidinium proton forms a hydrogen
bond to the carbonyl oxygen of the formate, with an NꢁꢁꢁO distance
of 2.78 0.01 Å. The values reported are the averages of the values
in the two independent ion pairs. The formate anion has an OCO
bond angle of 122.29 0.20° and is bound to the borane with a
d(B–O) of 1.53 0.01 Å.
While these results were being prepared for publication, Ashley,
Thompson and O’Hare reported a similar reaction using the FLP
containing 2,2,6,6-tetramethylpiperidine (TMP) and B(C6F5)3 [20].
They reported that at 100 °C, CO2 reacts with [TMPH+][HB(C6F5)3
]
ꢀ
to give a boro-formate salt with a [TMPH]+ counterion analogous to
our lutidinium salt, as revealed by spectroscopic and crystallo-
graphic analyses. Upon heating [TMPH+][HC(@O)OB(C6F5)3ꢀ] in
toluene to 80 °C under N2, they observed loss of CO2 and regener-
ð1Þ
To further probe the production of the proposed boro-formate
product, ESI-MS of the reaction mixture, diluted in acetonitrile,
show a peak with m/z = 557 in the negative ionꢀmode, with the
appropriate isotope pattern for HC(@O)OB(C6F5)3
.
19F NMR spec-
tra of the reaction mixture show growth of a set of three reso-
nances for the boroformate anion, in addition to two sets of
ꢀ
ꢀ
resonances for HB(ꢀC6F5)3 and hydroxy-borate (HOB(C6F5)3
)
(Fig. S2). HOB(C6F5)3 is almost always present in small amounts
despite subliming the B(C6F5)3 reactant, drying the solvents, and
flame-drying the glassware. The hydroxy-borate appears to arise
at least in part from the water adduct H2O–B(C6F5)3 in samples
purchased from Strem Chemical, as indicated by 19F NMR spectra
that are broad at room temperature and show both H2O–B(C6F5)3
and free B(C6F5)3 resonances at low temperatures [22].
[LutH+][HC(@O)OB(C6F5)3ꢀ] is formed directly from CO2 and H2
by the Lut/B(C6F5)3, as shown in Scheme 1. It has also been inde-
pendently synthesized from Lut, B(C6F5)3, and formic acid. The
air-stable white solid obtained is spectroscopically identical to that
obtained from CO2 via Eq. (1), and has also been characterized by
elemental analysis. Crystals of [LutH+][HC(@O)OB(C6F5)3ꢀ] suitable
for X-ray crystallography were grown from a dichloromethane
solution by vapor diffusion of heptanes. The X-ray crystal structure
2
Depending on the water concentration in the solvent, this N-H proton signal can
Fig. 1. ORTEP [23] drawing of [LutH+][HC(@O)B(C6F5)3ꢀ] with thermal ellipsoids at
shift and broaden out into the baseline due to exchange.
their 50% probability level.

128
S.D. Tran et al. / Inorganica Chimica Acta 369 (2011) 126–132
ation of [TMPH+][HB(C6F5)3ꢀ], indicating that formation of the for-
mate is reversible. More forcing conditions (144 h at 160 °C in ben-
zene) yielded a number of species including B(C6F5)3, TMP, C6F5H,
and CH3OB(C6F5)3, and that vacuum distillation of this solution at
100 °C yielded CH3OH in 17–25% yield. In contrast, we find that
heating our lutidinium salt at 80 °C results only in decomposition.
The apparently greater reactivity of the lutidine FLP than the
piperidine analog – room temperature versus 100 °C – is surprising
in light of calculations reported by Pápai et al. [24]. They find that
heterolytic cleavage of H2 is thermodynamically less favorable for
the methylborate at d 0.8 ppm (Fig. S8). Heating this lutidinium-
methylborate salt at 80 °C for 4 days resulted in decomposition
of the reagents without formation of methane. There thus appears
to be a significant kinetic barrier to interconvert Lut/B(C6F5)3 + CH4
with [LutH+][MeB(C6F5)3ꢀ] (Eq. (2)).
ð3Þ
Lut/B(C6F5)3
than
for
TMP/B(C6F5)3
by
approximately
The free energy barrier for the activation of methane by the
10 kcal molꢀ1. Perhaps the higher reactivity of Lut/B(C6F5)3 for
CO2 reduction is due ꢀto a mechanism involving both hydride dona-
tion from HB(C6F5)3 and Lewis-acid activation of CO2 by free
B(C6F5)3. However, preliminary studies show that the addition of
excess B(C6F5)3 to Eq. (1) does not significantly accelerate the reac-
tion. Similarly, the addition of excess lutidine does not cause a sub-
stantial change in rate.
tBu3P/B(C6F5)3 FLP has been calculated by Wang et al. to be
D
t
Gà = 38.3 kcal molꢀ1 [21]. Bu3P/B(C6F5)3 is thermodynamically a
more reactive FLP than Lut/B(C6F5)3, according to calculations of
Pápai et al. by approximately 8 kcal molꢀ1 [24], so the barrier for
Lut/B(C6F5)3 should be higher than 38 kcal molꢀ1. The observation
that Lut/B(C6F5)3 does not react with CH4 over 9 days at 80 °C (by
1H NMR spectroscopy) implies, using a pseudo-first order model,
that the minimum free energy barrier is
D .
Gà > 27.1 kcal molꢀ1
2.2. Hydrocarbons plus 2,6-lutidine/B(C6F5)3
Thus the lack of reaction of methane with the Lut/B(C6F5)3 FLP is
consistent with the computational studies.
Since Lut/B(C6F5)3 activates H2 under very mild conditions, we
have explored the activation of methane by this FLP. An NMR tube
was charged with Lut/B(C6F5)3, 3 atm 13CH4, and toluene-d8 and
flame sealed. The reaction was monitored by 1-D Heteronuclear
Multiple Quantum Coherence spectroscopy (HMQC), which ob-
serves 1H–13C coupling [25]. After heating at 80 °C for approxi-
mately 3 days, several new resonances were observed, including
peaks at d 5.54 ppm and d 3.07 ppm, and a weak doublet at d
2.2 ppm by HMQC (this signal was observed only with difficulty
in the 1H NMR spectra due to overlap with the lutidine and toluene
methyl groups being in the same region; Figs. S6 and S7). However,
similar peaks are observed in reactions of Lut/B(C6F5)3 with unla-
beled CH4 and upon heating Lut/B(C6F5)3 in toluene-d8 at 80 °C in
the absence of methane (Fig. S11). On the basis of this and other
experiments, we have no evidence for methane activation (Eq. (2)).
2.3. Activation of acidic C–H bonds
A number of organic molecules with heteroatoms and acidic C–
H bonds have been reacted with Lut/B(C6F5)3. The molecules tested
and their corresponding pKa values are listed in Table 1. The oxy-
gen or nitrogen atoms in these molecules can bind to the Lewis
acid, which makes their C–H bonds more acidic and facilitate
deprotonation by the Lewis base.
2.3.1. Nitrotoluenes
Addition of one equivalent of 3-nitrotoluene or 4-nitrotoluene
to a toluene-d8 solution of Lut/B(C6F5)3 caused an immediate color
change to bright yellow in a J. Young NMR tube. In both cases, 1H
NMR spectra showed small upfield shifts of the nitrotoluene reso-
nances (Figs. S12 and S13). The same yellow color and upfield shift
were observed in a reaction of substrate and Lewis acid without
the lutidine. Combining the Lewis base with nitrotoluene, however,
resulted in no color change or shift in the 1H spectrum. Dissolution
of 4-nitrotoluene in oleum also forms a yellow species [29]. These
data indicate complexation between the nitrotoluenes and
B(C6F5)3, presumably through one or more of the oxygen atoms
of the nitro group, but the resulting adduct is not sufficiently acidic
to be deprotonated by the lutidine base. This contrasts with the re-
port of C–H activation of toluene by an alane/lutidine FLP [26].
ð2Þ
The possible activation of the toluene was probed by treating
stoichiometric amounts of FLP and protio-toluene in benzene-d6
in a flame-sealed NMR tube. After heating to 80 °C for 4 days, the
aforementioned signals at d 5.5 and d 3.1 ppm were observed in
the 1H NMR spectrum. Also, 1-D HMQC of FLP + toluene in ben-
zene-d6 shows the same doublet centered at d 2.2 ppm. Heating
the FLP alone in C6D6 resulted in the same resonances in the 1H
NMR spectrum as when heated in the presence of toluene. 19F
NMR spectra of these reactions showed multiple species in solu-
tion. Thus the FLP decomposes under these reaction conditions
and there is no evidence for toluene activation. In contrast,
Chakraborty and Chen have suggested, based on partial NMR spec-
tra, that the FLP of 2,6-di-t-butylpyridine and the aluminum Lewis
acid Al(C6F5)3 react with toluene to give pyridinium aryl-aluminate
salts [26].
2.3.2. Acetonitrile
Stoichiometric amounts of acetonitrile and Lut/B(C6F5)3 were
dissolved in toluene-d8 in an NMR tube. A 1H NMR spectrum col-
lected after a few minutes showed a new signal at d 0.4 ppm
(Fig. S14). It has been reported that nitriles bind tightly to
B(C6F5)3 and the B–N adducts have been characterized by NMR,
IR, and X-ray crystallography [30]. For MeCN–B(C6F5)3, the re-
Table 1
To determine whether the lack of reaction with methane is due
to kinetic or thermodynamic factors, we independently prepared
the possible product of methane activation, the lutidinium-meth-
ylborate salt. Li[MeB(C6F5)3] was prepared from halide-free methyl
lithium (MeLi) and B(C6F5)3 [27] and reacted with lutidinium
chloride (from HCl gas + lutidine). Combining Li[MeB(C6F5)3] and
pKa values (in DMSO [28]) of C–H bonds
studied.
Substrate
Methane
3-Nitro- and 4-Nitro-
toluenes
Acetonitrile
Acetone
pKa
56
not
available
31.3
26.5
17.2
[LutH]Cl in chloroform gave
a
precipitate of LiCl and
[LutH+][MeB(C6F5)3ꢀ], as indicated by 1H NMR spectroscopy (Eq.
Nitromethane
(3)). In toluene-d8, [LutH+][MeB(C6F5)3ꢀ] has a 1H NMR signal for

S.D. Tran et al. / Inorganica Chimica Acta 369 (2011) 126–132
129
ported data include NMR spectra in benzene-d6, 1H: d 0.32 and
11B: d ꢀ10.3, and a CN stretching frequency at 2367 cmꢀ1. Our
reaction in toluene-d8 has a 11B NMR resonance at d ꢀ11.3 ppm,
above. The 13C NMR spectrum contained a triplet at the aforemen-
tioned chemical shift, confirming the presence of a CH2 group. The
2H NMR spectrum of a reaction of nitromethane-d3 with the FLP in
toluene also had a singlet as d 5.8 ppm. These results confirm that
the d 5.8 ppm and d 121 ppm resonances derive from MeNO2. ESI-
MS spectra were obtained of products derived from both CH3NO2
and CD3NO2, after dilution with acetonitrile. The protio sample
showed a peak at m/z = 572 in the negative ion mode, and the deu-
tero analog showed a peak two units higher. High resolution mass
and the derived white solid obtained in KBr has m .
CN = 2370 cmꢀ1
The adduct is unreactive with lutidine in toluene even after heating
to 80 °C over a day, as indicated by no change in the 1H NMR spec-
trum. Thus acetonitrile binds to the Lewis acid to form an adduct
but is not deprotonated, even under forcing conditions.
spectrometry confirmed the composition of the anion as
ꢀ
2.3.3. Acetone
H2C = NO2B(C6F5)3
.
Lut/B(C6F5)3 was treated with a stoichiometric amount of ace-
tone in toluene-d8. Within minutes the 1H NMR spectrum showed
two new singlets at d 1.8 ppm and d 3.7 ppm (br), integrating in a
3:2 ratio (Fig. S16). After 12 h, 1H NMR spectra showed additional
peaks indicating decomposition of the initial product. To avoid any
further reactions such as aldol condensations, half an equivalent of
acetone was reacted with the FLP. However, 1H NMR spectra of this
reaction still showed free acetone, possibly indicating that an equi-
librium with incomplete conversion may be reached. 19F NMR
spectra of both reactions showed multiple products. Attempts to
isolate solids from the reaction have been unsuccessful.
The analogous reaction of nitroethane gave a new primary
product with a quartet at d 6.2 ppm and a doublet at d 1.4 ppm
in its 1H NMR spectrum, as well as another set of quartet/doublet
signals of smaller intensity (ca. 17%) 0.5 ppm downfield from the
larger signal (Fig. S23). These suggested the formation of an a-car-
bon C–H activated nitrone product, as shown in Eq. (6). Low tem-
perature studies showed no substantial change in the NMR spectra
down to ꢀ80 °C, indicating that the nitrone isomers are not rapidly
interconverting. The 19F NMR spectrum shows a set of three major
peaks, similar to the nitromethane case, and some minor peaks
that are difficult to discern but could be due to the second isomer.
ð4Þ
ð6Þ
We tentatively assign the initial acetone product as a boro-eno-
late salt, analogous to the boro-formate described above (Eq. (4)).
The 1H NMR spectrum is consistent with this formulation if the
two downfield enolate protons are coincidentally degenerate at d
3.7 ppm. The peak at d 1.8 ppm is then due to the methyl group.
The 11B NMR spectrum showing a sharp singlet at d ꢀ13.6 ppm
indicates a four-coordinate boron atom, consistent with this
assignment [31].
All of the experimental results indicate that nitromethane and
nitroethane undergo C–H activation upon reaction with Lut/
B(C6F5)3. Presumably this occurs by coordination of the nitro group
to the Lewis acid, analogous to what is observed with the nitrotol-
uenes. Coordination makes the a-C–H protons more acidic and sus-
ceptible to deprotonation by the base. An alternative mechanism
would be initial tautomerization of the nitroalkane to the aci form
(nitronic acid): CH3NO2 ꢀ CH2@N(O)(OH) [32]. The aci form is a
weak acid and would certainly react with the FLP by O–H activa-
tion to give the boro-nitrone anion. This tautomerization to aci-
nitromethane typically requires catalysis by base. To test whether
it is occurring under the conditions of Eqs. (5) and (6), toluene-d8
solutions of nitromethane, lutidine, and excess methanol-d4 were
prepared. After three days, no change in the nitromethane peak
was observed, indicating that no tautomerization is occurring. Tau-
tomerization would have led to H/D exchange via rapid exchange
of the aci-OH with the CD3OD. The presence of lutidine or luti-
dine/methanol, has no effect on the CH3NO2 chemical shift.
2.3.4. Reactions of nitroalkanes
Nitromethane (MeNO2) reacts rapidly with a stoichiometric
amount of Lut/B(C6F5)3 in toluene-d8 in a sealed NMR tube. The
first 1H NMR spectrum obtained at ambient temperatures showed
a new signal at d 5.8 ppm (Fig. S18). A 11B NMR spectrum of this
solution showed a broad singlet at d ꢀ4.5 ppm, which is within
the normal range for four-coordinate boron species [31]. The 19F
NMR spectrum has one major species in solution, with ca. 6% hy-
droxy-borate (HOB(C6F5)3ꢀ). Cross peaks in a 2-D HMQC spectrum
were suggestive of a C–H activated nitromethane product. The d
5.8 1H NMR signal correlates with
a d
13C NMR signal at
121 ppm, with a 13C–1H coupling constant of 194 Hz, indicating
the presence of an sp2 hybridized carbon bound to an electron
withdrawing group (Fig. S21). On the basis of these data and those
below, the product is assigned as a boro-nitrone anion, as illus-
trated in Eq. (5). This assignment requires that the two nitrone pro-
ton signals in the 1H NMR are coincidentally degenerate at d 5.8.
Alternatively, the B(C6F5)3 group could be moving between the
two oxygen atoms rapidly with respect to the NMR timescale,
although there is no evidence for such fluxionality from studies
of the nitroethane product described below.
3. Conclusions
Carbon dioxide is rapidly reduced to a boroformate derivative
by H2 and the frustrated Lewis pair (FLP) 2,6-lutidine/tris(pentaflu-
orophenyl)borane. This reaction is much faster than the related
chemistry with 2,2,6,6-tetramethylpiperidine [20], despite
computational results suggesting that the latter is a more reactive
FLP [24] The product boro-formate salt was fully characterized by
NMR spectroscopy, mass spectrometry, elemental analysis, and X-
ray crystallography.
No activation of methane reactant or toluene solvent has been
observed. This is consistent with the large barrier to methane acti-
vation calculated by Wang et al. [21]. Acetonitrile and nitrotoluene
only form adducts with the Lewis acid. These less acidic substrates
do not undergo deprotonation/C–H activation. However, more
acidic C–H bonds are readily activated by the Lut/B(C6F5)3 FLP.
Nitromethane and nitroethane are converted to boro-nitrone an-
ð5Þ
13C-labeled MeNO2 was reacted with one equivalent of the FLP
in toluene-d8 to confirm the product assignment. 1H NMR spectra
of this solution showed a doublet centered at d 5.8 ppm with
JCH = 194 Hz, the same as seen in the HMQC spectrum described

130
S.D. Tran et al. / Inorganica Chimica Acta 369 (2011) 126–132
ions, as indicated by NMR and mass spectral data, and NMR spectra
tentatively suggest that acetone is converted to an enolate.
Me2C5H3NH+] [HB(C6F5)3ꢀ] was charged in a medium walled
NMR tube with a 14/20 ground glass joint attached. A 180° adaptor
was connected and the assembly was removed from the glove box
and attached to a 7.2 mL volume bulb on a vacuum manifold. Tol-
uene-d8 was added via vacuum transfer, 4 atm CO2 was condensed
into the tube, and then the tube was flame sealed. The tube was
stored at 77 K until just before the first NMR spectrum was ob-
4. Experimental
4.1. General procedures
1
tained. 1H (300 MHz): 1.83 (s, 6H, CH3), 3.81 (q, 1H, JBH = 83 Hz,
All experiments were conducted under inert atmosphere (nitro-
gen) using a glove box or double manifold N2/vacuum line with
Schlenk-like techniques and glassware, unless otherwise noted.
Tris(perfluorophenyl)borane (Strem) was sublimed under dynamic
vacuum in a 90 °C oil bath. Formic acid (99 + %, Acros Organics)
was used as received. Deuterated NMR solvents and isotopically la-
beled reagents were purchased from Cambridge Isotopes. 2,6-Luti-
dine was stirred over KOH overnight under nitrogen, vacuum
distilled and stored in a glove box. Toluene was dried over so-
dium/benzophenone. Toluene-d8 was dried over Na/K and stored
in a sealed vessel. Acetone was dried over CaSO4 and kept in the
glove box. Nitromethane was dried over CaCl2 and kept in glove
box. Para-nitrotoluene (Matheson Coleman & Bell) was pure by
1H NMR, dried under vacuum for ca. 1 h and kept in glove box.
Meta-nitrotoluene (Sigma–Aldrich) was dried with P2O5 and kept
in glove box. Acetonitrile was purchased from Honeywell Burdick
and Jackson, sparged with argon, and plumbed directly into a nitro-
gen filled glove box with stainless steel piping. CO2 and H2 pur-
chased from Praxair were passed through a Drierite column.
NMR spectra were recorded on 300 MHz or 500 MHz Bruker
Avance spectrometers and referenced to the residual solvent signal
3
3
BH), 6.02 (d, 2H, JHH = 8 Hz, meta-CH), 6.72 (t, 1H, JHH = 7 Hz,
para-CH), 8.31 (s, 1H, HC(@O)), 10.96 (br s, 1H, NH); 19F NMR
3
(282 MHz): ꢀ136.86 (br d, 6F, JFF = 19 Hz, ortho-C6F5), ꢀ161.67
3
(t, 3F, JFF = 21 Hz, para-C6F5), -166.53 (m, 6F, meta-C6F5);
13C{1H}NMR (126 MHz): 19.47 (s, CH3), 122.92 (s, meta-CH),
137.53 (s, para-CH), 142.35 (s, ortho-CH), 154.33 (s, HC(@O)).
4.4. Reaction of B(C6F5)3 and 2,6-Me2C5H3N with 13CH4
In a glove box, 5.2 mg (10.1 lmol) B(C6F5)3, 1.2 lL (10.1 lmol)
2,6-Me2C5H3N, and 0.5 mL toluene-d8 were combined in a medium
walled NMR tube. A 180° adaptor fitted with a cajon was con-
nected, taken out of the glove box, and attached to a 7.2 mL volume
bulb on a vacuum manifold. 13CH4 (3 atm) was condensed into the
tube, and then the tube was flame sealed to a total volume of 2 mL.
The tube was kept at 77 K until immediately before the first NMR
spectrum was obtained. After recording an initial 1H NMR spec-
trum, the tube was placed in an 80 °C oil bath and monitored by
1H NMR spectroscopy.
4.5. Synthesis of [2,6-Me2C5H3NH+] [H3CB(C6F5)3ꢀ] using MeLi and HCl
(g)
(1H and 13C) or an external CF3COOH standard (19F: d ꢀ76.55). 11
B
NMR was calibrated by 1H NMR, and a subtraction method was
used to obtain spectra without the signal from the borosilicate
NMR tubes. Coupling constants are reported in Hz. Mass spectra
were collected using a Bruker Esquire Liquid Chromatograph-Ion
Trap mass spectrometer. High resolution mass spectra were col-
lected using a Sciex Qstar XL mass spectrometer. Infrared spectra
was obtained on a Bruker Tensor 27 FTIR. Samples were prepared
in air with dried KBr with a nut and bolt to press into a pellet.
To obtain [H3CB(C6F5)3]Li following [27], in a glove box, 0.08 g
B(C6F5)3 (0.2 mmol) was placed in a 3-neck flask with two necks
capped with septa, and the third capped with a 180° adaptor.
The flask was evacuated on a high vacuum line. Dry diethyl ether
(20 mL, Fischer Scientific, purged with argon) was transferred into
the flask. The reaction flask was cooled to ꢀ78 °C and 0.1 mL 1.6 M
halide-free MeLi in diethyl ether (Sigma–Aldrich) was added drop-
wise via syringe while stirring. The reaction was left to stir and
warm to room temperature overnight. The solvent was removed
in vacuo resulting in a colorless oil. [2,6-Me2C5H3NH]Cl was syn-
thesized by bubbling HCl (g) into an anhydrous solution of
0.2 mL 2,6-lutidine in 10 mL diethyl ether. The HCl was generated
by dropwise addition of sulfuric acid to a separate flask loaded
with sodium chloride, and passed from this flask to the lutidine
solution using Teflon tubing. A white solid immediately precipi-
tated. The solid was isolated by filtration over a glass frit. In a glove
box, a small amount of [H3CB(C6F5)3]Li was combined with 2.8 mg
4.2. Synthesis of [2,6-Me2C5H3NH+] [HC(@O)OB(C6F5)3ꢀ] using formic
acid
In a glove box, 0.49 g (0.97 mmol) B(C6F5)3, 113 lL (0.98 mmol)
2,6-lutidine, approximately 20 mL dry toluene, and a Teflon mag-
netic stir bar was combined in a flask and capped with a rubber
septum. The flask was removed from the glove box, uncapped
and 36.8 lL formic acid was added to the flask, upon which the
cloudy solution became clear. The septum was replaced and cov-
ered with parafilm. The reaction was left to stir for 1.5 h. The solu-
tion was pumped to dryness, pentane added, and the solid isolated
by filtration. Yield: 0.625 g (97%). Anal. Calc. for C26H11BF15NO2: C,
46.95; H, 1.67; N, 2.11. Found: C, 46.82; H, 1.65; N, 2.14%. X-ray
quality crystals were grown in dichloromethane by vapor diffusion
of heptane. 1H NMR (300 MHz, tol-d8): 1.92 (s, 6H, CH3), 5.92 (d,
of [2,6-Me2C5H3NH]Cl (19 lmol) in CHCl3. A white precipitate
immediately formed. This solution was filtered through Celite
and collected in an NMR tube. The solvent was removed and tolu-
ene-d8 was added via vacuum transfer. 1H NMR (300 MHz, tol-d8):
0.83 (s, 3H, CH3), 2.26 (s, 6H, lutidine-CH3), 6.12 (d, 2H, 3JHH = 8 Hz,
3
meta-CH), 6.73 (t, 1H, JHH = 8 Hz, para-CH); 19F NMR (282 MHz,
3
3JHH = 8 Hz, 2H, meta-CH), 6.63 (t, JHH = 8 Hz, 1H, para-CH), 8.31
3
tol-d8): ꢀ166.04 (t, 6F, JFF = 22 Hz, meta-C6F5), ꢀ163.50 (t, 3F,
3
3JFF = 20 Hz, para-C6F5), ꢀ130.92 (d, 6F, JFF = 22 Hz, ortho-C6F5);
(s, 1H, HC(@O)), the NH peak is not visible most likely due to ex-
change with water in the sample; 19F NMR (282 MHz, tol-d8):
11B NMR (160 MHz, tol-d8): ꢀ15.34 (s).
3
4
ꢀ134.41 (dd, JFF = 23 Hz, JFF = 7 Hz, 6F, ortho-C6F5), ꢀ158.45 (t,
3JFF = 20 Hz, 3F, para-C6F5), ꢀ164.40 (m, 6F, meta-C6F5); 13C{1H}
NMR (75 MHz, tol-d8): 18.17 (s, CH3), 123.60 (s, meta-CH), 137.47
(s, para-CH), 143.96 (s, ortho-CH), 153.27 (s, HC(@O)).
4.6. Reaction of B(C6F5)3 and 2,6-Me2C5H3N with para-nitrotoluene
In a glove box, 0.5 mL of a freshly made 0.02 M solution of
B(C6F5)3 and 2,6-Me2C5H3N in toluene-d8 was combined with
4.3. Reaction of [2,6-Me2C5H3NH+] [HB(C6F5)3ꢀ] with CO2
1.8 mg (13 l
mol) para-nitrotoluene in a J. Young NMR tube. 1H
NMR (500 MHz): 1.76 (s, 3H, p-nitrotoluene-CH3), 2.31 (br s, 6H,
3
[2,6-Me2C5H3NH+] [HB(C6F5)3ꢀ] was prepared following the re-
ported synthesis [9]. In a glove box, 13.6 mg (0.02 mmol) [2,6-
lutidine-CH3), 6.44 (d, 2H, JHH = 9 Hz, ortho-CH, nitrotoluene),
3
6.56 (br s, 2H, meta-CH, nitrotoluene), 7.67 (d, 2H, JHH = 8 Hz,

S.D. Tran et al. / Inorganica Chimica Acta 369 (2011) 126–132
131
4.12. X-ray structure of [LutH+][HC(@O)OB(C6F5)3
]
–
meta-CH, lutidine), (the lutidine para-CH overlaps with the residual
signals for the toluene-d8 solvent).
The crystal structure of [LutH+][HC(@O)OB(C6F5)3ꢀ] was estab-
lished with a Nonius Kappa CCD FR590 single crystal X-ray diffrac-
tometer. A colorless prism measuring 0.60 ꢂ 0.24 ꢂ 0.20 mm3 was
mounted on a glass capillary with paratone oil as adhesive. Data
were collected at 130(3) K. The crystal-to-detector distance was
set to 31.5 mm and the exposure time was 180 s for each collected
frame. The scan width was 1.8°. Data collection was 99.3% com-
plete to 25° in h. A total of 58 919 partial and complete reflections
were collected covering the indices, h = ꢀ47 to 47, k = ꢀ16 to 17,
l = ꢀ24 to 24. 9314 symmetry independent reflections resulted
from integration with hkl-SCALEPACK [33] which applies a multi-
4.7. Reaction of B(C6F5)3 and 2,6-Me2C5H3N with meta-nitrotoluene
In a glove box, 5.3 mg (10.4
2,6-Me2C5H3N, 1.2 L (10 mol) meta-nitrotoluene, and 0.5 mL tol-
uene-d8 was combined in
J. Young NMR tube. 1H NMR
lmol) B(C6F5)3, 1.2 lL (10 lmol)
l
l
a
(500 MHz): 1.77 (s, m-nitrotoluene-CH3), 2.34 (br s, lutidine-CH3),
6.53 (br s), 6.65 (t, 3JHH = 9 Hz), 6.71 (br d), 7.64 (m) (all m-nitrotol-
uene–CH and lutidine CH; some signals overlap with the residual
signals for the toluene-d8 solvent).
plicative correction factor (S) to the observed intensities (I) and
2 hÞ=k2
has the following form: S ¼ ðeꢀ2Bðsin
Þ=scale. S is calculated from
4.8. Reaction of B(C6F5)3 and 2,6-Me2C5H3N with acetonitrile
the scale and the B factor determined for each frame and is then
applied to I to give the corrected intensity (Icorr). The data in mono-
clinic space group C2/c was of average quality (Rint = 0.065). The
structure was solved with direct methods within SIR97 [34] and re-
fined with SHELXL97 [35]. Scattering factors are from Waasmaier and
Kirfel [36]. All hydrogen atoms were located using a riding model.
All non-hydrogen atoms were refined anisotropically by full-
matrix least-squares. Crystallographic data are given in the
Supporting Information. The asymmetric unit contains two sym-
metry-independent cations and anions. The fluorides tend to exhi-
bit a larger thermal motion than the carbon ring atoms. Some
fluorides are forced into close contact (F9–F18 F20–F200). N2 forms
a hydrogen bond to O4 within the asymmetric unit, and N1 forms a
hydrogen bond to O2 of a symmetry related anion.
In a glove box, 51 mg (0.1 mmol) B(C6F5)3, 11.4
lL (0.1 mmol)
2,6-Me2C5H3N, 5.1 L (0.1 mmol) acetonitrile, and 4.9 mL toluene
l
was combined in a 25 mL flask with a magnetic stirbar. The reac-
tion was capped and stirred for ca. 20 min. The volatiles were re-
moved in vacuo, leaving a white solid. 1H NMR (500 MHz, tol-d8):
3
0.43 (s, CH3), 1.64 (s, Lut–CH3), 5.73 (d, JHH = 8 Hz, meta-CH),
3
6.42 (t, 1H, JHH = 9 Hz, para-CH); 11B NMR (160 MHz, tol-d8):
C„N = 2370 cmꢀ1
.
~
m
ꢀ11.1 (br s); IR (KBr)
4.9. Reaction of B(C6F5)3 and 2,6-Me2C5H3N with acetone
In a glove box, 5.5 mg (10.7
lmol) B(C6F5)3, 1.2 lL (10 lmol)
2,6-Me2C5H3N, 0.7 L (10 mol) acetone, and 0.5 mL toluene-d8
l
l
were combined in a J. Young NMR tube. 1H NMR (500 MHz): 1.81
(s, 3H, CH3), 2.26 (br, lutidine-CH3), 3.73 (s, 2H, CH2), 6.50 (br d, lu-
tidine meta-CH) (the lutidine para-CH overlaps with the residual
signals for the toluene-d8 solvent); 11B NMR (160 MHz, tol-d8):
ꢀ13.64 (s).
Acknowledgments
We acknowledge primary financial support from the Center for
Enabling New Technologies through Catalysis, a US National Sci-
ence Foundation Phase II Center for Chemical Innovation. We thank
the current members of the Mayer group, and Joseph M. Meredith,
Dr. Alexander J. M. Miller, Dr. Steven L. Matthews, Dr. Jeffery J.
Warren, and Dr. Lisa S. Park-Gehrke for helpful comments.
4.10. Reaction of B(C6F5)3 and 2,6-Me2C5H3N with nitromethane
In
a
glove box, 0.05 g B(C6F5)3 (0.1 mmol), 11.4
lL 2,6-
Me2C5H3N (0.1 mmol), 5.2
lL
nitromethane (0.1 mmol), and
Appendix A. Supplementary material
5.2 mL toluene were combined in a 20 mL round bottom flask with
magnetic stir bar. The reaction was capped with a 180° adaptor, re-
moved from the glove box, and left to stir for 20 min. The flask was
cooled with liquid nitrogen and lyophilized overnight, resulting in
a white powder. 1H NMR (300 MHz, tol-d8): 1.96 (s, 6H, CH3), 5.82
CCDC 788048 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif. Supplementary data associated with this article
can be found, in the online version, at doi:10.1016/
j.ica.2010.12.022.
3
(s, 2H, CH2), 6.20 (d, 2H, JHH = 8 Hz, meta-CH), 6.83 (t, 1H,
3JHH = 8 Hz, para-CH); 19F NMR (282 MHz, tol-d8): ꢀ162.37 (br t,
3
3
6F, JFF = 21 Hz, meta-C6F5), ꢀ155.42 (t, 3F, JFF = 21 Hz, para-
C6F5), ꢀ132.20 (d, 6F, JFF = 20 Hz, ortho-C6F5); 11B NMR
3
References
(160 MHz, tol-d8): ꢀ4.53 (s); 13C NMR (75 MHz, tol-d8, obtained
using 13CH3NO2): 121 (JCH = 194 Hz, CꢀH2NO2). High resolution
[1] S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor,
H.L. Miller (Eds.), Climate Change 2007: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, Cambridge University Press, New
York, 2007.
MS: 571.9957 m/z [CH2NO2B(C6F5)3
]
and 528.9895 m/z
[HOB(C6F5)3ꢀ].
[2] W.B. Tolman (Ed.), Activation of Small Molecules: Organometallic and
Bioinorganic Perspectives, Wiley-VCH Weinheim, Germany, 2006.
[3] G.A. Olah, A. Goeppert, G.K. Surya Prakash, J. Org. Chem. 74 (2009) 487.
[4] (a) W. Leitner, Angew. Chem., Int. Ed. Engl. 34 (1995) 2207;
(b) T. Sakakura, J.-C. Choi, H. Yasuda, Chem. Rev. 107 (2007) 2365;
(c) Y. Kayaki, M. Yamamoto, T. Ikariya, Angew. Chem., Int. Ed. 48 (2009) 4194.
[5] T.J. Blasing, K. Hand, Tellus 59 (2007) 15.
4.11. Reaction of B(C6F5)3 and 2,6-Me2C5H3N with nitroethane
In a glove box, 4.9 mg (9.6 mol) B(C6F5)3, 10 L of 1 M 2,6-
l
l
Me2C5H3N in toluene-d8 (10 lmol), 0.8 lL (11.2 lmol) nitroethane,
and 0.5 mL toluene-d8 was combined in a J. Young NMR tube. 1H
3
[6] R.G. Bergman, Nature 446 (2007) 391.
(500 MHz): major isomer: 1.40 (d, 3H, JHH = 7 Hz, CH3), 6.17 (q,
[7] K.I. Goldberg, A.S. Goldman (Eds.), Activation and Functionalization of C–H
Bonds, ACS Symposium Series, vol. 885, Washington, DC, 2004.
[8] A.E. Shilov, G.B. Shul’pin, Activation and Catalytic Reactions of Saturated
Hydrocarbons in the Presence of Metal Complexes, Kluwer Academic Press,
2000.
3
3
1H, JHH = 6 Hz, CH); minor isomer: 1.75 (d, JHH = 6 Hz, CH3), 6.61
3
(q, JHH = 6.5 Hz, CH); lutidinium: 2.07 (s, 6H, CH3), 6.09 (d, 2H,
3JHH = 7 Hz, meta-CH), 6.73 (t, 1H, JHH = 8 Hz, para-CH); 19F NMR
3
(282 MHz): -163.60 (br t, 6F, meta-C6F5), ꢀ157.72 (t, 3F,
[9] S.J. Geier, D.W. Stephan, J. Am. Chem. Soc. 131 (2009) 3476.
[10] D.W. Stephan, Dalton Trans. (2009) 3129.
3
3JFF = 20 Hz, para-C6F5), -132.16 (d, 6F, JFF = 19 Hz, ortho-C6F5).

132
S.D. Tran et al. / Inorganica Chimica Acta 369 (2011) 126–132
[11] D.W. Stephan, Org. Biomol. Chem. 6 (2008) 1535.
[27] J.-F. Carpentier, Z. Wu, C.W. Lee, S. Strömberg, J.N. Christopher, R.F. Jordan, J.
Am. Chem. Soc. 122 (2000) 7750.
[28] F.G. Bordwell, Acc. Chem. Res. 21 (1988) 456.
[29] J. Take, T. Tsuruya, T. Sato, Y. Yoneda, Bull. Chem. Soc. Japan 45 (1972) 3409.
[30] H. Jacobsen, H. Berke, S. Doring, G. Kehr, G. Erker, R. Fröhlich, O. Meyer,
Organometallics 18 (1999) 1724.
[12] M.A. Dureen, D.W. Stephan, J. Am. Chem. Soc. 131 (2009) 8396.
[13] A. Ramos, A.J. Lough, D.W. Stephan, Chem. Commun. (2009) 1118.
[14] D.W. Stephan, G. Erker, Angew. Chem., Int. Ed. 49 (2010) 46.
[15] E. Otten, R.D. Neu, D.W. Stephan, J. Am. Chem. Soc. 131 (2009) 9918.
[16] C.M. Mömming, E. Otten, G. Kehr, R. Frölich, S. Grimme, D.W. Stephan, G. Erker,
Angew. Chem., Int. Ed. 48 (2009) 6643.
ˇ
[31] Hermánek, Stanislav, Main factors governing the chemical shift of skeletal
[17] A. Berkefelt, W.E. Piers, M. Parvez, J. Am. Chem. Soc. 132 (2010) 10660.
[18] S.N. Riduan, T. Zhang, J.Y. Ying, Angew. Chem., Int. Ed. 48 (2009) 3322.
[19] S.D. Tran, T.A. Tronic, J.M. Mayer, These Studies of CO2 Reactions were
Reported at the Spring 2010 (239th) ACS National Meeting in San Francisco
(Abstract INOR 47).
atoms in boranes, heteroboranes and substituted derivatives, boron chemistry,
in: Proceedings of the 6th International Meeting on Boron Chemistry. New
Jersey, World Scientific Publishing Co., Inc., 1987.
[32] Feuer, Henry, Nitronic Acids and Esters, The Chemistry of the Nitro and Nitroso
Groups, New York, Interscience, 1969.
[20] A.E. Ashley, A.L. Thompson, D. O’Hare, Angew. Chem., Int. Ed. 48 (2009) 9839.
[21] G. Lu, L. Zhao, H. Li, F. Huang, Z. Wang, Eur. J. Inorg. Chem. 15 (2010) 2254.
[22] T. Beringhelli, D. Maggioni, G. D’Alfonso, Organometallics 20 (2001) 4927.
[23] L. Farrugia, J. Appl. Crystallogr. 30 (1997) 565.
[24] T.A. Rokob, A. Hamza, I. Pápai, J. Am. Chem. Soc. 131 (2009) 10701.
[25] Friebolin, Horst, Basic One- and Two-Dimensional NMR Spectroscopy,
Weinheim, VCH, 1991.
[33] (a) Z. Otwinowsky, W. Minor, Methods Enzymol. 276 (1997) 307;
(b) S. Mackay, C. Edwards, A. Henderson, C. Gilmore, N. Stewart, K. Shankland,
A. Donald, University of Glasgow, Scotland, 1997.
[34] A. Altomare, C. Burla, M. Camalli, L. Cascarano, C. Giacovazzo, A. Guagliardi,
A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr. 32 (1999) 115.
[35] G.M. Sheldrick, SHELXTL97, University of Gottingen, Germany, 1997.
[36] D. Waasmaier, A. Kirfel, Acta Crystallogr., Sect. A 51 (1995) 416.
[26] D. Chakraborty, E.Y.-X. Chen, Macromolecules 35 (2002) 13.