N-Methylacridinium Salts: Carbon Lewis Acids in Frustrated Lewis Pairs for σ-Bond Activation and Catalytic Reductions

Article abstract of DOI:10.1002/anie.201406122

N-methylacridinium salts are Lewis acids with high hydride ion affinity but low oxophilicity. The cation forms a Lewis adduct with 4-(N,N-dimethylamino)pyridine but a frustrated Lewis pair (FLP) with the weaker base 2,6-lutidine which activates H₂, even in the presence of H₂O. Anion effects dominate reactivity, with both solubility and rate of H₂ cleavage showing marked anion dependency. With the optimal anion, a N-methylacridinium salt catalyzes the reductive transfer hydrogenation and hydrosilylation of aldimines through amine-boranes and silanes, respectively. Furthermore, the same salt is active for the catalytic dehydrosilylation of alcohols (primary, secondary, tertiary, and ArOH) by silanes with no observable over-reduction to the alkanes.

Full text of DOI:10.1002/anie.201406122

Angewandte
Chemie
DOI: 10.1002/anie.201406122
Synthetic Methods
N-Methylacridinium Salts: Carbon Lewis Acids in Frustrated Lewis
Pairs for s-Bond Activation and Catalytic Reductions**
Ewan R. Clark and Michael J. Ingleson*
Abstract: N-methylacridinium salts are Lewis acids with high
hydride ion affinity but low oxophilicity. The cation forms
a Lewis adduct with 4-(N,N-dimethylamino)pyridine but
a frustrated Lewis pair (FLP) with the weaker base 2,6-
lutidine which activates H2, even in the presence of H2O. Anion
effects dominate reactivity, with both solubility and rate of H2
cleavage showing marked anion dependency. With the optimal
anion, a N-methylacridinium salt catalyzes the reductive
transfer hydrogenation and hydrosilylation of aldimines
through amine–boranes and silanes, respectively. Furthermore,
the same salt is active for the catalytic dehydrosilylation of
alcohols (primary, secondary, tertiary, and ArOH) by silanes
with no observable over-reduction to the alkanes.
acid based FLPs for catalyzing reductions was first reported
by Stephan and co-workers using [((Ph2PC6H4)2B(h6-
C6H5))RuCl][B(C6F5)4].[12] This compound, whilst able to
activate H2 with an appropriate base, still contains a precious
metal. Thus the goal of utilizing a metal-free, inexpensive
carbon Lewis acid for FLP-based reductions remains to be
realized.
In our prior work, the borocation 1+ (Scheme 1) was
found to act as a Lewis acid either at boron or at the C9-
position of the acridine moiety, depending upon the reaction
F
rustrated Lewis pairs (FLPs), pioneered by Stephan and
co-workers,[1] represent a versatile new method for small-
molecule activation, and have been successfully applied to the
catalytic hydrogenation of a range of substrates.[2] Related
Scheme 1. Hydride ion affinities of 1+ (relative to Et3B).
À
systems also activate the Si H bond in silanes, thus enabling
conditions.[13] Computational determination of HIAs con-
firmed that the C-centered HIA is greater than that at boron
by 13.9 kcalmolÀ1. The high HIA of 1+ at the carbon atom is
not surprising, as N-alkyl acridinium species have been
investigated as model compounds for the biological hydride
transfer system NADH/NAD+.[14,15] N-Methylacridinium salts
(2+; see Table 1 for structure) are particularly attractive Lewis
acids as they: a) are easy to synthesize, b) are indefinitely air
and moisture stable,[14] and c) show little propensity to
catalytic (de)hydrosilylation.[3,4] Fluoroaryl boranes, typified
by B(C6F5)3, are the most commonly studied Lewis acids
within the field. Despite their clear utility, these boranes are
not without drawbacks, the principal ones being cost and high
oxophilicity which can limit their utility and stability in wet
solvents and tolerance to functional groups.[5] Other main-
group Lewis acids, including aluminum,[6] silicon,[7] and
phosphorus[8] systems, have been exploited in FLPs, but
these remain extremely oxophilic and in many cases the H2-
activation products are not amenable to further catalytic
application. Thus there is a demand for cheaper, less oxophilic
Lewis acids for FLP applications.
coordinate H2O, thus indicating low oxophilicity. Herein we
+
À
report the incorporation of 2 into FLPs which activate H H,
À
À
Si H, and B H bonds and are catalysts for the reduction of
imines, as well as the dehydrosilylation of alcohols.
Softer carbon-centered Lewis acids were shown by the
groups of Bertrand[9] and Arduengo[10] to activate H2, but the
high hydride ion affinities (HIA) of these compounds
preclude application in reduction processes. Alcarazo and
co-workers have used electron-poor allenes as weaker carbon
Initially the HIA of 2+ was quantified[16] and computa-
tionally determined to be À53.3 kcalmolÀ1 (Table 1),
20.5 kcalmolÀ1 less than the C-centered value for 1+. The
marked difference is ascribed to the additional stabilization
=
afforded by significant B N bond character in 1-HC
Lewis acids,[11] which do activate RS SR bonds but are
incapable of H2 activation. The realization of carbon Lewis
(Scheme 1). The HIA of 2+ was nevertheless found to
exceed that of the model compounds of the conjugate Lewis
acids of known hydride donors, that is, Hantzsch ester (3+)
and a NADH model (4+). Significantly, 2+ has a considerably
lower HIA than Ph3C+ (consistent with the experimental
observation of hydride abstraction from N-methylacridane by
Ph3C+),[14b] essential for transferring a hydride to substrates
post H2 activation. It is however, still 12.3 kcalmolÀ1 greater
than that of B(C6F5)3, thus indicating that H2 activation in
a FLP with an appropriate base will be thermodynamically
favored.[17]
À
[*] Dr. E. R. Clark, Dr. M. J. Ingleson
School of Chemistry, University of Manchester
Oxford Road, Manchester M13 9PL (UK)
E-mail: Michael.ingleson@manchester.ac.uk
[**] The Leverhulme Trust (ERC) and the Royal Society (for a University
Research Fellowship to MJI) are acknowledged for funding. This
work was also funded by the EPSRC (grant number EP/K039547/1).
The authors would like to acknowledge the use of the EPSRC UK
National Service for Computational Chemistry Software (NSCCS) at
Imperial College London in carrying out this work.
A range of [2]X salts [X = I, SbF6, BPh4, tetra(3,5-
dichlorophenyl)borate (hereafter BArCl)][18] were readily
available in excellent yield by methylation of acridine with
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201406122.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Br, I) exist as well-separated ion pairs[21] (closest C9-X contact
of 3.896(3) ꢁ for [2]Cl·H2O), indicates that the Lewis acidity
of these species may be regarded as soft and orbital
controlled, and thus hydride selective.
Table 1: HIAs (relative to Et3B) at the M06-2X/6-311G(d,p), PCM(DCM)
level.
Lewis Acid
[Ph3C]+
HIA [kcalmolÀ1
]
À75.3
[2]BArCl forms FLP systems with oxidation-resistant
nitrogen-donor bases with moderate steric demand (e.g. 2,6-
lutidine). In contrast, a 1:1 admixture of [2]BArCl and 4-
DMAP results in adduct formation between acid and base in
solution, typified by the upfield shift of the N-methyl
resonance in the 1H NMR spectra from d1H 4.50 ppm (for
free acridinium cation) to d1H 3.67 ppm (d1H = 3.35 ppm in 5).
The adduct [2-(4-DMAP)]BArCl (6) can be isolated in good
yield and crystallizes with two metrically similar ion pairs in
the asymmetric unit (thus only one is discussed herein). C9 in
6 is strongly pyramidalized (the sums of non-N bond angles
about C9 are 328.208) and the acridinium moiety folds along
the C(1)-N(1)-C(9) axis by 20.998 (trans-annular folding of
À53.3
À43.1
À41.0
B(C6F5)3
À39.6
À
1.228 in [2]SbF6). The donor-acceptor N C bonds are com-
parable to those of other alkylated 4-DMAP compounds
consistent with strong dative bonding.[23]
The FLPs of 2[anion] and 2,6-lutidine were exposed to
four atmospheres of H2 and slow H2 bond cleavage occurred
at 608C, with significant dependency of the rate of the
reaction upon the anion (Table 2). [2]SbF6 was found to
undergo minimal H2 activation (entry 1) as a result of anion
decomposition, thus leading to a complex and intractable
mixture of degradation products. Whilst [2]BPh4 is almost
completely insoluble in CH2Cl2, it nevertheless displayed the
greatest rate of H2 activation at 608C (entry 2), albeit still
taking over five days to approach completion. From this data
we conclude that the rate of activation is in fact rapid
compared to that of [2]BArCl (entry 3), but severely limited by
solubility. [2]BArCl was chosen for further experiments by
virtue of its improved solubility, and overall anion and
methyliodide and subsequent anion exchange with the
appropriate metathesis reagent. [2]SbF6, [2]BPh4, and
[2]BArCl were crystallographically characterized as well-
separated ion pairs, and show good correlation with calcu-
lated structural metrics of 2+. The crystal structure of [2]SbF6
is shown in Figure 1 as an example.
Figure 1. a) Molecular structure of 2[SbF6] and b) molecular structure
of [2-(4-DMAP)]BArCl, 6 (only one molecule from the asymmetric unit
is shown, and counterions and disordered solvent are omitted for
clarity). Thermal ellipsoids are drawn at 50% probability. Selected
bond lengths [ꢀ]: For 2[SbF6]: C(1)-N(1) 1.487(14), closest C-FSbF5
distance C(9)-F(3) 3.348(15). For [2-(4-DMAP)]BArCl, 6: C(1)-N(1)
1.471(4), C(9)-N(2) 1.525(4), C(17)-N(3) 1.335(3).
Table 2: FLP activation of H2.[a]
Entry
Lewis acid
Lewis base
T [8C]
Completion (t)[b]
Experimental confirmation for the predicted higher HIA
relative to B(C6F5)3 was obtained by the abstraction of
hydride from [(2,6-lutidine)H][HB(C6F5)3] by [2]BArCl to
yield N-methylacridane (5) and B(C6F5)3. Upon combination
of 3 equivalents of [2]BArCl with Et3PO, to determine Lewis
acidity by the Gutmann–Beckett method,[19] a Dd31P of
4.3 ppm was determined, which is much lower than that of
B(C6F5)3 (at Dd31P 26.8 ppm).[20] The addition of crotonalde-
hyde to [2]BArCl in CH2Cl2 (the Childsꢀ method for assessing
Lewis acidity)[21] resulted in a minimal downfield shift of the
H3 proton with a Dd1H of 0.02 ppm. Thus 2+ is a significantly
weaker Lewis acid towards Et3PO and crotonaldehyde than
B(C6F5)3, in marked contrast to the ordering of the HIAs.
These remarkable differences, coupled with the observations
that [2]BArCl exhibits no observable H2O coordination (by
1H NMR spectroscopy) and that the halide salts [2]X (X = Cl,
1
2
3
4
5
6
[2]SbF6
[2]BPh4
[2]BArCl
[2]BArCl
[2]BArCl
[2]BArCl
2,6-lutidine
2,6-lutidine
2,6-lutidine
2,6-lutidine
2,6-lutidine
7
60[c]
13% (135 h)[f]
92% (135 h)[g]
97% (234 h)
98% (23 h)
60[c]
60[c]
100[d]
100[e]
100[d]
62% (72 h)
25% (72 h)[h]
[a] Reaction conditions: 0.1 mmol each of acid and base in 0.8 cm3
solvent sealed under about 4 atmospheres H2. [b] Reaction progress
assigned by relative intensities of the 1H NMR resonances of the
acridinium and acridane N-methyl groups, except where noted. No
products were isolated. [c] Reaction performed in CH2Cl2. [d] Reaction
performed in oDCB. [e] Reaction performed in undried oDCB. [f] Anion
degradation to give intractable insoluble material. [g] Because of low
solubility of [2]BPh4 in CH2Cl2, completion was assessed by relative
intensity of the N-methylacridane resonance to that of lutidine.
[h] Calculated by relative intensities of imine CH and amine CH2
resonances.
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Table 3: Catalytic reduction of aldimines.
thermal stability. The carbon-centered Lewis acidity was
unambiguously confirmed by studies with D2, with incorpo-
ration of 2D into the C9-position of the resultant 5 as
monitored by 1H and 2D NMR spectroscopy. H2 activation by
[2]BArCl/2,6-lutidine proceeds more rapidly at 1008C in
ortho-dichlorobenzene (oDCB; entry 4), thus indicating a sig-
nificant kinetic barrier to H2 bond cleavage, which in light of
the anion dependence observed, is attributed to anion/cation
interactions in solution. Importantly, there is no decomposi-
tion of [2]BArCl at 1008C after 2 days, thus precluding H2
activation via a B(3,5-Cl2C6H3)3/2,6-lutidine FLP.[24] Given the
utility of N-alkyl acridinium salts as photoredox catalysts in
a wide range of transformations,[25] [2]BArCl/2,6-lutidine was
exposed to dihydrogen and heated in the absence of light,
with dihydrogen activation still proceeding in the dark.
Pleasingly, the FLP systems are stable to water at room
temperature, although slow H2O activation is observed at
higher temperatures (608C) to form N-Me-9-OH acridane
and the 2,6-lutidinium cation. Performing H2 activation in wet
oDCB with [2]BArCl/2,6-lutidine (entry 5) remarkably
Entry
Substrate
Reductant
T [8C]
Conv. [%]
t [h]
1[b]
7
7
7
8
9
10
Me2NHBH3
PhMe2SiH
PhMe2SiH
PhMe2SiH
PhMe2SiH
PhMe2SiH
60
60
100
100
100
100
>99
98
18
296
4
4
24
24
2[a]
3[a,b]
4[a,b]
5[a,b]
6[a,b]
>98[c]
>98[c]
>98[d]
45[d]
[a] Reactions were performed on a 0.2 mmol scale with 50% excess
silane in 0.8 cm3 dry CH2Cl2 except where noted. [b] Reaction in oDCB
1
[c] Calculated by integration of H NMR peaks using cyclohexane as an
internal standard. [d] Consumption of starting imine determined by
integration of 1H NMR peaks using cyclohexane as an internal standard.
Concomitant transimination occurred under reaction conditions con-
suming BnN(SiMe2Ph)R to form a range of other amines; see the
Supporting Information for details.
À
resulted in H2 activation, however heterolytic O H cleavage
of H2O was also observed to give a minor product.
hydrosilylation of the unhindered imine 10, catalyzed by
[2]BArCl (entry 6). 10 is incompatible with B(C6F5)3-catalyzed
hydrosilylation because of the formation of a strong Lewis
adduct.[3b,27] In contrast, [2]BArCl shows no significant pro-
pensity to bind 10 in CH2Cl2 (the N-Me resonance of [2]+
remains at d1H = 4.50 ppm post addition of excess imine 10).
Finally, we investigated the utility of this species in other
B(C6F5)3-catalyzed reactions. [2]BArCl catalyzes the dehydro-
silylation of aromatic, primary, secondary, and tertiary
alcohols (Table 4). A range of silanes can be utilized as
reductants, though steric bulk precludes the use of triisopro-
pylsilane, and triphenylsilane causes anion degradation as
a minor competitive pathway (entry 2). Dehydrosilylation
proceeds unimpeded in the absence of light (entry 3 versus 4).
The reaction evolves dihydrogen gas using R3SiH (observed
by 1H NMR spectroscopy). When Et3SiD is used, a mixture of
With H2 activation unequivocally demonstrated, the
reduction of the unsaturated substrate N-benzylidene-tert-
butylamine (7) was explored in a FLP with [2]BArCl. This FLP
slowly activated H2 with reduction of the imine to the
corresponding amine after heating to 1008C (Table 2,
entry 6). To improve reduction kinetics, the ability of 2+ to
activate the inexpensive dihydrogen surrogate Me2NHBH3
was investigated for transfer-hydrogenation applications.
[2]BArCl reacts with Me2NHBH3 by rapid hydride transfer
to generate 5 and a cationic boron species. Repeating the
reaction in the presence of 7 resulted in formation of
(Me2NBH2)2, 5 and the protonated imine [7H]+, which
upon heating abstracted hydride from 5 to regenerate
[2]BArCl and lead to overall reduction of the imine to an
amine. The observation of 5 and [7H]+ indicates that it is the
hydride-transfer step that is the rate-limiting step. Catalyst-
free, direct reduction of 7 with Me2NHBH3 does occur,[26] but
this background reaction is slow in CH2Cl2 at 608C, thus
taking 69 h to reach only 78% conversion. In comparison,
a 5 mol% loading of [2]BArCl gave complete reduction of 7 in
the presence of Me2NH-BH3 after 18 hours at 608C (Table 3,
entry 1), thus confirming catalysis of the transfer hydro-
genation by the carbon Lewis acid [2]BArCl.
Table 4: Catalytic dehydrosilylation of alcohols.[a]
Entry
Substrate
Silane
Catalyst
t [h]
Conv. [%]
Loading [%]
1
2
3
4
5
6
7
8
9
iPrOH
iPrOH
BnOH
BnOH
BnOH
BnOH
CyOH
1-AdOH
Phenol
Et3SiH
Ph3SiH
5
10
5
<2
16
<1
<1
72
86[b]
75[c]
The applicability of [2]BArCl for the activation of Si H
À
bonds was next investigated. The direct 1:1 combination of
PhMe2SiH
PhMe2SiH
PhMe2SiH
PhMe2SiH
PhMe2SiH
PhMe2SiH
PhMe2SiH
>99
>99[d]
80
[2]BArCl and PhMe2SiH resulted in no observable reaction
5
and no loss of 3JH-H coupling between the Si H fragment and
À
0.5
0.5
5
5
5
2
93[e]
the adjacent methyl groups, as observed for analogous
systems with B(C6F5)3.[3] Nevertheless, mixing Ph3SiH and
Et3SiD with 5 mol% [2]BArCl resulted in H–D exchange at
<1
<1
<1
>99
>99
>99
À
room temperature, thus confirming activation of the Si H
[a] Reactions were performed at room temperature on 0.2 mmol scale
with 5% excess silane in 0.8 cm3 dry DCM except where noted. Yields
assessed by 1H NMR spectroscopy. [b] Total consumption of silane
observed, but competitive siloxane formation from the presence of H2O
in iPrOH. [c] Heated to 608C for reaction duration; total decomposition
of anion observed. [d] Identical reaction conditions to those in entry 4
but performed in total darkness. [e] Heated to 608C. Ad=adamantyl.
bond. Consistent with this, the hydrosilylation of a number of
aldimines was achieved using catalytic [2]BArCl. Whilst
hydrosilylation of 7 is slow at 608C it is complete within
4 hours at 1008C, giving the desired amine post hydrolysis.
Hydrosilylation was also observed for the N-Ph (8) and N-Bn
(9) imines (Table 3, entries 3 and 4). More remarkable, is the
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
2
[2] a) For a recent review article on FLP-based hydrogenation, see: L.
Hounjet, D. W. Stephan, Org. Process Res. Dev. 2014, 18, 385; For
select examples see: b) P. A. Chase, T. Jurca, D. W. Stephan, Chem.
Commun. 2008, 1701; c) B. Inꢂs, D. Palomas, S. Holle, S. Steinberg,
J. A. Nicasio, M. Alcarazo, Angew. Chem. Int. Ed. 2012, 51, 12367;
Angew. Chem. 2012, 124, 12533; d) H. Wang, R. Frçhlich, G. Kehr,
G. Erker, Chem. Commun. 2008, 5966; e) T. Mahdi, J. N.
del Castillo, D. W. Stephan, Organometallics 2013, 32, 1971; f) Y.
Segawa, D. W. Stephan, Chem. Commun. 2012, 48, 11963.
[3] For FLP silane activation, see: a) D. J. Parks, W. E. Piers, J. Am.
Chem. Soc. 1996, 118, 9440; b) J. M. Blackwell, E. R. Sonmor, T.
Scoccitti, W. E. Piers, Org. Lett. 2000, 2, 3921; c) D. J. Parks, J. M.
Blackwell, W. E. Piers, J. Org. Chem. 2000, 65, 3090; d) S.
Rendler, M. Oestreich, Angew. Chem. Int. Ed. 2008, 47, 5997;
Angew. Chem. 2008, 120, 6086; e) D. Chen, V. Lecih, F. Pang, J.
Klankermeye, Chem. Eur. J. 2012, 18, 5184; f) M. Alcarazo, C.
Gomez, S. Holle, R. Goddard, Angew. Chem. Int. Ed. 2010, 49,
5788; Angew. Chem. 2010, 122, 5924; g) For an overview of silane
activation by B(C6F5)3, see: W. E. Piers, A. J. V. Marwitz, L. G.
Mercier, Inorg. Chem. 2011, 50, 12252.
[4] For dehydrosilylation of alcohols, see: J. M. Blackwell, K. F.
Foster, V. H. Beck, W. E. Piers, J. Org. Chem. 1999, 64, 4887.
[5] J. W. Thomson, J. A. Hatnean, J. J. Hastie, A. Pasternak, D. W.
Stephan, P. A. Chase, Org. Process Res. Dev. 2013, 17, 1287.
[6] G. Mꢂnard, L. Tran, D. W. Stephan, Dalton Trans. 2013, 42,
13685.
[7] A. Schꢃfer, M. Reißmann, A. Schꢃfer, W. Saak, D. Haase, T.
Mꢄller, Angew. Chem. Int. Ed. 2011, 50, 12636; Angew. Chem.
2011, 123, 12845.
[8] C. B. Caputo, L. J. Hounjet, R. Dobrovetsky, D. W. Stephan,
Science 2013, 341, 1374.
[9] G. D. Frey, V. Lavallo, B. Donnadieu, W. W. Schoeller, G.
Bertrand, Science 2007, 316, 439.
[10] J. W. E. Runyon, O. Steinhof, H. V. Rasika Dias, J. C. Calabrese,
W. J. Marshall, A. J. Arduengo, Aust. J. Chem. 2011, 64, 1165.
[11] D. Palomas, S. Holle, B. Inꢂs, H. Bruns, R. Goddard, M.
Alcarazo, Dalton Trans. 2012, 41, 9073.
[12] M. P. Boone, D. W. Stephan, J. Am. Chem. Soc. 2013, 135, 8508.
[13] E. R. Clark, M. J. Ingleson, Organometallics 2013, 32, 6712.
[14] a) Y. Lu, D. Endicott, W. Kuester, Tetrahedron Lett. 2007, 48,
6356; b) W. Sliwa, Heterocycles 1994, 38, 897.
H2 and HD is evolved, with no D incorporation into the
product observed, thus demonstrating that no carbonyl
intermediates derived from alcohol dehydrogenation are
involved. This data is further supported by the facility with
which phenol is silylated, thus implicating an analogous
mechanism to that with B(C6F5)3 involving heterolytic
activation of Si H to form 5 and [RO(H)SiR3] , which
+
À
undergoes dehydrocoupling to release H2 (or mixtures of H2
and HD when Et3SiD is used and [D]-5 is formed).[3] Unlike
B(C6F5)3-catalyzed dehydrosilylation of alcohols,[3] no over-
reduction to alkanes with concomitant siloxane formation is
observed, even in the presence of a large excess of silane and
with prolonged heating, presumably as a result of the lower
reducing power of N-methylacridane versus that of [HB-
(C6F5)3]À. This reactivity allows the reaction to be performed
under ambient atmosphere, with no need to pre-dry solvents.
The catalyst rapidly converts H2O into the appropriate
siloxane under the reaction conditions (confirmed by delib-
erate siloxane synthesis), and any excess silane poses no
À
threat of over-reducing the R3Si OR’ product. This reaction
was demonstrated in the bulk synthesis of BnOSiPh3 using
unpurified solvents in air with 64% (unoptimized) conversion
despite the use of the challenging silane Ph3SiH (from anion
decomposition side reactions).
In summary, N-alkylated acridinium salts are introduced
as simple carbon Lewis acids for FLP-based s-bond activa-
tions. They were shown computationally and experimentally
to have an appropriate HIA to be useful carbocationic Lewis
acids in FLPs for H2 activation and for reduction (hydride
transfer) chemistry. The former was confirmed by N-methyl-
acridinium salts being effective in H2 activation, dehydrogen-
ation of amine–boranes, and silane activation, subject to
anion dependence upon reactivity. Their application in proof-
of-concept catalytic transfer hydrogenation and catalytic
hydrosilylation of imines has been demonstrated. Further-
more, the low Lewis acidity of [2]+ towards hard Lewis bases
enables the catalytic hydrosilylation of unhindered imines
which are incompatible with hydrogenation/hydrosilylation
catalyzed by B(C6F5)3. [2]BArCl is also a cheap, air- and
moisture-stable catalyst for the dehydrosilylation of alcohols,
thus functioning with excellent turnover and good (unopti-
mized) yield in bench-grade solvents. Current work involves
extending this family of carbon Lewis acids by developing
FLPs containing other N-alkylated pyridinium salts with
lower HIAs and different anionic components.
[15] a) X. Zhu, Y. Liu, J. Cheng, J. Org. Chem. 1999, 64, 8980; b) C.
Zheng, S.-L. You, Chem. Soc. Rev. 2012, 41, 2498.
[16] By the method described in: E. R. Clark, A. Del Grosso, M. J.
Ingleson, Chem. Eur. J. 2013, 19, 2462.
[17] T. A. Rokob, A. Hamza, I. Papai, J. Am. Chem. Soc. 2009, 131,
10701.
[18] A. B. Chaplin, A. S. Weller, Eur. J. Inorg. Chem. 2010, 5124.
[19] M. A. Beckett, G. C. Strickland, J. R. Holland, K. S. Varma,
Polymer 1996, 37, 4629.
[20] M. A. Beckett, D. S. Brassington, S. J. Coles, M. B. Hursthouse,
Inorg. Chem. Commun. 2000, 3, 530.
[21] R. F. Childs, D. L. Mulholland, A. Nixon, Can. J. Chem. 1981, 60,
809.
[22] P. Storoniak, K. Krzyminski, P. Dokurno, A. Konitz, J. Blaz-
ejowski, Aust. J. Chem. 2000, 53, 627.
Received: June 11, 2014
Revised: July 21, 2014
Published online: && &&, &&&&
[23] M. J. Corr, M. D. Roydhouse, K. F. Gibson, S. Zhou, A. R.
Kennedy, J. A. Murphy, J. Am. Chem. Soc. 2009, 131, 17890.
[24] S. J. Geier, D. W. Stephan, J. Am. Chem. Soc. 2009, 131, 3476.
[25] See for example: S. Fukuzumi, K. Ohkubo, Chem. Sci. 2013, 4, 561.
[26] X. Yang, L. Zhao, T. Fox, Z.-X. Wang, H. Berke, Angew. Chem.
Int. Ed. 2010, 49, 2058; Angew. Chem. 2010, 122, 2102.
[27] For the binding of aldimines to B(C6F5)3, see: J. M. Blackwell,
W. E. Piers, M. Parvez, R. McDonald, Organometallics 2002, 21,
1400.
Keywords: frustrated Lewis pairs · Lewis acids · reduction ·
silanes · synthetic methods
.
[1] a) G. C. Welch, R. R. San-Juan, J. D. Masuda, D. W. Stephan,
Science 2006, 314, 1124; For recent overviews of the FLP field
see: b) Frustrated Lewis Pairs: Uncovering and Understanding
(Ed.: G. Erker, D. W. Stephan), Springer, Berlin, 2013.
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Synthetic Methods
E. R. Clark,
M. J. Ingleson*
&&&& — &&&&
N-Methylacridinium Salts: Carbon Lewis
Acids in Frustrated Lewis Pairs for s-Bond
Activation and Catalytic Reductions
Softly does it: N-methylacridinium salts
are shown to be versatile Lewis acids
within frustrated Lewis pairs for stoi-
chiometric and catalytic transformations
by through H2 activation.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!