Long-range metal-ligand bifunctional catalysis: Cyclometallated iridium catalysts for the mild and rapid dehydrogenation of formic acid

Article abstract of DOI:10.1039/c2sc21923a

Formic acid (HCO₂H) is an important potential hydrogen storage material, which, in the presence of appropriate catalysts can be selectively dehydrogenated to give H₂ and CO₂. In this work, well defined N^C cyclometallated

Full text of DOI:10.1039/c2sc21923a

Chemical Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Long-range metal–ligand bifunctional catalysis:
cyclometallated iridium catalysts for the mild and rapid
Cite this: Chem. Sci., 2013, 4, 1234
dehydrogenation of formic acid†
a
b
a
a
Jonathan H. Barnard, Chao Wang, Neil G. Berry and Jianliang Xiao*
Formic acid (HCO
appropriate catalysts can be selectively dehydrogenated to give H
N^C cyclometallated iridium(III) complexes based on 2-aryl imidazoline ligands are found to be excellent
2
H) is an important potential hydrogen storage material, which, in the presence of
2
and CO . In this work, well defined
2
catalysts for the decomposition of HCO
2
H–NEt
3
1
2 2
mixtures to give H and CO under mild conditions with
ꢀ
ꢁ
high turnover frequencies (up to 147 000 h at 40 C) and essentially no CO formation. The modular
structures of these catalysts have allowed for the construction of structure–activity relationships for the
complexes, leading to the rational optimisation of the catalyst structure with respect to both the rate of
2
H production and catalyst lifetime. In particular, the presence of the remote g-NH unit in the ligand is
Received 6th November 2012
Accepted 19th December 2012
shown to be essential for catalytic activity, without which no reaction occurs. Mechanistic studies
suggest that the dehydrogenation is rate-limited by the step of hydride protonation, which is made
feasible by the g-NH unit via an unusual form of long-range metal–ligand bifunctional catalysis
involving formic acid-assisted proton hopping.
DOI: 10.1039/c2sc21923a
www.rsc.org/chemicalscience
However, if the H
any dehydration is undesirable, as CO impurities are not well
In the quest for the non-polluting fuels of tomorrow, much tolerated by fuel cells, especially for proton exchange
2
is to be used for the generation of electricity,
1
Introduction
1
15
attention is directed towards the use of hydrogen. However, membrane fuel cells.
despite the high energy density of H , its low densities, both in
A variety of transition metal catalysts have been reported for
the liquid and compressed gas forms, in conjunction with its the selective decomposition of HCO H to H and CO2
2
2
2
high ammability pose considerable barriers to its widespread
2
use. The sequestering of H2 within stable molecular and
supramolecular systems and its subsequent release may over-
come these barriers and is an important goal for the realisation
of using H
the use of formic acid (4.4 wt% H
Formic acid can be obtained by processing biomass, is an
easily handled liquid and importantly, the side-product of for-
2
as a fuel for transportation. Of these storage media,
3
2
) is particularly attractive.
4
mic acid dehydrogenation, CO , can easily be recycled via
2
5
6
electrochemical methods or by catalytic hydrogenation,
allowing CO to be used as a viable H storage medium. Formic
2
2
acid may be decomposed via a variety of methods, including
7
–13
transition metal catalysed dehydrogenation to H
2
and CO
2
or
1
4
by dehydration which is typically promoted by heat or acids.
a
Department of Chemistry, University of Liverpool, Liverpool, L69 7ZD, UK. E-mail:
jxiao@liv.ac.uk; Fax: +44 (0)151-7943588
b
Key Laboratory of Applied Surface and Colloid Chemistry, Ministry of Education,
Department of Chemistry & Chemical Engineering, Shaanxi Normal University,
Xi'an, 710062, China
†
Electronic supplementary information (ESI) available: Experimental details.
CCDC 889583, 889584. For ESI and crystallographic data in CIF or other
electronic format see DOI: 10.1039/c2sc21923a
Fig. 1 Homogeneous catalysts for the dehydrogenation of formic acid.
This journal is ª The Royal Society of Chemistry 2013
1
234 | Chem. Sci., 2013, 4, 1234–1244

View Article Online
Edge Article
Chemical Science
7
ꢀ13
7
(Fig. 1).
Puddephatt et al. rst reported the use of binuclear rational development of active catalysts. In particular, the
ruthenium phosphine complexes for formic acid dehydroge- modular structure of these complexes suggested to us that
nation, achieving a turnover frequency (TOF, mol H per mol optimisation of each of the constituent parts of the structure,
2
ꢀ
1
catalyst per h) of approximately 500 h with [Ru (m-CO)(CO) (m- independent of one another, might be possible. To facilitate
2
4
2
dppm) ] aer 0.25 h at room temperature. Later work by Wills this, the catalyst structure was considered as four distinct units,
8
et al. has detailed the use of binuclear Ru complexes formed i.e. the cyclometallated aryl ring, the neutral donor group, the
in situ from Ru(DMSO) Cl and triphenyl phosphine. Both the substituents on the donor group and the central metal ion
groups of Beller and Laurenczy have utilised ruthenium (Fig. 3), which were then examined in greater detail.
phosphine complexes for the decomposition of HCO H. Beller 2.1.1 Synthesis of initial cyclometallated complexes. A
and coworkers also reported a catalyst system composed of range of [Cp*IrCl(N^C)] complexes (1–13) bearing different
RuCl (benzene)]2 and 1,2-bis(diphenylphosphino)ethane nitrogen donor groups were rst synthesised from their parent
which gives high TOFs and turnover numbers (TONs) under the ligands (L1–L13) (Fig. 4) and [Cp*IrCl ] via the acetate-assisted
4
2
9
10
2
[
2
2
2
16a
2
continuous addition of HCO H. More recently, the same group cyclometallation protocol reported by Davies et al. (eqn (1)).
11
reported the use of in situ generated iron catalysts, including The Ir complexes were formed in good to excellent yields,
an Fe(II) tetraphos systems which is active in the absence of requiring only simple work up procedures and minimal puri-
11c
amine additives. Fukuzumi has reported the use of several cation. They also possess excellent stability, necessitating no
water-soluble complexes including [Cp*Rh(bpy)(H O)][SO protection from air or moisture.
and [Cp*Ir(H O)(bpm)Ru(bpy) ][SO ] for the dehydrogenation
2
4 2
]
2
2
4 4
ꢀ
1
of HCO H in water, with the latter giving a TOF of 426 h under
ambient conditions. Himeda et al. have reported the use of
Cp*Ir(4,4 -dihydroxy-2,2 -bipyridine)] and related complexes
2
12
0
0
[
(
1)
for the dehydrogenation of aqueous HCO H–HCOONa solu-
2
ꢁ
ꢀ
1
6k,6q,13
tions with TOFs of up to 228 000 h at 90 C achieved.
16
Cyclometallated Ru, Rh and Ir complexes have been shown to
be excellent catalysts for a range of novel redox processes
17
including amine and alcohol racematisation, dehydrogenation
18
19
reactions and transfer hydrogenation. Recently, we reported
2.1.2 The effect of the donor group. The complexes 1–13
that cyclometallated Cp*IrCl imido complexes derived from ace- were then screened as potential catalysts for the decomposition
ꢁ
tophenone imines (Fig. 2) are exceptionally active catalysts for the of the 5 : 2 F/T azeotrope at ambient temperature (25 C). The
2 3
reduction of imines using the 5 : 2 HCO H–NEt azeotrope (F/T) relative effectiveness of each complex was judged by the
20
as the hydrogen source. During our studies of these and related comparison of both the initial TOF and the total volume of gas
complexes, we oen noted varying degrees of gas evolution. collected in 2 h reaction time (Table 1). As can be seen, HCO H
2
Prompted by these observations and the high modularity of the dehydrogenation does occur; however, the nature of the donor
imido complexes, we were interested as to whether the intercep- group has a dramatic effect on the catalytic activity. Complexes
tion of the Ir–hydride intermediates by protons, rather than imi- with imidamide-based donor groups bearing an NH proton,
2
0,21
nium cations,
could lead to fast hydrogen evolution. Herein, such as 2-imidazolyl or 2-imidazolinyl, showed good initial
ꢀ
1
we report that a rationally arrived complex shows extremely high TOFs of ca. 500–1000 h
(Table 1, entries 1–5). In stark
activity for formicaciddecompositionundermildconditions, and contrast, complexes bearing other N-heterocyclic donors, such
most interestingly, the catalysis involves not only the metal center as 2-oxazolinyl (precatalyst 6) or 2-pyridyl (precatalyst 7), dis-
but also a NH functionality g to iridium. It appears that H
formation is facilitated by HCO H-mediated proton hopping.
2
played no observable activity at all. Similarly, the use of imine
(precatalysts 8–9) or saturated benzylamine (precatalyst 10)
derived ligands led to a complete loss of activity. No activity was
observed when [Cp*IrCl ] was used, nor did the reaction occur
2
2
2
2
Results and discussion
in the absence of any catalyst (entry 8).
Interestingly, replacing the NH proton in the 2-imidazolinyl
ligand L1 with either electron donating (precatalysts 11–12) or
electron withdrawing (precatalyst 13) substituents led to
2
.1 Catalyst development
In contrast to the great majority of catalysts for HCO H dehy-
drogenation, the well-dened nature of the cyclometallated
complexes, as exemplied in Fig. 2, allows for the detailed and
2
Fig. 2 Cyclometallated Ir(III) complexes used for transfer hydrogenation.
This journal is ª The Royal Society of Chemistry 2013
Fig. 3 The modular structure of cyclometallated Cp*MCl complexes.
Chem. Sci., 2013, 4, 1234–1244 | 1235

View Article Online
Chemical Science
Edge Article
2.1.3 The effect of the cyclometallated group. Modication
of the cyclometallated aryl ring in complex 1, affording pre-
catalysts 14–19, shows that the presence of electron donating
substituents para to the donor group (meta to Ir) enhances the
catalytic activity, whilst electron decient ones severely retard
22
the activity (Table 2). A r value of ꢀ1.8 was observed in the
Hammett plot (Fig. 5), suggesting that positive charge is
developed in the transition state of the rate-determining step of
the dehydrogenation reaction (vide infra). The precatalysts 14–
19 were prepared using the ligands L14–19 under the same
conditions as shown in eqn (1).
Among 14–19, the precatalyst 19, which bears the 3,4-meth-
ylenedioxy disubstituted ligand L19, is most active. Having two
electron-donating groups, complex 19 was synthesised in an
attempt to further enhance the rate of H2 formation. In the
reaction of L19 with [Cp*IrCl ] , the cyclometallation occurred
2
2
solely on the 2-position, affording complex 19. Since the imi-
dazolinyl group and the iridium are each in direct conjugation
to the electron-donating O-alkyl substituent (Fig. 6), the result,
alongside those above, suggests that the dehydrogenation
reaction benets from an electron rich iridium centre.
Fig. 4 Ligands examined during initial catalyst development.
2
.1.4 The effect of the metal ion. In addition to the Cp*IrCl
16a
complexes 1–19, the rhodium analogue Rh1 was prepared
a
from L1 and [Cp*RhCl ] and its structure conrmed by X-ray
2 2
2
Table 1 The effect of ligand structure on the rate of H production
crystallographic analysis (see the ESI†). The synthesis of the
6
Ru(h ꢀp-cymene)Cl analogue was attempted but proved
unsuccessful. Comparison of 1 and Rh1 in the dehydrogenation
reaction shows that Rh1 is not stable under the reaction
conditions, with rapid catalyst decomposition and negligible
activity observed.
Initial
TOF h
Vol H
2
(2 h)/mL
ꢀ1
Entry
Precatalyst
TON (2 h)
2.1.5 Arriving at the optimal catalyst. In our initial
screening, complex 4 and the related 5 proved to be the two
most active of those screened (Table 1). However, analysis by H
1
2
3
4
5
6
7
8
1
2
3
4
1090
536
536
517
926
0
49
200
182
188c
258
298
0
1
44.5
46
NMR spectroscopy showed them both to be inseparable
c
63
23
mixtures of two regioisomers 4a/4b and 5a/5b, resulting from
5
6–13
[Cp*IrCl
None
73
0
0
competing cyclometallation of the C
NaOAc-mediated cyclometallation reaction (eqn (2)).
2 4
and C phenyl rings in the
16a
b
2
]
2
0
0
0
0
0
a
ꢁ
2
Reactions were performed under an N atmosphere at 25 C using 10
mmol of catalyst precursor and 1.5 mL of 5 : 2 F/T for 2 h. Initial TOF
values were calculated from the volume of gas collected in the rst 3
b
c
min (see the ESI† for further details). 5 mmol of dimer used. Aer 1 h.
inactive catalysts, suggesting that the g-NH functionality plays a
critical role in the activity of these catalysts (1–5). This is not
caused by the steric bulk (or lack of) of the N-substituent, as
both N-methyl and N-benzyl complexes were inactive.
The lifetime of the active catalysts appears to depend on the
substituents of the donor group. It was noted that during the
course of the dehydrogenation reaction, complexes 1–3 (entries
(
2)
1
–3) displayed high initial TOFs; but this activity decreased aer
ca. 0.5 h with an associated darkening of the catalyst solution. isomers, model compounds were synthesised in order to mimic
However, complexes bearing the bulky 4,5-diphenyl-2-imidazo- the selective cyclometallation at the C and C phenyl rings of
linyl group (4 and 5) displayed prolonged catalytic activity L4. Attempts to synthesise the analogue of 4 with a mesityl
entries 4 and 5), with no sign of darkening observed even aer group at the C position failed due to an unexpected benzylic
To gain insight into the relative catalytic activities of these
2
4
(
2
2
3
weeks in neat 5 : 2 F/T solution.
sp cyclometallation of the mesityl methyl group. We then
1
236 | Chem. Sci., 2013, 4, 1234–1244
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
a
2
Table 2 The effect of differing para-substituents on rate of H production by analogues of 1
ꢀ1
Entry
1
Ligand
Precatalyst
Initial TOF h
1090
Vol H
2
(2 h)/mL
TON (2 h)
200
1
49
2
3
4
5
6
14
15
16
17
18
1690
1390
460
110
30
76.5
30
47
3
312
122
192
12
1
4
7
19
1960
120
490
a
ꢁ
Reactions were performed under an N
were calculated as for Table 1.
2
atmosphere at 25 C using 10 mmol of catalyst precursor and 1.5 mL of 5 : 2 F/T for 2 h. Initial TOF values
attempted the cyclometallation of the related (S,S)-4,5-dimesi-
tyl-2-phenyl-imidazoline ligand with [Cp*IrCl , which unfor-
2 2
]
tunately displayed a pronounced lack of reactivity, even under
prolonged and forcing conditions, preventing the synthesis of
24
2
the desired iridium complex. However, substitution of the C -
phenyl group by cyclohexyl allowed for the formation of the
desired C -cyclometallation product 20 (eqn (3)), whilst the use
4
of 3,5-di-tert-butylphenyl groups at the C and C positions (L21)
4
5
resulted in selective cyclometallation on the C
2
phenyl ring to
give 21 in high yield (eqn (4)). Most interestingly, comparison of
the complexes 20 and 21 showed that 20 exhibited no observ-
ꢁ
able activity for dehydrogenation of 5 : 2 F/T at 25 C, whilst 21
ꢀ1
Fig. 5 Hammett plot for substituents para to the donor group based on initial gave an initial TOF of 980 h , far exceeding that of 4, revealing
ꢁ
TOFs. Reactions were performed under an N
2
atmosphere at 25 C using 10 mmol
the importance of cyclometallation position to the catalytic
activity.
of precatalyst and 1.5 mL of 5 : 2 F/T. Data points are an average of 3 or 4
measurements. Error bars show the standard deviation from the mean.
Armed with the structure activity relationships for each part
of the catalyst structure, i.e. the R group attached to the NH-
bearing dative ligand, the dative ligand group L, the cyclo-
metallated aryl ring and the metal ion (see Fig. 3), we reasoned
that the combination of the optimised structures of each of the
individual subunits would lead to an “optimised” precatalyst
structure, 22, with the key features shown in eqn (5). Synthesis
of complex 22 was achieved again using the method of Davies
16a
et al. and the structure was unambiguously determined by
a combination of multinuclear NMR, HRMS and single crystal
X-ray diffraction studies (Fig. 7). As with complex 19,
III
Fig. 6 Conjugation of the bis-alkoxy substituent to the Ir centre in 19.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 1234–1244 | 1237

View Article Online
Chemical Science
Edge Article
15
cyclometallation occurred selectively at the more hindered CO. A comparison of the activities of 4, 20, 21, and 22 is shown
position ortho to the oxygen substituent.
in Fig. 8. In particular, the initial TOF of 22 is 5 times that of 4,
the starting point of our optimisation process. We noted,
however, that the rate of dehydrogenation decreases with time.
(
3) 2.2 Continuous dehydrogenation of formic acid
In contrast to complex 1, the more bulky complexes 4, 5, 21 and
22 show excellent stability in neat 5 : 2 F/T. Solutions stored
under N remained a bright yellow colour for up to 2 weeks and
2
maintained their catalytic activities (if recharged with added
HCO
2
H). Thus, the decrease in the rates observed with time
(Fig. 8) may stem from the consumption of formic acid.
Reasoning that the reaction could be slowed by a decrease in
formic acid concentration as the reaction progresses and/or
possible coordination of NEt3 to the single active site, we
(
4)
ꢁ
explored further reactions with 22 at 40 C, in which the cata-
lytic system was recharged by addition of fresh formic acid at
regular intervals. Fig. 9 shows that the catalyst indeed main-
tains its activity if more formic acid is added; without this the
dehydrogenation becomes slower aer a few minutes.
To further demonstrate the durability of the catalyst, fresh
formic acid was added over a period of 2 h (Fig. 10). This
ꢀ
1
allowed high TOFs to be observed, with average TOF ¼ 3080 h
ꢀ
1
over the rst 1 h and 3340 h over 2 h, and the catalyst showed
no apparent decrease in the rate of H formation. The almost
2
2
linear dependence of the H volume on time shows the superb
stability of the catalyst and indicates that the dehydrogenation
is not rate-limited by the formation of Ir–H hydride under the
(
5)
2
recharging conditions or at high [HCO H] (vide infra).
During the course of studying the continuous dehydroge-
nation of formic acid, it was noted that exceptionally high
activity was observed immediately aer (1–2 minutes) the
reaction was recharged with additional formic acid. For
example, addition of 0.1 mL of neat formic acid to a mixture of
Gratifyingly, complex 22 indeed showed the highest activity
of any of the precatalysts examined in the dehydrogenation of
ꢁ
ꢀ1
5
: 2 F/T at 25 C, with the initial TOF being 2570 h under the
1
conditions given in Table 1. GC and H NMR analysis of the gas
evolved conrmed the formation of H , while analysis with FT-
2
IR showed that the gas is essentially free of CO (see ESI†), which
is crucial for its use in fuel cells that are oen poorly tolerant of
Fig. 8 The effect of cyclometallation position and ligand structure on catalytic
ꢁ
Fig. 7 X-ray structure of 22. Thermal ellipsoids are drawn at 50% probability. A
molecule of n-hexane is omitted for clarity and the disorder of a t-butyl group
activity. Reactions were performed under an N
2
atmosphere at 25 C using 10
mmol of precatalyst and 1.5 mL of 5 : 2 F/T. For clarity only the major regisomer of
complex 4 is shown (regioisomer 4a).
(
C28, C29, C30, C31) is not shown.
1
238 | Chem. Sci., 2013, 4, 1234–1244
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
as a simplied model of complex 22. As in transfer hydroge-
nation reactions using 5 : 2 F/T as the H source, coordination
2
of formate to the metal centre and subsequent b-hydride elim-
25
ination furnishes the metal hydride and CO2. This was
demonstrated by stoichiometric reactions of 1 with tetra-n-
26
butylammonium formate in CD
give solutions of the corresponding hydride complex 23 (eqn
6)). Alternatively, 23 could be obtained as a pure solid by
3
CN at room temperature to
(
reaction of 1 with sodium formate in a biphasic DCM–water
mixture with tetra-n-butylammonium formate as a phase
transfer catalyst (see ESI†). Complex 23 showed a single reso-
nance in the hydride region (d ꢀ15.5 ppm) and an NH reso-
nance (d 5.8 ppm) and remained stable both in the solid state
Fig. 9 Comparison of the dehydrogenation of 5 : 2 F/T by 22 with and without
HCO
2
H recharging (0.05 mL of formic acid at 7.5 min intervals). Reactions were
ꢁ
and in solution (CD
The long-lived nature of 23 in the absence of exogenous acid is
at odds with the rapid H evolution observed under catalytic
conditions, suggesting that spontaneous formation of H , via,
3 2
CN) for prolonged periods (>3 days, N , rt).
performed under an N
2
atmosphere at 40 C using 10 mmol of precatalyst and
initiated with 1.5 mL of 5 : 2 F/T.
2
2
for example, long distance protonation of the hydride by the NH
proton, does not occur from this species alone.
(
6)
When 1 or 23 was treated with an excess of 5 : 2 F/T in CD CN
3
at ambient temperature, visible H
of the Cp*, aliphatic and aromatic regions of the H NMR
sphere at 40 C using 10 mmol of precatalyst and initiated with 1.5 mL of 5 : 2 F/T. spectrum of the solution showed no peaks other than those
2
evolution occurred. Analysis
2
Fig. 10 Dehydrogenation of 5 : 2 F/T by 22 with HCO H recharging (0.05 mL of
1
formic acid at 7.5 min intervals). Reactions were performed under an N
2
atmo-
ꢁ
2
corresponding to 23, F/T or H , showing 23 to be both a plau-
sible intermediate in the catalytic cycle and the catalyst resting
state. The latter is consistent with the dehydrogenation not
ꢁ
1
0 mmol of 22 in 1.5 mL of azeotrope at 40 C led to a large
being controlled by hydride formation and suggests H forma-
tion to be the rate limiting step. A small second hydride peak
increase in rate, with the reproducible formation of 100 mL of
2
ꢀ1
gas within 60 s, corresponding to a TOF of 24 500 h for that
period. Even more dramatic was the addition of 0.5 mL of neat
formic acid to an identical mixture, which led to the repro-
ducible formation of 100 mL of gas within 10 s, corresponding
(
<10%) was also observed but NOESY experiments showed no
correlations between it and any other proton resonances. In
2 2
addition, it does not appear to be derived from the [Cp*IrCl ]
ꢀ
1
27
dimer or breakdown Cp*Ir products, consistent with the clean
high-eld H NMR spectrum of 23 obtained under catalytic
to a TOF of 147 000 h for that period (see ESI† for the calcu-
lations). However, as the additional formic acid was consumed
the rate of gas evolution returned to the previous level. These
results suggest that formic acid is likely to be involved in the
rate-determined step of the dehydrogenation. It must be noted,
however, that the addition of formic acid to solutions of F/T
containing inactive complexes, or no catalyst, did not result in
any gas evolution.
1
conditions. As yet, the identity of this minor species is
unknown.
1
Notably, the NH proton of 23 was not observed in the H
NMR spectrum under the catalytic conditions. Similarly, the
addition of soluble acetate or dimethylphosphate salts to pure
3
solutions of 23 in CD CN led to the disappearance of its NH
1
proton in the H NMR spectrum (see the ESI†, Section 3.3). This
is most likely a result of the NH unit engaging in hydrogen
bonding with the oxygen anions, suggesting that under catalytic
2.3 Mechanism of H formation
2
2
.3.1 Formation of hydrides. The high catalytic activity of conditions, 23 hydrogen-bonds with formate. However, the NH
2 and other complexes bearing an N]C–NH unit, and the lack resonance remained unchanged when the neutral hydrogen
or methyl 4-methoxybenzoate were added
feature prompted us to investigate the mechanism of H2 to pure solutions of 23 in CD CN.
formation and the role of the apparently crucial remote NH
2.3.2 Reactions of hydrides. In order to gain insight into
2
of activity of closely related complexes without this structural bond acceptors NEt
3
3
functionality. Initially we examined hydride formation using 1 the mechanism of H
2
formation, the reactivity of 23 was further
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 1234–1244 | 1239

View Article Online
Chemical Science
Edge Article
investigated. In addition, the closely related hydride 24, which protonated with either triphenylphosphonium tetra-
29
30
was readily formed from complex 6 (as with the case of 23 from uoroborate or 2,6-lutidinium tetrauoroborate to release
), was studied for comparison (Fig. 11), as its precursor 6 had H and form the phosphine and MeCN ligated cationic species,
1
2
been shown to be inactive for H2 formation. Both hydride respectively (Fig. 11). However, it must be noted that triphe-
complexes were found to be stable in dry CD CN or in the nylphosphonium and 2,6-lutidinium are approximately 16 and
presence of added H O for prolonged periods of time (>3 days, 9 pK units more acidic than HOAc and related carboxylic acids
, rt). Addition of a slight excess of HOAc or 1 equivalent of in MeCN, respectively. These results show that whilst both
Ph CCH CO H to solutions of 23 led to the instant disappear- hydrides 23 and 24 can be protonated with strong acids, they
ance of the hydride resonance and the formation of H , as display a distinct difference in their reactivity toward carboxylic
3
2
a
28
N
2
3
2
2
2
1
observed by H NMR spectroscopy. In sharp contrast, no reac- acids, providing an explanation as to why 23 is active, whilst 24
tion occurred with 24 (Fig. 11). As both 23 and 24 are formed is not, in the dehydrogenation of HCOOH.
rapidly when their parent chlorides are treated with formate,
this suggests that the feasibility of protonation of the hydride to importance of the NH functionality and the carboxylic acid to
form H may explain the differing reactivities of 1 and 6 for formation, which hydrogen-bond with each other under
formic acid dehydrogenation. catalytic conditions, we considered the two following possibil-
The NH functionality is crucial for any observable activity in ities for hydride protonation:
both the hydride protonation reaction and catalytic H forma- (1) Analogous to the Grotthuss mechanism for proton
2.3.3 Mechanistic possibilities. Bearing in mind the
2
H
2
31
2
tion. As substituting the NH for electron-donating (NMe, NBn) transfer in H O and biological systems, formic acid could
2
and electron-withdrawing groups (NAc) or oxygen results in a participate in a proton-hopping process, hydrogen-bonding to
total loss of catalyst activity (Table 1), the effects conferred by the NH proton while enabling protonation of the hydride. Two
the NH unit must stem from the presence of the proton on pathways could be envisioned, one involving transfer of the
nitrogen, rather than any steric or electronic effects. Thus, any formic acid proton directly to the hydride via one or more
potential mechanism in which the Ir–hydride complex is molecules of hydrogen-bonding formic acid to generate a
directly protonated without involvement of the NH unit would dihydrogen complex 25 (eqn (7)), and the other involving an
appear to be unlikely.
initial proton transfer to the iridium-bound nitrogen, forming
32
The structure of the acid was also found to be crucial. Whilst 27, followed by H formation via a Noyori type mechanism
2
carboxylic acids rapidly protonated 23, despite their exceedingly (eqn (8)). In both cases, the proton from the distal nitrogen is
28
low acidities in anhydrous MeCN, other Brønsted acids did transferred, regenerating the formic acid. Subsequent loss of H2
not. Thus, despite possessing acidities close to, or greater than, and addition of formic acid would result in the protonation of
that of acetic acid in MeCN, no reaction was observed between the distal nitrogen and formation of a metal formato complex
either 2-nitrophenol, 4-cyanophenol or 9-uorene-9-carboxylic 26, which regenerates the hydride 23 upon decarboxylation.
acid methyl ester and either 23 or 24, showing that both the Studies by Casey and coworkers on the mechanism of H
2
loss
5
structure of the acid and the hydride are crucial for efficient H
2
from [(2,5-Ph -3,4-Tol (h -C COH))Ru(CO) H] at high tempera-
2 2 4 2
formation. On the other hand, both 23 and 24 could be tures have highlighted the role of H O or alcohols in catalysing
2
the transfer of the acidic proton to the hydride to form a dihy-
33
drogen complex, which subsequently releases H2.
(
7)
Fig. 11 Reactions of 23 and 24 with a variety of proton sources. Unless specified,
reactions were performed under N with 1.0 equivalent of the acid in anhydrous
CD CN at room temperature. pK 's given are in MeCN.
Trt is trityl (triphenylmethyl).
2
(
8)
28
a
3
a
2
; Excess H O was used.
1
240 | Chem. Sci., 2013, 4, 1234–1244
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
The formation of 26 is likely to involve the intermediacy of a the NH proton of 23 was indicated by the disappearance of the
3
7
neutral 16-electron species generated from H dissociation from NH resonance, although the hydride resonance remained
2
2
1
5. This species, 28, was observed by H NMR on reacting 1 with unchanged. However, addition of 1 equivalent of 4-cyanophenol
1
CD
.05 equivalents of potassium tert-butoxide in anhydrous (see also Fig. 11) to this solution did not result in the loss of the
1
3
CN, along with a colour change from bright yellow to deep hydride signal or the formation of hydrogen, as observed by H
red/brown (eqn (9)). Interestingly, treatment of 28 with NMR, even aer 24 h. Although the protonation of 23 by
hydrogen in CD CN led to the ready formation of 23 in the carboxylic acids may be more favourable than with phenols or
3
absence of a carboxylic acid. The analogous reaction with tri- carbon-based acids due to the intermediacy of 4-membered
phenylphosphonium tetrauoroborate gave the phosphine transition states in the latter cases, the ability of carboxylic acids
ligated cation (eqn (10)), which was conrmed by its indepen- to protonate 23 in the absence of any additives makes an acid-
dent synthesis (vide supra), thus supporting the identity of 28. assisted proton transfer pathway a more likely one, in particular
The fact that 23 can be generated from 28 via hydrogenation eqn (8).
lends support to the pathway shown in eqn (8), as the hydro-
genation is expected to proceed via the intermediacy of 25 and dehydrogenation of formic acid by 22 and related complexes
then 27, i.e. the reverse reaction of H formation (eqn (8)). proceeds via the mechanism shown in Fig. 12. The key feature
Hydrogenation of related neutral complexes bearing a basic of the mechanism is that the catalytic turnover is rate-limited
amide ligand with H to generate metal hydrides is known, by the protonation of the Ir–H hydride, which is the resting
albeit slow, as shown by the work of Noyori, Ikariya and state of the catalyst, and that its protonation involves partici-
Taken together, we tentatively suggest that the catalytic
2
2
3
4
Rauchfuss. Although long-range activation of H by the Ir and pation of both formic acid and the distal NH functionality,
2
35
distal N atom cannot be ruled out, the stability of 23 is at odds without which catalysis does not occur. Most likely, the
with this possibility.
protonation proceeds via formic acid-assisted proton hopping
from the distal to the proximal nitrogen, whereupon proton-
ation takes place, forming the dihydrogen complex and even-
2
tually H .
3
Conclusions
(
9)
This study shows that well dened bifunctional Cp*IrCl(N^C)
complexes containing cyclometallated 2-aryl-imidazoline
ligands are excellent precatalysts for the dehydrogenation of
2 3 2 2
azeotropic HCO H–NEt mixtures, producing H and CO with
high turnover rates under exceptionally mild conditions. In
contrast to most other systems for this process, the modular
nature of the catalyst precursors allowed for the independent
optimisation of the constituent “modules”, which when
(
10)
(
2) As the NH proton is strongly hydrogen bonded under
catalytic conditions, it was considered that this hydrogen
bonding of the NH proton could increase the electron
density at the Ir centre, imparting increased hydricity to the
hydride ligand and thus aiding its intermolecular proton-
ation by formic acid. This is a process which may be
considered analogous to that operational in the catalytic
triad present in trypsin and similar proteases, in which
hydrogen bonding of an aspartate residue to a histidine
residue increases the basicity (electron density) of the histi-
dine imidazole, allowing it to activate a nearby serine as a
3
6
powerful nucleophile.
To investigate whether hydrogen bonding between 23 and
formate could play a role in the hydrogen forming step under
catalytic conditions, CD CN solutions of 23 were rst treated
3
with a slight excess of dimethylimidazolium dimethylphos-
phate or tetra-n-butylammonium acetate. Hydrogen bonding to
2
Fig. 12 Proposed catalytic cycle for the dehydrogenation of HCO H.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 1234–1244 | 1241

View Article Online
Chemical Science
Edge Article
combined, led to a rationally designed “optimal” metal–ligand
bifunctional catalyst that showed excellent activity, out-
performing all its predecessors.
6 (a) P. G. Jessop, T. Ikariya and R. Noyori, Nature, 1994, 368,
231; (b) P. G. Jessop, Y. Hsiao, T. Ikariya and R. Noyori, J.
Am. Chem. Soc., 1996, 118, 344; (c) P. Munshi, A. D. Main,
J. C. Linehan, C. C. Tai and P. G. Jessop, J. Am. Chem. Soc.,
2002, 124, 7963; (d) P. G. Jessop, in The Handbook of
Homogeneous Hydrogenation, ed. J. G. de Vries and C. J.
Elsevier, Wiley-VCH, Weinheim, 2007, p. 489; (e)
G. Laurenczy, F. Joo and L. Nadasdi, Inorg. Chem., 2000, 39,
5083; (f) C. Q. Yin, Z. T. Xu, S. Y. Yang, S. M. Ng,
K. Y. Wong, Z. Y. Lin and C. P. Lau, Organometallics, 2001,
20, 1216; (g) C. C. Tai, J. Pitts, J. C. Linehan, A. D. Main,
P. Munshi and P. G. Jessop, Inorg. Chem., 2002, 41, 1606;
(h) Y. Y. Ohnishi, T. Matsunaga, Y. Nakao, H. Sato and
S. Sakaki, J. Am. Chem. Soc., 2005, 127, 4021; (i)
A. Urakawa, F. Jutz, G. Laurenczy and A. Baiker, Chem.–Eur.
J., 2007, 13, 3886; (j) S. Ogo, R. Kabe, H. Hayashi,
R. Harada and S. Fukuzumi, Dalton Trans., 2006, 4657; (k)
Y. Himeda, N. Onozawa-Komatsuzaki, H. Sugihara and
K. Kasuga, Organometallics, 2007, 26, 702; (l) R. Tanaka,
M. Yamashita and K. Nozaki, J. Am. Chem. Soc., 2009, 131,
Evidence is provided, which shows that the remote NH
functionality is crucial to the catalysis, without which there is
no dehydrogenation, and suggests that formic acid plays a
dual role, acting both as a hydride and a proton source and as
a proton shuttle. The NH proton does not directly protonate
the Ir–H hydride, nor does formic acid. Instead, a formic acid
assisted-proton hopping may occur, resulting in proton
transfer from the remote to the proximal nitrogen atom,
whereupon protonation of the hydride and subsequent release
2
of H takes place. This process constitutes a rare example of
bifunctional catalysis, in which unusual long-range metal–
ligand cooperation effects the catalysis through a conventional
“short-range” metal–ligand bifunctional mechanism. Further
studies of hydrogen release from these bifunctional catalysts
and applications in hydrogen transfer reactions and H2
formation from other substrates will be reported in due
course.
14168; (m) C. Federsel, A. Boddien, R. Jackstell,
R. Jennerjahn, P. J. Dyson, R. Scopelliti, G. Laurenczy and
M. Beller, Angew. Chem., Int. Ed., 2010, 49, 9777; (n)
D. Preti, C. Resta, S. Squarcialupi and G. Fachinetti, Angew.
Chem., Int. Ed., 2011, 50, 12551; (o) R. Tanaka,
M. Yamashita, L. W. Chung, K. Morokuma and K. Nozaki,
Organometallics, 2011, 30, 6724; (p) T. J. Schmeier,
G. E. Dobereiner, R. H. Crabtree and N. Hazari, J. Am.
Chem. Soc., 2011, 133, 9274; (q) J. F. Hull, Y. Himeda,
W.-H. Wang, B. Hashiguchi, R. Periana, D. J. Szalda,
J. T. Muckerman and E. Fujita, Nat. Chem., 2012, 4, 383.
7 (a) Y. Gao, J. Kuncheria, R. J. Puddephatt and G. P. A. Yap,
Chem. Commun., 1998, 2365; (b) Y. Gao, J. K. Kuncheria,
H. A. Jenkins, R. J. Puddephatt and G. P. A. Yap, J. Chem.
Soc., Dalton Trans., 2000, 3212.
Acknowledgements
We thank the University of Liverpool for support, Dr John Bacsa
for X-ray structural analysis and the EPSRC National Mass
Spectrometry Service Centre for mass analysis.
Notes and references
1
(a) N. S. Lewis and D. Nocera, Proc. Natl. Acad. Sci. U. S. A.,
006, 103, 15729; (b) J. A. Turner, Science, 1999, 285, 687;
c) J. A. Turner, Science, 2004, 305, 972.
2
(
2
(a) L. Schlapbach and A. Zuttel, Nature, 2001, 414, 353; (b)
S. Fukuzumi, Eur. J. Inorg. Chem., 2008, 1351; (c)
P. Makowski, A. Thomas, P. Kuhn and F. Goettmann,
Energy Environ. Sci., 2009, 2, 480.
8 (a) D. J. Morris, G. J. Clarkson and M. Wills, Organometallics,
2009, 28, 4133; (b) A. Majewski, D. J. Morris, K. Kendall and
M. Wills, ChemSusChem, 2010, 3, 431.
3
(a) F. Jo ´o , ChemSusChem, 2008, 1, 805; (b) S. Enthaler,
ChemSusChem, 2008, 1, 801; (c) B. Loges, A. Boddien,
F. G ¨a rtner, H. Junge and M. Beller, Top. Catal., 2010, 53,
9 (a) B. Loges, A. Boddien, H. Junge and M. Beller, Angew.
Chem., Int. Ed., 2008, 47, 3962; (b) A. Boddien, B. Loges,
H. Junge and M. Beller, ChemSusChem, 2008, 1, 751; (c)
H. Junge, A. Boddien, F. Capitta, B. Loges, J. R. Noyes,
S. Gladiali and M. Beller, Tetrahedron Lett., 2009, 50, 1603;
(d) A. Boddien, B. Loges, H. Junge, F. G ¨a rtner, J. R. Noyes
and M. Beller, Adv. Synth. Catal., 2009, 351, 2517; (e)
A. Boddien, F. G ¨a rtner, C. Federsel, P. Sponholz,
D. Mellmann, R. Jackstell, H. Junge and M. Beller, Angew.
Chem., Int. Ed., 2011, 50, 6411.
902; (d) T. C. Johnson, D. J. Morris and M. Wills, Chem.
Soc. Rev., 2010, 39, 81; (e) H.-L. Jiang, S. K. Singh,
J.-M. Yan, X.-B. Zhang and Q. Xu, ChemSusChem, 2010, 3,
541; (f) M. Grasemann and G. Laurenczy, Energy Environ.
Sci., 2012, 5, 8171.
4
5
F. Jin, J. Yun, G. Li, A. Kishita, K. Tohji and H. Enomoto,
Green Chem., 2008, 10, 612.
(a) E. E. Benson, C. P. Kubiak, A. J. Sathrum and J. M. Smieja,
Chem. Soc. Rev., 2009, 38, 89; (b) H. Arakawa, M. Aresta,
J. N. Armor, M. A. Barteau, E. J. Beckman, A. T. Bell, 10 (a) C. Fellay, P. J. Dyson and G. Laurenczy, Angew. Chem., Int.
J. E. Bercaw, C. Creutz, E. Dinjus, D. A. Dixon, K. Domen,
D. L. DuBois, J. Eckert, E. Fujita, D. H. Gibson,
W. A. Goddard, D. W. Goodman, J. Keller, G. J. Kubas,
H. H. Kung, J. E. Lyons, L. E. Manzer, T. J. Marks,
K. Morokuma, K. M. Nicholas, R. Periana, L. Que Jr,
Ed., 2008, 47, 3966; (b) C. Fellay, N. Yan, P. J. Dyson and
G. Laurenczy, Chem.–Eur. J., 2009, 15, 3752; (c) W. Gan,
P. J. Dyson and G. Laurenczy, React. Kinet. Catal. Lett.,
2009, 98, 205; (d) G. Papp, J. Csorba, G. Laurenczy and
F. Jo ´o , Angew. Chem., Int. Ed., 2011, 50, 10433.
J. Rostrup-Nielson, W. M. H. Sachtler, L. D. Schmidt, 11 (a) A. Boddien, B. Loges, F. G ¨a rtner, C. Torborg, K. Fumino,
A. Sen, G. A. Somorjai, P. C. Stair, B. R. Stults and
W. Tumas, Chem. Rev., 2001, 101, 953.
H. Junge, R. Ludwig and M. Beller, J. Am. Chem. Soc., 2010,
132, 8924; (b) A. Boddien, F. G ¨a rtner, R. Jackstell, H. Junge,
1
242 | Chem. Sci., 2013, 4, 1234–1244
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
A. Spannenberg, W. Baumann, R. Ludwig and M. Beller,
Angew. Chem., Int. Ed., 2010, 49, 8993; (c) A. Boddien,
L. Barloy, C. Sirlin and M. Pfeffer, Organometallics, 2008,
27, 5852.
D. Mellmann, F. G ¨a rtner, R. Jackstell, H. Junge, 20 (a) C. Wang, A. Pettman, J. Basca and J. Xiao, Angew. Chem.,
P. J. Dyson, G. Laurenczy, R. Ludwig and M. Beller, Science,
011, 333, 1733.
2 (a) S. Fukuzumi, T. Kobayashi and T. Suenobu,
ChemSusChem, 2008, 1, 827; (b) S. Fukuzumi, T. Kobayashi
and T. Suenobu, J. Am. Chem. Soc., 2010, 132, 1496; (c)
Int. Ed., 2010, 49, 7548; (b) J.-X. Jiang, C. Wang, A. Laybourn,
T. Hasell, R. Clowes, Y. Z. Khimyak, J. Xiao, S. J. Higgins,
D. J. Adams and A. I. Cooper, Angew. Chem. Int. Ed., 2011,
50, 1072; (c) C. Wang, B. Villa-Marcos and J. Xiao, Chem.
Commun., 2011, 47, 9773.
2
1
Y. Maenaka, T. Suenobu and S. Fukuzumi, Energy Environ. 21 (a) M. P. Magee and J. R. Norton, J. Am. Chem. Soc., 2001, 123,
Sci., 2012, 5, 7360.
3 (a) Y. Himeda, Green Chem., 2009, 11, 2018; (b) Y. Himeda,
S. Miyazawa and T. Hirose, ChemSusChem, 2011, 4,
1778; (b) H. Guan, M. Iimura, M. P. Magee, J. R. Norton and
˚
1
G. Zhu, J. Am. Chem. Soc., 2005, 127, 7805; (c) J. B. Aberg,
J. S. M. Samec and J. E. B ¨a ckvall, Chem. Commun., 2006,
487.
2771.
14 (a) H. Koch and W. Haaf, Org. Synth., 1973, 5, 20; (b) 22 The effect of single substituents meta to the donor group
G. H. Coleman and D. Craig, Org. Synth., 1943, 2,
83.
5 J. J. Baschuk and X. Li, Int. J. Energy Res., 2001, 25, 695.
6 (a) D. L. Davies, O. Al-Duaij, J. Fawcett, M. Giardiello,
S. T. Hilton and D. R. Russell, Dalton Trans., 2003, 4132;
(para to the Ir centre) could not be thoroughly assessed as
the majority of ligands bearing a meta substituent gave rise
to mixtures of complexes with the substituent ortho and
para to the Ir centre. For reasons as yet unknown, only the
methyl substituted ligand gave a single cyclometallation
product with the methyl group para to the Ir centre. The
effect of placing an electron donating methyl substituent
meta to the donor group (para to Ir) was found to be
similarly benecial to the rate of catalysis (initial TOF 1226
5
1
1
(
b) D. L. Davies, S. M. A. Donald, O. Al-Duaij,
S. A. Macgregor and M. P ¨o lleth, J. Am. Chem. Soc., 2006,
28, 4210; (c) Y. Boutadla, D. L. Davies, R. C. Jones and
1
K. Singh, Chem.–Eur. J., 2011, 17, 3438; (d) L. Li,
W. W. Brennessel and W. D. Jones, Organometallics, 2009,
ꢀ1
h
) as the presence of a methyl group para to the donor
ꢀ1
28, 3492; (e) M. E. van der Boom and D. Milstein, Chem.
group (meta to Ir) (initial TOF 1390 h ).
Rev., 2003, 103, 1759.
23 A total of 7 resonances were observed in the case of 4 in the
range 5.0–6.5 ppm corresponding to the following: the NH
proton, the two methylidyne protons of each of a pair of
diastereoisomeric complexes resulting from cyclometallation
1
1
7 (a) T. Jerphagnon, A. J. A. Gayet, F. Berthiol, V. Ritleng,
N. Mr ˇs i ´c , A. Meetsma, M. Pfeffer, A. J. Minnaard,
B. L. Feringa and J. G. de Vries, Chem.–Eur. J., 2009, 15,
1
2780; (b) S. Arita, T. Koike, Y. Kayaki and T. Ikariya,
on the C
2
ring, and the pair of methylidyne protons from the
aryl ring.
Angew. Chem., Int. Ed., 2008, 47, 2447; (c) S. Arita, T. Koike,
Y. Kayaki and T. Ikariya, Organometallics, 2008, 27, 24 No reaction observed by H NMR spectroscopy aer 4 days at
complex resulting from cyclometallation on the C
4
1
ꢁ
2795.
40 C.
8 (a) G. E. Dobereiner and R. H. Crabtree, Chem. Rev., 2010, 25 (a) T. Koike and T. Ikariya, Adv. Synth. Catal., 2004, 346, 37;
1
10, 681; (b) J. Choi, A. H. R. MacArthur, M. Brookhart and
(b) T. Abura, S. Ogo, Y. Watanabe and S. Fukuzumi, J. Am.
Chem. Soc., 2003, 125, 4149; (c) N. Makihara, S. Ogo and
Y. Watanabe, Organometallics, 2001, 20, 497.
A. S. Goldman, Chem. Rev., 2011, 111, 1761; (c) M. Gupta,
C. Hagen, R. J. Flesher, W. C. Kaska and C. M. Jensen,
Chem. Commun., 1996, 2083; (d) M. Gupta, C. Hagen, 26 See ESI† for details of preparation of tetra-n-
W. C. Kaska, R. E. Cramer and C. M. Jensen, J. Am. Chem. butylammonium formate.
2
Soc., 1997, 119, 840; (e) M. Gupta, W. C. Kaska and 27 Analysis of the reactions of [Cp*IrCl and
2
]
2
, [Ir(Cl)COD]
CN under the same
NMR spectroscopy did not show
C. M. Jensen, Chem. Commun., 1997, 461; (f) W.-w. Xu,
G. P. Rosini, K. Krogh-Jespersen, A. S. Goldman, M. Gupta,
C. M. Jensen and W. C. Kaska, Chem. Commun., 1997,
IrCl
conditions by
3 2 3
$nH O with excess 5 : 2 F/T in CD
1
H
resonances which were consistent with that of the minor
hydride species (d ꢀ16.33 ppm). In light of this, and the
lack of any species that would be derived from the
breakdown of 23 (i.e. free ligand), the minor hydride
species remains unassigned.
2273; (g) F. Liu, E. B. Pak, B. Singh, C. M. Jensen and
A. S. Goldman, J. Am. Chem. Soc., 1999, 121, 4086; (h)
F. Liu and A. S. Goldman, Chem. Commun., 1999, 655; (i)
K. Zhu, P. D. Achord, X. Zhang, K. Krogh-Jespersen and
A. S. Goldman, J. Am. Chem. Soc., 2004, 126, 13044; (j) 28 (a) A. K u¨ tt, I. Leito, I. Kaljurand, L. Soov ¨a li, V. M. Vlasov,
G. C. Fortman, A. M. Z. Slawin and S. P. Nolan,
Organometallics, 2011, 30, 5487; (k) T. W. Graham,
C.-W. Tsang, X. Chen, R. Guo, W. Jia, S.-M. Lu, C. Sui-Seng,
C. B. Ewart, A. Lough, D. Amaroso and K. Abdur-Rashid,
Angew. Chem., Int. Ed., 2010, 49, 8708.
9 (a) J. B. Sortais, V. Ritleng, A. Voelklin, A. Holuigue, H. Smail,
L. Barloy, C. Sirlin, G. K. M. Verzijl, J. A. F. Boogers,
A. H. M. de Vries, J. G. de Vries and M. Pfeffer, Org. Lett.,
L. M. Yagupolskii and I. A. Koppel, J. Org. Chem., 2006, 71,
2829; (b) I. Kaljurand, A. K u¨ tt, L. Soov ¨a li, T. Rodima,
V. M ¨a emets, I. Leito and I. A. Koppel, J. Org. Chem., 2005,
70, 1019.
29 Triphenylphosphonium tetrauoroborate was prepared
using the method described in P. J. C. Hausoul,
A. N. Parvulescu, M. Lutz, A. L. Spek, P. C. A. Bruijnincx,
B. M. Weckhuysen and R. J. M. K. Gebbink, Angew. Chem.,
Int. Ed., 2010, 49, 7972.
1
2005, 7, 1247; (b) N. Pannetier, J.-B. Sortais, P. S. Dieng,
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 1234–1244 | 1243

View Article Online
Chemical Science
Edge Article
3
0 2,6-Lutidinium tetrauoroborate was prepared using
the method described in F. Santoro, M. Althaus,
C. Bonaccorsi, S. Gischig and A. Mezzetti, Organometallics,
C. A. Sandoval, T. Ohkuma, N. Utsumi, K. Tsutsumi,
K. Murata and R. Noyori, Chem.–Asian J., 2006, 1, 102; (c)
Z. M. Heiden and T. B. Rauchfuss, J. Am. Chem. Soc., 2006,
128, 13048.
2008, 27, 3866.
3
3
1 C. J. T. von Grotthuss, Ann. Chim., 1806, 58, 54.
2 (a) T. Ikariya, K. Murata and R. Noyori, Org. Biomol. Chem.,
35 C. Gunanathan, B. Gnanaprakasam, M. A. Iron,
L. J. W. Shimon and D. Milstein, J. Am. Chem. Soc., 2010,
132, 14763.
2006, 4, 393; (b) R. Noyori, M. Yamakawa and
S. Hashiguchi, J. Org. Chem., 2001, 66, 7931.
36 L. Polg ´a r, Cell. Mol. Life Sci., 2005, 62, 2161.
33 C. P. Casey, J. B. Johnson, S. W. Singer and Q. Cui, J. Am. 37 For an example of the observation of hydrogen bonding
Chem. Soc., 2005, 127, 3100.
4 (a) K. J. Haack, S. Hashiguchi, A. Fujii, T. Ikariya and
R. Noyori, Angew. Chem., Int. Ed. Engl., 1997, 36, 285; (b)
between an N-heterocylic H-bond donor moiety (indolic
NH) and basic anions by H NMR spectroscopy, see Q. Li,
Y. Guo and S. Shao, Analyst, 2012, 137, 4497.
1
3
1
244 | Chem. Sci., 2013, 4, 1234–1244
This journal is ª The Royal Society of Chemistry 2013