A simple iridicycle catalyst for efficient transfer hydrogenation of n-heterocycles in water

Article abstract of DOI:10.1002/chem.201500016

A cyclometalated iridium complex is shown to catalyse the transfer hydrogenation of various nitrogen heterocycles, including but not limited to quinolines, isoquinolines, indoles and pyridinium salts, in an aqueous solution of HCO₂H/HCO₂Na under mild conditions. The catalyst shows excellent functional-group compatibility and high turnover number (up to 7500), with catalyst loadings as low as 0.01 mol % being feasible. Mechanistic investigation of the quinoline reduction suggests that the transfer hydrogenation proceeds via both 1,2- and 1,4-addition pathways, with the catalytic turnover being limited by the step of hydride transfer. An easily accessible iridicycle catalyst effects the transfer hydrogenation of a wide variety of N-heterocycles in water, including quinolines, isoquinolines, indoles, quinoxalines, and pyridines. The catalyst shows excellent functional-group compatibility and high turnover number (up to 7500), even with low catalyst loadings.

Full text of DOI:10.1002/chem.201500016

DOI: 10.1002/chem.201500016
Full Paper
&
Homogeneous Catalysis
A Simple Iridicycle Catalyst for Efficient Transfer Hydrogenation of
N-Heterocycles in Water
Dinesh Talwar, Ho Yin Li, Emma Durham, and Jianliang Xiao*[a]
Chem. Eur. J. 2015, 21, 1 – 11
1
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Abstract: A cyclometalated iridium complex is shown to cat-
alyse the transfer hydrogenation of various nitrogen hetero-
cycles, including but not limited to quinolines, isoquinolines,
indoles and pyridinium salts, in an aqueous solution of
HCO2H/HCO2Na under mild conditions. The catalyst shows
excellent functional-group compatibility and high turnover
number (up to 7500), with catalyst loadings as low as
0.01 mol% being feasible. Mechanistic investigation of the
quinoline reduction suggests that the transfer hydrogena-
tion proceeds via both 1,2- and 1,4-addition pathways, with
the catalytic turnover being limited by the step of hydride
transfer.
Introduction
ductants and have very limited functional-group compatibili-
ties. Whereas heterogeneous catalysts containing Pd, Pt, Ni or
Rh on supported materials can reduce a range of heterocycles
even under atmospheric pressure of hydrogen, they often
have limited selectivity and the potential of over reduction.
Homogeneous catalysis has attracted much attention, due to
the easily controllable selectivities and reactivities through
ligand modification. Nevertheless, there are still significant
challenges in this area, including the improvement in turnover
number (TON) and turnover frequency (TOF), reduction in cost,
and the expansion of the reaction scope.
Saturated nitrogen heterocycles are frequently found in drug
and biologically active molecules (Scheme 1), such as oxamni-
quine (a schistosomicide),[1] paroxetine (a CCRI-type antidepres-
sant),[2] salsolinol (an endogenous monoamine oxidase inhibi-
Transfer hydrogenation (TH) is a reaction of great interest
due to its operational simplicity. However, in contrast to that
of ketones, the TH of heterocycles has been much less ex-
plored. Yamaguchi demonstrated that by using [{Cp*IrCl2}2] in
a mixture of iPrOH and H2O under refluxing conditions, a series
of quinolines could be fully reduced to tetrahydroquinolines.[11]
The presence of an acid considerably enhanced the reduction
rate, presumably by activating the quinoline through protona-
tion to form a quinolinium salt, which is easier to reduce.[10]
Frediani and co-workers reported a Rh–bipyridine catalyst that
could reduce quinoline and pyridines with a moderate conver-
sion by using iPrOH as the hydride source.[12] Crabtree identi-
fied a cationic IrI–N-heterocyclic carbene (NHC) catalyst (1a)
that could reduce quinolines to tetrahydroquinolines in iPrOH
with moderate yields (Scheme 2). Pyrazine showed complete
conversion to piperazine under the reported reaction condi-
Scheme 1. Bioactive molecules that contain a saturated nitrogen heterocy-
cle.
tor)[3] and CEPC (a serotonin 5-HT2C antagonist).[4] The most ob-
vious route to access these types of molecules is via the reduc-
tion of the corresponding unsaturated parent heterocycles,
which can be efficiently synthesised by cross-coupling and
classic heterocyclic chemistry. Nonetheless, this method only
has 0.8% occurrence rate among the medical chemist’s tool-
box, despite the fact that approximately 43% of the total phar-
maceutical compounds contain aliphatic amines.[5]
This must be a reflection of either limited supply of
the building blocks from commercial sources or sig-
nificant challenges at the late-stage reduction step.[5]
Reduction of nitrogen heterocycles has traditional-
ly been carried out by using heterogeneous hydro-
genation catalysis (e.g., Pd/C, Rh/C, Adams’s catalyst,
Raney nickel),[6,7] Birch reduction,[8] and more recent-
ly with homogenous hydrogenation catalysis.[9,10] Al-
though there are many examples in the literature,
Scheme 2. Examples of catalysts studied for TH of quinolines.
one or more significant limitations are always found under
those reaction conditions. For example, Birch and metal hy-
dride reduction require stoichiometric amounts of metallic re-
tions.[13] Other N-heterocycles, such as isoquinoline, indole and
N-methylated pyridinium iodide, were found to be inactive in
this system. A versatile, simple and yet highly active system
was recently reported by our group. By using [{Cp*RhCl2}2]
with KI as an additive, a range of N-heterocycles, including qui-
nolines, isoquinolines, quinoxalines and pyridinium salts, were
reduced in the HCO2H/Et3N (F/T) azeotrope with high yields
using just 0.01–0.2 mol% catalyst.[14,15] TH of indoles did not
proceed under the protocol, however, and the reduction of 4-
substituted quinoline was rather sluggish. For example, at
[a] D. Talwar, Dr. H. Y. Li, E. Durham, Prof. J. Xiao
Department of Chemistry, University of Liverpool
Liverpool, L69 7ZD (UK)
E-mail: jxiao@liv.ac.uk
Homepage: http://pcwww.liv.ac.uk/~jxiao
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500016.
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
2
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
a higher catalyst loading of 2 mol% and 50% KI, only 23%
conversion of 4-methyl quinoline was obtained after 24 h.
Asymmetric transfer hydrogenation (ATH) of N-heterocycles
has also been investigated, with organocatalysts[16] and, to
a lesser degree, homogeneous catalysts.[17] However, Hantzsch
esters (HEH) are predominantly used in organocatalysis, which
are expensive hydrogen donors compared with others that are
commercially available (e.g., tBuÀHEH=£70.5 per gram versus
HCO2H=£30 per litre).
Table 1. Optimisation of conditions for the TH of 2-methylquinoline.[a]
Entry
Catalyst
Hydride source
Solvent
Conv. [%][b]
1[c]
2[d]
3
4
5
6
7
8
1c
1c
1c
1c
1c
1d
1a
1b
HCO2H/HCO2Na (pH 4.5) H2O
HCO2H/HCO2Na (pH 4.5) H2O
HCO2H/HCO2Na (pH 4.5) H2O
HCO2H/HCO2Na (pH 2.5) H2O
HCO2H/HCO2Na (pH 6.5) H2O
HCO2H/HCO2Na (pH 4.5) H2O
HCO2H/HCO2Na (pH 4.5) H2O
HCO2H/HCO2Na (pH 4.5) H2O
>99
>99
70
20
<5
12
<1
8
The reaction conditions for both TH and ATH of N-heterocy-
cles are not yet ideal, as high catalyst loadings, high reaction
temperatures and/or a limited substrate scope are limitations
often encountered. Moreover, organic solvents are normally
used, which imposes an environmental impact. In addition, an
active, versatile catalyst capable of either hydrogenation or TH
of various N-heterocycles, such as quinolines, isoquinolines,
quinoxalines, indoles and pyridines, remains to be developed.
Recently, we[18] and other groups[19] have reported a series of
cyclometalated iridium complexes (iridicycles), some of which
have been successfully applied to the reduction of carbonyls
and imines and reductive amination of ketones with various
hydrogen sources.[18a–d,g,h,j,l,19d,i] Of particular relevance is that an
iridicycle complex also catalyses the hydrogenation with H2
(1 bar) of various quinolines, quinoxalines and indoles.[18c] The
hydrogenation appears to work well only in 2,2,2-trifluoroetha-
nol (TFE), albeit with a catalyst loading of ꢀ1 mol%. The cata-
lyst was not capable of hydrogenating more inert N-heterocy-
cles such as isoquinolines or pyridines. We recently developed
the cyclometalated iridium complex 1c (Scheme 2), which was
found to be robust for the TH of a range of a-substituted ke-
tones and for the reductive amination of ketones in water.[18h,l]
Herein, we report a TH protocol with this catalyst, which ena-
bles the efficient reduction of an ample variety of N-heterocy-
cles, including isoquinolines and pyridinium salts, at a low cat-
alyst loading in water without the addition of any organic sol-
vent.
9
[{Cp*RhCl2}2] HCO2H/HCO2Na (pH 4.5) H2O
[{Cp*RhCl2}2] HCO2H/HCO2Na (pH 4.5) H2O
5
4
38
64
68
1
3
13
4
15
n.r.
10[e,f]
11
[{Cp*IrCl2}2]
HCO2H/HCO2Na (pH 4.5) H2O
12[g]
13[g]
14[g]
15[g]
16[g]
17[g]
18
1c
1c
1c
1c
1c
1c
1c
1c
HCO2H/Et3N (5:2)
HCO2H/Et3N (5:2)
HCO2H/Et3N (5:2)
HCO2H/Et3N (5:2)
HCO2H/Et3N (5:2)
HCO2H/Et3N (5:2)
0.1m KOH/iPrOH
Et3SiH
MeOH
TFE
THF
Toluene
DCM
DMF
iPrOH
H2O
19[h]
[a] See the experimental section for reaction conditions. [b] Conversion
determined by 1H NMR spectroscopy. [c] Reaction conducted at 808C.
[d] Reaction conducted at 608C. [e] With 10 mol% KI. [f] This reaction pro-
ceeds well when run in HCO2H/Et3N (5:2) azeotrope in the absence of
water and in presence of KI as an additive (see ref. [14]). [g] 0.5 mL of
HCO2H/Et3N (5:2). [h] Et3SiH (20 equiv); n.r.=no reaction observed.
showed that it exhibited very limited substrate scope (see the
Supporting Information). For instance, TH of 3-methylquinoline
led to its tetrahydro variant only in 4% conversion after 20 h.
To establish whether aqueous conditions were optimum, other
hydride sources and solvents were also tested using 1c for the
TH of 2-methylquinoline. Apart from water, TH also worked in
MeOH and TFE with F/T as the hydrogen source, with slightly
lower or comparable conversions (64% and 68%, respectively;
Table 1, entries 12 and 13). Much lower conversions were re-
corded in nonprotic solvents, such as THF or DMF (<5% con-
version in 3 h; Table 1, entries 14 and 17). Other commonly
used hydride sources such as iPrOH and Et3SiH gave sluggish
rates under the reaction conditions (Table 1, entries 18 and 19).
Once the optimal TH condition for 2-methylquinoline had
been established, an array of 26 diversely substituted quino-
lines (2a–z) was hydrogenated in the aqueous formate solu-
tion of pH 4.5 (Table 2). The Ir-based catalyst 1c exhibited high
reactivity for all of the quinoline substrates examined. Thus,
unsubstituted quinoline 2d, 2-substituted quinoline 2a and 3-
substitued quinoline 2b were all effectively reduced at 308C
with excellent yields (Table 2, entries 1, 2 and 4). Increasing the
steric bulkiness at the 2-postion led to a decrease in conver-
sion, which could be compensated by increasing the reaction
temperature to reflux (e.g., 3e; Table 2, entry 5). Challenging 4-
substitued quinolines 2c and 2z were also reduced in high
yields, albeit requiring high temperatures (Table 2, entries 3
and 26).
Results and Discussion
In our previous study on the TH of a-substituted ketones in
water, the catalyst 1c (Scheme 2) exhibited the highest activity
at pH 4.5;[18l] therefore, the same condition was adopted for
the optimisation study here. 2-Methylquinoline (2a) was
chosen as a model substrate. TH of 2a gave full conversion
within 3 h with only 0.1 mol% loading of 1c at 808C or 608C,
in an aqueous formate solution of pH 4.5 (Table 1, entries 1
and 2). Gratifyingly, lowering the temperature to 308C also led
to a 70% conversion within 3 h (Table 1, entry 3). Screening of
the solution pH with 1c revealed that the reaction occurs only
within a certain window of acidic conditions (Table 1, entries 4
and 5). Thus, pH 4.5 was adopted for subsequent studies. In
contrast, the analogous Rh complex 1d only gave a 12% con-
version (Table 1, entry 6). Other catalysts that are known to be
active for the TH of quinolines (Scheme 2) showed much lower
activities under the reaction conditions employed (Table 1, en-
tries 7–11). Although the dimeric [{Cp*IrCl2}2] also led to moder-
ate conversion (38% in 3 h, Table 1, entry 11), further testing
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
3
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Table 2. TH of quinolines with 1c in water.[a]
Entry
1
Substrate
Product
Yield [%][b]
96
Entry
14
Substrate
Product
Yield [%][b]
3a
3b
3n
3o
94
2
93
90
90
84
15
93
98
90
91
3[c]
3c
3d
3e
16
3p
3q
3r
4
17[c,d]
5[c]
18[c]
6
7
3 f
94
97
19
20
3s
3t
92
82
3g
8
3h
97
21[c]
3u
62
9
3i
3j
95
98
22
23
3v
95
96
10
3w
11
3k
92
24[c]
3x
82
12
13
3l
95
96
25
3y
3z
95
84
3m
26[c,e]
[a] Reaction conditions (unless otherwise stated): 2 (2.5 mmol), 1c (0.1 mol%), HCO2H/HCO2Na aqueous solution (pH 4.5; 3 mL; prepared using 14.0 mmol
of HCO2H and 29.4 mmol of HCO2Na in 2.8 mL of H2O), 308C, stirred in a carousel tube for 14 h; [b] Yield of isolated product. [c] Reaction was carried out
1
at reflux. [d] Yield determined by H NMR spectroscopy. [e] 0.5 mol% 1c used.
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
4
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Functional groups including halogen (2g–k), ether (2m–o),
amine (2q), amide (2r), ester (2s), carboxylic acid (2t), hetero-
cycles (2v–x) and a trifluoromethyl group (2p) were all tolerat-
ed under the reaction conditions. They appear to have exhibit-
ed insignificant effect on the yields, with products isolated in
average yields of >90% (Table 2, entries 7–20 and 22–24).
Even a product bearing a highly sensitive functional group,
that is, boronic acid pinacol ester 3u, was isolated in 62%
(Table 2, entry 21), together with 30% of the deboronated
product 3a.
To demonstrate the potential usefulness of this method in
process chemistry, 2a was used as the model substrate for
a larger scale reduction. As shown in Scheme 3, 35.8 g
(0.25 mol) of 2a was effectively reduced with just 0.01 mol%
1c at 308C (75% conversion in 24 h; TON=7500). The product
was separated from the reaction mixture by a simple phase
separation and purified by fractional distillation. No special
equipment was required for this reaction, nor was an inert at-
mosphere necessary. In addition, no organic solvent was re-
Scheme 3. Large-scale TH of 2-methylquinoline in water.
Table 3. TH of isoquinolinium and pyridinium salts in water.[a]
Entry
1
Substrate
Product
Yield [%][b]
90
Entry
9
Substrate
Product
Yield [%][b]
90
5a
5b
7c
2
95
98
97
10
7d
94
81
82
3[c]
5c
11
7e
7 f
4[c]
5d
12[c]
5
6
5e
5 f
99
91
13[c]
7g
7h
95
92
14
7
7a
7b
90
72
15
16
7i
7j
82
80
8[d]
[a] Reaction conditions: isoquinolinium or pyridinium salt (2.5 mmol), 1c (1 mol%), HCO2H/HCO2Na aqueous solution (pH 4.5; 3 mL; prepared using
14.0 mmol of HCO2H and 29.4 mmol of HCO2Na in 2.8 mL of H2O), reflux, stirred in a carousel tube for 24 h (isoquinolinium) or 36 h (pyridinium). [b] Yield
1
of isolated product. [c] Yield determined by H NMR spectroscopy. [d] Isolated as debenzylated product after the column.
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
quired for the entire operation, and only minimal waste was
generated, showing the protocol to be greener than traditional
methods that use NaBH3CN in acetic acid or Raney nickel or
Pd/Al2O3 under high pressure of H2.[20,21]
Table 4. TH of indoles in water.[a]
Based on the successful results obtained for the TH of qui-
nolines with 1c, the substrate scope was expanded to more
challenging isoquinolines and pyridines. However, reduction of
isoquinoline and 2-phenylpyridine led to the recovery of the
starting material under the reaction conditions used in Table 2
(see the Supporting Information). It was thought that activat-
ing the substrate by quaternising the nitrogen atom would
lead to a higher activity.[10] This was indeed the case, and the
results obtained for isoquinoline and pyridine reduction are
shown in Table 3. Using optimised conditions, an array of six
isoquinolinium (4a–f) and ten pyridinium (6a–j) salts were re-
duced. Unsubstituted isoquinolinium, 1-methyl, 3-methyl and
6-methyl isoquinolinium salts gave the highest yields (>95%;
Table 3, entries 2–5). Increasing the steric bulk at the 1-position
by replacing the hydrogen or methyl group with a phenyl did
not affect the yield significantly (Table 3, entries 1–3). A func-
tional group, for example, bromide, was well tolerated under
the reaction conditions (Table 3, entry 6). Likewise, 2-substitut-
ed pyridinium salts (6a–e) were all reduced with good yields,
regardless of the nature of the functional groups (Table 3, en-
tries 7–11). Interestingly, substrates bearing an electron-with-
drawing group at the 4-position gave exclusively the fully re-
duced piperidines, whereas those having an electron-donating
group led to the partially reduced 3,4-unsaturated piperidines
(Table 3, entries 13 and 14 versus 15 and 16). This phenomen-
on could be explained by a competitive 1,2-hydride addition
versus 1,4-hydride addition (see below). Having an electron-
withdrawing substituent probably renders the 4-position more
electrophilic, favouring the 1,4-addition.
Entry
1
Substrate
Product
Yield [%][b]
9a
9b
9c
9d
9e
96
2
94
30
92
78
3[c,d]
4
5[c]
6[c,d]
9 f
33
7
9g
n.r.
[a] Reaction conditions (unless otherwise stated): Indole (2.5 mmol), 1c
(0.1 mol%), HCO2H/HCO2Na aqueous solution (pH 4.5; 3 mL; 14.0 mmol
of HCO2H and 29.4 mmol of HCO2Na in 2.8 mL of H2O), 308C, stirred in
a carousel tube for 16 h. [b] Yield of isolated product; n.r.=no reaction
observed. [c] Using 0.5 mol% 1c, at reflux and with the addition of
1
MeOH (1 mL). [d] Yield determined by H NMR spectroscopy.
The substrate scope of indoles was examined next with 1c
under the conditions of Table 2. A range of indoles with both
electron-donating and electron-withdrawing groups were re-
duced to the corresponding indolines in good yields (Table 4).
However, TH of 5-bromoindole 8c gave a lower yield (Table 4,
entry 3). For this substrate, a thick layer of coating was always
observed on the reaction vessel above the solvent level, even
at reflux and with the addition of MeOH as a co-solvent. This
reflects that the solubility of 8c was an issue under the reac-
tion conditions employed, which might have led to the lower
conversion. Disappointingly, TH of the sterically hindered 2-
phenylindole failed to proceed under the present reaction con-
ditions, and 3-methylindole gave a low yield (Table 4, entries 6
and 7). One of the explanations could be the unfavourable tau-
tomerisation of 8 f or difficulty in its protonation at the 3-posi-
tion due to steric effects, following which 1,2-hydride addition
could occur.
former was isolated exclusively as its mono N-formyl derivative
(Table 5, entries 1–3). Interestingly, 1H-cyclopenta[b]pyridine
10d was reduced at the carbocycle ring to give the pyridine
11 d (Table 5, entry 4). Both the cyclic and acyclic imines were
fully reduced to give the corresponding amines 11 e and 11 f,
respectively, with good yields (Table 5, entries 5 and 6). Salsoli-
dine (11 e) is isolated from the plants of the genus salsola and
is a stereoselective competitive inhibitor of the enzyme mono-
amine oxidase.[22]
The reduction of quinolines in an acidic medium has been
suggested to proceed by an ionic or outer-sphere pathway.[23]
The initial hydride delivery to the protonated quinoline may
occur in a 1,4-addition fashion; subsequent isomerisation and
further reduction via 1,2-addition would afford the product
(Scheme 4). If the reaction is initiated by 1,2-hydride addition,
the resulting 1,2-dihydroquinoline may not be further reduced;
but it may undergo dehydrogenation to go back to the start-
ing material or disproportionate.[24] To gain more insight into
the reaction mechanism, a combination of reactions of inter-
mediates, isotope labelling and stoichiometric reactions was
explored.
To further demonstrate the potential of 1c as a versatile cat-
alyst for the TH of a range of heterocycles, rather than a speci-
alised one for a particular class of substrates, a range of diverse
substrates, including cyclic and acyclic imines and other fused
heterocycles, were examined (Table 5). Quinoxaline (10a), acri-
dine (10b) and neocuproine (10c) were all reduced to their
corresponding products in excellent yields, although the
Deuterium labelling reactions were carried out on the model
substrate 2a with 1c at 308C in water. By using fully deuterat-
ed reagents and solvent, 87%, 94% and 100% deuterium in-
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 5. TH of other N-heterocycles and imines in water.[a]
Entry
1[c]
Substrate
Product
Yield [%][b]
11 a 90
2
11 b 82
One possible explanation for deuterium [Eq. (2)] and hydro-
gen [Eq. (3)] incorporation at the 2- and 3-positions is that fol-
lowing the formation of the iridium hydride/deuteride, the
transfer of the hydride/deuteride to the substrate is the rate-
limiting step. This would allow the iridium hydride/deuteride
to be scrambled with the solvent [Eqs. (4) and (5)], producing
a mixture of iridium hydride and deuteride and leading to the
partial incorporation of deuterium into the product.[25] The re-
action shown in Eq. (3) also reveals that when H2O was used as
the solvent, no deuterium was incorporated at the 3-position.
This is consistent with the assumption that there is an acid-
mediated enamine–imine isomerisation reaction between the
hydride addition steps (Scheme 4).
3
11 c
96
4[c]
11 d 99
11 e 98
5
6[d]
11 f
90
[a] Reaction conditions (unless otherwise stated): 10 (2.5 mmol), 1c
(0.1 mol%), HCO2H/HCO2Na aqueous solution (pH 4.5; 3 mL; 14.0 mmol
of HCO2H and 29.4 mmol of HCO2Na in 2.8 mL of H2O), reflux, stirred in
a carousel tube for 16 h. [b] Yield of isolated product. [c] Reaction con-
ducted at reflux. [d] Reaction conducted at 808C and yield determined by
1H NMR spectroscopy.
Further support to the hydride transfer being rate limiting
was gained by monitoring the reduction of the protonated 2a
1
(pKa 5.4) with 1c using in situ H NMR spectroscopy (see the
Supporting Information). As noted before, neutral 2a could
not be reduced with an isolated, closely-related iridium hydri-
de.[18c] The reaction was carried out in an NMR tube equipped
with a Young’s tap, containing 1 equivalent of 1c and 5 equiv-
alents of 2a.HBF4 in [D3]MeCN. After the addition of 5 equiva-
lents of F/T, a hydride signal was immediately detected at d=
À15.8 ppm.[26] Whereas the signal of the product tetrahydro-
quinoline gradually increased in intensity over time, signals
corresponding to the potential intermediates 2a1, 2a2 and 2a3
were not observed. However, the hydride signal remained,
which, together with its rapid formation, is consistent with the
assumption that the transfer hydrogenation in question is turn-
over-limited by the hydride-transfer step.
Scheme 4. Possible reaction pathways for the TH of quinolines.
corporation was observed at the 2-, 3- and 4-positions of the
product, respectively [Eq. (1)]. When HCO2Na and HCO2H were
used together with D2O, 52%, 49%, and 55% deuterium incor-
poration took place at the 2-, 3- and 4-position of the product,
respectively [Eq. (2)]. However, when DCO2D and DCO2Na were
used in H2O, only 18%, 0%, and 14% deuterium incorporation
occurred at these positions, respectively [Eq. (3)].
The TH of 2a may yield two distinct intermediates, namely
1,2-dihydroquinoline (via 1,2-addition) and 3,4-dihydroquino-
line (via 1,4-addition; Scheme 4). To gain evidence into their
possible involvement in the reduction, a 1:1 mixture of dihy-
droquinoline 2l1 and quinoline 2l was subjected to the stan-
dard reduction conditions [Eq. (6)]. In the presence of 1c, only
the fully reduced tetrahydroquinoline 3l (100%) was obtained
after the reaction. However, in the absence of 1c, 23% of 3l
and 77% of 2l were obtained. In both cases, no starting dihy-
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
7
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
droquinoline 2l1 remained after the reaction. These results
supports the hypothesis that the 1,2-addition product (e.g.,
2l1) is consumed via a disproportionation pathway instead of
being reduced by the catalysis of 1c. Further evidence sup-
porting this hypothesis comes from the observation that when
2l2 was used as a substrate, no reaction occurred [Eq. (7)].
Scheme 5. TH of N-cinnamylidene aniline with 1c in water.
3a in the same way as in Pathway 1. While the pathways
1 and 2 are competitive, the rate of each is likely to be affected
by both steric and electronic effects arising from the catalyst
as well as substrate.
Conclusions
In summary, a wide variety of N-heterocycles, including, but
not limited to, quinolines, isoquinolines, indoles, quinoxalines
and pyridinium salts, have been effectively reduced using an
iridicycle catalyst in water. This reaction is applicable to large-
scale synthesis with no need for specialised equipment. The
use of environmentally benign solvent, renewable hydride
donor and easy workup and purification provides a significant
advantage for practical applications. To our knowledge, this
work constitutes the first example of a versatile homogeneous
To probe whether the 1,4-addition precedes the 1,2-addition
in the TH or vice versa, a model substrate, N-cinnamylidene
aniline 12a, was subjected to
the 1c-catalysed reduction.
Under the standard reaction
conditions, a mixture of 13a
and 13b (43:57) was obtained
(Scheme 5), indicating that both
1,2- and 1,4-hydride additions
are likely to happen for quino-
line-type substrates.
On the basis of the experi-
mental results presented above,
a simplified mechanism is pro-
posed for the TH of quinolines
(Scheme 6). Catalyst 1c reacts
with formate to generate the
active Ir-H species that can
react with the 2ax in two differ-
ent pathways. In Pathway 1,
2ax undergoes 1,4-addition to
give the 1,4-dihydroquinoline
2a2, which equilibrates with 2a3
under the acidic conditions. Fur-
ther reduction by a new hydride
via 1,2-addition then yields 3a.
Pathway 2 involves the 1,2-hy-
dride addition as the first step
to give 1,2-dihydroquinoline
2a1, which acts as a hydrogen
donor, reducing 2ax to 2a2
with itself turning into 2ax. The
intermediate 2a2 is reduced to
Scheme 6. Proposed mechanism for the TH of quinolines catalysed by 1c.
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
8
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
catalyst that can reduce a range of N-heterocycles in water
under the TH conditions. In addition, the iridicycle has been
shown to be highly effective in TH of various carbonyl com-
pounds in water and in reductive amination to afford primary
amines.[18a,l] Further work including preparation of a chiral ver-
sion of this and related iridicycles is ongoing in our laboratory.
Acknowledgements
We thank EPSRC and University of Liverpool for a DTA student-
ship (D.T.) and EPSRC/TSB for a postdoctoral fellowship (H.Y.L.).
Keywords: green chemistry · iridium · N-heterocycles · transfer
hydrogenation · water
[1] L. Pica-Mattoccia, D. Cioli, Am. J. Trop. Med. Hyg. 1985, 34, 112–118.
[2] M. A. Katzman, CNS drugs 2009, 23, 103–120.
Experimental Section
[3] M. Naoi, W. Maruyama, G. M. Nagy, NeuroToxicology 2004, 25, 193–204.
[4] J. D. McConvy, A. A. Harland, R. Maglathlin, D. E. Nichols, Neurosci. Lett.
2011, 505, 3–10.
Typical procedure for the TH of quinolines
Quinoline (2.5 mmol) and 1c (1.6 mg, 2.5ꢁ10À3 mmol) were placed
in a carousel reaction tube. HCO2H/HCO2Na aqueous solution of
pH 4.5 (3 mL; prepared using 14.0 mmol of HCO2H and 29.4 mmol
of HCO2Na in 2.8 mL of H2O) was then introduced and the mixture
was stirred at 308C for 14 h. The reaction mixture was quenched
with saturated sodium bicarbonate solution. The aqueous layer
was extracted with ethyl acetate (3ꢁ10 mL) and the combined or-
ganic layers were washed with brine (20 mL). The organic layer
was collected and dried over anhydrous sodium sulfate. Filtration,
followed by evaporation of the solvent under reduced pressure,
gave the crude mixture that was purified with flash column chro-
matography to afford the desired 1,2,3,4-tetrahydroquinoline.
[5] a) T. W. J. Cooper, I. B. Campbell, S. J. F. Macdonald, Angew. Chem. Int. Ed.
2010, 49, 8082–8091; Angew. Chem. 2010, 122, 8258–8267; b) S. D.
Roughley, A. M. Jordan, J. Med. Chem. 2011, 54, 3451–3479.
[6] a) R. A. W. Johnstone, A. H. Wilby, I. D. Entwistle, Chem. Rev. 1985, 85,
129–170; b) R. L. Augustine, Heterogeneous Catalysis for the Synthetic
Chemist, CRC Press, Boca Raton, FL, 1995; c) F. Glorius, N. Spielkamp, S.
Holle, R. Goddard, C. W. Lehmann, Angew. Chem. Int. Ed. 2004, 43,
2850–2852; Angew. Chem. 2004, 116, 2910–2913.
[7] Science of Synthesis: Houben–Weyl Methods of Molecular Transformation,
Vol. 15 (Ed.: D. St. C. Black), Georg Thieme Verlag, Stuttgart, 2005.
[8] J. T. Kim, V. Gevorgyan, J. Org. Chem. 2005, 70, 2054–2059.
[9] a) The Handbook of Homogeneous Hydrogenation (Eds.: J. G. de Vries,
C. J. Elsevier), Wiley-VCH, Weinheim, 2007; b) C. Bianchini, P. Barbaro, M.
Macchi, A. Meli, F. Vizza, Helv. Chim. Acta 2001, 84, 2895–2923; c) F. Glo-
rius, Org. Biomol. Chem. 2005, 3, 4171–4175; d) D. S. Wang, Q. A. Chen,
S. M. Lu, Y. G. Zhou, Chem. Rev. 2012, 112, 2557–2590; e) G. E. Doberein-
er, A. Nova, N. D. Schley, N. Hazari, S. J. Miller, O. Eisenstein, R. H. Crab-
tree, J. Am. Chem. Soc. 2011, 133, 7547–7562; f) E. Baralt, S. J. Smith, J.
Hurwitz, I. T. Horvath, R. H. Fish, J. Am. Chem. Soc. 1992, 114, 5187–
5196; g) T. Wang, L.-G. Zhuo, Z. Li, F. Chen, Z. Ding, Y. He, Q.-H. Fan, J.
Xiang, Z.-X. Yu, A. S. C. Chan, J. Am. Chem. Soc. 2011, 133, 9878–9891;
h) W.-B. Wang, S.-M. Lu, P.-Y. Yang, X.-W. Han, Y.-G. Zhou, J. Am. Chem.
Soc. 2003, 125, 10536–10537; i) G. Zhu, K. Pang, G. Parkin, J. Am. Chem.
Soc. 2008, 130, 1564–1565; j) W.-J. Tang, J. Tan, L.-J. Xu, K.-H. Lam, Q.-H.
Fan, A. S. C. Chan, Adv. Synth. Catal. 2010, 352, 1055–1062.
Typical procedure for the TH of isoquinolinium and pyridini-
um salts
Isoquinolinium or pyridinium (2.5 mmol) and 1c (16 mg, 2.5ꢁ
10À2 mmol) were placed in a carousel reaction tube. HCO2H/
HCO2Na aqueous solution of pH 4.5 (3 mL; 14.0 mmol of HCO2H
and 29.4 mmol of HCO2Na in 2.8 mL of H2O) was then introduced
and the mixture was stirred at reflux temperature for 24 h (36 h for
pyridinium). The reaction mixture was quenched with saturated
sodium bicarbonate solution. The aqueous layer was extracted
with ethyl acetate (3ꢁ10 mL) and the combined organic layers
were washed with brine (20 mL). The organic layer was collected
and dried over anhydrous sodium sulfate. Filtration, followed by
evaporation of the solvent under reduced pressure, gave the crude
mixture that was purified with flash column chromatography to
afford the desired 1,2,3,4-tetrahydroisoquinoline or piperidine.
[10] Y.-G. Zhou, Acc. Chem. Res. 2007, 40, 1357–1366.
[11] K. I. Fujita, C. Kitatsuji, S. Furukawa, R. Yamaguchi, Tetrahedron Lett.
2004, 45, 3215–3217.
[12] P. Frediani, L. Rosi, L. Cetarini, M. Frediani, Inorg. Chim. Acta 2006, 359,
2650–2657.
[13] A. M. Voutchkova, D. Gnanamgari, C. E. Jakobsche, C. Butler, S. J. Miller,
J. Parr, R. H. Crabtree, J. Organomet. Chem. 2008, 693, 1815–1821.
[14] J. Wu, C. Wang, W. Tang, A. Pettman, J. Xiao, Chem. Eur. J. 2012, 18,
9525–9529.
[15] J. Wu, W. Tang, A. Pettman, J. Xiao, Adv. Synth. Catal. 2013, 355, 35–40.
[16] a) M. Rueping, A. R. Antonchick, T. Theissmann, Angew. Chem. Int. Ed.
2006, 45, 3683–3686; Angew. Chem. 2006, 118, 3765–3768; b) M. Ruep-
ing, T. Theissmann, Chem. Sci. 2010, 1, 473–476; c) M. Rueping, T.
Theissmann, M. Stoeckel, A. P. Antonchick, Org. Biomol. Chem. 2011, 9,
6844–6850; d) M. Rueping, F. Tato, F. R. Schoepke, Chem. Eur. J. 2010,
16, 2688–2691; e) M. Rueping, A. P. Antonchick, Angew. Chem. Int. Ed.
2007, 46, 4562–4565; Angew. Chem. 2007, 119, 4646–4649; f) C. Metal-
linos, F. B. Barrett, S. Xu, Synlett 2008, 720–724; g) Q. S. Guo, D. M. Du,
J. Xu, Angew. Chem. Int. Ed. 2008, 47, 759–762; Angew. Chem. 2008,
120, 771–774.
[17] a) D. W. Wang, W. Zeng, Y. G. Zhou, Tetrahedron: Asymmetry 2007, 18,
1103–1107; b) X. F. Tu, L. Z. Gong, Angew. Chem. Int. Ed. 2012, 51,
11346–11349; Angew. Chem. 2012, 124, 11508–11511; c) D. W. Wang,
D. S. Wang, Q. A. Chen, Y. G. Zhou, Chem. Eur. J. 2010, 16, 1133–1136;
d) V. Parekh, J. A. Ramsden, M. Wills, Tetrahedron: Asymmetry 2010, 21,
1549–1556; e) C. Wang, C. Q. Li, X. F. Wu, A. Pettman, J. L. Xiao, Angew.
Chem. Int. Ed. 2009, 48, 6524–6528; Angew. Chem. 2009, 121, 6646–
6650.
Typical procedure for the TH of indoles, imines and other N-
heterocycles
Imine or N-heterocycle (2.5 mmol) and 1c (1.6 mg, 2.5ꢁ10À3 mmol)
were placed in a carousel reaction tube. HCO2H/HCO2Na aqueous
solution of pH 4.5 (3 mL; 14.0 mmol of HCO2H and 29.4 mmol of
HCO2Na in 2.8 mL of H2O) was then introduced (for substrate 8c,
8e and 8 f, 1 mL of MeOH was added) and the mixture was stirred
at 308C (or at reflux; refer to Tables 4 and 5 for the reaction tem-
perature) for 16 h. The reaction mixture was quenched with satu-
rated sodium bicarbonate solution. The aqueous layer was extract-
ed with ethyl acetate (3ꢁ10 mL) and the combined organic layers
were washed with brine (20 mL). The organic layer was collected
and dried over anhydrous sodium sulfate. Filtration, followed by
evaporation of the solvent under reduced pressure, gave the crude
mixture that was purified with flash column chromatography to
afford the desired product.
[18] a) D. Talwar, N. P. Salguero, C. M. Robertson, J. L. Xiao, Chem. Eur. J.
2014, 20, 245–252; b) W. J. Tang, C. H. Lau, X. F. Wu, J. L. Xiao, Synlett
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
9
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
2014, 25, 81–84; c) J. Wu, J. H. Barnard, Y. Zhang, D. Talwar, C. M. Rob-
ertson, J. L. Xiao, Chem. Commun. 2013, 49, 7052–7054; d) Y. W. Wei, D.
Xue, Q. Lei, C. Wang, J. L. Xiao, Green Chem. 2013, 15, 629–634; e) Y. W.
Wei, C. Wang, X. Jiang, D. Xue, J. Li, J. L. Xiao, Chem. Commun. 2013, 49,
5408–5410; f) C. Wang, H. Y. T. Chen, J. Bacsa, C. R. A. Catlow, J. L. Xiao,
Dalton Trans. 2013, 42, 935–940; g) B. Villa-Marcos, W. J. Tang, X. F. Wu,
J. L. Xiao, Org. Biomol. Chem. 2013, 11, 6934–6939; h) Q. Lei, Y. W. Wei,
D. Talwar, C. Wang, D. Xue, J. L. Xiao, Chem. Eur. J. 2013, 19, 4021–
4029; i) J. H. Barnard, C. Wang, N. G. Berry, J. L. Xiao, Chem. Sci. 2013, 4,
1234–1244; j) C. Wang, A. Pettman, J. Basca, J. L. Xiao, Angew. Chem.
Int. Ed. 2010, 49, 7548–7552; Angew. Chem. 2010, 122, 7710–7714; k) J.
Wu, D. Talwar, S. Johnston, M. Yan, J. Xiao, Angew. Chem. Int. Ed. 2013,
52, 6983–6987; Angew. Chem. 2013, 125, 7121–7125; l) D. Talwar, X.
Wu, O. Saidi, N. P. Salguero, J. L. Xiao, Chem. Eur. J. 2014, 20, 12835–
12842.
D. Bellus), Georg Thieme Verlag, Stuttgart, 2006; c) C. Metallinos, F. B.
Barrett, Y. Wang, S. Xu, N. J. Taylor, Tetrahedron 2006, 62, 11145–11157.
[21] a) Comprehensive Organic Synthesis, Vol. 8: Reduction (Eds.: B. M. Trost, I.
Fleming), Pergamon, Oxford, 1991; b) J. A. Joule, K. Mills, Heterocyclic
Chemistry, 5th ed., Wiley-Blackwell, Chichester, 2010; c) P. Stanetty, O. Ka-
semann, K. Mereiter, M. D. Mihovilovic, ARKIVOC 2005, 2005, 83–95.
[22] T. F. Kaufman, Tetrahedron: Asymmetry 2004, 15, 1203–1237.
[23] a) H. F. Zhou, Z. W. Li, Z. J. Wang, T. L. Wang, L. J. Xu, Y. He, Q. H. Fan, J.
Pan, L. Q. Gu, A. S. C. Chan, Angew. Chem. Int. Ed. 2008, 47, 8464–8467;
Angew. Chem. 2008, 120, 8592–8595; b) R. M. Bullock, Chem. Eur. J.
2004, 10, 2366–2374; c) C. Wang, B. Villa-Marcos, J. L. Xiao, Chem.
Commun. 2011, 47, 9773–9785; d) Y. M. He, F. T. Song, Q. H. Fan, Top.
Curr. Chem. 2013, 343, 145–190.
˝
[24] a) G. Eros, K. Nagy, H. Mehdi, I. Pꢂpai, P. Nagy, P. Kirꢂly, G. Tꢂrkꢂnyi, T.
Soꢃs, Chem. Eur. J. 2012, 18, 574–585; b) T. P. Forrest, G. A. Dauphinee,
S. A. Deraniyagala, Can. J. Chem. 1985, 63, 412–417; c) N. S. Mani, P.
Chen, T. K. Jones, J. Org. Chem. 1999, 64, 6911–6914; d) D. E. Minter,
P. L. Stotter, J. Org. Chem. 1981, 46, 3965–3970; e) M. Brasse, J. A.
Ellman, R. G. Bergman, Chem. Commun. 2011, 47, 5019–5021.
[25] a) M. C. Carriꢃn, M. R. CastaÇeda, G. Espino, C. Aliende, L. Santos, A. M.
Rodrꢄguez, B. R. Manzano, F. A. Jalꢃn, A. Lledꢃs, ACS Catal. 2014, 4,
1040–1053; b) Y. Himeda, S. Miyazawa, N. O. Komatsuzaki, T. Hirose, K.
Kasuga, Dalton Trans. 2009, 6286–6288.
[19] a) B. Li, T. Roisnel, C. Darcel, P. H. Dixneuf, Dalton Trans. 2012, 41,
10934–10937; b) E. Kumaran, W. K. Leong, Organometallics 2012, 31,
4849–4853; c) A. J. Howarth, D. L. Davies, F. Lelj, M. O. Wolf, B. O. Pat-
rick, Dalton Trans. 2012, 41, 10150–10152; d) N. Pannetier, J. B. Sortais,
J. T. Issenhuth, L. Barloy, C. Sirlin, A. Holuigue, L. Lefort, L. Panella, J. G.
de Vries, M. Pfeffer, Adv. Synth. Catal. 2011, 353, 2844–2852; e) D. L.
Davies, M. P. Lowe, K. S. Ryder, K. Singh, S. Singh, Dalton Trans. 2011, 40,
1028–1030; f) Y. Boutadla, D. L. Davies, R. C. Jones, K. Singh, Chem. Eur.
J. 2011, 17, 3438–3448; g) Y. Boutadla, D. L. Davies, O. Al-Duaij, J. Faw-
cett, R. C. Jones, K. Singh, Dalton Trans. 2010, 39, 10447–10457; h) D. L.
Davies, O. Al-Duaij, J. Fawcett, M. Giardiello, S. T. Hilton, D. R. Russell,
Dalton Trans. 2003, 4132–4138; i) S. Arita, T. Koike, Y. Kayaki, T. Ikariya,
Organometallics 2008, 27, 2795–2802.
[26] The active hydride catalyst derived from 1c has been isolated and char-
acterised: see reference [18a] for details.
Received: January 3, 2015
[20] a) G. W. Gribble, Chem. Soc. Rev. 1998, 27, 395–404; b) Science of Syn-
thesis: Houben–Weyl Methods of Molecular Transformation, Vol. 31b (Ed.:
Published online on && &&, 0000
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
10
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
An easily accessible iridicycle catalyst
effects the transfer hydrogenation of
a wide variety of N-heterocycles in
water, including quinolines, isoquino-
lines, indoles, quinoxalines, and pyri-
dines. The catalyst shows excellent func-
tional-group compatibility and high
turnover number (up to 7500), even
with low catalyst loadings.
&
Homogeneous Catalysis
D. Talwar, H. Y. Li, E. Durham, J. Xiao*
&& – &&
A Simple Iridicycle Catalyst for
Efficient Transfer Hydrogenation of
N-Heterocycles in Water
Transfer hydrogenation
Transfer hydrogenation is a simple safe way for reduction
reactions and that using water as solvent has the additional
benefit of being “green”. In their Full paper on page && ff.,
J. Xiao et al. report that a cyclometalated iridium complex
catalyzes the transfer hydrogenation of various nitrogen
heterocycles in an aqueous solution of formate under mild
conditions. The catalyst shows excellent functional-group
compatibility and high turnover number, with loadings as
low as 0.01 mol% being feasible. Mechanistic study of
quinoline reduction suggests that the reaction proceeds via
both 1,2- and 1,4-addition pathways, with the catalytic
turnover limited by hydride transfer.
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
11
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ