This is the peer reviewed version of the following article: Chemistry - A European Journal (2016), 22(30), 1060710613 which has been published in final form at
http://onlinelibrary.wiley.com/doi/10.1002/chem.201601464/abstract
This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for SelfArchiving.

Asymmetric Hydrogenation of Seven-membered CN-containing
Heterocycles and Rationalization of the Enantioselectivity
Bugga Balakrishna,[a],[d] Antonio Bauzá,[b] Antonio Frontera,*[b] Anton Vidal-Ferran*[a],[c],[d]

Abstract: Iridium(I) complexes of phosphine-phosphite ligands
efficiently catalyze the enantioselective hydrogenation of diverse
seven-membered CN-containing heterocyclic compounds (eleven
examples; up to 97% ee). POP ligand L3, which incorporates an
ortho-diphenyl substituted octahydrobinol phosphite fragment,
provided the highest enantioselectivities in the hydrogenation of
most of the heterocyclic compounds studied. The observed sense of
stereoselection was rationalized by means of DFT calculations.

Introduction
Many biologically active compounds contain a chiral
heterocyclic structural motif.[1] The development of methods to
access partly or fully reduced enantiopure heterocycles has
increased in scope and importance in recent years. The
enantioselective reduction of heterocyclic compounds is
becoming more important in the preparation of their partly or
fully reduced analogues, as this strategy benefits from a great
diversity of starting materials and minimizes the need for
manipulation of functional groups during the preparation of
target compounds.[ 2 ] Transition metal-catalyzed asymmetric
hydrogenation has been employed to reduce a large number of
nitrogenated heterocyclic compounds (for instance pyridines and
other
monocyclic
nitrogenated
derivatives,
quinolines,
isoquinolines,
quinoxalines
and
related
compounds,
benzoxazines and related derivatives, indoles, etc.) with high

[a]

[b]

[c]

[d]

B. Balakrishna, Prof. Dr. A. Vidal-Ferran
Institute of Chemical Research of Catalonia (ICIQ)
Avinguda Països Catalans 16,
E-43007, Tarragona, Spain.
E-mail: avidal@iciq.cat
Antonio Bauzá, Prof. Dr. Antonio Frontera
Departament de Química, Universitat de les Illes Balears (UIB)
Cra. de Valldemossa, km 7.5. Palma
07122 Palma de Mallorca, Spain.
Prof. Dr. A. Vidal-Ferran
Catalan Institution for Research and Advanced Studies (ICREA)
Passeig Lluís Companys 23,
E-08010 Barcelona, Spain.
B. Balakrishna, Prof. Dr. A. Vidal-Ferran
The Barcelona Institute of Science and Technology,
Avinguda Països Catalans 16,
E-43007, Tarragona, Spain.

catalytic efficiencies and enantioselectivities. Although
nitrogenated seven-membered heterocyclic motifs with
stereogenic centers constitute an important pharmacophore, [ 3 ]
examples of the asymmetric reduction of seven-membered
heterocyclic derivatives are scarce in the literature.[4]
We recently reported the hydrogenation of an array of
structurally diverse five- and six-membered heterocyclic
compounds 1 mediated by iridium(I) complexes of phosphinephosphite ligands[5] [Ir(Cl)(cod)(POP)] with high catalytic activity
(Scheme 1).[6] Ligand L1, which incorporates an (Ra)-configured
[1,1'-binaphthalene]-2,2'-diol phosphite group, provided the
highest enantioselectivities in the asymmetric hydrogenation of
heterocyclic compounds 1. Interestingly, we described that the
addition of a Brønsted acid (cat. amounts of HCl for 2-alkyl
substituted quinolines[6a] and quinoxalines[6a] and stoichiometric
amounts of rac-camphorsulfonic acid for indoles [6c]) increased
the conversion of the hydrogenation reaction and even positively
affected the enantioselectivity.[6a] Herein we report on the
development of efficient Ir-(POP) complexes as catalysts for
the enantioselective asymmetric hydrogenation of diversely
substituted seven-membered CN-containing heterocyclic
compounds (see structures 3, 5, 7, 9 and 11 in Scheme 1).

Table 1). The hydrogenation of 3a was also run at lower temperature (0 ºC instead of rt), based on the premise that lowering
Table 1. Asymmetric hydrogenation

[a]

of 3a mediated by [Ir(Cl)(cod)(L1L4)].

[c]

Entry

Ligand

Observations on the
reaction conditions

1

L1

No additive

99

42 (S)

2

L1

10 mol% HCl

99

53 (S)

3

L2

No additive

99

29 (S)

4

L2

10 mol% HCl

99

57 (S)

5

L3

No additive

99

79 (R)

6

L3

10 mol% HCl

99

86 (R)

7

L4

No additive

99

16 (S)

8

L4

10 mol% HCl

99[f]

8 (S)

[b]

Conv

ee (%)
(config.)[d]

9

L3

20 mol% HCl

99

86 (R)

10

L3

10 mol% HCl; 0 ºC

90

87 (R)

11

L3

10 mol% HCl, DCM

99

91 (R)

12

L3

10 mol% HCl, toluene

99

91 (R)

99

91 (R)

13

Scheme 1. Enantioselective partial hydrogenation of nitrogenated
heterocyclic compounds.

Results and Discussion
Ir-catalyzed asymmetric hydrogenation. Initial screening
and optimization
The present work began with the enantioselective hydrogenation
of oxazepine 3a as the model substrate using well-established[6a]
in situ prepared iridium complexes of POP ligands L1L4 as
pre-catalysts. The reaction conditions employed (1 mol% cat.;
[3a] = 0.2 M in THF, 80 bar H2, rt) efficiently led to the
hydrogenated product 4a with complete conversion in the
absence of HCl as additive (see entries 1, 3, 5 and 7 in Table 1)
with enantioselectivities ranging from 16% ee for L4 to 79% ee
for L3.[7] Enantioselectivities were improved by using 10 mol%
HCl in the case of ligands L1, L2 and L3 (compare entries 1, 3
and 5 with entries 2, 4 and 6 in Table 1), with ligand L3 providing
the highest enantioselectivities (86% ee, entry 6 in Table 1). The
effects of a set of achiral and enantiomerically pure additives on
the hydrogenation of substrate 3a were also studied.[8] With the
exception of HCl, other acid additives scarcely affected the
conversion and enantioselectivity of the hydrogenation of this
substrate (see Table SI2).[8] Further attempts to optimize the
hydrogenation conditions involved varying the amount of HCl
(see entry 9 in Table 1) and reducing the temperature (see entry
10 in Table 1). Increasing the amount of additive did not have
any effect on the conversion and ee (compare entries 6 and 9 in

L3

[e]

10 mol% HCl, MeTHF

[a] Reaction conditions: [{Ir(Cl)(cod)}2]/POP ligand/substrate = 0.5:1.1:100
for pre-catalyst levels of 1 mol%, respectively, at rt, 20 h and a substrate
concentration of 0.20 M in THF. If additive was present, the indicated amount
of additive with respect to 3a was added to a solution of the substrate before
adding the catalyst. The values shown are the average of at least two runs. [b]
Conversions were determined by 1H NMR. [c] Determined by HPLC analysis
using chiral stationary phases. [d] Absolute configuration was assigned by
comparison with literature data (see ref. [4c]). [e] MeTHF = 2methyltetrahydrofuran. [f] The selectivity of the reaction towards 4a was 35%.

the temperature normally offers higher ee. As listed in entry 10
of Table 1, the ee at 0 ºC was 1% higher than at rt. However,
further hydrogenation studies at this temperature were not
considered, as conversion was incomplete. The enantioselectivity of the hydrogenation of 3a was solvent dependent,
and 2-methyltetrahydrofuran (MeTHF), dichloromethane (DCM)
and toluene led to a noticeable increase in the ee of the reaction
(compare entries 11, 12 and 13 with entry 6 in Table 1).
Expanding the substrate scope
Once the optimal hydrogenation conditions for 3a had been
established, the hydrogenation of an array of seven-membered
heterocycles was studied. 2-Methyltetrahydrofuran was chosen
as the optimal solvent for further studies, given the good results
in the hydrogenation of 3a. The heterocycles studied and the
optimal ligand (L1 or L3) for achieving the highest ee’s are
summarized in Table 2. The results using 10 mol% HCl are only
indicated in Table 2 when this additive provided higher
conversions and/or ee’s. The reader is referred to the
Supplementary Information (Table SI3) for complete
hydrogenation results employing L1 and L3 as ligands and in
the absence or presence of anhydrous HCl as additive.
Table 2. Asymmetric hydrogenation[a] of seven-membered N-Heterocyclic
compounds mediated by [Ir(Cl)(cod)(L1 or L3)].

Entry

Substrate
(X, R)
[e]

Ligand

Additive

[b]

Conv

ee (%)[c]
[d]
((config.)

1

3a (O, Me)

L3

10 mol% HCl

99

2

3b (O, Ph)

L1

none

97

83 (S)

3

3c (O, iPr)

L3

none

99

73 (R)

4

3d (O, Bn)

L3

10 mol% HCl

99[f]

87 (R)

5

5a (CH2, Me)

L3

none

99

substituents at the 3 and 3’ positions of the [1,1'-biaryl]-2,2'-diol
group did not change the configuration of the final product in Rhmediated hydrogenations.[10] Interestingly, ligands L1 (or L2) and
L3, whose main difference is the presence or absence of
substituents at the 3 and 3’ positions of the binaphthyl motif, led
to opposite enantiomers of 4, 6 and 8 depending on the nature
of the R substituent (alkyl or aryl groups; as an example, see
Scheme 2 for the hydrogenation reactions leading to 4).

70 (R)

[g]

91 (R)

6

5b (CH2, Ph)

L3

10 mol% HCl

7

7a (NMe, Me)

L3

10 mol% HCl

99

8

7b (NMe, Ph)

L1

none

24[g]

36 (S)

9

9a (S, Me)

L3

10 mol% HCl

99

91 (R)

10

9b (S, Ph)

L3

none

72[f]

97 (S)

[f]

77 (R)

11

11a (SO2, Me)

L3

10 mol% HCl

33

99

9 (S)
84 (R)

[a] Reaction conditions: [{Ir(Cl)(cod)}2]/POP ligand/substrate = 0.5:1.1:100
for pre-catalyst levels of 1 mol%, respectively, at rt, 20 h and a substrate
concentration of 0.20 M in MeTHF. If additive was present, the indicated
amount of additive with respect to substrate was added to a solution of the
substrate before adding the catalyst. The values shown are the average of at
least two runs. [b] Conversions were determined by 1H NMR. Isolated yields
after chromatography were > 95%, unless otherwise stated. [c] Determined by
HPLC analysis using chiral stationary phases. [d] Absolute configurations of
4a, 4b, 4d, 10a and 10b were assigned by comparison with literature data
(see ref. [4c] for the oxazepines and ref. [4h] for the thiazepines). The
configurations of 4c, 6a, 6b, 8a and 8b were assumed by analogy. The
absolute configuration of 12a was determined by X-Ray analysis (see SI for
details). [e] These results have been already summarized in Table 1, but are
included here for comparison. [f] Isolated yields for 4d, 10b and 12a were,
32%, 53% and 59%, respectively. [g] Isolation was not attempted due to low
conversions and ee’s.

Several trends can be extracted from the results listed in Table
2. Firstly, L3 was the ligand of choice for heterocycles with R =
alkyl group (entries 1, 35, 7, 9 and 11 in Table 2). For these
substrates, ee’s ranged from 70 to 91%, with the highest ee’s
being obtained for oxazepine 4a and thiazepine 9a (91% ee; see
entries 1 and 9 in Table 2). Secondly, the use of HCl as additive
in the hydrogenation of substrates with R being an alkyl group
had, with the exception of substrates 3c and 5a (see entries 3
and 5 in Table 2), a positive effect on the ee’s.[9] Examples of the
use of Brønsted acids as substrate activators in iridiummediated hydrogenations are numerous [2j] however, despite the
improvement in catalyst activity and/or selectivity induced by
these additives, their role remains in many cases unclear.[2j] The
third and last trend that can be seen in the data summarized in
Table 2, is that the hydrogenation of the substrates with R = Ph
(compounds 3b, 5b, 7b and 9b) was more complicated than that
of their alkyl substituted analogues. Though the hydrogenation
of the O- and S-containing substrates took place efficiently in the
absence of HCl (conversions up to 97%, ee’s up to 97%, entries
2 and 10 in Table 2), the hydrogenation of carbon and nitrogenanalogues 5b and 7b proceeded with low conversions and ee’s
(entries 6 and 8 in Table 2).
Previous work from our group on asymmetric hydrogenations
mediated by Ir-[6] or Rh- complexes[10] of POP ligands revealed
that the phosphite group was the principal stereochemical
director in the reaction (opposite configurations for the resulting
hydrogenated products are obtained when the configuration of
the phosphite moiety is inverted). Moreover, the introduction of

Scheme 2. Hydrogenation reactions of 3 leading to products 4.

Rationalization of the stereochemical outcome of the
hydrogenations by DFT calculations
To shed light on the correlation between the features of the
POP ligands and enantioselection in the hydrogenation
towards alkyl-substituted products, we performed a theoretical
investigation into the reactivity of the catalytic systems derived
from ligands L1 and L3 with substrate 3a. We considered this
substrate to be suited to our purposes, as the configuration of
the final product depends on whether the substituents at the 3
and 3’ positions of the phosphite group are H (L1; S-configured
product, see entries 1 and 2 in Table 1 and Scheme 2) or Ph
groups (L3; R-configured product; see entries 5 and 6 in Table 1
and Scheme 2) and the absolute stereochemistry of its
hydrogenated product was unequivocally assigned. Theoretical
studies of enantioselective processes usually focus on the
stereo-determining step and compare the energy of transition
states (TS’s) for the paths leading to the R and S products.[11]
The mechanism for the hydrogenation of iminic bonds is
complex but has been explored at the experimental and
theoretical level by a number of groups. [12] Mechanistic studies
from Pfaltz and co-workers have revealed that the employed
iridium complexes react with acyclic CN-containing substrates
to form stable cyclometallated iridium complexes that are the
real hydrogenation catalysts.[ 13 ] Besides, a number of
mechanistic studies from other research groups have revealed
that the hydrogenation process takes place by transfer of proton
and hydride to the CN bond of the heterocyclic derivative being
non-coordinated to the metal center.[14]
We first explored the possibility of the hydrogenation pathway
involving the formation of cyclometallated iridium complexes
derived from 3a. The stabilities of the plausible four-membered
iridacycles derived from 3a were computed at the BP86/def2SVP level of theory (see Figure SI68), which is a good
compromise between the size of the system (up to 139 atoms
for the iridacycle involving L3) and the accuracy of the results
(see the SI for a comparison of this functional with MP2 and
M06-2X methods). The energy content of the resulting fourmembered iridacycles was very high indicating the highly

strained nature of these compounds.[ 15 ] Therefore this
hydrogenation pathway via the formation of such intermediates
derived from 3a was not further explored.[ 16 ] As regards the
hydrogenation pathway involving proton and hydride transfers to
the CN bond of heterocyclic derivatives, Crabtree and
Eisenstein[14b] identified octahedral dihydrido mono-dihydrogen
iridium complexes as crucial intermediates in the hydrogenation
process. This is because hydrogen transfer from H2 to the CN
bond starts with proton migration from the dihydrogen ligand to
the nitrogen atom. Once the CN bond is protonated,[ 17 ] the
position alpha to nitrogen is activated for the subsequent hydride
transfer, which is the stereo-determining step (see Scheme 2).
Our own studies[6a] and those of others[5b] have demonstrated
that the complexation of POP ligands with [{Ir(-Cl)(cod)}2]
quantitatively leads to compounds [Ir(Cl)(cod)(POP)], which
correspond to the expected neutral pentacoordinated iridium(I)
complexes. Removal of the cod ligand under hydrogenative
conditions led to a complex mixture of POP-iridium complexes.
Unfortunately, neither NMR nor X-Ray analysis allowed us to
unequivocally establish the structure of the iridium complexes
present in solution (no crystals suitable for X-Ray analysis could
be isolated from this mixture). For this reason, the relative
stabilities of the plausible [Ir(Cl)(H)2(HH)(L1 or L3)] complexes
were computed at the BP86/def2-SVP level of theory. From
among all the possible isomers in an octahedral iridium complex
with one bidentate (L1 or L3) and one chlorido ligand, only those
with the two hydrido and dihydrogen ligands in a fac[18] (facial)
geometry were considered.[19] This is because hydrogen transfer
from [Ir(Cl)(H)2(HH)(L1 or L3)] complexes will lead to a
trihydrido iridium complex and metal trihydrides have an intrinsic
preference[14b] for a fac geometry to avoid hydrido ligands that
are mutually placed in a trans fashion. Interestingly, in
[Ir(Cl)(H)2(HH)(L1)] the favorable fac isomers (see Figure 1)
contain chlorido ligands pointing in the same direction and
perpendicular to the plane that contains the POP and Ir atoms
(see Figure 1a,b). The slightly more favored complex (difference
0.3 kcal·mol–1) has the HH ligand cis to the phosphite group.
With regard to [Ir(Cl)(H)2(HH)(L3)], the lowest energy isomers
also contain chlorido ligands perpendicular to the same plane
but pointing in the opposite direction with respect to the
[Ir(Cl)(H)2(HH)(L1)] complexes (see Figure 1c,d). This is due to
the formation of intramolecular CH···Cl bonds (see Figure 1
and SI70). This differentiating feature is very important for
rationalizing the opposite enantioselectivity observed for L1 and
L3 with methyl substituted substrates such as 3a (vide infra).

Figure 1. (a-d) Optimized geometries of the most stable isomers of
[Ir(Cl)(H)2(HH)(L1 or L3)] (some H atoms omitted for clarity; distances in Å;
energies in kcal·mol–1).

Protonation of 3a by [Ir(Cl)(H)2(HH)(L1 or L3)][17] leads to the
[H3a][Ir(Cl)(H)3(L1 or L3)] assembly (see Figure 2a,b). Both
isomers of complex [Ir(Cl)(H)2(HH)(L1)] (see Figure 1a,b) yield
the same fac trihydrido iridium complex upon proton transfer
(same behavior for L3), which simplifies the study. Proton
transfer is not the stereo-determining step and the configuration
of the final product is determined at the later stages of the
catalytic cycle. Therefore, we employed calculations for
understanding the stereochemical outcome of the reaction from
the protonated substrate (H3a). Beginning from the initial
geometry after proton transfer, where the NH group points to
the IrH motif (see Figure 2a,b), we examined different
orientations of the protonated substrate (H3a) interacting with
[Ir(Cl)(H)3(L1 or L3)]. Remarkably, we found a pre-TS complex
for each ligand (see Figure 2c,d) that was lower in energy than
the initial assembly due to the formation of favorable noncovalent interactions. In the case of L1, the preferred
arrangement is governed by two interactions that fix the
geometry of the substrate (H-bonds and CH3··· interactions,
see Figure 2c). This pre-organized complex facilitates the
nucleophilic attack of the hydrido group that is located 3.0 Å
away from the C atom in the CN group (pro-(S) attack). In the
case of L3, the presence of the chlorido ligand at the position
opposite the POP containing plane with respect to L1 and the
formation of a strong NH···Cl interaction fixes the substrate in a

Figure 2. (a,b) Optimized geometries of [H3a][Ir(Cl)(H)3(L1 or L3)] assemblies.
(c,d) Distances in Å of pre-TS complexes found for [H3a][Ir(Cl)(H)3(L1 or L3)]
(some H atoms omitted for clarity; distances in Å).

different arrangement compared to L1. Moreover the formation
of a CH···HIr non-covalent interaction[20] (see Figure 2d) fixes
the position of the substrate, facilitating the pro-(R) attack of the
hydrido ligand (located at 3.1 Å). The pre-TS complexes that
would organize the protonated substrate towards the minor
enantiomer (i.e. pro-(R) attack for L1 and pro-(S) attack for L3)
were not found in the potential hypersurface. The geometries of
the TS’s are shown in Figure 3. It is important to note the
existence of strong NH···Cl hydrogen-bonds[21] in the favored
TS’s. These interactions are crucial in rationalizing the observed
sense of stereoselection. The difference in energy between the
two TS’s states derived from L1 (G# = 2.2 kcal mol-1) is mainly
governed by the different strength of two hydrogen-bond
interactions: an NH···Cl hydrogen-bond for the TS leading to
the major enantiomer (TSS; see Figure 3a) and an NH···O
hydrogen-bond[22] for the TS leading to the minor enantiomer
(TSR; see Figure 3c). Since the NH···Cl interaction involves an
anionic ligand, it is electrostatically favored with respect to the
NH···O interaction. As regards L3, the transition state leading
to the major enantiomer is also stabilized by an NH···Cl
hydrogen-bond (TSR; see Figure 3b), whilst that leading to the
minor enantiomer is only stabilized by a weaker NH···
interaction involving a phenyl group (see Figure 3c). Since this
NH··· interaction involving TSS derived from L3 is also
weaker than the NH···O hydrogen bond in TSR derived from L1
(both leading to the minor enantiomers of the hydrogenation
product of 3a), the ΔΔG# value for L3 (4.3 kcal mol1) is higher
than that for L1 (2.2 kcal mol1). This observation is in
agreement with the higher enantioselectivity observed
experimentally in the hydrogenation of 3a with L3 than that with
L1 (compare entry 5 with entry 1 in Table 1, respectively).

Figure 3. Optimized geometries of the transition states of L1 (a) and L3 (b)
and their energetic profiles in kcal·mol–1 (some H atoms omitted for clarity). (c)
optimized structures of the high energy TS’s (distances in Å).

Conclusions
In conclusion, catalytic screening in enantioselective
hydrogenation reactions indicate that the iridium complexes of
chiral POP ligands L1 and L3 are excellent catalysts in the
hydrogenation of various seven-membered heterocycles that
contain CN bonds. The “lead” pre-catalyst for alkyl-substituted
seven-membered heterocycles (derived from ligand L3) in
combination with catalytic amounts of HCl exhibits excellent
catalytic properties in this transformation. The hydrogenation of
aryl-substituted seven-membered heterocycles was more
complicated and highly efficient hydrogenation conditions could
only be developed for phenyl substituted oxa- and thia-azepines
employing L1 without any additive. The enantioselectivity has
been rationalized by means of DFT calculations, which identified
the position of the Cl-ligand in catalytically relevant iridium
structures and a number of non-covalent interactions (i.e.
NH…Cl, CH… and CH…HIr interactions[23]) as key features in
rationalizing the stereochemical outcome of the reactions with

ligands L1 and L3. We are currently expanding the scope of the
hydrogenation reaction mediated by these ligands to new
heterocycles and will report on this work in due course.

Experimental Section
General procedure for the Ir-catalyzed asymmetric hydrogenation
A solution of the required amount of [{Ir(μCl)(cod)}2] (0.005 mmol)
and the POP ligand (0.011 mmol) in the corresponding dry and
deoxygenated solvent (5.0 mL) was loaded into an autoclave under N2, in
which the required amounts of substrate (1 mmol) and anhydrous HCl
(0.1 M solution in the required solvent), if necessary, were placed
beforehand. The concentration of the substrate was adjusted to a final
0.20 M concentration. The autoclave was purged three times with H2 (at a
pressure not higher than the one selected) and finally, the autoclave was
pressurized with H2 to the desired pressure. The reaction mixture was
stirred at the desired temperature for the stated reaction time. The
autoclave was subsequently depressurized, the reaction mixture passed
through a short pad of SiO2 and further eluted with EtOAc (2 x 1 mL). The
resulting solution was evaporated in vacuo. The conversion was
determined by 1H NMR and enantioselectivities were determined by
HPLC analysis on chiral stationary phases.

[5]

[6]

[7]

[8]
[9]

Acknowledgements
The authors would like to thank MINECO (CTQ2014-60256-P,
CTQ2014-57393-C2-1-P and Severo Ochoa Excellence
Accreditation 2014-2018 SEV-2013-0319) and the ICIQ
Foundation for financial support. B. B. thanks the AGAUR for a
pre-doctoral fellowship (2013FI-B 00545). Dr. José Luis NúñezRico and Dr. Héctor Fernández-Pérez are gratefully
acknowledged for proof-reading the manuscript and Dr. J.
Benet-Buchholz for the X-Ray crystallographic data.

[10]

[11]
[12]

Keywords:
Enantiopure
heterocycles
•
Asymmetric
hydrogenation • Substrate activation • Iridium • Phosphinephosphites
[13]

References:
[14]
[1] See, for example: a) J. W. Daly, T. F. Spande, H. M. Garraffo, J. Nat.
Prod. 2005, 68, 1556-1575; b) J. P. Michael, Nat. Prod. Rep. 2005, 22,
603-626; c) Pharmaceutical Substances, 5th ed.; (Eds.: A. Kleemann, J.
Engel, B Kutscher, D. Reichert) Georg Thieme Verlag, Stuttgart, 2008.
[2] For general reviews on the enantioselective reduction of C=N-containing
heterocyclic compounds, see for example: a) F. Glorius, Org. Biomol.
Chem. 2005, 3, 4171-4175; b) Y.-G. Zhou, Acc. Chem. Res. 2007, 40,
1357-1366; c) R. Kuwano, Heterocycles 2008, 76, 909-922; d) M.
Rueping, J. Dufour, F. R. Schoepke, Green Chem. 2011, 13, 1084-1105;
e) J. G. de Vries, N. Mršic, Catal. Sci. Technol. 2011, 1, 51-59; f) D.-S.
Wang, Q.-A. Chen, S.-M. Lu, Y.-G. Zhou, Chem. Rev. 2012, 112, 25572590; g) Z. Yu, W. Jin, Q. Jiang, Angew. Chem., Int. Ed. 2012, 51, 60606072; h) Q.-A. Chen, Z.-S. Ye, Y. Duan, Y.-G. Zhou, Chem. Soc. Rev.
2013, 42, 497-511; i) T. Nagano, A. Iimuro, K. Yamaji, Y. Kita, K.
Mashima, Heterocycles 2014, 88, 103-127; j) B. Balakrishna, J. L. NúñezRico, A. Vidal-Ferran, Eur. J. Org. Chem. 2015, 2015, 5293-5303.
[3] For general reviews, see for example: a) J. A. Smith, P. P. Molesworth, J.
H. Ryan in Seven-Membered Rings, Vol. 21 (Eds.: G. W. Gribble and J.
A. Joule), Elsevier Ltd., Oxford, Amsterdam, 2009, pp. 491-530; b) J. H.
Ryan, J. A. Smith, C. Hyland, A. G. Meyer, C. C. Williams, A. C.
Bissember, J. Just in Seven-Membered Rings, Vol. 26 (Eds.: G. W.
Gribble and J. A. Joule), Elsevier, Oxford, Amsterdam, 2014, pp. 521-571.
[4] a) Z. Y. Han, H. Xiao, L. Z. Gong, Bioorg. Med. Chem. Lett. 2009, 19,
3729-3732; b) M. Rueping, E. Merino, R. M. Koenigs, Adv. Synth. Catal.

[15]

[16]

[17]

2010, 352, 2629-2634; c) K. Gao, C.-B. Yu, W. Li, Y.-G. Zhou, X.-M.
Zhang, Chem. Commun. 2011, 47, 7845-7847; d) X. Chen, Y. Zheng, C.
Shu, W. Yuan, B. Liu, X. Zhang, J. Org. Chem. 2011, 76, 9109-9115; e)
Z.-Y. Ding, F. Chen, J. Qin, Y.-M. He, Q.-H. Fan, Angew. Chem., Int. Ed.
2012, 51, 5706-5710; f) K. Gao, B. Wu, C.-B. Yu, Q.-A. Chen, Z.-S. Ye,
Y.-G. Zhou, Org. Lett. 2012, 14, 3890-3893; g) B. Ma, Z. Ding, J. Liu, Y.
He, Q.-H. Fan, Chem. Asian J. 2013, 8, 1101-1104; h) J. Wang,
Tetrahedron Lett. 2013, 54, 5956-5959; i) R.-N. Guo, K. Gao, Z.-S. Ye, L.
Shi, Y. Li, Y.-G. Zhou, Pure Appl. Chem. 2013, 85, 843-849; j) D. J.
Cowan, J. L. Collins, M. B. Mitchell, J. A. Ray, P. W. Sutton, A. A.
Sarjeant, E. E. Boros, J. Org. Chem. 2013, 78, 12726-12734; k) C. Guo,
D.-W. Sun, S. Yang, S.-J. Mao, X.-H. Xu, S.-F. Zhu, Q.-L. Zhou, J. Am.
Chem. Soc. 2015, 137, 90-93; l) W. Li, C. Schlepphorst, C. Daniliuc, F.
Glorius, Angew. Chem. Int. Ed. 2016, 55, 3300-3303.
Phosphine-phosphites ligands are also referred to as POP ligands (see
for example: a) T. Robert, J. Velder, H.-G. Schmalz, Angew. Chem. Int.
Ed. 2008, 47, 7718-7721; b) A. Suárez, M. A. Méndez-Rojas, A. Pizzano,
Organometallics 2002, 21, 4611-4621 and reference 6).
a) J. L. Núñez-Rico, H. Fernández-Pérez, J. Benet-Buchholz, A. VidalFerran, Organometallics 2010, 29, 6627-6631; b) J. L. Núñez-Rico, A.
Vidal-Ferran, Org. Lett. 2013, 15, 2066-2069; c) J. L. Núñez-Rico, H.
Fernández-Pérez, A. Vidal-Ferran, Green Chem. 2014, 16, 1153-1157.
The ligands with opposite configuration at the phosphite fragment were
also studied in this transformation and led in all cases to lower
enantioselectivities than those obtained with the corresponding
diastereomeric ligand L1L4. See the SI for details (Table SI1).
For a complete summary on the effects of acid additives, see Table SI2 in
the Supporting Information.
See Table SI3 from the Supporting Information for the hydrogenation
conditions assayed.
a) H. Fernández-Pérez, S. M. A. Donald, I. J. Munslow, J. BenetBuchholz, F. Maseras, A. Vidal-Ferran, Chem. Eur. J. 2010, 16, 64956508; b) H. Fernández-Pérez, J. Benet-Buchholz, A. Vidal-Ferran, Chem.
Eur. J. 2014, 20, 15375-15384; c) J. R. Lao, J. Benet-Buchholz, A. VidalFerran, Organometallics 2014, 33, 2960-2963.
a) D. Balcells, F. Maseras, New J. Chem. 2007, 31, 333-343; b) J. M.
Brown, R. J. Deeth, Angew. Chem. Int. Ed. 2009, 48, 4476-4479.
For recent general reviews on the hydrogenation of imines, see for
example: a) N. Fleury-Brégeot, V. de la Fuente, S. Castillón, C. Claver,
ChemCatChem 2010, 2, 1346-1371; b) K. H. Hopmann, A. Bayer, Coord.
Chem. Rev. 2014, 268, 59-82. For representative examples on the
hydrogenation of cyclic imines, see for example: c) T. Morimoto, N.
Suzuki, K. Achiwa, Tetrahedron: Asymmetry 1998, 9, 183-187; d) T.
Yamagata, H. Tadaoka, M. Nagata, T. Hirao, Y. Kataoka, V.
Ratovelomanana-Vidal, J. P. Genet, K. Mashima, Organometallics 2006,
25, 2505-2513; e) C. Li, J. Xiao, J. Am. Chem. Soc. 2008, 130, 1320813209; f) J.-H. Xie, P.-C. Yan, Q.-Q. Zhang, K.-X. Yuan, Q.-L. Zhou, ACS
Catal. 2012, 2, 561-564.
Y. Schramm, F. Barrios-Landeros, A. Pfaltz, Chem. Sci. 2013, 4, 27602766.
For seminal work on the concept of stepwise proton and hydride transfer
to iminic CN bonds see: a) M. Martin, E. Sola, S. Tejero, J. L. Andres, L.
A. Oro, Chem. Eur. J. 2006, 12, 4043-4056. For experimental and
theoretical mechanistic studies on the iridium-mediated hydrogenation of
cyclic iminic bonds by an outer-sphere coordination pathway, see: b) G.
E. Dobereiner, A. Nova, N. D. Schley, N. Hazari, S. J. Miller, O.
Eisenstein, R. H. Crabtree, J. Am. Chem. Soc. 2011, 133, 7547-7562; c)
T. Nagano, A. Iimuro, R. Schwenk, T. Ohshima, Y. Kita, A. Togni, K.
Mashima, Chem. Eur. J. 2012, 18, 11578-11592. For experimental and
theoretical mechanistic studies on the ruthenium-mediated hydrogenation
of cyclic iminic bonds by an outer-sphere coordination pathway, see: d) 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.
For the calculated energy contents of the four-membered iridacyles
derived from 3a and their structures, see the Supporting Information
(Figure SI68).
The Pfaltz mechanism (reference 13) cannot be excluded for substrates
where cyclometallation can be expected to be favored, such as for the
arylsubstituted substrates studied in this work.
The use of catalytic amounts of HCl as additive translates to partial
protonation of the substrate (HCl is added to the substrate before the
catalyst). However, it should be recalled at this point that dihydrogen
ligands in iridium complexes might certainly play a role in the protonation
of the substrate: most of the heterocycles studied are fully hydrogenated

[18]

[19]

[20]
[21]

[22]
[23]

with high enantioselectivities employing substoichiometric amounts of HCl,
or even in the absence of HCl (see Tables 1, SI1, SI2 and SI3).
Nomenclature of Inorganic Chemistry, IUPAC recommendations 2005
(Eds.: N. G. Connelly, T. Damhus, R. M. Hartshorn, A. T. Hutton), RSC
Publishing, Northampton, 2005.
In fact, for each ligand we have optimized six possible isomeric iridium
complexes (including mer complexes, see Figures SI69 and SI70) and the
most favorable ones are fac isomers.
See, for example: R. Custelcean, J. E. Jackson, Chem. Rev. 2001, 101,
1963-1980.
For the strength of NH···Cl interactions, see for example: a) D. K.
Kumar, A. Ballabh, D. A. Jose, P. Dastidar, A. Das, Cryst. Growth Des.
2005, 5, 651-660; b) F. F. Said, T.-G. Ong, G. P. A. Yap, D. Richeson,
Cryst. Growth Des. 2005, 5, 1881-1888.
For the strength of NH···O interactions, see for example: S. J.
Grabowski, Chem. Rev. 2011, 111, 2597-2625.
See, for example: Y. Zhao, S. Benz, N. Sakai, S. Matile, Chem. Sci. 2015,
6, 6219-6223.

Entry for the Table of Contents

FULL PAPER
Bugga Balakrishna, Antonio Bauzá,
Antonio Frontera,* Anton Vidal-Ferran*
Page No. – Page No.

Cl-Switch: Efficient enantioselective hydrogenation of seven-membered Nheterocycles mediated by Ir-(POP) complexes is described (11 examples, up to
97% ee). The position of the Cl ligand in catalytically relevant Ir-species (amongst
other factors) is key for rationalizing the stereochemical outcome.

Asymmetric Hydrogenation of Sevenmembered CN-containing
Heterocycles and Rationalization of
the Enantioselectivity