A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium salts

Article abstract of DOI:10.1016/j.tet.2018.03.031

NMR studies of the catalyst, deuteration experiments, mass spectrometry, and isolation and characterization of intermediates, allow us to propose an outer-sphere mechanism for the Iridium-catalyzed asymmetric hydrogenation of N-alkyl-2-arylpyridinium salts.

Full text of DOI:10.1016/j.tet.2018.03.031

Tetrahedron xxx (2018) 1e9
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
A mechanistic investigation of an Iridium-catalyzed asymmetric
hydrogenation of pyridinium salts
,
b
,
*
,
,
,
Yuhua Huang a
1, Shaodong Liu a 1, Yizhou Liu b **, Yonggang Chen b, Mark Weisel b,
R. Thomas Williamson b, Ian W. Davies b, Xumu Zhang a
,
***
a Department of Chemistry, Rutgers, The State University of New Jersey, 610 Taylor Road, Piscataway, NJ 08854, USA
b Department of Process Chemistry and Department of Discovery Chemistry, Merck & Co., Inc., P.O. Box 2000, Rahway, NJ 07065, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
NMR studies of the catalyst, deuteration experiments, mass spectrometry, and isolation and character-
ization of intermediates, allow us to propose an outer-sphere mechanism for the Iridium-catalyzed
asymmetric hydrogenation of N-alkyl-2-arylpyridinium salts.
Received 13 January 2018
Received in revised form
10 March 2018
Accepted 13 March 2018
Available online xxx
© 2018 Published by Elsevier Ltd.
Keywords:
Asymmetric hydrogenation
Iridium
Pyridinium salt
Chiral piperidine
Outer-sphere
1. Introduction
Over the past decades the mechanism of asymmetric hydroge-
nation of imine, alkene and N-heteroarenes has been investigated
The chiral piperidine core is ubiquitous in many natural prod-
ucts and is also a key pharmacophore in many medicinal agents.1
Given the significance of this structural class, synthesis of these
piperidines has attracted tremendous attention in academic and
industrial laboratories.2 From the perspective of atom economy,
asymmetric hydrogenation of readily available substituted pyri-
dines is the most direct and important approach to access chiral
piperidines.3
A number of successful examples of pyridine reduction using
either heterogeneous or homogeneous catalysts have been re-
ported, including a variety of examples utilizing enantioselective Ir-
catalyzed hydrogenation.4 In 2014 we reported a highly enantio-
selective Ir-catalyzed asymmetric hydrogenation of simple N-ben-
zylpyridinium salts using a unique phosphole MP2-SEGPHOS as
catalyst (Scheme 1).5
extensively.6 In contrast, the mechanistic studies of Ir-catalyzed
enantioselective hydrogenation of pyridine substrates have been
scarce. In 2013, Xiao and co-workers reported an isotopic labeling
experiment for the non-enantioselective, iodide anion promoted,
rhodium-catalyzed transfer hydrogenation of N-benzylated pyr-
idinium salts.7a In 2016, Lefort and co-worker reported the
rhodium-catalyzed asymmetric hydrogenation of N-benzyl 3-
substituted pyridinium salts.7b In this report, the authors eluci-
dated the reaction pathway with deuterium labeling experiments.
Soon afterwards Lefort and co-workers also reported the hydro-
genation of 2-substituted pyridinium salts by iridium complexed to
a mixture of ligands and identified an iminium intermediate prior
to the final enantioselective reduction step.7c In the meanwhile, we
also independently investigated the reaction mechanism of our
previously reported enantioselective reduction of 2-aryl-pyr-
idinium salts with the Ir-MP2-SEGPHOS catalyst through a series of
deuteration experiments. We then structurally characterized the
catalyst by NMR and identified reaction intermediates to gain
further mechanistic insight. Through these studies we propose a
probable multi-step hydrogenation cascade that involves multiple
tautomerizations followed by an enantioselective addition via an
outer-sphere mechanism.
* Corresponding author. Department of Process Chemistry and Department of
Discovery Chemistry, Merck & Coo., Inc., P.O. Box 2000, Rahway, NJ 07065, USA.
** Corresponding author.
*** Corresponding author.
E-mail address: yuhua.huang@merck.com (Y. Huang).
Yuhua Huang and Shaodong Liu contributed equally to this work.
1
https://doi.org/10.1016/j.tet.2018.03.031
0040-4020/© 2018 Published by Elsevier Ltd.
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

2
Y. Huang et al. / Tetrahedron xxx (2018) 1e9
number of hydride signals lack apparent coupling to 31P, as 31P-
decoupling has minimal effects on their 1H line-shapes. Therefore,
these hydrides unlikely reside on the catalytic complex of interest.
Region “c” contains a number of broad signals that are substantially
sharpened upon 31P-decoupling. The observation suggested these
hydrides have small couplings with more than one 31P atoms, but
the exact coupling constants could not be reliably extracted due to
linewidth. Region “d” contains two sharp triplet peaks, which arise
from hydrides that are coupled to two 31P atoms through constants
of 14e18 Hz. The hydrides in “d” bind Iridium perpendicularly to
the MP2-SEGPHOS ligand plane and thereby have small “cis” cou-
plings to both phosphorus atoms.
Scheme 1. Ir/MP2-SEGPHOS catalyzed hydrogenation of 2-phenyl-pyridinium salt.
2. Results and discussion
2.1. NMR studies of the catalyst
The NOESY experiment further suggests that hydrides in regions
“a” and “c” form a di-hydride complex, as shown in Fig. 4. As
mentioned earlier, region “a” hydrides are co-planar with MP2-
SEGPHOS with both a large and a small 31P coupling, so region “c”
hydrides, which bear two small 31P couplings, should bind
perpendicularly to MP2-SegPhos. This disposition also places the
two hydrides in proximity, which can readily lead to NOE signals, as
observed in Fig. 3d. The chemical shifts of regions “a” and “c”
around ꢀ12 ppm and ꢀ22pm, and the large “trans” and small “cis”
1H-31P couplings of 150 and 17 Hz, are consistent with values re-
ported for a different Ir-diphosphine-dihydride complex of a
similar coordination geometry.9 The “X1” position is occupied by a
counter-ion (Brꢀ or Clꢀ), whereas the “X2” position is occupied by a
solvent molecule. Note that “X1” is unlikely occupied by a solvent,
because a hydride on the anti-position of a solvent generally has a
chemical shift of ca. ꢀ30 ppm or even more upfield,10 whereas the
“c” hydrides display chemical shifts of ꢀ22 ppm. The presence of
several distinct di-hydride species, as indicated by multiple signals
in each region, may arise from different stereoisomers and/or from
binding of different counter-ions. The formation of oligomeric Ir-
complexes may also contribute to the observed spectral hetero-
geneity.8d,10c Note that Fig. 4 only shows only one possible stereo-
isomer. A different stereoisomer, which arises by swapping the
positions of the “a” hydride with “X1”, is also consistent with both
1H-31P couplings and NOESY data. This stereoisomer and the one
displayed in Fig. 4 are diastereomers, because MP2-SEGPHOS is
atropchiral. Hydrides in Region “b”, which are not part of the cat-
alytic complex of interest, in fact undergo exchange at the ms-s
time scale with hydrides “a”, as seen by exchange peaks of nega-
tive correlation in NOESY spectrum, which also shows indirect
NOESY correlations to hydrides “c”. The chemistry relevance of this
hydride pool is not clear to us. For the two hydrides in region “d”,
lack of NOESY cross peaks supports that they likely belong to two
different mono-hydride species. They unlikely reside on a single di-
hydride species with two “trans” disposed protons, due to unequal
peak integrals. We cannot rule out the possibility that they are from
two different di-hydride species both with two equivalent protons
of “trans” configuration, but 1H-31P HMBC shows two distinct 31P
signals coupled to each hydride (data not shown), which indicates
the corresponding complex lacks rigorous symmetry.
First, we studied the structure of the precatalyst by NMR.
[Ir(COD)Cl]2 and MP2-SEGPHOS were mixed at the 1:2 M ratio in
the solvent of acetone-d6. The proposed structure is shown in Fig. 1.
The NMR data, shown in Fig. 2, support a structure with pseudo
C2 symmetry. Specifically, the protons and carbons on equivalent
positions of MP2-SEGPHOS and the COD moiety exhibited identical
NMR chemical shifts (Fig. 2a). Interestingly, the 31P NMR spectrum,
revealed broad (~130 Hz linewidth) but distinct resonances
(9.3 ppm and 4.8 ppm) for the two phosphorus atoms on MP2-
SEGPHOS(Fig. 2b), revealing that this complex lacks rigorous
symmetry Therefore, we propose that a Clꢀ counter-ion binds Irþ to
make a charge-neutral complex and by doing so, it breaks C2
symmetry. The J coupling between the two 31P atoms was not
observed, presumably due to the small 2JPP cis coupling of less than
30 Hz in magnitude8 and broad lines of ~130 Hz. Proton-
phosphorus J correlations were observed for both MP2-SEGPHOS
and COD in the 1H-31P HMBC experiment (Fig. 2c), and intermo-
lecular NOEs were observed between MP2-SEGPHOSand COD (dash
circled in Fig. 2d), suggesting the formation of an Ir(COD) (MP2-
SEGPHOS) complex.
The activated catalyst was obtained by applying 600 psi H2
pressure to the precatalysts for 16 h. Due to apparent instability of
the complex, the potential stabilizing effects of various counter-
ions were also evaluated. It was found that including 2.5 fold
molar equivalents of NH4Br provided appreciable stabilization,
although significant decomposition still occurred after 20 h, as
indicated by decaying hydride.
NMR signal intensities. As shown in Fig. 3, the resulting acti-
vated catalyst lacks a single major constituent but instead contains
a complex mixture of different species and possibly stereoisomers.
Based on 1H chemical shift distribution and differential effects of
31P-decoupling on 1H resonances, we classified the hydride spec-
trum into four regions labeled as “a”, “b”, “c”, and “d” in Fig. 3a.
Region “a” contains four major peaks with a “doublet of doublet”
31P coupling pattern. The primary and secondary couplings are ca.
150 Hz and 17 Hz, which are consistent with “trans” and “cis”
couplings to the two different phosphorus atoms on MP2-SEGPHOS,
respectively. Therefore, hydrides in region “a” should be in the
coordination plane with MP2-SEGPHOS. Region “b” includes a
2.2. Deuteration experiments
Although the pyridine 2-phenyl position is known to undergo
asymmetric hydrogenation as described earlier, the reaction sym-
metry on other pyridine positions is unknown due to a lack of
chirality on those positions in the final product. However, one can
actually introduce chirality into these positions by reducing pyri-
dine with deuterium gas, and by determining the configuration of
these eCHD centers to gain mechanistic insight into the hydroge-
nation process. In the deuteration experiment, N-benzyl-2-
phenylpyridinium bromide 1 was exposed to 99.9% D2 and upon
reaction completion the crude was subjected to NMR analysis for
Fig. 1. Proposed structure of Ir(COD) (MP2-SEGPHOS) precatalyst. Protons of identical
chemical shifts are indicated by the same numbers.
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

Y. Huang et al. / Tetrahedron xxx (2018) 1e9
3
Fig. 2. Key NMR data supporting the proposed structure of Ir(COD)(MP2-SegPhos) precatalyst. (a) 1H spectrum of the precatalyst. Assignment is based on the atomic numbering
in Fig. 1. (b) 31P spectrum of the precatalyst. (c) 1H-13 P HMBC spectrum optimized for 8 Hz J coupling. The 1H and 31P spectra in (a) and (b) are also displayed as horizontal and
vertical traces, respectively. (d) NOESY spectrum of the precatalyst, with a mixing time of 300 ms. The intermolecular NOE signals between MP2-SegPhos and COD are enclosed in
dashed circles.
Fig. 3. Key NMR data of the activated catalyst. (a) 1H spectra showing the hydride protons. The four distinct hydride regions are aphetically indicated as “a-d”. The top and bottom
traces show the 1H and 1H{31P} spectra, respectively. (b) Blown-up view of region “a” in (a). (c) Blown-up view of region “d” in (a). (d) NOESY spectrum with 300 ms mixing time,
showing cross peaks in the hydride regions. Note that the exchange peaks (red) between regions “a” and “b” have an opposite sign from the real NOE peaks (blue).
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

4
Y. Huang et al. / Tetrahedron xxx (2018) 1e9
expected level of ~50% as observed at C3, C4, and C6 (Equation (1)).
(Fig. 5). These results suggested the presence of a proton source
other than D2. Acetone enolation can potentially act as a proton
donor but control experiments using deuterated solvents under H2
revealed no deuterium content in the product, ruling out the pos-
sibility that the solvent serves as the proton source. Therefore, the
residual H2O in solvents stood out as the most likely proton source
at C5 in the deuteration experiment.
To further confirm the involvement of water, we performed the
reaction in Equation (2) under strictly anhydrous conditions. We
introduced D2O as an additive to the otherwise dry solvent system
(Fig. 6). With two equivalents of D2O, deuterium atoms were
observed only at C3- and C5-position (40% and 34%, respectively).
This demonstrated that water can indeed be a proton source for
reductions at these two carbons. In fact, it is quite conceivable to
have solvent exchangeable protons at C3 and C5 positions of the
pyridinium starting material. For example, a probable reaction
pathway involves an initial 1, 4 addition of hydrogen, a solvent
exchange on the ammonium cation, and a subsequent tautomeri-
zation as shown in Scheme 2. This path would lead to a water
proton translocating onto the C5 position.
Fig. 4. Proposed structure of the dihydride complex. The blue arrow indicates a large
1H-13 P coupling, the red arrow indicates small 1H-13 P couplings, and the green arrows
indicates the presence of NOE. Note that only one possible stereoisomer is displayed.
The other stereoisomer, which exchanges the positions of the proton and X1 within the
horizontal plane, is also consistent with the observed 1H-31P couplings and NOESY
data.
We further evaluated the effect of adding different equivalents
of D2O in the reaction. Deuterium incorporation at both C3- and C5-
position increased with more water, but the deuterium content at
C3 remained higher than that at the C5-position (Fig. 6). This result
seemed to contradict the original observation that only C5 dis-
played a lower than expected deuteration level while C3 did not
seem to receive any water proton (Fig. 5). This “discrepancy” is
likely because the availability of water was quite different in these
two experiments. In the original deuteration experiment, water
content was presumably quite low in the commercial dry solvents
and could easily be depleted at an early stage of hydrogenation. The
fact that C5 but not C3 received water protons supports a mecha-
nism where C5 hydrogenation occurs more readily and C3 hydro-
genation occurs at a later stage in the reaction cascade. In this case,
the trace amounts of residual water in the system would have been
consumed during C5 hydrogenation and most water could end up
being deuterated after isotopic exchange. During the subsequent
quantification of deuterium incorporation at each carbon site. We
found that deuterium was equally incorporated onto both “axial”
and “equatorial” positions at all methylene groups of the product,
indicating symmetric hydrogenation, except for C5 whose stereo-
chemistry was not determined due to isochronous 1H NMR reso-
nances of germinal protons (See SI). Interestingly, NMR detected
only 6% deuterium content at the C5-position instead of an
a
a
Fig. 5. Deuteration experiment under D2 (600psi). Estimated value; see supporting
information for further details.
overlapped.
Fig. 6. Deuteration experiment using D2O as an additive. Estimated value; see sup-
b
b
Axial and equatorial proton chemical shifts are
porting information for further details. Axial and equatorial proton chemical shifts
are overlapped.
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

Y. Huang et al. / Tetrahedron xxx (2018) 1e9
5
outcome of the entire transformation. The results of this experi-
ment provide further support for our proposed mechanism.
2.3. Identification of intermediates
Naturally, we sought to characterize the presence of in-
termediates generated during the hydrogenation reaction, which
could provide direct proof to support the mechanism. Initially, we
carried out the reaction with 1 as model substrate over a period of
3 h and observed new peak corresponding to tetrahydropyridine
(Mþ4) by HPLC-MS. However, not surprisingly, this intermediate
decomposed upon attempted chromatographic separation and
could not be further characterized.
Scheme 2. Proton equilibrated through the intermediate species over the reaction
process.
1e3 tautomerization, exchange with water no longer affects the
deuteration on C-3, leading to the expected deuterium level of 50%
as observed.
Through metal and solvent screenings, we found that hydro-
genation with Rh metal in dichloromethane provided a significant
amount of tetrahydropyridine intermediate 4 (34%) (Scheme 5) but
it appeared that a Rh catalyst could not effectively catalyze the last
reduction transformation under present conditions. Subsequent
electrospray ionization mass spectroscopic analysis of the mixture
revealed a peak at m/zþ 250.1606. This result was consistent with
the presence of enamine intermediate 4. The NMR analysis of the
solution also confirmed the identity of enamine intermediate 4. The
solvent of this solution was removed in a glove box, and Ir(COD)Cl/
MP2-SEGPHOS DCE/acetone were added as described in previous
hydrogenation reports5 (Scheme 5). As expected, the reaction
proceeded smoothly to form the desired piperidine product with
95% ee and 92% isolated yield, which confirms that this enamine 4
was one of the relatively stable intermediates generated over the
hydrogenation process.
It is worth noting that when the reaction was carried out with
commercial anhydrous solvent acetone and DCE, a different inter-
mediate was detected, HRMS showed a peak at m/zþ 268.1694
(calculated m/z ¼ 250.1596), which implied an addition of water to
enamine 4 (Scheme 6). From 13C NMR, we identified a carbonyl
group resonating at 200 ppm lending support to structure 10. NMR
samples were also taken at various time points and proved the
existence of an intermediate ketone, starting material and product
(Fig. 7).
To further confirm that ketone 10 is a viable intermediate
generated in the hydrogenation process, we synthesized this spe-
cies by conveniently reacting amide 11 with phenylmagnesium
bromide. UPLC, HRMS, and NMR of the crude reaction mixture all
support that reaction product is the same as the ketone interme-
diate found in the hydrogenation reaction. After removing solvent
in a glove box, the crude material was subjected to the original
hydrogenation conditions and the corresponding product was
indeed detected by UPLC (Scheme 7). Therefore, we confirm that
ketone 10 exists as one of the intermediates in the hydrogenation of
1, especially when water is present.
A sequence of events as depicted in Scheme 3 seems to reconcile
the results from these two experiments. We propose that tauto-
merizations between the intermediate enaminium and iminium
are involved in this reaction. Enamine 2, generated via probable 1,
4-hydrogen addition is tautomerized to iminium 3. After the sec-
ond hydrogen addition, the corresponding enamine 4 can be
generated. The second tautomerization of 4 results in the formation
of iminium 5. Subsequent chiral reduction of 5 affords the final
chiral piperidine. Generally, the reduction rate of tetra-substituted
iminium 5 was slower compared to the tautomerization rate of
enamine 4. Therefore, the second tautomerization had longer time
to undergo a reversible process of protonation and deprotonation,
which reconciles the aforementioned higher deuterium content at
C3 with excess amounts of D2O as additive in the hydrogenation of
1 (Fig. 6).
To further explore the origin of selectivity in the asymmetric
hydrogenation, we investigated the hydrogenation of disubstituted
5-methyl-2-phenyl-pyridium salt. Upon exposure to the previously
optimized conditions, we obtained the corresponding hydroge-
nated products with 93% ee and 83% ee, respectively (dr ¼ 2.3). As
shown in Scheme 4, the stereochemistry observed in the asym-
metric hydrogenation of 5-methyl-2-phenyl-pyridium salt can be
rationalized. Here again, we propose 1, 4-addition as the first
reduction step in the hydrogenation process. The tautomerization
of the enamine 6 provides a mixture of enantiomers 7a and 7b,
whose ratio determines the low product diastereoselectivity. Note
that the high ee of 93%/84%, defined by the ratio of (10a-10b’)/
(10aþ10b’) or (10b-10a’)/(10b þ 10a’), is due to the high enantio-
selectivity of the final hydrogenation of emanine 9, whereas the
dr of 2.3, defined by (10aþ10b’)/(10b þ 10a’), is largely determined
by the ee of 7. By setting 10a’ to 1, one can simply solve the equa-
tions and obtain: 10a ¼ 27.6, 10b ¼ 11.4, 10b’ ¼ 1. The enantiomer
ratio of 7 is (10aþ10a’)/(10b þ 10b’) ¼ 2.3, which is the product dr,
because of the high enantio-selectivity of the final reduction step.
The high enantioselectivity suggests the hydride delivery proceeds
through an outer-sphere mechanism, thus the selectivity relies on
the geometrical fit between catalyst and substrate.7d Therefore, the
final reduction of the emamine 9 determines stereochemical
2.4. Outer-sphere mechanism
It was commonly believed that one challenging aspect of the
asymmetric hydrogenation of pyridines is that these substrates
possess strong coordination ability, presumably leading to the
deactivation or “poisoning” of catalysts. With this generally so-
called inner sphere pathway involving direct coordination of un-
saturated bonds to the metal, one might suggest that the lone
electron pair of pyridine will occupy coordination sites of a catalyst
such as a
with the catalyst as a
d
donor that will leave no space for a substrate to bind
donor. To test the inner sphere hypothesis,
p
pyridine, 2, 6-lutidine and t-Bu-lutidine were added to the pyr-
idinium hydrogenation reaction (See SI). Strikingly, all reactions
proceeded smoothly to the product with no significant drop in
either conversion or enantioselectivity. Even when one equivalent
Scheme 3. Tautomerization.
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

6
Y. Huang et al. / Tetrahedron xxx (2018) 1e9
Scheme 4. Hydrogenation of di-substituted pyridinium salt.
Fig. 7. Time course for the reaction (3 h).
Scheme 5. Hydrogenation of enamine 4 to piperidine.
Scheme 7. Hydrogenation of the synthesized ketone 10 to Piperidine.
2.5. Mechanistic proposal
Based on the combination of NMR studies, deuteration experi-
ments, spectroscopic and structural data together we propose a
catalytic reaction mechanism illustrated in Scheme 8. Under the
high pressure of hydrogen and in the presence of the pyridinium
salt 1 the ligand complex 1a forms the active “cis” di-hydride
complex 1b. A probable initial 1, 4-hydride transfer (not ruling
out other possible initial pathway) generates enamine 2. Enamine 2
rapidly isomerizes to iminium 3. Through an outer-sphere pathway,
the active ligand complex 1b delivers hydride to iminium 3 to
afford enamine 4. In a similar manner, enamine 4 isomerizes to
iminium 5, which exists in its ketone form when encountering
water. Hydride delivery and chiral reduction of 5 takes place
through another outer-sphere pathway to release the final chiral
piperidine to complete the catalytic cycle. The high enantiose-
lectivity can then be ascribed to the formation of the anionic
complex 1b that dominantly selected the enantioface of the pyri-
dium salt.
Scheme 6. Intermediate ketone 10 formed in the presence of water.
of pyridine was added (additive: catalyst ¼ 400:1), chiral piperidine
was obtained in excellent conversion and selectivity (97% and 93%
ee). These observations suggested a hydride transfer mechanism is
more plausible in this catalytic systemPrevious reports have sug-
gested an outer-sphere mechanism for transition metal catalyzed
reduction on indoles,6j imines6h,9 and quinolines.11 Most recently
Qu and Kozlowski reported their DFT (density functional theory)
calculations, which support an outer-sphere dissociative mecha-
nism for an Iridium/MeO-BoQPhos reduction of pyridinium salts.7d
Our experimental evidences suggested that our Iridium/MP2-
SEGPHOS hydrogenation of pyridinium salt might go through a
similar step-wise outer-sphere pathway.
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

Y. Huang et al. / Tetrahedron xxx (2018) 1e9
7
Scheme 8. Proposed Outer-sphere mechanism for the asymmetric hydrogenation of 1.
3. Conclusion
desired products.
In summary, we propose a mechanism for the hydrogenation of
N-benzyl-2-phenylpyridium bromide salt when using an MP2-
SEGPHOS/Iridium catalyst. Experimental data suggest that this
iridium hydrogenation proceeds through an outer-sphere pathway
involving sequential proton and hydride transfer. Two tautomeri-
zations were proposed based on the evidence for deprotonation
and protonation processes at the C3 and C5 positions. Additionally,
two key intermediates in this proposed mechanism were isolated
or spectroscopically characterized. Stoichiometric and catalytic
experiments with the isolated intermediates provided further ev-
idence for the mechanistic proposal.
4.1.1. 1-Benzyl-2-phenylpiperidine (2a)
White solid; 97% yield; 96% ee. The compound data were in good
accordance with the literature.13 1H NMR (400 MHz, CDCl3):
d
1.30e1.43 (m, 1H),1.52e1.67 (m, 3H),1.70e1.83 (m, 2H), 1.87e1.98
(dt, J1 ¼3.4 Hz, J2 ¼ 11.5 Hz, 1H), 2.80 (d, J ¼ 13.2 Hz, 1H), 2.92e3.00
(m, 1H), 3.10 (dt, J1 ¼ 2.6 Hz, J2 ¼ 11.2 Hz, 1H), 3.76 (d, J ¼ 13.8 Hz,
1H), 7.16e7.28 (m, 6H), 7.32 (t, J ¼ 7.7 Hz, 2H), 7.45 (d, J ¼ 7.5 Hz,
2H). 13C NMR (400 MHz, CDCl3):
d 25.3, 26.0, 37.0, 53.4, 59.8, 69.2,
126.5, 126.9, 127.5, 128.0, 128.5, 128.7, 139.9, 145.7. Enantiomeric
excess was determined by SFC: OJ-3 column (150 ꢁ 4.6 mm),
200 bar, 40 ꢂC, CO2/MeOH with 25 mM IBA ¼ 95:5 (0e4 min), 60:40
(4e6.4 min), 35:65 (6.4e6.5 min), 3 ml/min. HRMS calcd for
C18H22Nþ: 252.1747, found: 252.1751.
4. Experimental section
4.1. Typical procedure for the synthesis of 2-arylpyridine12
4.2. Asymmetric pyridinium salt reductions with D2
A Schlenk flask containing a magnetic stir bar was charged
sequentially with Pd(OAc)2 (0.074 g, 0.33 mmol), dppf (0.19 g,
0.34 mmol), 2-bromopyridine (0.79 g, 5 mmol), aryl boronic acid
(6 mmol), and degassed dioxane (20 ml). The mixture was stirred at
In
a
nitrogen-filled glovebox, MP2-SEGPHOS(4.58 mg,
0.0099 mmol) and [Ir(COD)Cl]2 (3.07 mg, 0.00457 mmol) were
placed into a vial and stirring for 30 min in acetone (7.2 ml). Pyr-
idinium salt (0.05 mmol) was placed into 4 ml hydrogenation vials.
0.2 ml catalyst solution and remaining solvent were added. The
vials were placed in a parallel hydrogenation block and following
three D2 purges, pressurized to 600 psi at 30 ꢂC for 20 h. After
carefully releasing the hydrogen, selectivities were determined by
direct sampling of the reaction mixture on SFC or HPLC. Then
saturated sodium carbonate was added and the mixture was stirred
room temperature for 15 min.
A solution of Cs2CO3 (3.26 g,
10 mmol) in 5 ml of degassed H2O was added. The mixture was
heated to reflux for 3 h. The organic layer was separated and
extracted with EtOAc twice, and the combined organic extracts
were dried over Na2SO4 and concentrated in vacuo. Purification was
performed by a silica gel column, eluted with hexane/EtOAc to give
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

8
Y. Huang et al. / Tetrahedron xxx (2018) 1e9
(d) Wang D-S, Chen Q-A, Lu S-M, Zhou Y-G. Chem Rev. 2012;112:2557e2590;
for 15e30 min. The organic layer was separated and extracted with
CH2Cl2 twice, and the combined organic extracts were dried over
Na2SO4 and concentrated in vacuo. Purification was performed by a
silica gel column, eluted with hexane/EtOAc to give desired
product.
€
€
(e) Kallstrom S, Leino R. Bioorg Med Chem. 2008;16:601e635.
3. (a) Rylander PN. Catalytic Hydrogenation in Organic Synthesis. New York: Aca-
demic Press; 1979;
(b) Augustine RL. Heterogeneous Catalysis for the Synthetic Chemist. New York:
Marcel Dekker; 1995;
(c) Freifelder M. Adv Catal. 1963;14:203;
(d) Tang W, Zhang X. Chem Rev. 2003;103:3029e3070;
(e) Cui X, Burgess K. Chem Rev. 2005;105:3272e3296;
(f) Church TL, Andersson PG. Coord Chem Rev. 2008;252:513e531;
(g) Gridnev ID, Imamoto T. Chem Commun. 2009:7447e7464;
(h) Xie J-H, Zhu S-F, Zhou Q-L. Chem Rev. 2011;111:1713e1760;
(i) Zhou Y-G. Acc Chem Res. 2007;40:1357e1366;
(j) Carroll MP, Guiry PJ. Chem Soc Rev. 2014;43:819e833;
(k) Verendel JJ, Pamies O, Dieguez M, Andersson PG. Chem Rev. 2014;114:
2130e2169;
(l) Zhang Z, Butt NA, Zhang W. Chem Rev. 2016;116:14769e14827;
(m) Zhao D, Candish L, Paul D, Glorius F. ACS Catal. 2016;6:5978e5988;
(n) Zhu S-F, Zhou A-L. Acc Chem Res. 2017;50:988e1001;
(o) Yu Z, Jin W, Jiang Q. Angew Chem Int Ed. 2012;51:6060e6072;
(p) Limuro A, Hijashida K, Nagae H, Mashima K. Heterocyles. 2017;95:63e80.
4. For selected examples of pyridine reductions, see: (a) Studer M, Wedemeyer-
Exl C, Spindler F, Blaser H-U. Monatsh Chem. 2000;131:1335e1343;
(b) Legault CY, Charette AB. J Am Chem Soc. 2005;127:8966e8967;
(c) Wang X, Zeng W, Zhou Y-G. Tetrahedron Lett. 2008;49:4922e4924;
(d) Ye Z-S, Chen M-W, Chen Q-A, Shi L, Duan Y, Zhou Y-G. Angew Chem Int Ed.
2012;51:10181e10184;
(e) Liu Y, Du H. J Am Chem Soc. 2013;135:12968e12971;
(f) Wei X, Qu B, Zeng X, et al. J Am Chem Soc. 2016;138:15473e15481. For
selected examples of quinoline and isoquinoline reductions, see:;
(g) Wang W-B, Lu S-M, Yang P-Y, Han X-W, Zhou YeG. J Am Chem Soc.
2003;125:10536e10537;
4.3. NMR sample preparation of Ir(COD)(MP2-SEGPHOS)
precatalyst and Ir(MP2-SEGPHOS) hydride catalyst
The precatalyst was formed by mixing (Ir[COD]Cl)2 and MP2-
SEGPHOSat 1:2 M ratio in acetone-d6 in the glove-box, which
yielded an NMR sample of 2.5 mM Ir(COD)(MP2-SEGPHOS). A 5 mm
NMR tube with a low-pressure/vacuum valve (Wilmad-LabGlass,
Inc) was employed to maintain a N2 atmosphere. To obtain the
activated catalytic complex, we dissolved the Ir(COD)(MP2-SEG-
PHOS) complex in acetone-d6 under a H2 pressure of 600psi for
16 h. The sample was then quickly transferred to a low-pressure
NMR tube under N2 atmosphere in a glove box. In order to eval-
uate the potential stabilizing effects of different counter-ions, we
prepared three other NMR samples of the charged catalyst but with
2.5 molar equivalence of (NH4)Br, (NH4)BF4, and (NH4)PF6, respec-
tively. The N2 gas in the headspace of the sample tubes was then
flushed out immediately with 1.5 bar of H2 on a pressure manifold
following sample transfer.
(h) Lu S-M, Wang Y-Q, Han X-W, Zhou Y-G. Angew Chem Int Ed. 2006;45:
2260e2263;
(i) Shi L, Ye Z-S, Cao L-L, Guo R-N, Hu Y, Zhou Y-G. Angew Chem Int Ed. 2012;51:
8286e8289;
(j) Wang Y, Xia J, Yang G, Zhang W. Tetrahedron. 2018;74:477e482.
5. Chang M, Huang Y, Liu S, et al. Angew Chem Int Ed. 2014;53:12761e12764.
6. (a) Limuro A, Yamaji K, Kandula S, Nagano T, Kita Y, Mashima K. Angew Chem
Int Ed. 2013;52:2046e2050;
(b) Shi L, Ye Z, Cao L, Guo R, Hu Y, Zhou Y-G. Angew Chem. 2012;124:
8411e8414;
4.3.1. Deuterium content quantification by NMR
Deuterium level quantification was calculated by integrating 1H
spectrum and qualitatively verified by multiplicity-edited HSQC
spectrum (See Supporting information for more details).
4.3.2. NMR experiments
All NMR experiments were performed at 25 ꢂC. Experiments for
(c) Wang D-W, Wang X-B, Wang D-S, Lu S-M, Zhou Y-G, Li Y-X. J Org Chem.
2009;74:2780e2787;
deuterium level quantification were performed on
a Bruker
500 MHz spectrometer equipped with a room-temperature probe.
Experiments involving 31P NMR were conducted on a Bruker
500 MHz spectrometer equipped with a Bruker SmartProbe™.
Other NMR experiments, including those for the characterization of
the Ir(COD)(MP2-SEGPHOS) precatalyst and the activated Ir(MP2-
SegPhos) hydride catalyst, and the enamine and ketone in-
termediates, were conducted on a Bruker 500 MHz spectrometer
equipped with a Prodigy™ probe.
(d) Cai X-F, Guo R-N, Chen M-W, Shi L, Zhou Y-G. Chem Eur J. 2014;20:
7245e7248;
(e) Wang T, Zhuo L-G, Li Z, et al. J Am Chem Soc. 2011;133:9878e9891;
(f) Kuwano R, Hashiguchi Y, Ikeda R, Ishizuka K. Angew Chem Int Ed. 2015;54:
2393e2396;
(g) Ye Z-S, Guo R-N, Cai X-F, Chen M-W, Shi L, Zhou Y-G. Angew Chem Int Ed.
2013;52:3685e3689;
(h) Tutkowski B, Kerdphon S, Lime E, et al. ACS Catal. 2018;8:615e623;
(i) Schramm Y, Barrios-Landeros F, Pfaltz A. Chem Sci. 2013;4:2760e2766;
(j) Duan Y, Li L, Chen M-W, Yu C-B, Fan H-J, Zhou Y-G. J Am Chem Soc. 2014;136:
7688e7700;
(k) Xie J-B, Xie J-H, Liu X-Y, Zhang Q-Q, Zhou Q-L. Chem Asian J. 2011;6:
899e908;
Acknowledgements
(l) Yuan Q, Liu D, Zhang W. Org Lett. 2017;19:1886e1889;
(m) Cui X, Fan Y, Hall M-B, Burgess K. Chem Eur J. 2005;11:6859e6868.
7. (a) An isotopic labeling experiment has been reported for the non-
enantioselective Rhodium-catalyzed transfer hydrogenation of N-benzylated
pyridinium salts: Wu J, Tang W, Pettman A, Xiao J Adv Synth Catal. 2013;355:
35e40;
(b) During the period of time we worked on the extensive mechanism studies
and put the work together in writing, Lefort and co-workers reported their
mechanism study on rhodium catalyzed asymmetric hydrogenation of N-
benzyl 3-substituted pyridinium salts: Renom-Carrasco M, Gajewski P,
Pignataro L, de Vries JG, Piarulli U, Gennari C Lefort L Chem Eur J. 2016;22:
9528e9532;
(c) Renom-Carrasco M, Gajewski P, Pignataro L, et al. Adv Synth Catal. 2016;358:
2589e2593. We acknowledge the work by Lefort and coworkers on the
mechanism studies, of which our work is independent but shares considerable
similarity in the experimental design and converges in the general conclusion
as well;
The authors thank the following scientists at Merck for experi-
mental support and helpful discussions: Mr. Jerry Hill and Mr.
Mathew Maust (High Pressure Laboratory) and Mr. Rong-Sheng
Yang (HRMS analysis), Dr. YingNi Ji Chen, Dr. George Zhou, Dr.
Xianhai Huang and Dr. Frank Bennett.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.tet.2018.03.031.
References
(d) Most recently, a DFT computational studies have been reported for Ir-
catalyzed AH of alfa-(Hetero)aryl pyridinium salts using MeO-BoQPhos: Qu B,
Mangunuru HPR, Kozlowski MC, Senanayake CH, et al Org Lett. 2018;20:
1333e1337.
8. (a) Crabtree RH. The Organometallic Chemistry of the Transition Metals. sixth ed.
John Wiley & Sons, Inc; 2014;
(b) Goodfellow RJ, Taylor BF. J Chem Soc Dalton Trans. 1974;0:1676e1684;
(c) Deeming AJ, Doherty S, Marshall JE. Polyhedron. 1991;10:1857e1864;
(d) Kita Y, Yamaji K, Higashida K, Sathaiah K, Limuro A, Mashima K. Chem Eur J.
2015;21:1915e1927.
1. (a) Lovering F, Bikker J, Humblet C. J Med Chem. 2009;52:6752e6756;
(b) Petukhov PA, Zhang J, Wang CZ, Ye YP, Johnson KM, Kozikowski AP. J Med
Chem. 2004;47:3009e3018;
(c) Zhang Y, Liu Y-B, Li Y, et al. J Nat Prod. 2013;76:1058e1063;
(d) Peng Z, Wong J-W, Hansen EC, Puchlopek-Dermenci ALA, Clarke HJ. Org Lett.
2014;16:860e863.
2. For selected reviews, see: (a) Nugent T-C. Chiral Amine Synthesis. Weinheim:
Wiley-VCH; 2010;
(b) Bailey PD, Millwood PA, Smith PD. Chem Commun. 1998:633e640;
(c) Glorius F. Org Biomol Chem. 2005;3:4171e4175;
9. Hopmann KH, Bayer A. Organometallics. 2011;30:2483e2497.
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031

Y. Huang et al. / Tetrahedron xxx (2018) 1e9
9
10. (a) Scott NM, Pons V, Stevens ED, Heinekey DM, Nolan SP. Angew Chem Int Ed.
11. (a) Dobereiner GE, Nova A, Schley ND, et al.
7547e7562;
J
Am Chem Soc. 2011;133:
2005;44:2512e2515;
(b) Castillo MR, Martín M, Fraile JM, Mayoral JA. Angew Chem Int Ed. 2011;50:
3240e3243;
(c) Gruber S, Pfaltz A. Angew Chem Int Ed. 2014;53:1896e1900.
(b) Zhou H, Li Z, Wang Z, et al. Angew Chem Int Ed. 2008;47:8464e8467.
12. Dai X, Chen Y, Garrell S, et al. J Org Chem. 2013;78:7758e7763.
13. Takahashi M, Micalizio GC. J Am Chem Soc. 2007;129:7514e7516.
Please cite this article in press as: Huang Y, et al., A mechanistic investigation of an Iridium-catalyzed asymmetric hydrogenation of pyridinium
salts, Tetrahedron (2018), https://doi.org/10.1016/j.tet.2018.03.031