Pyranoside phosphite-oxazoline ligands for the highly versatile and enantioselective Ir-catalyzed hydrogenation of minimally functionalized olefins. A combined theoretical and experimental study

Article abstract of DOI:10.1021/ja204948k

A modular set of phosphite-oxazoline (P,N) ligands has been applied to the title reaction. Excellent ligands have been identified for a range of substrates, including previously challenging terminally disubstituted olefins, where we now have reached enantioselectivities of 99% for a range of substrates. The selectivity is best for minimally functionalized substrates with at least a moderate size difference between geminal groups. A DFT study has allowed identification of the preferred pathway. Computational prediction of enantioselectivities gave very good accuracy.

Full text of DOI:10.1021/ja204948k

ARTICLE
pubs.acs.org/JACS
Pyranoside PhosphiteꢀOxazoline Ligands for the Highly Versatile
and Enantioselective Ir-Catalyzed Hydrogenation of Minimally
Functionalized Olefins. A Combined Theoretical and Experimental
Study
Javier Mazuela,† Per-Ola Norrby,*,‡ Pher G. Andersson,*,§,|| Oscar Pꢀamies,† and Montserrat Diꢁeguez*,†
†Departament de Química Física i Inorgꢀanica, Universitat Rovira i Virgili, Campus Sescelades, C/Marcel lí Domingo, s/n. 43007
3
Tarragona, Spain
ꢁ
‡Department of Chemistry, University of Gothenburg, Kemigarden 4, SE-412 96 G€oteborg, Sweden
§Department of Biochemistry and Organic Chemistry, Uppsala University, Box 576, 751 23 Uppsala, Sweden
School of Chemistry, University of KwaZulu-Natal, Durban, South Africa
S Supporting Information
b
ABSTRACT: A modular set of phosphiteꢀoxazoline (P,N) ligands has been
applied to the title reaction. Excellent ligands have been identified for a range of
substrates, including previously challenging terminally disubstituted olefins,
where we now have reached enantioselectivities of 99% for a range of
substrates. The selectivity is best for minimally functionalized substrates with
at least a moderate size difference between geminal groups. A DFT study has
allowed identification of the preferred pathway. Computational prediction of
enantioselectivities gave very good accuracy.
in 2008, we communicated the successful application of pyrano-
side phosphiteꢀoxazoline ligands in the Ir-catalyzed asymmetric
hydrogenation of model trisubstituted and terminal substrates.5a
To fully investigate the potential of these ligands, in this Article
we expand the 2008 study to other pyranoside phosphiteꢀ
oxazoline ligands (Figure 1) and to other types of more challen-
ging substrates. We have therefore extended the Irꢀphosphiteꢀ
oxazoline catalyst library by including pyranoside ligands with
new substituents at the oxazoline (ligands L4) and at the biaryl
phosphite group (a, b, f, g, and j). Ligands L1ꢀL5aꢀk combine
the advantages of phosphite and sugar cores: that is to say, they
are readily available from cheap feedstocks and have high resis-
tance to oxidation, and a straightforward modular construction.6
With this ligand library, we can investigate the effect of system-
atically varying the electronic and steric properties of the oxazo-
line substituent (L1ꢀL5) and different substituents and
configurations in the biaryl phosphite moiety (aꢀk) with the
aim to maximize catalyst performance. By judicious choice of the
ligand components, we achieved high enantioselectivities and
activities in a wide range of E- and Z-trisubstituted and 1,
1-disubstituted alkenes.
1. INTRODUCTION
Pharmaceuticals, agrochemicals, fragrances, fine chemicals,
and natural chemicals all rely on the preparation of enantiomeri-
cally enriched compounds.1 Because of its high efficiency, atom
economy, and operational simplicity, the asymmetric hydro-
genation of properly selected prochiral starting materials can
be a sustainable and direct synthetic tool for preparing these
compounds.1
Whereas the reduction of olefins containing an adjacent polar
group (i.e., dehydroaminoacids) by Rh- and Ru-catalyst precur-
sors modified with phosphorus ligands has a long history, the
asymmetric hydrogenation of minimally functionalized olefins is
less developed because these substrates do not have an adjacent
polar group to direct the reaction.1 Iridium complexes with chiral
P,N ligands have become established as one of the most efficient
catalyst types for the hydrogenation of minimally functionalized
olefins, and they complement Rhꢀ and Ruꢀdiphosphine
complexes.2,3 The first successful P,N ligands4 contained a
phosphine or phosphinite moiety as P-donor group and either
an oxazoline,4b,g,j oxazole,4d thiazole,4i or pyridine4c as N-donor
group. However, these iridiumꢀphosphine/phosphinite,N cat-
alysts were still highly substrate-dependent, and the development
of efficient chiral ligands that tolerate a broader range of
substrates remained a challenge. In our efforts to expand the
range of ligands and improve performance, we recently discov-
ered that the presence of biarylꢀphosphite moieties in these P,
N-ligands provides greater substrate versatility than previous
Irꢀphosphine/phosphinite,N catalyst systems.5 In this context,
Despite the recent success of Ir/phosphiteꢀnitrogen catalyst
systems, no mechanistic studies have been carried out on the
hydrogenation of minimally functionalized alkenes using phos-
phite ligands. In this context, the mechanistic aspects of these
Received: May 30, 2011
Published: July 15, 2011
r
2011 American Chemical Society
13634
dx.doi.org/10.1021/ja204948k J. Am. Chem. Soc. 2011, 133, 13634–13645
|

Journal of the American Chemical Society
ARTICLE
Figure 1. Pyranoside phosphiteꢀoxazoline ligand library L1ꢀL5aꢀk. In this work, the ligand library has been expanded by including new substituents
at the oxazoline (L4) and at the biaryl phosphite moiety (a, b, f, g, and i).
Scheme 1. Synthesis of New Pyranoside Ligands
Scheme 2. Synthesis of Catalyst Precursors [Ir(cod)(PꢀN)]
BArF (PꢀN = L1ꢀL5aꢀk)
(1 equiv), in the presence of water (Scheme 2).9 All complexes
were isolated as air-stable orange solids and were used without
further purification.
ligands are still not understood well enough for the a priori
prediction of the type of ligand needed for high selectivity.
To address this important point, in this Article we have also
performed DFT calculations to explain the origin of enantios-
electivity for these highly versatile pyranoside Ir/phosphiteꢀ
oxazoline catalytic systems. It should be noted that we have also
elucidated a computational model that can explain the enantios-
electivities obtained with 1,1-disubstituted substrates, which was
lacking in the literature.
The complexes were characterized by elemental analysis and
1H, 13C, and 31P NMR spectroscopy. The spectral assignments
1
1
were based on information from 1Hꢀ H and 13Cꢀ H correla-
tion measurements and were as expected for these C1 iridium
complexes. The VT-NMR spectra indicate that only one isomer
is present in solution. One singlet in the 31P{1H} NMR spectra
was obtained in all cases.10
2.3. Asymmetric Hydrogenation of Trisubstituted Ole-
fins. 2.3.1. Asymmetric Hydrogenation of Minimally Functional-
ized Trisubstituted Olefins. In a first set of experiments, we used
the Ir-catalyzed hydrogenation of substrates trans-R-methylstil-
bene S1 and Z-2-(4-methoxyphenyl)-2-butene S2 to study the
potential of ligands L1ꢀL5aꢀk. Substrate S1 was chosen as a
model for the hydrogenation of E-isomers because it has been
reduced with a wide range of ligands, which enable the efficiency
of the various ligand systems to be compared directly.2 To assess
the potential of the ligand library L1ꢀL5aꢀk for the more
demanding Z-isomers, which are usually hydrogenated less
enantioselectively than the corresponding E-isomers, we chose
substrate S2 as a model. The reactions proceeded smoothly at
room temperature. Excellent activities and enantioselectivities
(up to >99% for S1 and up to 95% for S2) were obtained. The
results, which are summarized in Table 1, indicate that activity is
mainly affected by the steric properties of the oxazoline sub-
stituent and by the substituents at the ortho positions of the
biaryl phosphite moiety. Bulky substituents need to be present in
the biaryl phosphite and less sterically demanding substituents in
the oxazoline if activities are to be high. Enantioselectivity, on the
other hand, is affected by the electronic and steric properties of
the substituents in the oxazoline moiety and by the substituents/
configuration in the biaryl phosphite moiety. However, the effect
2. RESULTS AND DISCUSSION
2.1. Synthesis of Ligands. The synthesis of the new pyrano-
side phosphite-oxazoline ligands (L4c, L5a, L5b, L5f, and L5g) is
straightforward following the procedure previously described for
ligands L1ꢀL3c, L5aꢀe, and L5hꢀk (Scheme 1).5a,7 They were
therefore efficiently synthesized in one step by reacting the
corresponding sugar oxazolineꢀalcohols (1 and 2) with 1 equiv
of the corresponding biaryl phosphorochloridite (ClP(OR)2;
(OR)2 = aꢀc, f, g) in the presence of pyridine, in a parallel way
(see Experimental Section for details). Oxazolineꢀalcohols 1
and 2 are easily prepared from inexpensive D-glucosamine on a
large scale.8 All of the ligands were stable during purification on
neutral alumina under an atmosphere of argon and isolated in
moderate-to-good yields as white solids. They were stable at
room temperature and very stable to hydrolysis. The elemental
analyses were in agreement with the assigned structure. The 1H,
13C, and 31P NMR spectra were as expected for these C1 ligands.
2.2. Synthesis of the Ir-Catalyst Precursors. The catalyst
precursors were made by refluxing a dichloromethane solution of
the appropriate ligand (L1ꢀL5aꢀk) in the presence of 0.5 equiv
of [Ir(μ-Cl)cod]2 for 1 h and then exchanging the counterion with
sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate (NaBArF)
13635
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
Table 1. Ir-Catalyzed Asymmetric Hydrogenation of S1 and
Scheme 3. Selected Hydrogenation Results of Trisubstituted
Olefins Using [Ir(cod)(L1ꢀL5aꢀk)]BArF Catalyst
Precursorsa
S2 Using Ligands L1ꢀL5aꢀka
a Reaction conditions: 1 mol % catalyst precursor, CH2Cl2 as solvent, 50
bar H2, 2 h.
phosphiteꢀoxazoline ligand library L1ꢀL5aꢀk. The most note-
worthy results are shown in Scheme 3. The enantioselectivities
are among the best observed for these substrates.2,4 In general,
the IrꢀL5e catalyst precursor provides the best enantioselec-
tivities for the hydrogenation of E-trisubstituted olefins (S1,
S3ꢀS5), while ligand L5c gives the best enantioselectivities for
Z-trisubstituted olefins (S2, S6ꢀS8). Our results also indicated
that enantioselectivity (ee values up to >99%) is relatively
insensitive to the electronic nature of the substrate phenyl ring
(i.e., substrates S1, S5, and S6 versus S3, S4, and S2, re-
spectively). Notably the [Ir(cod)(L1ꢀL5aꢀk)]BArF catalyst
precursors also proved to be highly active and enantioselective
in the reduction of triarylsubstituted substrates S9 and S10
(Scheme 3, ee’s ranging from 98% to >99%). This latter substrate
class provides an easy entry point to diarylmethine chiral centers,
which are present in several important drugs (such as (R)-tolte-
rodine and sertraline) and natural products (i.e., podohyllotoxin).11
Despite this, only one previous study has been made using Ir-
catalysts modified with chiral phosphine/thiazole ligands (ee’s
up to 99%).12
2.3.2. Asymmetric Hydrogenation of Trisubstituted Olefins
Containing a Neighboring Polar Group. To further study the
potential of the pyranoside phosphiteꢀoxazoline ligand library
L1ꢀL5aꢀk in the reduction of minimally functionalized trisub-
stituted olefins, we screened it in the Ir-catalyzed hydrogenation
of trisubstituted alkenes containing a neighboring polar group.
These substrates are interesting because they allow for further
functionalization, and they are important intermediates for the
synthesis of high-value chemicals. The results are summarized in
Scheme 4. Again, excellent enantioselectivities (ee values up
to >99%) for a range of substrates were obtained under mild
reaction conditions. The reduction of several R,β-unsaturated
esters (S11ꢀS14) followed the same trends as those observed
for the previous E-trisubstituted substrates. Therefore, enantios-
electivities were best using ligand L5e. It should be noted that
ee’s are highly independent of the electronic nature of the
substrate phenyl ring and the substituent in the ester function-
ality. As expected, the hydrogenation of vinylsilane S15 was also
best using ligand L5e. However, for trisubstituted allylic alcohol
of these ligand parameters on enantioselectivity depends on the
substrate type (E- or Z-isomers). Thus, while for the E-substrate
S1 the enantioselectivity was best with ligand L5e (>99% ee),
enantioselectivities for the more demanding Z-substrate S2 were
best with ligand L5c (95% ee). For both types of substrates, we
also found a cooperative effect between the configuration of the
biaryl phosphite moiety and the configuration of the sugar back-
bone on enantioselectivity. This led to a matched combination
for ligands L5g, L5i, and L5k, which contain an R-biaryl moiety
(entries 11, 13, and 15). In addition, a comparison of the ab-
solute stereochemistry obtained by using tropoisomeric biphenyl
ligands L5aꢀe with those obtained with the related atropo-
isomeric biaryl ligands (L5fꢀk) shows that the tropoisomeric
biphenyl moiety in ligands L5aꢀe adopts an R-configuration
when complexed with iridium.10
Next, we used the ligands that provided the best results (ligands
L5c and L5e) to study these reactions at a low catalyst loading
(0.2 mol %). In these conditions, the excellent enantioselectivities
and activities were maintained (Table 1, entries 16 and 17).
We then studied the asymmetric hydrogenation of other
E- and Z-trisubstituted olefins (S3ꢀS10) by using the pyranoside
13636
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
Scheme 4. Selected Hydrogenation Results of Trisubstituted
Olefins Bearing a Neighboring Polar Group Using
[Ir(cod)(L1ꢀL5aꢀk)]BArF Catalyst Precursorsa
Scheme 5
Table 2. Selected Results for the Ir-Catalyzed Hydrogenation
of S26 Using the Ligands L1ꢀL5aꢀka
entry
ligand
% convb
% eec
1
2
L1c
L2c
L3c
L4c
L5a
L5b
L5c
L5d
L5e
L5f
100
100
100
100
99
76 (S)
84 (S)
89 (S)
87 (S)
97 (S)
ꢀ
92 (S)
97 (S)
99 (S)
39 (S)
94 (S)
34 (S)
97 (S)
23 (S)
98 (S)
3
4
5
6
<5
7
100
100
100
100
100
95
a Reaction conditions: 1 mol % catalyst precursor, CH2Cl2 as solvent, 50
bar H2, 2 h.
8
9
10
11
12
13
14
15
L5g
L5h
L5i
S16, allylic acetate S17, several R,β-unsaturated ketones S18ꢀ
S22, and vinylboronates S23ꢀS25, enantioselectivities were best
with ligand L5c. It should be pointed out the unprecedented
excellent enantioselectivities obtained in the hydrogenation of
vinylboronates (ee’s ranging from 92% to >99%). The hydro-
genation of vinylboronates provides an easy access to chiral
borane compounds, which are useful building blocks in organic
synthesis because the CꢀB bond can be readily converted to
CꢀO, CꢀN, and CꢀC bonds with retention of the chirality. For
R,β-unsaturated ketones and vinylboronates, ee’s are again highly
independent of the electronic properties of the phenyl substrate
ring. Moreover, our results indicate that for enones S18ꢀS22 the
substituent linked to the ketone functionality has little effect on
enantioselectivity. In summary, Ir/L1ꢀL5aꢀk complexes have
therefore emerged as one of the few catalytic systems that can
hydrogenate a wide range of minimally functionalized E- and
Z-trisubstituted olefins (including those containing a neighbor-
ing polar group) in high activities and enantioselectivities.4,5
2.4. Asymmetric Hydrogenation of 1,1-Disubstituted
Terminal Olefins. We next screened ligands L1ꢀL5aꢀk in
the asymmetric hydrogenation of more demanding terminal
olefins. Enantioselectivity is more difficult to control in these sub-
strates than in trisubstituted olefins. There are two main reasons
for this:2d,e (a) the two substituents in the substrate can easily
exchange positions in the chiral environment formed by the
catalysts, thus reversing the face selectivity (Scheme 5a), and (b)
the terminal double bond can isomerize to form the more stable
internal alkene, which usually leads to the predominant forma-
tion of the opposite enantiomer of the hydrogenated product
100
100
100
L5j
L5k
a Reactions carried out using 1 mmol of S26 and 0.2 mol % of Ir-catalyst
precursor at 1 bar of H2. b Conversion measured by GC after 2 h.
c Enantiomeric excesses determined by chiral GC.
(Scheme 5b). Few known catalytic systems provide high enantios-
electivities for these substrates, and those that do are usually
limited in substrate scope.2e,13,14 In contrast to the hydrogena-
tion of trisubstituted olefins, the enantioselectivity in the reduc-
tion of terminal alkenes is highly pressure dependent. Therefore,
hydrogenation at an atmospheric pressure of H2 gave, in general,
significantly higher ee values than at higher pressures.4b,13
2.4.1. Asymmetric Hydrogenation of Unfunctionalized 1,
1-Disubstituted Terminal Olefins. In a first set of experiments,
we used the Ir-catalyzed asymmetric hydrogenation of 2-phe-
nylbut-1-ene S26. The results obtained using the ligand library
L1ꢀL5aꢀk in optimized conditions are shown in Table 2. We
were again able to fine-tune the ligand parameters to produce
high activities and enantioselectivities (ee’s up to 99%) in the
hydrogenation of this substrate using low catalyst loadings
(0.2 mol %) and hydrogen pressures (1 bar). Enantioselectivities
were affected by the electronic and steric properties of the
oxazoline moiety and by the substituents/configurations of the
biaryl phosphite group. In general, the effect of these ligand
parameters on enantioselectivity followed the same trend as for
13637
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
Table 3. Selected Results for the Ir-Catalyzed Hydrogenation of Minimally Functionalized 1,1-Disubstituted Terminal Olefins
Using Ligands L1ꢀL5aꢀka
a Reactions carried out using 1 mmol of substrate and 0.2 mol % of Ir-catalyst precursor at 1 bar of H2. b Conversion measured by 1H NMR or GC.
c Enantiomeric excesses determined by chiral GC (except for entries 12ꢀ14 that were measured by HPLC). d Reaction carried out at 50 bar of H2.
the reduction of trisubstituted olefins. However, the effect of the
substituents at the para positions of the biaryl phosphite moiety
is different. Therefore, in contrast to the previous trisubstituted
olefins, enantioselectivities increased if the steric bulk of the
substituent at the para position of the biaryl phosphite moiety
decreased (i.e., L5e (H)> L5d (OMe) > L5c (tBu)). In summary,
enantioselectivities were best (99% ee) when phosphiteꢀoxazo-
line ligand L5e, which contains a phenyl group in the oxazoline
moiety and bulky ortho trimethylsilyl groups at the biphenyl
phosphite moiety, was used. This result, which again clearly shows
the efficiency of using modular scaffolds in ligand design, is among
the best that have been reported for this demanding substrate.2e
We then studied the asymmetric hydrogenation of other 1,
1-disubstituted arylꢀalkyl substrates (S27ꢀS34), 1,1-disubstit-
uted heteroarylꢀalkyl olefins (S35ꢀS37), and 1,1-diaryl term-
inal alkenes (S38ꢀS40) by using the phosphiteꢀoxazoline
ligand library L1ꢀL5aꢀk. The most noteworthy results are
shown in Table 3. Again, these results are among the best
reported for these substrates.2e For arylꢀalkyl and heteroar-
ylꢀalkyl substrates (S27ꢀS37), the results follow the same
trends as the hydrogenation of S26. Again, the catalyst precursor
containing the phosphiteꢀoxazoline ligand L5e provided the
best enantioselectivities (ee’s up to 99%). However, for the
hydrogenation of the diaryl terminal alkenes, enantioselectivities
were best using ligand L5k, which differs from L5e in the biaryl
phosphite group.
Our results with several 1,1-disubstituted arylꢀalkyl substrates
(S26ꢀS32) indicated that enantioselectivity is affected by the
nature of the alkyl chain (ee’s ranging from 83% to 99%, Table 2,
entry 9 and Table 3, entries 1ꢀ6). One plausible explanation for
this can be found in the competition between direct hydrogena-
tion versus isomerization for the different substrates. This is
supported by the fact that the hydrogenation of substrate S32
bearing a tert-butyl group, for which isomerization cannot occur,
provides high levels of enantioselectivity (ee’s up to 97%; Table 3,
entry 6), while the lowest enantioselectivity of the series (ee’s up
to 84%; Table 3, entries 4,5) is found for substrates S30 and S31,
which form the most stable isomerized tetrasubstituted olefins.
The hydrogenations of several para-substituted 2-phenylbut-
2-enes with different electronic properties (S26, S33, and S34) all
gave similar highactivities and enantioselectivities (full conversion,
ee’s up to 99%; Table 2, entry 9 and Table 3, entries 7 and 8).
We next decided to apply this ligand library in the asymmetric
hydrogenation of 1,1-heteroaromatic alkenes (S35ꢀS37,
Table 3, entries 9ꢀ11). Even though heterocycles are used in
industry and the heterocyclic part can be modified posthydro-
genation, very few previous studies have been made.5c,d Under
standard conditions, our catalyst systems were also able to
13638
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
hydrogenate this type of substrate with excellent activities and
enantioselectivities (ee’s up to 99%).
enantioselectivities (71% ee for S43 and 96% ee for S44) were
best with ligands L5k and L3c, respectively. These results, which
again clearly show the efficiency of using modular scaffolds in
ligand design, are among the best that have been reported for
these demanding substrate classes.2e
Encouraged by the excellent results, we decided to study the
hydrogenation of several diaryl terminal alkenes (S38ꢀS40;
Table 3, entries 12ꢀ14). Enantiopure diarylalkanes are impor-
tant intermediates for the preparation of drugs and research
materials.15 They have traditionally been prepared using rather
laborious approaches.15,16 We have recently demonstrated that
they can be prepared more efficiently using enantioselective
hydrogenation.5c Interestingly, both substrate types differing
electronically (S38) and sterically (S39, S40) were hydrogenated
with fair enantioselectivities (ee’s up to 70%) using ligand L5k.
2.4.2. Asymmetric Hydrogenation of 1,1-Disubstituted Term-
inal Olefins Containing a Neighboring Polar Group. To further
determine the scope of this ligand library, we also examined the
asymmetric hydrogenation of 1,1-disubstituted terminal olefins
containing a polar neighboring group (S41ꢀS44). The results
are summarized in Scheme 6.
We initially tested the ligand library in the hydrogenation of
the allylic alcohol S41 and allylic acetate S42. Derivatives of the
hydrogenation of these products are important intermediates for
the synthesis of high-value cosmetics, natural products, and
drugs.17 Enantioselectivities up to 70% were obtained using
ligand L5c. We then turned our attention to the asymmetric
reduction of the trifluoromethyl olefin S43 and allylic silane
S44. The hydrogenation of these compounds gave rise to
important organic intermediates, and a number of innovative
new organofluorine18 and organosilicon19 drugs are now being
developed. Despite this, only one family of ligands has been
successfully applied in the reduction of these substrates.5c The
2.5. Origin of Enantioselectivity. Computational Studies.
Although the mechanism of olefin hydrogenation (and conse-
quently of stereocontrol) by Rh catalysts is well understood,20
the mechanism when chiral iridium catalysts are used is not,
despite having been investigated both experimentally and
computationally.1,2 In the first case, there is enough evidence
to support a RhI/RhIII mechanism in which substrate chelation to
metal plays a pivotal role in stereodiscrimination, but in the
second four different mechanisms have been proposed (two of
them involving IrI/IrIII intermediates21 and the other two IrIII/IrV
species22). Andersson and co-workers have recently used DFT
calculations and a full, experimentally tested combination of
ligands (mainly phosphine/phosphinite,N) and substrates to
study all of the possible diastereomeric routes of the four
different mechanisms.23 Their studies agree with the two already
proposed catalytic cycles passing through IrIII/IrV intermediates;22
however, they failed to distinguish the two IrIII/IrV mechanisms.24
One of the mechanisms involves an IrIII/IrV migratory-insertion/
reductive-elimination pathway (labeled 3/5-MI in Scheme 7),22c
whereas the second mechanism goes through an IrIII/IrV σ-metathesis/
reductive-elimination pathway (labeled 3/5-Meta in Scheme 7).22a,b
From these cycles, it has been demonstrated that the π-olefin
complex C and the transition states for the migratory-insertion
in 3/5-MI (TS) and the σ-metathesis in 3/5-Meta (TS0) are
responsible for the enantiocontrol in the iridium hydrogena-
tion.23 It was demonstrated that the enantioselectivity could be
reliably obtained from calculated relative energies of migratory
insertion transition states.23
Scheme 6. Selected Hydrogenation Results of 1,1-Disubsti-
tuted Terminal Olefins Containing a Neighboring Polar
Group Using [Ir(cod)(L1ꢀL5aꢀk)]BArF Catalyst
Precursorsa
On the basis of these previous studies and in an attempt to
rationalize the enantioselectivity obtained with the iridium/
phosphiteꢀoxazoline catalyst library reported in this Article,
we performed a computational study of the complexes C and
transition states (TS and TS0) most commonly involved in the
enantiocontrol of the iridium-catalyzed hydrogenation reaction
(Scheme 7).
All geometries were optimized in the gas phase using the Jaguar
program25 and both the B3LYP hybrid density functional26
and the LACVP* basis set. Termochemical data were computed
at 298.15 K for all optimized geometries. Energies in CH2Cl2
a Reaction conditions: 0.5 mol % catalyst precursor, CH2Cl2 as solvent,
50 bar H2, 2 h.
Scheme 7. 3/5-MI and 3/5-Meta Catalytic Cycles for the Ir-Hydrogenation
13639
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
olefins coordinated on the si face are shown in green and are
hydrogenated to (R)-product, whereas those coordinated on the
re face are shown in red and are hydrogenated to the (S)-product.
The results show that the most stable transition state (C13/5-MI
,
Table 4) matches the major product obtained experimentally
(R product, Table 1, entry 9). We also found that the hydro-
genation products are formed through the 3/5-MI mechanism,
with the best 3/5-Meta path significantly higher in energy for
both enantiomers. Therefore, the energies for hydrogenation
pathway 3/5-MI, in both the major and the minor configurations,
are around 15 kJ/mol more stable than those for the 3/5-Meta
pathway (see Table 4). In addition, the difference in energy
between the most stable transition state for the major and the
minor product is 15.6 kJ/mol, which agrees with the excellent
enantioselectivity obtained experimentally using the Ir/L5e
catalyst system (Table 1, entry 9).
Figure 2. Ligand truncation used for preliminary DFT calculations.
Table 4. Calculated Energies for the Most Stable Isomers of
Complex C and Transition States TS and TS0 with Substrate
S1 Using the Truncated Ligand System
Because the method gives satisfactory results, we next applied
it to the reaction of S1 with the full ligand structure. The
difference in energy of 22.9 kJ/mol between the most stable
transition states (Figure 3) matches well with the excellent
enantioselectivities obtained experimentally (>99% ee, Table 1,
entry 9).
The optimized DFT structure shows a quadrant diagram
(depicted in Figure 4), a model already used by Andersson and
co-workers to describe the steroselectivity in the Ir-catalyzed
hydrogenation of trisubstituted olefins.4d In this quadrant model,
we found that the phenyl group of the oxazoline substituent
blocks the upper left quadrant, and one of the aryls of the biaryl
phosphite group partly occupies the lower right quadrant making
it semihindered (Figure 4a). The other two quadrants, which are
free from bulky groups, are open (Figure 4a). Therefore, the
calculated structure clearly shows a chiral pocket that is well
suited to olefins with large trans-substituents (E-olefins). In the
case of the trisubstituted E-olefin S1, the smallest substituent
(H atom of S1) will face the steric bulk of the ligand, and the
olefin will preferentially coordinate to the catalyst from the si side
(Figure 4b), as the opposite coordination mode will result in
unfavorable interactions between one of the large trans-substit-
uents and the steric bulk of the ligand.
This model also explains the lack of enantioselectivity ob-
served when ligands L5f, L5h, and L5j were used (Table 1,
entries 10, 12, and 14). In all tof hese ligands, the biaryl phosphite
group has an S-configuration, which is opposite to the config-
uration adopted by ligand L5e when coordinated to the Ir, as
shown by the catalytic results and the DFT calculation (Figures 3
and 4a). This change in the configuration moves the previously
found semihindered quadrant from the lower right to the upper
right. Therefore, the favorable chiral pocket generated by the
Ir/L5e catalyst, which can accommodate large trans-substituents,
is lost, and consequently Ir/L5f, Ir/L5h, and Ir/L5j catalysts fail
to control the face coordination preference of the olefin.
Using this quadrant model, we can also predict the change in
the sense of enantioselectivity observed experimentally when
E-trisubstituted olefins (i.e., S1) change to Z-olefins (i.e., S2).
The lowest energy transition state, then, will be achieved with the
Z-olefin coordinated through the re face, the opposite face to that
of the E-olefin, with the hydrogen atom positioned in the hindered
upper left quadrant and the aryl substituent in the semihindered
lower right quadrant (Figure 4c). The catalytic results also show
that to obtain high enantioselectivity in the reduction of Z-olefins
we have to switch to ligand L5c. This is not unexpected because
the DFT calculations have shown that the catalytic system Ir/L5e
solution were calculated as single-point energies from optimized
structures at the B3LYP/LACVP* level of theory using a
PoissonꢀBoltzmann continuous field. Final energies were re-
trieved from single-point calculations at the B3LYP/LACV3P++*
level of theory and corrected by inclusion of the van der Waals
repulsion energy calculated by DFT-D3.27 Reported energies
(in kJ/mol) are the Gibbs free energy, including thermodynamic
and solvation contributions. For more computational details, see
the Experimental Section and the Supporting Information.
This computational study has been performed with two
different types of substrate: S1 as a model substrate for trisub-
stituted olefins and S26 as a model for 1,1-disubstituted olefins.
We chose to study the catalyst precursor Ir/L5e because for both
substrates L5e proved to have the best combination of ligand
parameters for high enantioselectivity. For preliminary calcula-
tions, the ligand was simplified by removing the benzylidene
protecting group (Figure 2) in an attempt to accelerate the DFT
calculation and verify whether the method gave satisfactory
results. The method then was used to explore the reactions
involving the complete ligand.
Initially, we applied the calculation for the trisubstituted olefin
S1. Both 3/5-MI and 3/5-Meta mechanisms were investigated
for all approach vectors of the alkene. Table 4 shows the
calculated energies for the most stable isomers of complex C
and for the most stable isomers of the transition states (TS and
TS0) with the truncated ligand. These key isomers are the result
of varying the relative coordination of the substrate (si face or re
face) and the relative position of the hydride (up or down).28 The
13640
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
Figure 3. Calculated transition states (TS) for the major and the minor pathways with the full ligand. Hydrogen atoms of the ligand L5e and substrate
S1 have been omitted for clarity.
Figure 4. Quadrant diagram describing the enantioselective substrateꢀligand interactions.
generated a pocket that is well suited to E-olefins. Ligand L5c
differs from the previous ligand L5e in the presence of bulky
substituents at the para position of the biphenyl group. These
substituents increase the dihedral angle of the biaryl group, which
results in lower occupancy of the lower right quadrant than with
the previous Ir/L5e catalytic system. So, the substituents of the
biphenyl moieties can tune the steric hindrance of the lower right
quadrant so that it can accommodate the phenyl substituent of
Z-substrates and lead to high enantioselectivity. The same
explanation may account for the excellent enantioselectivities
obtained with triarylsubstituted olefins S9 and S10.
Finally, we did the calculation for the 1,1-disubstituted olefin
S26, using first the truncated ligand structure of L5e without the
benzylidene protecting group (Figure 2). We found that the
results followed the same trend as for the trisubstituted olefin
(Table 5). Therefore, the most stable transition state observed
matches the major product, and the preferred pathway is the 3/5-
MI with an energy difference of 15 kJ/mol with respect to the
3/5-Meta pathway. Using the full structure of ligand L5e, we
observe that the difference in energy between the transition
states for the major and the minor products is 9.5 kJ/mol
(Figure 5). Thus, not only is the major isomer correctly identified
for this challenging system, but the quantitative prediction of
96% ee at ambient temperature matches well with the enantios-
electivity obtained experimentally (99% ee, Table 2, entry 9).
In summary, the substrate versatility of our Ir/phosphiteꢀ
oxazoline catalytic system is higher than that of related Irꢀpho-
sphine/phosphiniteꢀnitrogen catalysts because the biaryl phos-
phite moiety makes the catalyst more flexible. Therefore, subtle
variations in the dihedral angle of the biaryl phosphite moiety,
which may arise from changing the biaryl substituents and/or be
imposed by the substrate itself, lead to a different occupancy of
the semihindered quadrant. As a result, the catalyst can adapt its
chiral environment to a range of substrate types.
3. CONCLUSIONS
The modular ligand design has been shown to be highly
successful, both in finding highly selective ligands for each
substrate and in identifying two general ligands (L5c and L5e)
with good performance over the entire range of substrates. The
good performance extends even to the very challenging class of
terminally disubstituted olefins. The enantioselectivity was below
90% for olefins with two similarly sized substituents, such as aryl
13641
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
versus aryl or s-alkyl, but even a moderate size difference like aryl
versus n-alkyl allowed good enantioselectivities in the range
90ꢀ99%. It should be pointed out that these catalysts are also
very tolerant to the presence of neighboring polar group. Thus, a
range of allylic alcohols, acetates, R,β-unsaturated ketones, R,
β-unsaturated esters, and vinylboronates were hydrogenated in
high enantioselectivities (ee’s up to >99%).
The computational study allowed identification of the pre-
ferred reaction path, an Ir(III/V) cycle with migratory insertion
of a hydride as the selectivity-determining step.23 The alternative
metathesis mechanism22a,b was consistently higher in energy. For
selected models, the DFT calculations allowed computational
determination of the observed selectivities with high accuracy.
Both the favored enantiomer and the effect of ligand modifica-
tions could be rationalized in terms of the previously proposed
quadrant model.
4. EXPERIMENTAL SECTION
4.1. General Considerations. All reactions were carried out using
standard Schlenk techniques under an argon atmosphere. Solvents were
purified and dried by standard procedures. Phosphorochloridites are
easily prepared in one step from the corresponding biaryls.29 Ligands
L1ꢀL3c, L5aꢀe, and L5hꢀk were prepared as previously described.5a,7
[Ir(cod)(L)]BArF (L = L1ꢀL3c, L5cꢀe, L5h,i, and L5k) were
prepared previously.5a 1H, 13C{1H}, and 31P{1H} NMR spectra were
recorded using a 400 MHz spectrometer. Chemical shifts are relative to
that of SiMe4 (1H and 13C) as internal standard or H3PO4 (31P) as
external standard. 1H, 13C, and 31P assignments were made on the basis
Table 5. Calculated Energies for the Most Stable Isomers of
Complex C and Transition States TS and TS0 with Substrate
S26 Using the Truncated Ligand System
1
1
1
13
1
31
of Hꢀ H gCOSY, Hꢀ C gHSQC, and Hꢀ P gHMBC experi-
ments. All catalytic experiments were performed three times.
4.2. Computational Details. Geometries of all substrates were
optimized using the Jaguar program25 and both the B3LYP hybrid
density functional26 and the LACVP* basis sets. The complexes were
treated with charge = +1 and in the singlet state. In some cases, structures
were first optimized using geometric constraints to generate starting
structures that were subsequently optimized without geometric con-
straints. No symmetry constraints were applied. Normal-mode analysis
of stable structure revealed no imaginary frequencies or a single
imaginary frequency with negligibly low frequency (ν < 100 cmꢀ1);
those of transition states had a single imaginary frequency of higher
negative frequency (usually ν < ꢀ500 cmꢀ1). Reported energies are the
Gibbs free energy at 298.15 K. LACVP in Jaguar defines a combination
of the LANL2DZ basis set30 for iridium and the 6-31G basis set for other
atoms. Final energies were retrieved from single-point calculations at the
B3LYP/LACV3P++* level of theory. LACV3P++* differs from LACVP*
because it uses the 6-311++G* basis set instead of 6-31G*, one additional
p function, and two additional d functions on Ir. Energies in CH2Cl2
solution were calculated as single-point energies from optimized
Figure 5. Calculated transitions states (TS) for the major and the minor pathways with the full ligand. Hydrogen atoms of the ligand L5e and substrate
S26 have been omitted for clarity.
13642
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
structures at the B3LYP/LACVP* level of theory using a Poissonꢀ
Boltzmann continuous field with ε = 8.93 and F = 1.3266 g/mL to
calculate the solvent radius (2.33 Å).31 Final energies were corrected by
inclusion of the van der Waals repulsion energy calculated by DFT-D3.27
4.3. General Procedure for the Preparation of the Phos-
phiteꢀOxazoline Ligands. The corresponding phosphorochlori-
dite (3.0 mmol) produced in situ was dissolved in toluene (12.5 mL),
and pyridine (1.14 mL, 14 mmol) was added. The corresponding
hydroxylꢀoxazoline compound (2.8 mmol) was azeotropically dried
with toluene (3 ꢂ 2 mL) and then dissolved in toluene (12.5 mL) to
which pyridine (1.14 mL, 14 mmol) was added. The oxazoline solution
was transferred slowly at 0 °C to the solution of phosphorochloridite.
The reaction mixture was stirred overnight at 80 °C, and the pyridine
salts were removed by filtration. Evaporation of the solvent gave a white
foam, which was purified by flash chromatography in alumina to produce
the corresponding ligand as a white solid.
(d, 1H, H-1, 3J1ꢀ2= 7.6 Hz), 5.53 (s, 1H, H-7), 7.1ꢀ8.5 (m, 26H, CHd,
aromatics). 13C NMR (CDCl3), δ: 15.3 (CH2), 22.8 (b, CH2, cod), 28.9
(b, CH2, cod), 31.3 (CH3, tBu), 31.4 (CH3, tBu), 31.6 (CH3, tBu), 32.4
t
t
(b, CH2, cod), 32.9 (b, CH2, cod), 34.9 (C, Bu), 35.1 (C, Bu), 35.5
(C, tBu), 36.0 (C, tBu), 64.7 (C-5), 67.1 (CHd, cod), 70.4 (C-2), 70.7
(C-6), 72.3 (CHd, cod), 79.5 (d, C-3, 2JCꢀP = 20.5 Hz), 80.3 (C-4),
99.6 (b, CHd, cod), 103.0 (C-7), 103.2 (C-1), 104.1 (d, CHd, cod,
JCꢀP = 12.2 Hz), 117.7 (b, CHd, BArF), 119ꢀ134 (aromatic carbons),
135.0 (b, CHd, BArF), 136ꢀ149 (aromatic carbons), 161.9 (q, CꢀB,
BArF, 1JCꢀB = 49.6 Hz), 168.7 (CdN). Anal. Calcd for C89H84BF24Ir-
NO7P: C, 54.27; H, 4.30; N, 0.71. Found: C, 54.32; H, 4.35; N, 0.68.
[Ir(cod)(L5a)]BArF. Yield: 118 mg (92%). 31P (CDCl3), δ: 112.7 (s).
1H (CDCl3), δ: 1.7ꢀ2.4 (b, 8H, CH2, cod), 3.73 (m, 2H, H-5, H-60),
3.81 (m, 2H, CHd, cod and H-4), 4.24 (b, 1H, CHd, cod), 4.41
(m, 2H, H-6, H-2), 4.71 (b, 1H, CHd, cod), 4.77 (m, 1H, H-3), 5.33 (b, 1H,
CHd, cod), 5.69 (s, 1H, H-7), 6.03 (d, 1H, H-1, 3J1,2 = 7.2 Hz), 7.1ꢀ8.2
(m, 30H, CHd). 13C (CDCl3), δ: 22.4 (b, CH2, cod), 23.1 (b, CH2,
cod), 31.9 (b, CH2, cod), 32.3 (b, CH2, cod), 63.2 (C-5), 68.9 (C-6),
70.2 (C-2), 76.7 (C-3), 80.0 (C-4), 100.5 (d, CHd, cod, JCꢀP = 18 Hz),
101.6 (C-7), 102.7 (C-1), 103.3 (d, CHd, cod, JCꢀP = 22 Hz), 117.8
(b, CHd, BArF), 119ꢀ134 (aromatic carbons), 135.1 (b, CHd, BArF),
136ꢀ149 (aromatic carbons), 161.7 (q, CꢀB, BArF, 1JCꢀB = 49.6 Hz),
170.3 (CdN). Anal. Calcd for C72H50BF24IrNO7P: C, 49.95; H, 2.91;
N, 0.81. Found: C, 50.02; H, 2.93; N, 0.79.
L5f. Yield: 1.0 g (48%). 31P NMR (CDCl3), δ: 147.3 (s). 1H NMR
(C6D6), δ: 1.44 (s, 9H, CH3, tBu), 1.69 (s, 9H, CH3, tBu), 1.74 (s, 3H,
CH3ꢀAr), 1.81 (s, 3H, CH3ꢀAr), 2.09 (s, 3H, CH3ꢀAr), 2.11 (s, 3H,
CH3ꢀAr), 3.52 (m, 1H, H-60), 3.67 (m, 1H, H-5), 3.78 (m, 1H, H-4),
4.11 (m, 1H, H-2), 4.27 (m, 1H, H-6), 4.92 (m, 1H, H-3), 5.52 (d, 1H,
H-1, 3J1ꢀ2= 7.6 Hz), 5.57 (s, 1H, H-7), 7.14ꢀ7.83 (m, 12H, CHd). 13
C
NMR (C6D6), δ: 16.1 (CH3ꢀAr), 16.4 (CH3ꢀAr), 19.8 (CH3ꢀAr),
20.0 (CH3ꢀAr), 32.9 (CH3, tBu), 33.1 (CH3, tBu), 35.9 (C, tBu), 36.0
[Ir(cod)(L5b)]BArF. Yield: 134 mg (93%). 31P NMR (CDCl3), δ:
113.5 (s). 1H NMR (CDCl3), δ: 1.8ꢀ2.1 (b, 6H, CH2, cod), 2.19 (s, 3H,
CH3), 2.28 (b, 5H, CH2, cod and CH3), 2.34 (s, 3H, CH3), 2.39 (s, 3H,
CH3), 3.54 (m, 2H, CHd, cod and H-60), 3.63 (m, 1H, H-4), 3.68
(m, 1H, CHd, cod and H-5), 4.21 (m, 1H, H-6), 4.33 (m, 1H, H-2),
4.39 (b, 1H, CHd, cod), 4.81 (m, 1H, H-3), 5.11 (b, 1H, CHd, cod),
5.46 (s, 1H, H-7), 5.71 (d, 1H, H-1, 2J1ꢀ2 = 7.5 Hz), 7.1ꢀ8.5 (m, 26H,
aromatics). 13C NMR (CDCl3), δ: 17.2 (CH3), 17.4 (CH3), 21.1
(CH3), 21.3 (CH3), 23.1 (b, CH2, cod), 25.4 (b, CH2, cod), 33.4
(b, CH2, cod), 33.6 (b, CH2, cod), 68.9 (C-6), 68.4 (CHd, cod), 69.5
(C-2), 70.5 (CHd, cod), 78.6 (d, C-3, 2Jc-p = 12 Hz), 78.4 (C-4), 99.9
(d, CHd, cod, JCꢀP = 22.2 Hz), 102.1 (C-7), 103.2 (C-1), 104.3
(d, CHd, cod, JCꢀP = 16 Hz), 117.6 (b, CHd, BArF), 119ꢀ134
(aromatic carbons), 135.1 (b, CHd, BArF), 136ꢀ149 (aromatic
carbons), 161.6 (q, CꢀB, BArF, 1JCꢀB = 49.6 Hz), 170.4 (CdN). Anal.
Calcd for C76H58BF24IrNO7P: C, 51.07; H, 3.27; N, 0.78. Found: C,
51.11; H, 3.32; N, 0.80.
t
(C, Bu), 64.5 (C-5), 70.3 (C-6), 70.5 (C-2), 79.1 (C-3), 81.0 (C-4),
102.9.0 (C-7), 103.2 (C-1), 125.7 (CHd), 125.8 (CHd), 127.0
(CHd), 128,1 (CHd), 128.4 (CHd), 128.6 (CHd), 129.1 (CHd),
129.4 (CHd), 129,6 (CHd), 129.7 (CHd), 129.9 (CHd), 130.4
(CHd), 130.7 (CHd), 130.9 (C), 131.8 (C), 132.1 (C), 132.2 (C),
132.5 (C), 134.4 (C), 135.2 (C), 137.3 (C), 137.6 (C), 138.0 (C), 144.6
(C), 145.3 (C), 165.7 (CdN). Anal. Calcd for C44H50NO7P: C, 71.82;
H, 6.85; N, 1.90. Found: C, 71.87; H, 6.93; N, 1.88.
L5g. Yield: 1.1 g (52%). 31P NMR (CDCl3), δ: 147.8 (s). 1H NMR
(C6D6), δ: 1.43 (s, 9H, CH3, tBu), 1.63 (s, 9H, CH3, tBu), 1.71 (s, 3H,
CH3ꢀAr), 1.83 (s, 3H, CH3ꢀAr), 2.11 (s, 6H, CH3ꢀAr), 3.53 (m, 1H,
H-60), 3.68 (m, 1H, H-5), 3.73 (m, 1H, H-4), 4.14 (m, 1H, H-2), 4.25
(m, 1H, H-6), 4.75 (m, 1H, H-3), 5.51 (d, 1H, H-1, 3J1ꢀ2= 7.6 Hz), 5.58
(s, 1H, H-7), 7.14ꢀ7.83 (m, 12H, CHd). 13C NMR (C6D6), δ: 16.1
t
(CH3ꢀAr), 16.4 (CH3ꢀAr), 19.9 (CH3ꢀAr), 33.1 (CH3, Bu), 35.9
(C, tBu), 36.0 (C, tBu), 64.6 (C-5), 70.2 (C-6), 71.2 (C-2), 79.8 (C-3),
80.5 (C-4), 102.9 (C-7), 103.1 (C-1), 125.8 (CHd), 125.9 (CHd),
126.8 (CHd), 128,0 (CHd), 128.6 (CHd), 128.9 (CHd), 129.5
(CHd), 129.7 (CHd), 129.8 (CHd), 130.7 (CHd), 130.9 (CHd),
131.7 (CHd), 131.9 (C), 132.3 (C), 132.5 (C), 132.8 (C), 134.5 (C),
135.2 (C), 137.3 (C), 137.6 (C), 138.0 (C), 144.6 (C), 145.3 (C), 165.9
(CdN). Anal. Calcd for C44H50NO7P: C, 71.82; H, 6.85; N, 1.90.
Found: C, 71.85; H, 6.89; N, 1.89.
[Ir(cod)(L5f)]BArF. Yield: 134 mg (94%). 31P NMR (CDCl3), δ:
1
108.4 (s). H NMR (CDCl3), δ: 1.41 (s, 9H, CH3, tBu), 1.57 (s, 9H,
t
CH3, Bu), 1.69 (s, 3H, CH3ꢀAr), 1.77 (s, 3H, CH3ꢀAr), 1.8ꢀ2.1
(b, 7H, CH2, cod and CH3ꢀAr), 2.14 (s, 3H, CH3ꢀAr), 2.2ꢀ2.4 (b, 4H,
CH2, cod), 3.52 (m, 1H, H-60), 3.65 (m, 1H, CHd, cod), 3.68 (m, 1H,
H-5), 3.82 (m, 1H, H-4), 4.08 (b, 2H, CHd, cod and H-2), 4.23(m, 1H,
H-6), 4.79 (b, 1H, CHd, cod), 4.83 (m, 1H, H-3), 5.32 (b, 1H, CHd,
cod), 5.49 (d, 1H, H-1, 3J1ꢀ2 = 7.6 Hz), 5.59 (s, 1H, H-7), 7.1ꢀ8.3 (m,
24H, aromatics). 13C NMR (CDCl3), δ: 16.1 (CH3ꢀAr), 16.4
(CH3ꢀAr), 17.8 (CH3ꢀAr), 19.1 (CH3ꢀAr), 26.8 (b, CH2, cod),
4.4. TypicalProcedureforthePreparationof[Ir(cod)(L)]BArF.
The corresponding ligand (0.074 mmol) was dissolved in CH2Cl2
(2 mL), and [Ir(μ-Cl)cod]2 (25 mg, 0.037 mmol) was added. The reac-
tion was refluxed at 50 °C for 1 h. After 5 min at room temperature,
NaBArF (77.1 mg, 0.082 mmol) and water (2 mL) were added, and the
reaction mixture was stirred vigorously for 30 min at room temperature.
The phases were separated, and the aqueous phase was extracted twice
with CH2Cl2. The combined organic phases were filtered through a Celite
plug, dried with MgSO4, and the solvent was evaporated to give the
product as an orange solid.
t
t
27.6 (b, CH2, cod), 31.3 (CH3, Bu), 31.4 (CH3, Bu), 33.7 (b, CH2,
cod), 34.9 (C, tBu), 35.1 (C, tBu), 64.7 (C-5), 68.8 (CHd, cod), 70.5
(C-2), 71.1 (CHd, cod), 79.2 (C-3), 81.3 (C-4), 100.4 (d, CHd, cod,
JCꢀP = 21.8 Hz), 102.7 (C-7), 103.1 (b, CHd, cod), 103.6 (C-1), 117.4
(b, CHd, BArF), 119ꢀ134 (aromatic carbons), 135.1 (b, CHd, BArF),
136ꢀ149 (aromatic carbons), 161.6 (q, CꢀB, BArF, 1JCꢀB = 49.6 Hz),
169.4 (CdN). Anal. Calcd for C82H74BF24IrNO7PSi2: C, 50.99; H,
3.86; N, 0.73. Found: C, 51.02; H, 3.92; N, 0.72.
[Ir(cod)(L4c)]BArF. Yield: 138 mg (95%). 31P NMR (CDCl3), δ:
1
t
104.3 (s). H NMR (CDCl3), δ: 1.36 (s, 9H, CH3, Bu), 1.41 (s, 9H,
CH3, tBu), 1.65 (s, 2H, CH2), 1.70 (s, 9H, CH3, tBu), 1.73 (s, 9H, CH3,
tBu), 1.8ꢀ2.2 (b, 8H, CH2, cod), 3.48 (m, 1H, H-60), 3.65 (m, 2H,
CHd, cod and H-5), 3.84 (m, 1H, H-4), 4.13 (m, 1H, H-2), 4.15 (b, 1H,
[Ir(cod)(L5g)]BArF. Yield: 142 mg (95%). 31P NMR (CDCl3), δ:
1
107.6 (s). H NMR (CDCl3), δ: 1.40 (s, 9H, CH3, tBu), 1.55 (s, 9H,
CH3, tBu), 1.78 (s, 3H, CH3ꢀAr), 1.81 (s, 3H, CH3ꢀAr), 1.8ꢀ2.1 (b,
7H, CH2, cod and CH3ꢀAr), 2.2ꢀ2.4 (b, 7H, CH2, cod and CH3ꢀAr),
3.57 (m, 2H, H-60 and CHd, cod), 3.69 (m, 1H, H-5), 3.95 (m, 1H,
CHd, cod), 4.24 (dd, 1H, H-6, 2J6ꢀ6 = 10.0 Hz, 3J6ꢀ5 = 4.8 Hz), 4.79
0
(b, 1H, CHd, cod), 4.92 (m, 1H, H-3), 5.18 (b, 1H, CHd, cod), 5.49
13643
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
H-4), 4.11 (b, 2H, CHd, cod and H-2), 4.19 (m, 1H, H-6), 4.27 (b, 1H,
CHd, cod), 4.79 (m, 1H, H-3), 5.28 (b, 1H, CHd, cod), 5.32 (d, 1H,
H-1, 3J1ꢀ2 = 7.6 Hz), 5.57 (s, 1H, H-7), 7.1ꢀ8.3 (m, 24H, aromatics).
13CNMR(CDCl3), δ:16.3(CH3ꢀAr), 16.5 (CH3ꢀAr), 18.4 (CH3ꢀAr),
18.9 (CH3ꢀAr), 27.2 (b, CH2, cod), 27.4 (b, CH2, cod), 31.6 (CH3,
(C, tBu), 35.1 (C, tBu), 64.9 (C-5), 71.3 (CHd, cod), 71.6 (CHd, cod),
71.9 (C-2), 79.4 (C-3), 80.4 (C-4), 100.9 (b, CHd, cod), 102.7 (C-7),
103.3 (b, CHd, cod), 103.7 (C-1), 117.5 (b, CHd, BArF), 119ꢀ134
(aromatic carbons), 135.1 (b, CHd, BArF), 136ꢀ149 (aromatic
carbons), 161.7 (q, CꢀB, BArF, 1JCꢀB = 49.6 Hz), 169.1 (CdN). Anal.
Calcd for C82H74BF24IrNO7PSi2: C, 50.99; H, 3.86; N, 0.73. Found: C,
51.09; H, 3.96; N, 0.71.
’ REFERENCES
(1) (a) Asymmetric Catalysis in Industrial Scale: Challenges, Ap-
proaches and Solutions; Blaser, H. U., Schmidt, E., Eds.; Wiley: Weinheim,
Germany, 2003. (b) Catalytic Asymmetric Synthesis; Ojima, I., Ed.; Wiley-
VCH: New York, 2000. (c) Brown, J. M. In Comprehensive Asymmetric
Catalysis; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-
Verlag: Berlin, 1999; Vol. I, pp 121ꢀ182. (d) Asymmetric Catalysis in
Organic Synthesis; Noyori, R., Ed.; Wiley: New York, 1994. (e) Applied
Homogeneous Catalysis with Organometallic Compounds, 2nd ed.; Cornils,
B., Herrmann, W. A., Eds.; Wiley-VCH: Weinheim, 2002.
(2) For recent reviews, see: (a) K€allstr€om, K.; Munslow, I.; Anders-
son, P. G. Chem.-Eur. J. 2006, 12, 3194. (b) Roseblade, S. J.; Pfaltz, A. Acc.
Chem. Res. 2007, 40, 1402. (c) Church, T. L.; Andersson, P. G. Coord.
Chem. Rev. 2008, 252, 513. (d) Cui, X.; Burgess, K. Chem. Rev. 2005,
105, 3272. (e) Pꢀamies, O.; Andersson, P. G.; Diꢁeguez, M. Chem.-Eur. J.
2010, 16, 14232.
(3) Chelating oxazolineꢀcarbene ligands, mainly developed by
Burgess and co-workers, have also been successfully used and also to a
less extent chiral diphosphines. See, for instance: (a) Perry, M. C.; Cui,
X.; Powell, M. T.; Hou, D. R.; Reibenspies, J. H.; Burgess, K. J. Am. Chem.
Soc. 2003, 125, 113. (b) Cui, X.; Burgess, K. J. Am. Chem. Soc. 2003,
125, 14212. (c) Cui, X.; Ogle, J. W.; Burgess, K. Chem. Commun.
2005, 672. (d) Zhao, J.; Burgess, K. J. Am. Chem. Soc. 2009, 131, 13236.
(e) Co, T. T.; Kim, T. J. Chem. Commun. 2006, 3537. (f) Forman, G. S.;
Ohkuma, T.; Hems, W. P.; Noyori, R. Tetrahedron Lett. 2000, 41,
9471.
(4) See, for instance: (a) Perry, M. C.; Cui, X.; Powell, M. T.; Hou,
D.-R.; Reibenspies, J. H.; Burgess, K. J. Am. Chem. Soc. 2003, 125, 5391.
(b) Blankestein, J.; Pfaltz, A. Angew. Chem., Int. Ed. 2001, 40, 4445. (c)
Kaiser, S.; Smidt, S. P.; Pfaltz, A. Angew. Chem., Int. Ed. 2006, 45, 5194.
(d) K€allstr€om, K.; Hedberg, C.; Brandt, P.; Bayer, P.; Andersson, P. G.
J. Am. Chem. Soc. 2004, 126, 14308. (e) Engman, M.; Diesen, J. S.;
Paptchikhine, A.; Andersson, P. G. J. Am. Chem. Soc. 2007, 129, 4536. (f)
Trifonova, A.; Diesen, J. S.; Andersson, P. G. Chem.-Eur. J. 2006,
12, 2318. (g) Menges, G.; Pfaltz, A. Adv. Synth. Catal. 2002,
334, 4044. (h) Drury, W. J., III; Zimmermann, N.; Keenan, M.; Hayashi,
M.; Kaiser, S.; Goddard, R.; Pfaltz, A. Angew. Chem., Int. Ed. 2004, 43, 70.
(i) Hedberg, C.; K€allstr€om, K.; Brandt, P.; Hansen, L. K.; Andersson, P. G.
J. Am. Chem. Soc. 2006, 128, 2995. (j) Franzke, A.; Pfaltz, A. Chem.-Eur.
J.2011, 17, 4131. (k) Tang, W.; Wang, W.; Zhang, X. Angew. Chem.,
Int. Ed. 2003, 42, 943. (l) Hou, D.-R.; Reibenspies, J.; Colacot, T. J.;
Burgess, K. Chem.-Eur. J. 2001, 7, 5391. (m) Cozzi, P. G.; Menges, F.;
Kaiser, S. Synlett 2003, 833. (n) Lightfoot, A.; Schnider, P.; Pfaltz, A.
Angew. Chem., Int. Ed. 1998, 37, 2897. (o) Menges, F.; Neuburger,
M.; Pfaltz, A. Org. Lett. 2002, 4, 4713. (p) Liu, D.; Tang, W.; Zhang,
X. Org. Lett. 2004, 6, 513. (q) Lu, S.-M.; Bolm, C. Angew. Chem., Int.
Ed. 2008, 47, 8920. (r) Lu, W.-J.; Chen, Y.-W.; Hou, X.-L. Adv. Synth.
Catal. 2010, 352, 103. (s) Paptchikhine, A.; Cheruku, P.; Engman,
M.; Andersson, P. G. Chem. Commun. 2009, 5996.
t
tBu), 31.9 (CH3, Bu), 32.4 (b, CH2, cod), 32.8 (b, CH2, cod), 34.9
[Ir(cod)(L5i)]BArF. This compound has been prepared following a
slight modification of the general procedure.9 Yield: 115 mg (91%). 31
P
NMR (CDCl3), δ: 107.7 (s). 1H NMR (CDCl3), δ: 1.6ꢀ2.5 (m, 8H,
CH2 cod), 3.85 (b, 3H, CHd, cod, H-5 and H-60), 3.91 (m, 1H, H-3),
3.98 (b, 2H, CHd, cod and H-4), 4.01 (b, 1H, CHd, cod), 4.33 (m, 1H,
H-6), 4.71 (m, 1H, H-2), 4.99 (b, 1H, CHd, cod), 5.45 (s, 1H, H-7),
3
5.51 (b, 1H, CHd, cod), 6.31 (d, 1H, H-1, J1ꢀ2 = 6.4 Hz), 7.1ꢀ8.3
(m, 32H, CHd, aromatics). 13C NMR (CDCl3), δ: 26.3 (b, CH2, cod),
28.5 (b, CH2, cod), 32.3 (b, CH2, cod), 34.5 (b, CH2, cod), 64.5 (CHd,
cod), 65.9 (C-5), 69.8 (C-6), 70.8 (CHd, cod), 70.9 (C-2), 74.3 (C-4),
80.3 (C-3), 98.4 (d, CHd, cod, JCꢀP = 22 Hz), 101.8 (s, C-7), 102.9
(d, CHd, cod, JCꢀP = 12.2 Hz), 103.8 (s, C-1), 117.7 (b, CHd, BArF),
120ꢀ134 (aromatic carbons), 135.0 (b, CHd, BArF), 135ꢀ137
1
(aromatic carbons), 161.7 (q, CꢀB BArF, JCꢀB = 49 Hz), 170.3
(s, CdN). Anal. Calcd for C72H52BF24IrNO5P: C, 50.83; H, 3.08; N,
0.82. Found: C, 51.02; H, 3.14; N, 0.80.
4.5. Typical Procedure for the Hydrogenation of Olefins.
The alkene (1 mmol) and Ir complex (0.2 mol %) were dissolved in
CH2Cl2 (2 mL) in a high-pressure autoclave, which was purged four
times with hydrogen. Next, it was pressurized at the desired pressure.
After the desired reaction time, the autoclave was depressurized and the
solvent evaporated off. The residue was dissolved in Et2O (1.5 mL) and
filtered through a short Celite plug. The enantiomeric excess was
determined by chiral GC or chiral HPLC, and conversions were deter-
mined by 1H NMR. The enantiomeric excesses of hydrogenated products
were determined using the conditions previously described.4d,q,s,5c
’ ASSOCIATED CONTENT
S
Supporting Information. All energies (electronic, nuclear
b
repulsion, free energies, ZPE, and DFT-D3 attraction correction)
and structures for all complexes and transition states. Full list of
catalytic results. This material is available free of charge via the
Internet at http://pubs.acs.org.
(5) (a) Diꢁeguez, M.; Mazuela, J.; Pꢀamies, O.; Verendel, J. J.;
Andersson, P. G. J. Am. Chem. Soc. 2008, 130, 7208. (b) Diꢁeguez, M.;
Mazuela, J.; Pꢀamies, O.; Verendel, J. J.; Andersson, P. G. Chem. Commun.
2008, 3888. (c) Mazuela, J.; Verendel, J. J.; Coll, M.; Sch€affner, B.;
B€orner, A.; Andersson, P. G.; Pꢀamies, O.; Diꢁeguez, M. M. J. Am. Chem.
Soc. 2009, 131, 12344. (d) Mazuela, J.; Paptchikhine, A.; Pꢀamies, O.;
Andersson, P. G.; Diꢁeguez, M. Chem.-Eur. J. 2010, 16, 4567.
’ AUTHOR INFORMATION
Corresponding Author
pon@chem.gu.se; pher.andersson@biorg.uu.se; montserrat.die-
guez@urv.cat
(6) One of the limitations of using sugars as precursors for ligands is
that often only one of the enantiomers is readily available. However, this
limitation can be overcome by the rational design of pseudo-enantio-
meric ligands. See, for example: (a) RajanBabu, T. V.; Ayers, T. A.;
Casalnuovo, A. L. J. Am. Chem. Soc. 1994, 116, 4101. (b) RajanBabu,
T. V.; Ayers, T. A.; Halliday, G. A.; You, K. K.; Calabrese, J. C. J. Org.
’ ACKNOWLEDGMENT
We would like to thank the COST D40, the Spanish Govern-
ment for providing grants Consolider Ingenio Intecat-CSD2006-
0003, CTQ2010-15835, 2008PGIR/07 to O.P. and 2008PGIR/
08 to M.D., the Catalan Government for grant 2009SGR116, and
the ICREA Foundation for providing M.D. and O.P. with
financial support through the ICREA Academia awards.
ꢁ
Chem. 1997, 62, 6012. (c) Khiar, N.; Navas, R.; Suꢁarez, B.; Alvarez, E.;
Fernꢁandez, I. Org. Lett. 2008, 10, 3697. (d) Nakano, H.; Yokohama, J.;
Okuyama, Y.; Fujita, R.; Hongo, H. Tetrahedron: Asymmetry 2003,
14, 2361. (e) Ruffo, F.; Del Lito, R.; De Roma, A.; D’Errico, A.;
Magnolia, S. Tetrahedron: Asymmetry 2006, 17, 2265.
13644
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645

Journal of the American Chemical Society
ARTICLE
(7) (a) Mata, Y.; Pꢀamies, O.; Diꢁeguez, M.; Claver, C. Adv. Synth.
Catal. 2005, 347, 1943. (b) Mata, Y.; Pꢀamies, O.; Diꢁeguez, M. Chem.-Eur.
J. 2007, 13, 3296. (c) Mata, Y.; Pꢀamies, O.; Diꢁeguez, M. Adv. Synth. Catal.
2009, 351, 3217.
(8) Yonehara, K.; Jashizume, T.; Mori, K.; Ohe, K.; Uemura, S. J. Org.
Chem. 1999, 64, 9374.
(9) For the preparation of [Ir(cod)(L5j)]BArF, the anion exchange
has been performed at low temperature (0 °C) without the presence
of water.
(23) Church, T. L.; Rasmussen, T.; Andersson, P. G. Organometallics
2010, 29, 6769.
(24) Recently, the group of Hopmann et al. has performed a
computationally study using a phosphineꢀoxazoline (PHOX)-based
iridium catalyst, which indicates that the hydrogenation of S1 follows the
3/5-Meta pathway. See: Hopmann, K. H.; Bayer, A. Organometallics
2011, 30, 2483.
(25) Jaguar, version 7.0; Schr€odinger, LLC: New York, 2010;
http://www.schrodinger.com/.
(10) The rapid ring inversions (tropoisomerization) in the biaryl
phosphite moiety are usually stopped upon coordination to the metal
center. See, for example: (a) Buisman, G. J. H.; van der Veen, L. A.;
Klootwijk, A.; de Lange, W. G. J.; Kamer, P. C. J.; van Leeuwen, P. W. N.
M.; Vogt, D. Organometallics 1997, 16, 2929. (b) Diꢁeguez, M.; Pꢀamies,
O.; Ruiz, A.; Claver, C.; Castillꢁon, S. Chem.-Eur. J. 2001, 7, 3086. (c)
Pꢀamies, O.; Diꢁeguez, M.; Net, G.; Ruiz, A.; Claver, C. Organometallics
2000, 19, 1488. (d) Pꢀamies, O.; Diꢁeguez, M.; Net, G.; Ruiz, A.; Claver, C.
J. Org. Chem. 2001, 66, 8364. (e) Pꢀamies, O.; Diꢁeguez, M. Chem.-Eur. J.
2008, 14, 944.
(26) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.;
Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(27) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
2010, 132, 154104.
(28) The rest of the possible isomers generated from considering the
hydride attack to the less substituted carbon of the olefin have not been
considered, because it has been previously demonstrated that they
proceed with higher energies. See ref 4i.
(29) Buisman, G. J. H.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.
Tetrahedron: Asymmetry 1993, 4, 1625.
(11) (a) Rovner, E. S.; Wein, A. J. Eur. Urol. 2002, 41, 6. (b) Wefer, J.;
Truss, M. C.; Jonas, U. World J. Urol. 2001, 19, 312. (c) Hills, C. J.;
Winter, S. A.; Balfour, J. A. Drugs 1998, 55, 813. (d) McRae, A. L.; Brady,
K. T. Expert Opin. Pharmacother. 2001, 2, 883. (e) Gordaliza, M.; García,
P. A.; Miguel del Corral, J. M.; Castro, M. A.; Gꢁomez-Zurita, M. A.
Toxicon 2004, 44, 441.
(30) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(31) (a) Tannor, D. J.; Marten, B.; Murphy, R.; Friesner, R. A.;
Sitkoff, D.; Nicholls, A.; Honig, B.; Ringnalda, M.; Goddard, W. A. J. Am.
Chem. Soc. 1994, 116, 11875. (b) Marten, B.; Kim, K.; Cortis, C.;
Friesner, R. A.; Murphy, R. B.; Ringnalda, M. N.; Sitkoff, D.; Honig, B.
J. Phys. Chem. 1996, 100, 11775.
(12) Tolstoy, P.; Engman, M.; Paptchikhine, A.; Bergquist, J.; Church,
T. L.; Leung, A. W.-M.; Andersson, P. G. J. Am. Chem. Soc. 2009, 131, 8855.
(13) For successful application of Ir-catalysts, see: (a) McIntyre, S.;
H€ormann, E.; Menges, F.; Smidt, S. P.; Pfaltz, A. Adv. Synth. Catal. 2005,
347, 282.(b) Reference 4b.
(14) For successful applications of Sm-, Ru-, and Rh-complexes, see:
(a) Conticello, V. P.; Brard, L.; Giardello, M. A.; Tsuji, Y.; Sabat, M.;
Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1992, 114, 2761. (b) Giardello,
M. A.; Conticello, V. P.; Brard, L.; Gagnꢁe, M. R.; Marks, T. J. J. Am. Chem.
Soc. 1994, 116, 10241.(c) Reference 3e. (d) Reference 3f.
(15) (a) Fessard, T. C.; Andrews, S. P.; Motoyohsi, H.; Carreira, E.
Angew. Chem., Int. Ed. 2007, 46, 9331. (b) Prat, L.; Dupas, G.; Duflos, J.;
Quꢁeguiner, G.; Bourguignon, J.; Levacher, V. Tetrahedron Lett. 2001,
42, 4515. (c) Wilkinson, J. A.; Rossington, S. B.; Ducki, S.; Leonard, J.;
Hussain, N. Tetrahedron 2006, 62, 1833.
(16) (a) Okamoto, K.; Nishibayashi, Y.; Uemura, S.; Toshimitsu, A.
Angew. Chem., Int. Ed. 2005, 44, 3588. (b) Hatanaka, Y.; Hiyama, T.
J. Am. Chem. Soc. 1990, 112, 7793.
(17) See, for instance: (a) Abate, A.; Brenna, E.; Fuganti, C.; Gatti,
G. G.; Givenzana, T.; Malpezzi, L.; Serra, S. J. Org. Chem. 2005, 70, 1281.
(b) Drescher, K.; Haupt, A.; Unger, L.; Rutner, S. C.; Braje, W.; Grandel,
R.; Henry, C.; Backfisch, G.; Beyerbach, A.; Bisch, W. WO Patent 2006/
040182 A1, 2006.
(18) See, for instance: (a) Schlosser, M. Angew. Chem., Int. Ed. 1998,
37, 1496. (b) Smart, B. E. J. Fluorine Chem. 2001, 109, 3.(c) Asymmetric
Fluoroorganic Chemistry: Synthesis, Application and Future Directions;
Ramachandran, P. V., Ed.; American Chemical Society: Washington,
DC, 2000. (d) A special issue has been devoted to “Fluorine in the Life
Sciences”. ChemBioChem 2004, 5, 559.
(19) See, for instance: (a) Fleming, I.; Barbero, A.; Walter, D. Chem.
Rev. 1997, 97, 2063. (b) Jones, G. R.; Landais, Y. Tetrahedron 1996,
52, 7599. (c) Bains, W.; Tacke, R. Curr. Opin. Drug Discovery Dev. 2003,
6, 526.
(20) (a) Feldgus, S.; Landis, C. R. J. Am. Chem. Soc. 2000, 122,
12714. (b) Donoghue, P. J.; Helquist, P.; Norrby, P.-O.; Wiest, O. J. Am.
Chem. Soc. 2009, 131, 410.
(21) (a) Roseblade, S. J.; Pfaltz, A. C. R. Chim. 2007, 10, 178. (b)
Vazquez-Serrano, L. D.; Owens, B. T.; Buriak, J. M. Chem. Commun.
2002, 2518.
(22) (a) Cui, X.; Fan, Y.; Hall, M. B.; Burgess, K. Chem.-Eur. J. 2005,
11, 6859. (b) Fan, Y.; Cui, X.; Burgess, K.; Hall, M. B. J. Am. Chem. Soc.
2004, 126, 16688. (c) Brandt, P.; Hedberg, C.; Andersson, P. G. Chem.-
Eur. J. 2003, 9, 339.
13645
dx.doi.org/10.1021/ja204948k |J. Am. Chem. Soc. 2011, 133, 13634–13645